BIG TALKER is the name for a new 13.560mHz Shortwave Transmitter Project growing out of the Talking Pixie2 SW Transmitter project. Talking Pixie2 is a 2-transistor Part 15 device most recently covered in the blog “13.560 Back In The News.”
BIG TALKER is being designed as a 3-transistor part 15 device with particular features, including:
Original Pixie2 crystal oscillator;
Buffer stage with low impedance output and extra “punch” to excite the Class C final stage;
Class C Final Stage with transformer driven amplitude modulation;
LM 317 IC user operated power output control.
No “first version” yet exists, but with your help the BIG TALKER will be put together part by part.
Questions or suggestions welcome.
|
|
Get your Quartz Crystals, Including 13.56 MHz for Short Wave and 1.0 MHz for Medium Wave – click here for a full selection! |
Been Working On It
Saturday was spent with pencil, paper, and reference material. Now we have a rough draft circuit to start with, very much a group effort as much of the work has already been contributed by members. Here is a description of the four modules comprising this 3-transistor circuit:
Module 1 – The Oscillator – Is the original Pixie2 crystal oscillator, since it has been seemingly reliable and required no changes at any point in our previous work with Talking Pixie2;
Module 2 – The Buffer – As proposed by Bruce MICRO1700, this circuit is taken from J.R.Cunningham’s buffer amp for his 100mW MW transmitter, Page 14, from his publication LOW POWER RADIO BROADCASTING available here within this site…
part15.us/files.p15/melprb.pdf
Module 3 – The Final – Is a Class C Amplifier, the base suggested by Radio8Z, the collector as taken from the Talking Pixie2, the output filter as submitted by SCWIS;
Module 4 – The Power Control – Is a LM 317 IC suggested by Radio8Z, and controls the power to the final amp.
This circuit has not been built yet by anyone and serves only as a starting point for the project.
The drawing needs to be improved… soon, soon…
http://kdxradio.com/bigtalker.html
Suggestions
Hi Carl,
Your schematic appears to be a good starting point but I think the buffer stage as shown will be a problem.
1. The bias resistors R6 and R7 are too high in value for an RF amplifier because there are capacitances from the base to emitter and base to collector which should be presented with low resistances to obtain high frequency operation.
2. The bias resistor R6 will feed back the output signal to the input and will reduce the stage gain. It would be better to use a voltage divider bias circuit with an emitter resistor. Use of two resistors in the emitter circuit, one bypassed, is recommended This will give design control over the Q point and gain. If you are interested in looking into this I can post the design equations.
3. The Q2 collector inductor labeled RFC should be replaced with a resistor. Use of an inductor in this type of circuit makes it prone to self oscillation at an unknown frequency.
These are just some things to watch out for if you use this circuit.
One of the unknowns (to me at least) is the output voltage from the oscillator. Knowing this would be a big aid in designing the buffer amplifier. Otherwise it would take an educated guess to establish the buffer stage gain. Is there a way you could measure this?
I believe you are on your way to building a very good transmitter.
Neil
On the Job
Thanks again, Neil for your review of this new effort.
Let me concentrate on getting the output voltage from the oscillator.
Even if I have to purchase something, like a real test device.
I’ve just posted a much better drawing of this first draft
Wait here, I’ll be right back.
Oscillator Output
Carl,
A simulation of the oscillator circuit with a 9V supply and a 1k ohm load coupled with 82pF yields an output voltage across the load of 5.6 volts peak to peak. The positive peak is clipped but the negative peak is a sine wave.
The simulation program doesn’t include a crystal so I used a parallel LC circuit in place of the crystal. This may affect the results but if it is reasonably accurate there is plenty of signal present to drive the buffer.
Neil
Growth
Version 1.0.1 of Big Talker is now being drafted, showing changes to Q2 as suggested, but not yet showing values of the parts involved, as they have not been calculated.
Changes:
L5 will be replaced by a resistor (Rx), part number to be assigned, supplying VCC+ to collector;
R6 & R7 will become a voltage divider between VCC+ and ground, biasing the base;
The emitter will have a resistor added (Ry), part number to be assigned.
Radio8Z, I have questions about the oscillator…. the reason it is set at 9-VDC is because the original Pixie2 circuit showed it at 9-VDC. Is there any reason to believe that this voltage should be changed?
Neil, you mentioned that “the positive peak is clipped” into a 1k load. two questions on that…
Is the “load” you are referring to R2 (1.2k)? Or, perhaps you are referring to a theoretical load.
And, is the clipping of the positive peak something we should correct?
Reply to Questions
Carl,
The load I referenced represents my guess at the input resistance of a buffer stage. If it differs from 1k ohms it will just result in a change of output voltage. It is necessary to evaluate a circuit accounting for loading and that is why I did this.
There is no need to correct the clipping since the class C final will produce a clipped waveform anyway. The output filter network will remove this distortion.
I chose 9V since that is what you had originally specified. The only problem with raising the voltage is it increases the power dissipated by the crystal and could cause it to fail. I don’t know what the limit is but you have already reported that it ran for days at 9V so this appears to be safe.
When you pick the resistors for the buffer I recommend keeping the collector and emitter resistors fairly low, such as a few hundred ohms total. This may not be necessary since you are intending to produce milliwatts from the final and it may cause excessive heating in the buffer transistor but RF circuits work best when the impedances are low. If the output of the oscillator is near what was simulated then the buffer doesn’t need much gain. A simple emitter follower (voltage gain = 1) may suffice.
Edited to Add: The simulation program I use is LTSPICE which is available for free on the net compliments of Linear Technologies. It has a nice, intuitive schematic capture GUI and a modest learning curve. The hardest thing I encountered was understanding the transient analysis setup and if you try the program I can give you numbers for this. Simulation allows easy changes in the circuit where the effects can be seen immediately. Your knowledge plate is probably full but you might like this program once you try it.
End Edit
Neil
Knowledge Plate
I love the metaphor “knowledge plate.”
My knowledge plate needs filling and is never full enough.
I’ll try LTSPICE. After all, what plate is complete without spice.
About ultimate gain…. I thought about that this morning…..
I propose adding the maximum gain achievable with these three transistors, such that the power knob can be set to approximate Part 15 output at around midway on its rotation.
This anticipates the wide range of antenna efficiencies that the transmitter may encounter, and leaves the joy of “calibration” to the operator.
Hi Guys!
Carl – when all is said and done – you’ll
probably want to hang a voltage regulator
circuit on this thing. It would be between
the power source and the transmitter
circuit. That way, if the supply voltage
was raised a few volts above 9 – every thing
would stay the same. (I hope.) :^)
Bruce, MICRO1690/1700
Voltage Regulation
Interesting point, Bruce, and something we can discuss right now.
The power supply is a Velleman K1823, which essentially is another LM 317 IC with an adjustment from 1.5V to 35V. I don’t know how stable it is, but it does claim to be a “voltage regulator,” so we might be covered.
Let me know if you think we need another regulator to lock us at 9V, or maybe a more specific power supply that is only 9V without the variability.
I also love sticking permanent meters on equipment, but that can come later after we solve the design.
Volts and Other Volts
Since it’s come up twice in the past few messages, let’s talk some more about the 9-Volts being assigned for the oscillator circuit.
I said that I stuck with 9V because that’s what the original Pixie2 specified. But why is it 9-Volts? Maybe it should be some other number.
Neil said something I had not noticed before, that the crystal was subjected to whatever the voltage is, and yes…. looking at the circuit, the crystal goes from base of Q1 to ground. Sure it’s taking voltage. However, I remember that a technician at the company who sent the crystal asked to see the intended circuit before sending anything, so he had a chance to note the 9V spec., and that suggests that we are on safe ground. Yet, I wouldn’t mind discussing these fine points over and over until we are completely in control of what is being built.
With all this in mind, perhaps Bertrand over in Paris is taking a big risk running his transmitter at 18-Volts. True, he has a different crystal, but was it selected based on intended voltage?
Of two crystals in my possession, each from a different supplier, neither one came with a spec sheet showing any of the tolerances that have been discussed for crystals.
The fact that it “works” is only one item on the checklist. There are also questions of other characteristics……. I’ll go back through threads and document what’s been said about crystals and include it with documentation.
Reply To Myself
If I keep linking notes to my notes with the “reply” button this could be a very one sided conversation.
But I did another drawing showing Module 2, which is being done from scratch according to input from Radio 8Z
http://www.kdxradio.com/am_files/bigtalk_mod2_1.0.1.JPG
What we are doing is determining values for R6, R7, R8 and R9.
Ready, get set,……BANG!
Change Notice
The above linked example circuit was intended to be emitter-follower….
move C12 from the collector to the emitter of Q2.
Looking Ahead
I don’t mean to overload with too many complications all at once, but here’s a concern regarding Q3 with its variable power control. As it stands, this is to be a 2N2222A driven by a LM 317 which has a range of 1.5 to 35-Volts. But I have already noticed that Q3 becomes hot to the touch at 12-Volts, so we will need to think of a safe way of managing the situation – either with a stronger transistor or less power range.
Q1 Volt Reading
Knowing it may not be very accurate since RF is hard to read with a common meter, I just looked at the AC output of oscillator Q1 and it said just under 1-Volt.
Keep Going Carl
I’m reading it all.
I just don’t have anything
to contribute right now.
Continue on your quest!
Bruce, MICRO1690/1700
Buffer R Values
Carl,
I simulated the voltage divider biased buffer as shown in your your post: http://www.kdxradio.com/am_files/bigtalk_mod2_1.0.1.JPG
The R values I used are:
R6 = 4.7k
R7 = 1k
R8 = 220
R9 = 39
C12 = 560 nF
I tried taking the output by connecting C12 to the emitter but it did not provide enough drive for the final. I recommend keeping C12 connected to the collector as shown in your drawing.
I also simulated a final stage connected with a 22 ohm resistor in series with C12 to the base of the final transistor and with a diode connected to the base to ground (banded end to base). The output from the whole circuit into a 50 ohm load resistor was surprisingly high with the supply at 9V.
Modulation was also simulated and showed distortion when the supply to the final is 9V. Keeping the oscillator and driver supply at 9V and lowering the final stage supply to 3.5V produced good modulation with 240 mW delivered to the 50 ohm load. The simulated efficiency of the final amplifier stage was 72% with the final transistor dissipating 93 mW. This would indicate that the final supply voltage should be variable and probably lower than 3.5 volts.
It appears that keeping the oscillator and driver at 9V and adjusting the final amp supply voltage is a good approach since this will keep the base drive to the final at a high level.
The buffer stage simulated at a dissipation of 220 mW and if this is too high the buffer resistors could be increased but this may affect the drive to the final.
It remains to be seen how accurately the simulation predicts the operation of a real circuit but it can serve as a guide.
Neil
Buffer R Values
Neil,
I look at this thread occasionally and think the design is progressing very nicely. The adjustable voltage regulator is one very important improvement, as is the buffer transistor.
I have a suggestion for possible improvement to the buffer circuit that you might try in simulation.
A buffer is a terrific idea to isolate the oscillator from the RF output for oscillator frequency stability. Much has been said about the lack of a buffer in the Talking House transmitter causing unwanted FM modulation.
I think a simple emitter-follow would provide better isolation by presenting a much higher impedance to the output of the oscillator:
– Connect C12 input to Q2 emitter
– Remove R8 and connect Q2 collector to +9V
– Change R6 and R7 to equal values (say 51k) for simple Q2 static bias to mid-voltage (4.5V)
– Change R9 to a value that allows Q2 emitter voltage to swing at full oscillator output amplitude (-.6V) without drawing excessive Q2 current (say 20mA?) and without wrecking the waveform too much. Try about 500 ohms and work up or down from there.
You might also try changing R3 from 100 ohms upward a bit. 100 seems kind of low. I think you might find equivalent waveforms with R3 up to about 800 ohms. Above that, things start to fall apart.
The emitter follower config will give the best isolation between the oscillator and RF amp.
Just my little contribution. Neil, you are doing a great job. Wish I had more time to contribute.
Transitors
Carl,
You can use the modern, plastic TO-92 version of the 2N2222 transistor for all 3 of transistors. It’s designated PN2222A and is available from mouser.com for 10 cents each in quantities of 1 – 25. Mouser part # 512-PN2222ABU.
Data sheet: http://www.fairchildsemi.com/ds/PN/PN2222A.pdf
Catching Up
Before responding to your newest post, Radio8Z, I’ll mention what I was doing while you posted it. I re-read your posting of 6/5 where you said, “Use of two resistors in the emitter circuit, one bypassed, is recommended.”
I updated my drawing to show the emitter as you described (C11 should say C13):
http://www.kdxradio.com/am_files/bigtalk_mod2_1.0.2.JPG
Now to your latest post… “And Radio 8Z takes the lead….!!!”
I will put a new drawing showing the proposed changes as v 1.0.3
Just a minute.
But What About…
From your description, Radio 8Z, the power adjustment to the final (Q3) will serve a different purpose than first imagined… it has become an adjustment for optimizing the performance of Q3.
That leaves open the question of whether we can have an RF power output control.
3rd Version of Q2 Buffer
Using the values presented by Radio8Z, here is what Q2, the buffer amp, looks like
http://kdxradio.com/am_files/bigtalk_mod2_1.0.3.JPG
Moving next toward the actual test phase, but parts need to be ordered.
Take a few days off.
Put to Sleep
The correction fluid in the Gulf and radiation in the jet stream have put everyone to sleep. Low power radio is sending the click at the end of a record with the DJ snoozing at the mixer.
Even this Big Talker project is in super slow mo as I fuss with the parts list and sort though little parts in the junk box to figure out what needs to be ordered online.
Many naps later I’ll be back with more…….snore…….
Coaxial Cable
From the Big Talker transmitter to the antenna will be a coaxial cable, which up until now has not been discussed.
Is RG58/U the best for SW radio communications?
What kind of connector should be used at the transmitter?
RG-58
Hi Carl! I think RG-58 is fine at 13 MHz.
And at that frequency, I think you can
use whatever kind of connector you want.
Best Wishes,
Bruce, MICRO1690/1700
Slow Step For Mankind
Thank you Bruce about RG58.
The latest projected version of the Big Talker, based on Radio8Z’s design for Q2 buffer amp, is now officially illustrated in a drawing for Version 1.0.3…
http://www.kdxradio.com/bigtalker.html
We’ve added an on-board dummy load and switches and fuse, although the value of the fuse needs to be calculated.
Am still organizing a parts order before building and testing.
Wonderful End Today
Today is ending very well with your further contribution to this project, PhilB!
All at once we are talking about Version 1.0.4 of the Big Talker SW transmitter.
Tomorrow the new diagram will appear….
And I recognize the transistor PN2222A because I just saw some in the AMT3000 SStran AM transmitter!
Emitter Follower Buffer
To reply to Phil’s suggestion, I had also simulated an emitter follower as a buffer but the follower has a voltage gain of 1 which did not provide enough drive to the output base. The follower does not necessarily provide a high input Z because the load in the emitter appears at the base multiplied by the hfe of the transistor. The same is true for the common emitter buffer I proposed but the emitter R is higher than would be expected from an emitter follower loaded with the low Z of the final base. Both simulations loaded the oscillator and it was necessary to set the buffer voltage gain (unloaded) to about 6 to provide sufficient base drive to the final.
It may well be that an emitter follower would work but the simulation indicated otherwise.
The rather low values of resistors were chosen to provide a reasonably low buffer output impedance and also to swamp parasitic effects at the operating frequency. It becomes a compromise between buffer power consumption and drive capability.
Neil
Two Trial Versions
This project is exceedingly fortunate to be receiving suggestions from a committee of professional design engineers, and at this juncture we have Versions 1.0.3 and 1.0.4, coming with slightly different sets of design logic. Therefore both versions now stand as the targets for actual test.
This link shows Version 1.0.3
http://www.kdxradio.com/bigtalker.html
This link shows the particular changes comprising Version 1.0.4
http://www.kdxradio.com/am_files/bigtalk_mod2_1.0.4.JPG
The Radio8Z Simulation
Neil at Radio 8Z has sent us the simulation made with LTSpice, the basis for Big Talker Version 1.0.3…. there is at least one part I left out which I’ll be sure to add….
http://kdxradio.com/am_files/Carls_Tx_Simul_Clip0001.jpg
The Differences
For the past 24-hours I’ve been studying the transmitter simulation from Radio8Z and spotting the differences with the drawing I posted for Version 1.0.3 of Big Talker project, and here is what I have found…
Add resistor R11 (22-ohms) from collector of Q2 to C12 to base Q3
(the part numbers I am using are from my diagram; the different numbers generated within the Simulation were generated by LTSPICE);
Change resistor C6 to .047uF junction T1/L1 to ground;
This morning a major difference registered in my brain with a bump!
The diode (D1) and 100-ohm resistor (R3) are GONE!
I’ll get corrected drawings online later.
Radio8Z Version Corrected
A close look at the picture of Radio8Z’s LTSPICE transmitter simulation, has resulted in a few parts corrections in Version 1.0.3, the drawings are now fixed. And this question for Radio8Z…. Will it be o.k. to change the transistor’s to the newer PN2222A types?
Now to discuss Version 1.0.4, based on PhilB’s suggestions… At the time the changes were submitted it was not yet realized that both the diode and R3 (100-ohms) were entirely omitted…. does that change matter to your design ideas?
http://www.kdxradio.com/bigtalker.html
http://www.kdxradio.com/am_files/bigtalk_mod2_1.0.4.JPG
PN2222A
And this question for Radio8Z…. Will it be o.k. to change the transistor’s to the newer PN2222A types?
Should work as well as the 2N2222. One thing to file away for future reference is that with a final supply voltage of 3.5 volts the simulation shows a peak collector to emitter voltage on the final of 45 volts. The maximum rating for the 2222 series is 40 volts. Not knowing how accurate the simulation is prevents a definite conclusion based on this but if you notice failures of the final amp transistor this may be a factor. It would be a good idea to keep the final supply voltage lower than 3.5 volts since there is more than enough output power available. I would try it and see what happens.
Also, Phil suggested raising R3 from 100 ohms to perhaps 800 ohms. I removed R3 and the simulation showed improved output but removing R3 may not be a good idea in case there is collector to base leakage or the coupling capacitor leaks. A simulation with R3 = 680 to 1000 ohms showed minimal effect on the output so it would be a good idea to use a convenient value in this range for R3.
The diode was removed because it wasn’t needed since the voltage from the buffer is high enough to forward bias the final base without it.
For the record, the 22 ohm resistor in series with the base circuit softens the current “punch” to the base to avoid over driving the base and also “softens” the load seen by the buffer. The change in the cap at the transformer to inductor junction to .047uF may not be needed but it seemed that a lower impedance at radio frequencies would be helpful. This larger value may cut the treble and if so it could be made smaller.
This is a great project and it is progressing well. There may be the need for final tweaks but I have found the simulation program to be pretty reliable in predicting performance. Those following the thread may wonder why all the changes but this is typical of the electronic design process because of the interaction of the various stages in the circuit. Change one and it affects the others so they have to be changed also.
Neil
PN2222A
Neil,
I’m not convinced that the 40 V rating for the 2222 transistors is a real concern.
I am more concerned about the simulation results we see for various filters feeding a 50 ohm resistive load. If you look at the output voltage and current waveforms without any filter into a 50 ohm resistor, the waveforms look great (nice square waves) and the max voltage is around 7 volts. With the recommended filter (and various other filters I tried from standard calculations) the waveform looks really crappy with high voltage spikes. I don’t know what’s going on here. I am very familiar with the standard series-resonant antenna circuit (without any added filter) used for Part 15 AM broadcast transmitters. It produces a much cleaner output-transistor collector waveform than what we see with a filter feeding a 50 ohm resistor.
Can you shed any light on why the collector waveform is so crappy with a filter feeding a 50 ohm resistor?
R3 is Now Shown
By mistake the drawing for Version 1.0.3 of Big Talker showed the base of Q3 directly grounded. It is highly unlikely the output would have put out.
The picture has been corrected and now shows R3 as 800-ohms, a value put forth by PhilB and Radio8Z.
http://www.kdxradio.com/bigtalker.html
About C6
I’ll coin a term here by saying that I have “an educated subjective observation” regarding capacitor C6 connecting the junction of T1/L1 to ground…
Being a recordist, sensitive to the upper musical octaves and voice sibilance and presence, I detected no roll-off when I first used .1uF as C6. Actual measuring equipment would be the final judge, but I think we are in safe territory with .047uF
Of course this observation comes from the 2-transistor circuit and things could be different with the 3-transistor Big Talker.
C6 and Phil’s Comments
Carl,
I agree about C6. It may be a matter of trial and listen with different values. (Also, don’t let the discussion between me and Phil as we examine the finer points of the simulation dissuade you from proceeding. The simulation predicts your design will work well.)
Phil,
I also am not too concerned about the max VCEO rating unless failures occur where this might be a factor to consider. I have had none but my circuits are running less power and the peak voltage was much lower (7 volts).
Yes, the simulated collector waveforms look very rough, in fact for this simulation there is a double peak each cycle. The simulations for the 1680 kHz circuits with an output filter produced voltage waveforms which were textbook perfect for a class E amp and were verified by scope measurements. For these circuits the RFC supplying power was carefully selected for optimum wave switching timing. This was not done by me for Carl’s design.
My speculation about the waveshapes in this case with the filters is that there are four inductors and three capacitances seen by the collector and there exists multiple resonanaces above (and possibly below) the cutoff frequency. Another way to think about this is that the waveform is the square wave you described but the fundamental has been removed (passed to the load) and the harmonics remain, thus the distortion. However, it would seem that this would also be the case with the no filter circuit driving the tuned antenna load as you described. The other possibility is that the simulation fails for this high order filter system at this frequency. (The software documentation cautions about artifacts which can result from the selection of the simulation parameters such as the number of steps used.) There was some evidence of the double peak on the signal at the final base and this might be due to Miller capacitance effects feeding back the collector signal.
It is encouraging that despite the distortion the output waveform is a very good sine wave with simulated harmonics down more than 45 dBc.
If there is interest, I can repeat the simulation without the filter using a tuned series resonant LCR load.
Neil
Still Suaded
Radio8Z wrote: “don’t let the discussion between me and Phil as we examine the finer points of the simulation dissuade you from proceeding.”
Heck no. The discussion is vital to having the very best result and does everything to suade the continuance of the Big Talker project.
This is probably the first and only time in the world a group of interested technicians tried to design a 13.560 transmitter of this small power.
The Output Filter
For the several levels of understanding needed for everyone following this project, let me review what the filter is meant to do, also to make sure my understanding is correct…
The output filter consisting of all the parts shown connected to the collector of Q3 (particularly L2,3,4 and C8,9) is intended to remove or reduce the harmonics and other spurious emissions EXCEPT the main frequency, which should pass mostly un-blemished to the antenna load.
Our member in France has found that his output is distinctly stronger without the filter. Maybe I am naive, but I thought the filter, if well designed, should leave the main frequency unscathed.
Phil’s words, “I am very familiar with the standard series-resonant antenna circuit (without any added filter) used for Part 15 AM transmitters…”
I take it that refers to a loading coil tuned for the short 3-meter antenna…
But in the case of the 13.560mHz project, there is no limit on antenna length, therefore a 1/4-wave antenna could be employed, for which I presume no loading coil would serve a purpose.
Have I more or less correctly described the different position we are in?
Yes, Neil, seeing the simulation minus the filter using a tuned series resonant LCR load would be very interesting.
Case for the Output Filter
Carl pondered Have I more or less correctly described the different position we are in?
I believe you have correctly described the situation. Allow me to make a case for using the output filter. As I began to simulate the system without the filter using a series resonant antenna loading coil the question arose “What R, L, and C do I use to simulate the antenna?” This is not only a concern for the simulation, it is a concern for a real antenna system. I conclude the parameters are not known without knowing exactly what antenna system will be used. For example, a self resonant half wave dipole will require no loading coil and the only harmonic suppression will be due to the antenna length and will be poor for multiples of the antenna operating frequency. In other words there will be little harmonic suppression from the antenna and an output low pass filter should be used to provide this attenuation.
It is true that a coil loaded 3m antenna system intended for use in the broadcast band can effectively be used to radiate a clean signal but this is because the parameters are known for typical installations and the L can be adjusted for optimum passing (and filtering) of the carrier.
A simulation of Carl’s transmitter output filter shows a loss of 1 dB at the operating frequency when driving a 50 ohm load. This means the output power will be 89% of the filter input power and will not be a factor because the transmitter is predicted to produce more than adequate power to achieve the legal field strength. The filter suppresses the harmonics by at least 45 dBc which complies with the out of band radiation requirements.
The risk of harmonic radiation without the filter compared to the small signal loss with the filter is the reason the filter should be used.
Neil
Antennas and Filter Talk
Here are simulation results for some antennas from EZNEC at 13.56 MHz:
Half-wave dipole elevated 20 ft:
Len: 34.7 ft
Impedance: 88.9 – J 0.01388 ohms
2:1 SWR bandwidth: 1.2 MHz
Gain: 7.14 dBi at 65 degree elevation angle
(sky wave), bidirectional
Field strength at ground level at 30 meters with .0025W antenna power: 661.36 uV/m RMS
Half-wave dipole elevated 30 ft:
Len: 35.58 ft
Impedance: 89.31 + J 0.1196 ohms
2:1 SWR bandwidth: 1.25 MHz
Gain: 7.32 dBi at 37 degree elevation angle
(sky wave), bidirectional
Field strength at ground level at 30 meters with .0025W antenna power: 3,857.04 uV/m RMS
1 quarter wave vertical at ground level:
Len: 17.6 ft
Impedance: 36.07 + J 0.2016 ohms
2:1 SWR bandwidth: .95 MHz
Gain: 5.15 dBi at zero degree elevation
(ground wave), omnidirectional
Field strength at ground level at 30 meters with .0025W antenna power: 16,237 uV/m RMS
COMMENTS:
The quarter wave vertical appears to be the clear winner for maximum ground wave signal in all directions. It was modeled with a #12 wire. The wire could be attached to a 17.6 ft length of PVC pipe. It might be good to construct it with several tapering sections of pipe to reduce overall mass while maintaining strength at the bottom. A ground rod and at least 4 17 ft radials are required for lightning protection and RF ground. Use 50 ohm coax with a vertical.
Even though the dipoles perform worse for ground wave propagation, they can be brought up to the FCC max 15,848 uV/m level by increasing the transmitter power to compensate. Choose a dipole if you want bidirectionality. Use 75 ohm coax with a dipole.
The antenna power level of .0025W was used in all models. It gives close to the FCC field strength limit for the quarter wave vertical. Actual transmitter input power will probably be around .005W (5 mW) to account for an estimated 50% transistor + filter efficiency. Results may vary.
FIELD STRENGTH AT DISTANCE:
For the vertical with 16,237 uV/m at 30 meters, the range for perfect ground will be as follows:
30 m: 16,237 uV/m
.5 mi: 615.62 uV/m
1 mi: 307.82 uV/m
1.5 mi: 205.21 uV/m
2 mi: 153.91 uV/m
2.5 mi: 123.13 uV/m
3 mi: 102.61 uV/m
Beyond .5 to 1 mi, the local ground conductivity will reduce the ground wave below the values shown. The FCC has “inverse distance” graphs that can be used to show the effect of ground conductivity on the ground wave field strength. It ain’t easy. Bottom line: expect less range depending on your local ground conductivity.
FILTER REQUIREMENT:
All antennas show resonance below 3:1 SWR at ODD HARMONICS, so a low-pass filter is mandatory.
Heat Sinks
Carl,
The simulation predicts the current drawn by the oscillator and buffer stages will be 33 mA at 9 volts. The power dissipated by the LM317 will be this multiplied by the voltage drop across the regulator. Unless the input DC voltage is much higher than 9V a heat sink probably will not be required.
The final draws 83 mA simulated with the supply set to 3.5 Volts. This reguator will probably require heat sinking but from experience the small approx. 1″ square sinks generally are adequate.
In both cases, if the regulators are not adequately heat sinked they will shut down with no damage. Watch for the output voltage to fall below the setting. Bear in mind that the metal tab on the ICs is connected to VOUT so make sure nothing touches the heat sinks.
It is a good idea to supply each regulator from the main DC supply rather than cascading them since this will reduce the power dissipated by the first IC.
Neil
Still On the Power Supply
The Velleman Power Supply kit K1823 does not include a transformer and comes with a skimpy little app sheet that contains little facts strewn about like spattered paint.
In fact the older 1823 now in use with my Talking Pixie2 SW 2-transistor transmitter has no mention at all of the transformer. I happened to have a 24-Volt transformer which I assigned to it.
The newer 1823 has a little chart showing that choice of transformer has a direct bearing on the actual DC voltage range that will result. I will replicate that chart now:
MAX. DC Output Voltage / Transformer Rating
3 to 5V / 9VAC/15VA
5 to 8V / 12VAC/30VA
8 to 13V / 15VAC/30VA
13 to 15V / 18VAC/30VA
15 to 18V / 22VAC/30VA
18 to 22V / 24AC/50VA
22 to 35V / 28VAC/50VA
What I like about this is that it gives us the opportunity to place a cap or a top limit on the Maximum Voltage, mainly in relation to Q3.
We already know that Q1 and Q2 will be set to 9VDC, so the number we need now is the absolute top Voltage ever needed for Q3. That number will tell us which transformer to select.
One Choice
Before opening the above question about limiting the voltage range through transformer selection, the transformer I’d chosen from a catalog has these characteristics
115VAC / 16VCT @ 3.5Amps
This choice would allow Voltage ranges on the LM317s of 1.5 to 14VDC
Power Supply Pic
Here is the approximate power supply with our modifications.
I forgot to mention that we also need to power the Velleman K4001 Audio Modulation Amplifier, rated 8-18VDC / 0.5A
http://kdxradio.com/am_files/pow_sup_1.JPG
So we probably need heat sink on both LM317s ??
Filter Talk
Thank you Neil for verifying that my point of view about the output situation is a fair one.
With all this in mind, last night and today I scanned the bands up above 13.560mHz looking for harmonic and spurious output, using my 2-transistor Talking Pixie and most important, using the exact filter we’re talking about, and every thing was totally clean up to the 2nd harmonic at 27.120mHz, which was noticeably squashed…in that the audio was garbled and suppressed.
Clean all the way up to 54mHz, but the 3rd harmonic could not be checked at 54.240mHz because the “all-band” radio only allows exact VHF TV frequencies and skips the guard bands. Analog TV equipment was unable to find the harmonic.
4th, 5th and 6th harmonics cannot be heard at all, so it is evident, corroborating what Neil just said, that the present filter, submitted by SCWIS, is doing its job.
Progress Today
Actually for many days it’s been a careful listing of needed parts, checking inventory on hand to see what pennies can be saved.
This afternoon I spoke with a tech at the supply company to get advice on heat sinks for the LM317, not all that easy to figure out.
There are two LM317’s in the Big Talker: one in the Velleman K1823 Power Supply circuit and another as regulator for Q3. The better heat sinked they are, the more current they can endure.
Then I got to wondering if the 2nd LM317 should pull its input from the output of the K1823, which will put added current demand on the 1st LM317, or perhaps the 2nd LM317 should tap its power from just after the full-wave rectifier of the K1823, to better distribute the load.
I’ll come back later with drawings to support this idea but input from the contributors is welcome at any time.
Power Math
According to Radio8Z’s estimates, Q1 & Q2 will draw an estimated 33mA, and I add to this 500mA to power the Velleman 7W Modulation Amplifier, for an estimated total of 533mA, well within the 1A capacity of the first (of 2) LM317T regulators.
Q3, the final stage, will pull an estimated 83mA from the 2nd LM317T, making the total power being drawn from the DC+ full-wave rectifier to be 616mA, still all underneath the 1Amp rating.
If this power consumption math is within reasonable bounds, the only remaining factoid I need is a “Maximum Voltage” to which Q3 can be raised. That will determine the AC power transformer secondary supply voltage.
Math Correction
Re-reading Neil’s power dissipation estimate I discover having overlooked an important step. Here’s the crucial line:
“…the current drawn by Q1 & Q2 will be 33mA at 9-Volts. The power dissipated by the LM317 will be this multiplied by the voltage drop across the regulator.”
Let me re-do this and get back to you.
Now Worried
Looking strictly at one of the LM317 regulators, serving Q1, Q2 and the Audio Amp, using the figures from my existing power supply….
33ma (Q1,Q2) + 500mA (Audio Amp) = 533mA
DC+ Supply (with 24-VAC xfmr) +27VDC
Voltage drop across LM317 to provide +9VDC = 18V
533mA * 18V = 9594mA ………uh, 9.5 Amps !!!!!
What is this, a window air conditioner?
Not to Worry
Carll,
The power dissipated by the first regulator will be PDISS = (VIN – VOUT) x I. Using your numbers PDISS = (27V – 9V) x 0.533 A = 9.56 Watts. This will run pretty hot.
Now the good news: You can lower VDC by choosing another transformer or by wiring the rectifier using two diodes and the center tap of the transformer which will halve the VDC. The DC voltage into the LM317 should be two volts higher than the output voltage at minimum. Allowing for some ripple make this a bit higher, say 4 volts so a good target will be 13 volts DC input.
The 500 mA rating of the audio amp is probably for maximum power output. The simulated RF power output at a RF final supply voltage of 3.5 volts is around 230 mW. The audio power needed to modulate this at 100% is 1/2 the carrier power which is 115 mW (actually is is 1/2 the DC input power but this will be close enough). Assuming the audio amp is 30% efficient this is a current of I = (Paudio/VDC)/0.30) = (0.115W/9V)/(0.30) = 43 mA drawn by the audio amplifier. So the total DC current for the first regulator should be around 43 mA + 33 mA = 76 mA and the power dissipated will be (13V-9V) x 76 mA = .3 Watts. You can rework this with your own numbers using this is as a ballpark estimate.
So, lower the DC input voltage to about 13 volts and a small heat sink will suffice. The DC voltage out from a full wave rectified supply using a center tap and two diodes is VDC = Vsecondary/2 x 1.4 – 0.7V. If you use four diodes and no center tap the voltage is VDC = 1.4 x Vsecondary -1.4V. Working backwards using the four diode circuit you showed, Vsecondary = (VDC + 1.4V)/1.4 = (13V + 1.4V)/1.4 = 10.3 VAC. If you use the closest standard value of 12 VAC for the transformer and allow for the AC voltage to be a bit higher than rated at low load the regulator shouldn’t overheat.
Neil
Power Management
Reading “Not to Worry” made me aware that the formula was producing a value in “Watts,” and I mistakenly thought I was coming up with power dissipation in “Amps.”
ALSO, I thought of the idea that the Audio Amp doesn’t need to be fed by a power regulator, because it has a stated voltage input range of 8 – 18VDC. By feeding it straight from the bridge rectifier the demand on the LM317T is reduced.
The bridge rectifier consists of 4 X 1N4007 diodes rated at 1-Amp apiece. Does that mean the bridge is only good for 1-Amp, or do those values add up to either 2-Amps or 4-Amps?
Power for Audio Amp
You could run the audio amp from the DC before the regulator if it has sufficient power supply hum rejection. Since you added the final amp current and the audio amp current I thought you intended to power them both from the regulator.
In a full wave rectifier with four diodes, each pair of diodes conducts for one half cycle. Thus the bridge can deliver a current with an average current of double the diode rating, i.e. 2 Amps.
Neil
Power Work
At first I envisioned all circuits being fed from LM317 regulators, but when it appeared we might be exceeding safe current limits I realized that maybe the audio amp could tap straight from the DC bridge rectifier.
I cannot test this on my present system because I have 25-Volts DC which is too much for the audio amp.
So instead, I am testing something else. I have attached the audio amp to the one LM317 regulator running the Talking Pixie2 SW transmitter. Since this is a 2-transistor circuit, it is the equivalent of driving Q1, Q2 and audio amp in the new circuit being designed. Everything has been running just fine for 30-minutes and will be on all day.
Also, I am now aligned with your suggestion of using a 12-Volt transformer, which will cool things way down.
On another matter, simply for fun, I have been trying tiny light bulbs in series with the antenna trying to get them to light up, but so far no detectable glimmer.
Project Status
Design of the Big Talker 3-Transistor Shortwave Transmitter has gone through a vigorous design period while many details have been discussed, everything except the size of the mounting hardware.
Today, only 15-minutes ago, I dispatched the parts order which was so cautiously gathered together.
Many hours were spent sorting through parts boxes on the philosophy that “if we’ve saved these parts, we should try to use some of them.” At the same time I kept thinking, “these resistors and capacitors cost only a few pennies, why should we spend so much time trying to find each one in a parts box?” Technical people qualify as nerds.
In review, we will be testing two versions of the new transmitter, the major difference being in the circuit for Q2, the buffer amplifier. We have design suggestions from PhilB and Radio8Z, and we will document both of them.
Antenna Tuning
Not far up the scroll PhilB left a very useful set of results for various antenna possibilities for the Big Talker SW transmitter. But each one of the descriptions comes from the stand point of having plenty of open space.
There will be users in small apartments, limited spaces or only experimenting, and the question becomes: can we invent an antenna tuning unit (ATU) to match the Big Talker to a shorter antenna?
This might amount to a coil and capacitor, a lot like is done for Part 15 medium wave.
Big Talking Fourth
The 4th is just in time for the arrival yesterday of all the PARTS for building BIG TALKER, the 3-transistor 13.560mHz Part 15 shortwave transmitter.
Don’t go far away… frequent reports will be posted during the building process.
Circuit Card
The first Big Talker circuit card has been built, and can be seen now
http://www.kdxradio.com/am_files/CC_bigtalkv1.0.3.JPG
This is the Radio 8Z circuit, with guidance from PhilB.
It has not yet been “fired up,” as I next need to carefully inspect it for accuracy and build the audio amp and power supply.
After it has been turned on and evaluated, The PhilB version will be tested, mostly by making changes to parts in the buffer amplifier circuit.
FIRST RESULTS
For many continuous hours I built the Velleman audio amp and power regulator kits and wired everything to the Big Talker circuit card, including the 12.6 VAC power transformer.
Test Number 1 is in progress right now, as DC supply is wired to Q1 & Q2 so I could set the 9VDC. That done, I clipped an antenna to the output of the buffer (Q2) and checked the radio. We are putting a good signal at 13.555mHz. That’s not quite right.
So let’s talk about crystals. The first crystal (of two) was purchased from a U.S. supplier, and is installed in the Talking Pixie 2 (2-transistor) transmitter, which has been operating at 13.560mHz.
This 2nd crystal came from an overseas supplier and is marked as “13.560” but must be mis-marked.
To solve the situation we will order another crystal from the U.S. source.
The remaining test, scheduled within the day, is to attach voltage to Q3, the Class C Final, set it for 3.5VDC and test modulation.
Thumbs seem to be in a mostly upward position.
SECOND RESULTS
We are in the midst of testing the Big Talker Shortwave 3-transistor Transmitter, the first of its kind. Stage 1 of the test checked 9VDC powering Q1 & Q2, which succeeded.
In the next phase we are turning on the Class C Final RF Amp, powered by a second LM317 regulator, allowing a separate voltage setting of the output stage.
There is a problem. The variable DC voltage range is only 1.5 to 2VDC. We are unable to reach the destination of 3.5VDC.
Perhaps the Q3 circuit is overloading the available current?
Or, there is a wiring error or defective part in the regulator circuit.
At 2VDC there was no trace of RF output from the final, even though Q1 & Q2 continued to produce RF.
NEXT… we will check the regulator circuit with the LOAD disconnected, to make sure it is functioning properly.
CLASS C WHERE ARE YOU
The regulator problem was a wire unsoldered.
Now with Q3 set at 3.5VDC we find there is RF from Q1 through Q2 but nothing from Q3.
The audio circuit checks out and the amplitude modulation is reaching the collector of Q3 but not modulating in the RF way.
Awaiting advice from the team.
Inching Forward
The circuit we have built is shown here as Ver. 1.0.3
http://www.kdxradio.com/bigtalker.html
Removed correct crystal from other transmitter and installed to put us on the right frequency.
Strong RF out of Q2 with antenna hung on C12 .56uF, but nothing out of Q3 at all.
Tried changing R3 from 800 to 1k but it did nothing.
In an attempt to get back on the air I installed the collector feedback bias resistor (22k) from collector to base of Q3, removed R3 (1k), but the only trace of audio modulation heard was low level and mushy. The RF out of Q3 in this state is poor, regardless of voltage setting.
Now we will modify Q2 to some of the values proposed by PhilB.
Inching Forward
Carl,
I recommend slowing down a bit.
I didn’t quite follow the progression from what you had working before to what you are reporting now, and what changes you have made in your new incarnation.
Since you don’t seem to be getting anything out of Q3, but do have a signal out of Q2, you should focus on the Q3 area before changing anything prior to Q3.
What transistor are you using for Q3? No output equates to either a cooked Q3 or a miswired connection to it. What DC voltages do you see on the emitter, base and collector of Q3.
Cold Solder Junction
Slowing down did help and by reviewing every square millimeter of the Q3 area I have found a “non-solder point,” classically known as a “cold solder.”
It was the route into the base of Q3, so it was the main problem all day.
Now I am getting modulation and RF output from Q3 and the circuit in use is that of you, sir, Mr. PhilB, and all further work for the day is wrapped up.
Tomorrow I will retrace and repeat so that Radio8Z’s Q2 circuit can be evaluated in this better arena.
JULY 4th REPORT
Hello Again:
Yesterday we found that with PhilB’s recommended circuit for Q2 (buffer amp) we achieved a working transmitter with true Class C Final and good amplitude modulation.
Today we changed the Q2 circuit as previously proposed by Radio8Z and found there was no amplitude modulation. There was a good, strong, RF carrier passing Class C through Q3, and we verified the presence of audio all the way out to the RF output. Turning the voltage level for Q3 all the way down to 1VDC, the lowest setting, a hint of scratchy audio became audible on the signal.
To compare the two ways of wiring Q2, find pictures at these links
Version 1.0.3 is the design of Radio8Z
http://kdxradio.com/am_files/bigtalker_1.0.3.JPG
Version 1.0.4 is the design of PhilB
http://kdxradio.com/am_files/bigtalk_mod2_1.0.4.JPG
By the way, in both cases R3 was 1k
Also, all transistors are PN2222A
If there is anything you want me to try next, I’ll watch for your note.
Modulation Situation
Carl,
The simulation of the circuit showed good modulation up to 100% with VCC = 3.5 volts but it did not account for the use of a transformer.
My suspicion is that there is an audio impedance match problem with the use of the step up transformer. The equivalent collector resistance can be calculated as VCC^2/(2XPout). Using the numbers from the simulation yields about 25 ohms. This is the collector resistance for RF but it indicates that the impedance to audio is low.
The “8Z” driver drives the final amp pretty hard and probably results in higher collector current and lower impedance than the emitter follower driver produces in the final. You could try a 1:1 transformer or you could just go with the circuit that works.
It would be interesting if you can measure the DC current in the final collector circuit (the current through the audio transformer) as a way to compare the operation of the two driver circuits.
Neil
Acknowledged
Neil, I will do those things you have detailed, as follows.
If I do not have a 1:1 transformer there is a correctly sized one available which is 600:600. Will that work?
Now I have a good reason to replace my unresponsive milliamp meter, which will also be part of the plan.
My intention is to end up with two transmitters. The Talking Pixie2 boards can be upgraded to one of these new versions, and I already have double of almost every part.
Here is a new question…..
You may have noticed earlier that one of my crystals comes in at 13.555mHz. Is there a way of bumping it up to the correct frequency so I don’t need to order a new crystal?
New Progress
This related thread talks about modulation transformers and provides useful information for this transmitter project
http://www.part15.us/node/2900
The transformer now in house and ready for testing is a Triad TY-146P, rated for 1-Watt as 150:150-ohms; 600:600-ohms; and several mixtures of the two.
Because the Big Talker is not restricted to 100mW or less, the extra power handling capacity may be beneficial.
The Velleman audio amp is nominally designed for speaker impedances, but it seems comfortable driving just about anything, and has a generous overhead of available wattage.
ALSO, there is now a BK 2712 Multi-meter on hand so we can grab the collector current from Q3.
A Kind of Update
Following this thread is complicated by the long intervals of time between entries and the disparate nature of matters being discussed.
Most recently the question of amplitude modulation transformers in the Part 15 domain has been the subject.
Upon sensing that the Radio Shack 8-to-1k ohm transformer might suffer from both improper impedance and insufficient power handling capacity, I swapped that transformer for a junk-box 70-Volt to 8-ohm matching transformer with unknown impedance values (outside of the 8-ohm side), and already hear a deeper modulation on the working version of Big Talker, which is the Version 1.0.4.
We are very close to the point where subjective evaluations give way to objective, as scientific measuring equipment is being moved into place.
Meter Readings and Other Things
Before I turned off Version 1.0.4, based on the Q2 circuit submitted by PhilB, I grabbed collector current readings from Q3, the C-Class RF Amplifier. Nine voltage settings were used:
3.5V = 4.656mA
4V = 5.202mA
4.5V = 6.002mA
5V = 6.996mA
5.5V = 7.855mA
6V = 8.442mA
6.5V = 9.405mA
7V = 9.860mA
7.5V = 10.664mA
A change has been made to all versions. The Radio Shack mini-transformer is removed, and in this Version 1.0.4, I am using a 70-Volt type audio transformer with the Velleman amplifier feeding the 8-ohm winding and the secondary is the red and blue wires, but I do not have a data sheet showing impedance of the secondary and its several color coded taps. The one criterion is that it “sounds good.”
The Radio8Z Version
After unwiring the PhilB version of Q2, I carefully installed parts for the Radio8Z version of Q2. For this description I will use the part numbers displayed on Radio8Z’s simulation schematic printout.
http://kdxradio.com/am_files/Carls_Tx_Simul_Clip0001.jpg
R3, R4, R5 and R6 are exact resistor values.
R8 calls for 22-ohms but I used 27-ohms.
My drawing on my website shows a part that was not shown in the final Radio8Z printout, 800-ohms from Q3 base to ground, which I tested both with and without. The values I tried for R3 include 1K, 910 and 780.
The new Triad TY-146P 1W transformer is installed 1:1 , 150/150-ohms.
Result: No RF carrier from Q3, no trace of audio modulation at any voltage level.
I will keep things as they are until I get further instructions about what to try.
EDITING TO ADD: I just found this statement by Radio8Z from June 16 in the note re: PN2222A Transistor
“Also, Phil suggested raising R3 from 100 ohms to perhaps 800 ohms. I removed R3 and the simulation showed improved output but removing R3 may not be a good idea in case there is collector to base leakage or the coupling capacitor leaks. A simulation with R3 = 680 to 1000 ohms showed minimal effect on the output so it would be a good idea to use a convenient value in this range for R3.”
Based on this I am restoring R3.
Interesting modulation transformer article
Four page reading on finding power transformers that will also serve as good modulation transformers, basic modulation transformer info
http://www.part15.us/files.p15/modxfmr.pdf
Odd
It is odd that you get no RF output from Q3. If you get a chance measure the E, B, and C DC voltages to ground on Q2 and Q3. This might give a clue as to what is happening.
I won’t be able to post again until Sunday with another response.
Neil
Productivity
Thanks Neil for your review of the situation and give me a little time to get the Volt readings you want. In this interim I’ve returned the one working transmitter to the PhilB version (1.0.4), but in the hours ahead I’m getting transmitter number 2 ready for dedication as Version 1.0.3, the Radio 8Z.
I tried the new Triad 1W transformer with the PhilB version, and note that it sounded alright but had no real “force.” First it was set to 150:150-ohms, then changed to 150:600, which was a little better, but returning to the 70-Volt(junk box)speaker system transformer the loud, punchy audio is back and so I reviewed the technology behind these type transformers.
There are some good papers at the Sound & Video Contractor Magazine website, explaining the (U.S.Standard) 70.7-Volt system for distributing sound to numerous loudspeakers in giant public address installations.
Looking closer at the transformer on hand, I guess it is about a 5-Watt transformer with the secondary side being 8-Ohms, and recall that it was originally attached to a 6″ or 8″ damaged speaker.
If my understanding is correct, these are essentially 1-to-1 transformers, with several windings at the primary for choosing the desired wattage for the particular speaker.
Faded print on the transformer shows labeling for the several primary tap-wires: 5, 4, 2, 1 and 1/2 plus “C” for “Common.” In the Big Talker the Velleman Audio Amp is driving the 8-ohm tap, and the 4-Watt side is modulating the final RF stage, although now that we have an idea what is happening it could probably be set on 1-Watt.
SCWIS, I right away printed out the excellent paper you linked, showing the way for measuring unknown transformers for their characteristics. A very good paper for everyone’s reference file.
Welcome to Part15.US Summer School!
Back in Action
We’ve been waiting almost a week for two different parts deliveries and another 13.560 crystal so I could have a separate build of the Radio8Z version of the buffer amp, and today the crystal arrived.
Neil, you wanted DC voltage readings from Q2 and Q3, so what I have is a set of readings from your version, and one from PhilB’s version.
First, DC Readings Version 1.0.3 – Radio8Z
Q2 C = 8.084
B = .3940
E = .1135
Q3 C = 2.8451
B = 0.000
E = 0.000
Second, Readings Version 1.0.4 – PhilB
Q2 C = 8.645
B = 3.607
E = 3.3680
Q3 C = 1.490
B = 0.0290
E = 0.000
Right now I have the 8Z version on the air.
The oscillator and buffer are definitely producing an RF carrier, because when I become an antenna by gripping C10 (using the 8Z part numbers from the simulation) the radio, 15-feet away, goes to full silence.
Q3 is sputtering like a badly tuned radio on sibilant audio sounds, and there is the slightest hint of a carrier.
Awaiting next ideas.
P.S. The modulation transformer is a 70.7V P.A. type, estimated to be 8-ohms to 8-0hms, 1:1….. It works very well on the PhilB version.
Q2 Voltages
Carl, there appears to be something wrong with the biasing on Q2. The base voltage is too low which would explain the lack of RF from Q3. The voltages should be close to:
C 4.4 V
B 1.5 V
E .76 V
Are you sure the resistor values are correct and are placed properly? Just a reminder the 4.7 k colors are Yellow, Violet, Red; and the 1 k colors are Brown, Black, Red. The 220 ohm is Red, Red, Brown and the 39 ohm is orange, white, black. If in doubt, measure them with the ohmmeter. ETA I just looked at the pic of the circuit you posted and the colors on the 1 k resistor from base to ground look like it is a 300 ohms resistor (orange, black, brown) and the 4.7 k colors look like a 47 k (yellow, violet, orange) resistor. It could be the photo colors are not true but it is worth a mention. A 300 ohm resistor in place of a 1 k would give about the voltages you reported. End Edit
Is the supply voltage 9 volts?
Using 27 ohms instead of 22 ohms should not make a difference in the DC values.
When my students analyzed and built this type of circuit the bias voltages were always within 10% of the calculated values.
It could also be that the transistor is bad but the voltages don’t indicate this as a problem.
Neil
TRIUMPH
Neil, you are on the air!
Your suggestion of checking resistor values revealed a major mistake at R3 (using part numbers from the simulation), where I’ve been using a resistor that looks deceptively like the familiar “1k,” because it has the same colors but in a different order.
red black brown = 200 ohms ……….Whooops!
This must have been the problem this whole time, because your circuit works exactly as designed.
Voltage readings B+ for Q1 & Q2 = 9V; B+ for Q3 = 3V
Q2 C = 3.4
B = 1.74
E = 1.06
Q3 C = 1.13
B = .0985
E = 0.000
Everyone wins!
This weekend the Big Talker web page will be updated to show the two great short wave 13.560 circuits.
Small matter —- R8 is now 22 ohms….. to get one resistor I was forced to buy 100, but now there will be 22 ohm resistors for the next eight life times.
Q3 MilliAmp Readings
Radio8Z I have taken a set of milliAmp readings for Q3 in the Radio8Z Version 1.0.3 13.56mHz SW transmitter.
1.5V = 3.600 mA
2.0V = 5.300
2.5V = 6.840
3.0V = 8.08
3.5V = 9.45
4.0V = 11.9
4.5V = 12.96
5.0V = 13.94
5.5V = 16.53
6.0V = 17.98
Shall I go higher?
SCHEMATIC
Both working versions of Big Talker have been squeezed into a single drawing, appearing like an ancient scroll on an 8.5″ X 14″ page, but I hope it’s easy to follow. You can always post any questions here in this thread.
http://www.kdxradio.com/bigtalker.html
Please make note that the Parts List has not yet been updated to match the schematic, something that will be done by sometime Sunday (7-24).
Also, there must soon be a huge CREDITS posting, like they have at the ending of major motion pictures, showing the many contributions to this project from members of part15.us
PARTS LIST IS UPDATED
The Parts List for both Versions of The Big Talker Shortwave Part 15 Transmitter is now updated
http://www.kdxradio.com/bigtalker.html
If you spot a mistake please report.
Output Current
Neil,
I see back in June you wrote, “The final draws 83mA simulated with the supply set to 3.5V.”
Yet I am seeing only 9.45mA with setting of 3.5V.
Right now I am running the final at 7.5V with current 21.5mA
What do you think of these readings?
The meter is inserted right at the B+ feed to the audio transformer.
Q1 & Q2 are set at 9V
CREDITS
This attribution to the many inspirations and contributors to the design of two two versions the Big Talker 13.560mHz Part 15 shortwave transmitter is a work in progress. Please add any names and achievements.
CREDITS
Anonymous – the original AM Talking Pixie2 2-transistor diagram and its predecessor, the CW key-code version;
Wilcom Labs – first known Part 15 shortwave broadcaster;
Bruce MICRO 1690/1700 – who told us we needed a buffer amplifier;
Neil Radio8Z for continuous design suggestions and a project simulation;
PhilB for invaluable input and an alternative buffer suggestion;
Ermi Roos for guidance on wiring the power transformer;
SCWIS for a tremendous guide on evaluating potential audio modulation transformers.
Current Readings
I have been out of town for a few days so I am just catching up here. Carl, congrats on getting two good designs up and running. It was my pleasure to provide some input to the project.
You wrote “I see back in June you wrote, “The final draws 83mA simulated with the supply set to 3.5V.”
Yet I am seeing only 9.45mA with setting of 3.5V. ”
I wonder about that too. The simulation was done with a 50 ohm resistive load. Were your readings taken with a load attached?
I still think the much discussed diode from base to ground (banded end on base) on Q3 would help boost the output signal but it is your call as to whether you want to bother with this.
Keep us posted on your future escapades with this transmitter.
Neil
Aha!
Glad you’re back Neil and ….
Oh, gosh, I checked the final current with a dipole antenna attached. Of course I should have used the 50-ohm load. I will check it that way tonight.
However, the oversight tells us something interesting… that the dipole (depending on how the current looks with the 50-ohm load) may not be well matched and is telling us by the weak current reading. Is that a reasonable interpretation?
As to the diode…. I will try it. I actually like the thought of having the diode. Makes it more techy.
50 OHM LOAD
The mA readings are about the same with a 50-ohm load at the output.
3.5V = 9.957mA
7V = 21.4mA
12V = 40.5mA
12-Volts is the maximum power setting.
DIODE BOOST
The diode with its band toward base Q3 and other side to ground
boosts the mA readings somewhat
3.5V = 13.9mA
7V = 27.4mA
12V = 52.8mA
Another Factor
During this trouble shooting stage I should also put forward the fact that the Radio8Z version of the transmitter still has older transistor types because I re-used the original board from the Pixie2 circuit….
Q1 = 2N2222
Q2 = 2N2222A
Should I install the newer PN2222 types?
Power Input
Carl,
You probably realize that the DC input power is VCC X IC and your data reveal a range from 49 to 634 milliwatts. This is a very nice range with which you should be able to get a suitable field strength with almost any antenna.
Regarding switching transistors, Q1 and Q2 operate with bias and gain set by the passive components so changing these is probably not worthwhile. It may change things if you replace Q3 but I don’t expect any dramatic difference.
Neil
Coupling Cap
Carl,
The coupling capacitor to the base of Q3 may be too big. It simulated OK and I didn’t try different values and it may be worth a shot to experiment with this in your circuit.
When I designed my BCB transmitter I tried different values for the coupling cap, both with simulation and prototype, and found an optimum value in terms of efficiency. More or less capacitance than optimum lowered the efficiency but this was at 1680 kHz and things are very much different at 13.56 MHz so this is offered only as food for thought.
There are design equations for coupling if impedance matching on the input is used but I haven’t seen any guides for selecting this cap for the circuit here other than to provide a low impedance at the operating frequency. It is also possibly a factor that larger capacitors have more stray inductance than small ones. It is counter-intuiitive but this stray inductance makes the capacitance appear larger at high frequencies. Trying different values might improve things.
Neil
Fine Tuning
These transmitter designs are a milestone in Part 15 development, and there will be no end of fun tinkering with them in many ways.
One thought is to build in permanent meters with external controls for setting all the key voltages and watching the current.
Because transmission lines can be run from indoors to out there will be many seasons of antenna science in action.
I’m also looking into circuits for SWR meter and field strength.
But getting back to the immediate, I will try several capacitor values to couple RF at the base of Q3.
By the way, all Q3s are pn2222, the newer type transistors.
Later tonight or tomorrow the schematic and parts list is being updated to show corrections and refinements.
No Sooner Done Then Said
As promised the Big Talker page is updated with corrections to the drawing and parts list.
http://www.kdxradio.com/bigtalker.html
Fine Points
While working on the Radio8Z version of Big Talker the diode D1 got put back in the circuit and increased the general efficiency of the class C Q3 final stage.
The PhilB version had been developed without mention of D1 so I returned to that version and tried D1, and it had an interesting limiting effect which caused the transmitter to modulate properly only within a range of 3V to 4V to Q3. Removing the diode allows modulation throughout the voltage range of 1.5V to 12V. So, for the time being, the PhilB version does not use D1.
Another matter that needs work in my system is the excessive audio gain of IC3, the audio amp. It will help to design a T-pad attenuator to bring the line level down.
This Mornings Big Talker Tests
With the Radio8Z version of Big Talker running well, we have returned to the PhilB version to try a few things.
The fully working version contained these situations:
R9 (Q2 emitter) was running at 680-ohms because the specified 500-ohm resistor was not available;
R3 (Q3 base) was running at 910-ohms because the specified 800-ohms was unavailable;
D1 was absent because at the time of the design it had been temporarily set aside from the ongoing discussions.
Today we installed a 510-ohm resistor at R9;
We installed 800-ohms of resistance at R3 (a 600 and 200 in series);
We installed D1 at base of Q3 to ground.
Result: Q3 current readings and modulation description
4.5V = 15.0mA audio slightly garbled
4V = 12.6mA audio louder and clear.
NEXT we will test the Big Talker with D1 removed.
Oscillator-to-RF out stage buffer
Hi Carl,
I’m not sure it is worth your time to do all the work involved with trying both the Radio8z and PhilB versions of the Q2 buffer circuit. I originally only suggested the emitter follower version as a possibly better solution for isolating the load of Q3 from the oscillator transistor, Q1.
The bottom line for performance of the buffer stage is whether it performs well enough to keep the oscillator stage stable through the power excursions of the Q3 output stage when modulated to the 100%extreme.
Theoretically, an emitter-follower buffer stage should isolate better because the base input impedance is increased to Beta * Rb. Beta is the current gain of the transistor and Rb is the impedance from base to ground. With a Beta of 100 and an Rb of 200 ohms (likely in this circuit), the input impedance of the Q2 buffer transistor will be 20,000 ohms which will be a very light load on the Q1 oscillator circuit.
Here is something you can try for verification of the buffer performance. Use the Radio8z circuit, which seems to be working for you. Use a receiver with a BFO. With no modulation on the TX, tune the receiver and BFO to produce a tone of about 1 kHz. Then increase the TX modulation. You will hear modulation artifacts, but the 1 kHz tone should remain fairly stable. If it warbles wildly, you will know you have an oscillator instability problem caused by the Q3 modulated load on the Q1 oscillator.
Some warbling is probably inevitable, but if it’s not too bad, you can stick with the radio8z circuit.
PhilB
Question About Beta*Rb
Maybe I am confused about which Rb Phil is referencing. The base input resistance for an emitter follower is the transistor beta times the emitter resistance, not the “base resistance”. The emitter resistance includes anything connected to the emitter including any resistances in parallel with the emitter resistor. Thus the base resistance of Q3 will be reflected back to the base of Q2 and perhaps this is the Rb Phil was referencing.
It is true that there is isolation between the oscillator and Q3 and the test Phil suggested may indicate if it is sufficient.
In the “Radio8Z” buffer the base of Q3 could affect the load on the oscillator but not by much since the feedback from Q2 collector to Q2 base is via the hre of Q2 and the Miller capacitance. Since the gain is small (8) the Miller capacitance effects should be small and would affect primarily the high frequency breakpoint of the Q2 stage. Unlike the emitter follower, this circuit does not reflect the Q3 base resistance back to the Q2 base so the isolation should be good. Phil’s test could confirm this.
If I misunderstood Phil’s post I invite him to clarify for me .
Neil
The Case For Two
I would suggest that having two variants of the Big Talker makes the entire venture more interesting for future builders of this project.
Both designs now extant work fully and my tinkerings at this stage have more to do with small fine tunings and back tracking to try parts values previously mentioned and re-visiting design statements previously made.
One rule I often break is “don’t make too many changes all at once.” For example, I’ve just removed the Audio IC Amplifier from the 9V regulated line and connected it straight to the D.C. 12V line, but that greatly increased the amplifier gain and make it so strong that I can’t go low enough on the volume control. Therefore a T-pad attenuator is today’s next step. Probably -15 or -20dB.
Further, I hope that time and interest will permit, speaking to Neil and PhilB, your continuing to follow this thread as there will be a few questions about the engineering details.
It would also be the right time for interested parties to jump aboard and build what’s been created, and send fresh commentaries about this project.
The antenna simulations previously posted by PhilB are very important to this project, and I will make them part of the online documentation.
Now it’s time to look up the T-pad formulas.
Oscillator Loading and Buffering
As I review the posts by Phil and myself I am not sure that we are talking about the same “loading” so here’s my input on the subject.
Both buffer circuits present a load to the oscillator and for the frequency to vary with the audio due to loading a change in the buffer Rin with audio would need to occur.. Using Rb to be the base input resistance of Q3, Beta the beta for Q2, and and || meaning in parallel with, and neglecting the dynamic emitter resistance (Q2), then the input resistance of the buffers (Rin) can be calculated.
For the emitter follower Rin = R6||R7||(Beta*(R9||R3||Rb))
For the common emitter buffer Rin = R6||R7||(Beta*R9)
These Rin loads can pull the frequency of the oscillator but in order to produce frequency variations with modulation the audio signal would have to affect the oscillator loading in some fashion since a constant load would result in a constant frequency pull. If we assume that the audio changes Rb then we can use the above equations to assess how this would change the buffer Rin and therefore the oscillator loading.
For the emitter follower, Rb appears in the equation for Rin and for the common emitter buffer (the “radio8z” circuit) Rb does not appear in the equation for Rin. The conclusion is that changes in Rb can affect oscillator loading for the emitter follower (probably negligibly small) but not for the common emitter buffer unless some other coupling is present.
Another mechanism for the audio changing the oscillator frequency is a change in the supply voltage with the audio but Carl’s circuit uses separate regulators for the oscillator/buffer and final which should isolate the audio from the oscillator.
If there is no observed frequency variation with the audio then this discussion becomes moot but if there is variation then some coupling mechanism is in play. Without the proper test equipment oscillator pulling with audio is difficult to assess. The listening test suggested by Phil could yield “warbling” which is caused by effects other than FM. If the audio from the receiver sounds good then there is no practical need to worry about this and both buffers are suitable to the task.
Neil
Test Station
With so much radio transmission activity it is time to establish a more serious array of testing capability. As this unfolds I’ll come to this website for bit by bit advice.
For example, I have an oscilloscope which needs a physical spot to exist, so I will be moving something to make way. So far as I know that will allow seeing the waveform envelope for a look at the modulation. What else will it provide?
For absolute observation of the frequency accuracy and deviation I guess a frequency meter is the key. There is a small one on the new DMM but its highest range is 500kHz. For now I will conduct the listening test described by PhilB and report the result.
It is likely we are in very good shape based on how excellent the audio sounds with both designs. Musical depth and harmonics are razor keen and voice is well focused and quite clear. I sent a high frequency tone, up maybe about 5kHz or more, and it was distinct.
A lingering problem that I consider to be a “back burner” issue is a polarized hum or buzz that is constant on reception, even on the portable, except that moving around finds many location where the hum cancels itself out while the audio remains strong. I do not think this hum is being generated by the transmitter, it seems plausible that perhaps AC wiring running parallel up in the attic is re-radiating the signal, or another thought is that my 1/2-wavelength dipole, with its two 1/4-wavelength sections, is receiving and re-radiating a station on a different frequency. This hum seems to be less strong on certain days, but the change in intensity seems to come and go very gradually over many hours. An outdoor antenna installation will change the scene entirely and maybe solve the hum problem. Plumbing cannot be ruled out. Reception on the portable is sensitive to my touching metal water faucets, causing the signal to raise and drop in received intensity. There is no path to earth ground at this point.
Collation
My brain tends to collate information as the day goes by. I think of something said in one posting, and an idea from another, and come out with an all new scrambled version that sometimes makes sense.
Just a short time back questions were asked about the audio possibly sneaking into the carrier at an earlier stage.
About that same time I explained how I had moved IC3, the audio amp, from the 9V regulated line shared by Q1 (oscillator) and Q2 (buffer), and connected it instead directly to the 12V B+ with no regulation. I only did that because I wanted to dedicate the 9V feed to Q1 and Q2.
But think of this… perhaps audio from the audio amp could have snuck across the 9V line into Q1 or Q2 ???
Now, as a revisionist historian, I can claim that I “knew what I was doing.”
Last Thoughts Late at Night
It’s just about midnight and the experimenting at 13.560 has got to be put away. A few notes…
Installed a 15dB T-pad at the input to IC3 audio amp, and the volume setting is better distributed on the potentiometer, but it would be even better with -20dB.
The very first working version of Big Talker was the PhilB v 1.0.4 and it was built with 680-ohm for R9 (in place of specified 500 which I did not have), 910-ohm for R3 (didn’t have 800), and no diode D1. The performance was excellent. Since then I have gone back and added those originally specified resistors and tried it with and without D1. Didn’t like it. Now it’s back to the original and it sounds wonderful. But it’s now past midnight.
Earlier I tried to learn more about that hum problem. I disconnected the dipole and replaced it with a CB whip which I’m sure has a loading coil. The range was terrible, but I was able to bring the portable radio right up to the whip and hear the cleanest most perfect carrier 100% hum-free. But as I moved away from the whip the induced hum on the receiver got louder and louder. Probably is in need of an earth ground.
See you Friday.
Hum
Keep in mind that things are different at 13.56 MHz compared to the AM BCB as I relate my hum experience.
On my monitor receiver next to the transmit antenna there is no hum and on the portable when it is near the antenna there is no hum. As I move the portable around the house and driveway there is a area where hum is present. I noticed that if I rotate the portable I could null out the hum and receive a strong signal. This clue hints that the hum is coming from a source other than the transmitter.
Using the portable as a “hum sniffer” I found the hum was coming from a wall mounted light dimmer at one end of the house and that the hum was only present when the dimmer was switched off. I suppose the rather long load side wire from the dimmer is acting as an antenna coupling the RF into the dimmer and re-radiating a signal modulated by the AC power voltage. It is strange that it only appears when the dimmer is off but even when off the dimmer circuit appears connected to the hot line on the load side.
Neil
Hum at 13.560
Your hum description is similar to the situation here, except that so far I haven’t traced an exact source, but the likely suspect is house wiring.
Walking in the house with portable in hand there are radiant hot spots where the hum is blasting loud and other null spots where it tunes out.
With the antenna being 34.7ft half-wave that’s the approximate length of everything else in the building: pipes and wiring, so there’s plenty of opportunity for signal transfers back and forth.
The distances that all these elements have between themselves and even the walls probably sets up cavities and resonances galore.
Outdoors we must go.
Something is Different
The first time I built the PhilB version of the Big Talker I seem to recall that it worked over a wide range of Q3 voltages while allowing audio modulation.
Having now returned to that original circuit, there is the sense that something is different. Well of course it is……. we changed audio transformers after that early success.
The original transformer was a carryover from the Talking Pixie2 and was a Radio Shack 8-ohm to 1k miniature, which got changed for two reasons:
Radio 8Z believed 1k was too high an impedance for his design;
I believed the power handling capacity of the transformer was below what we’d need for full modulation.
A 4-watt 70.7Volt type audio transformer made a perfect fit and provides 8-ohm to 8-ohm low impedance with plenty of power headroom.
That is what is different now, in re-testing the PhilB circuit.
I will guess that PhilB took the 1k transformer into account with his design, whereas Radio8Z was intentionally choosing to keep impedances low throughout his design to favor high frequency performance.
This is very good brain exercise and there could be another thought at any time.
Engineering Questions
As the builder of The Big Talker circuits but not the designer, I have some questions about design engineering issues.
Back on 6-15-11 “Buffer R Values” PhilB wrote, regarding R3 at the base of Q3, “…try changing R3 from 100 upward a bit… I think you might find equivalent waveforms with R3 up to about 800 ohms. Above that, things start falling apart.”
PhilB, what do you mean by “falling apart?” What would be the symptoms of that? I have been using R3 = 910-ohms which slightly raises the amount of mW dissipated by R3 compared to 800-ohms, and the audio modulation is stable.
Another comment to PhilB, not a question….RE: your 7-8-11 “OSCILLATOR-TO-RF OUT STAGE BUFFER,” you suggested the BFO method to test oscillator stability, but neither my Sangeon nor Grundig have BFO, and the Zenith Transoceanic which does have BFO has a gap at 13.560mHz and is therefore useless.
Finally a question for 8Z, which goes to the depths of my ignorance…RE: entry of 7-26-11 “POWER INPUT,” you said, You probably realize that the input power is VCC X IC…”
My response reveals that my mind is “out of context” and I wonder…
– The input to what? I’m thinking you mean Q3;
– VCC refers to the B+ DC voltage, and I suppose you mean the DC B+ for Q3. While we’re at it, what do the letters “VCC” stand for? V=Voltage, but what is “CC?”;
– Define “IC.” I am guessing it means the Q3 base “I=Input….”, “C = …? Current?” No.
Thank you.
Power Input
Carl,
In my post about the power input I used standard engineering notation for the voltages and currents and I should have realized that not all readers are familiar with this.
VCC refers to the power supply voltage which feeds the collector circuit of a transistor stage. IC is the collector current. The DC power input to Q3 is the collector to emitter voltage (VCE) times the collector current (IC). In your circuit VCE = VCC so the DC input power to Q3 is VCC X IC.
This is how we measure and calculate the final input power when we want to assure compliance with the 100 milliwatt rule for the broadcast band transmitters.
So, when you reported the supply voltage and current for Q3 you can take the voltage times the current to get the DC input power to the final. For example, from your data for the “8Z” circuit for a 3.5 volt supply, the power input is 3.5V x 9.45 mA = 33 mW. You can likewise calculate the power input for other voltages and for both transmitters.
Since you are operating under the field strength limits this power is just a point of interest.
Neil
Welcome Information
Just after posting my engineering notation questions I started rummaging in technical reference books, but your welcome explanation is more succinct than what I was (so far) able to find.
Also, your vivid explanation will help others who dig through this thread and tackle the Big Talker.
In fact the other evening I started printing much of what’s posted here to assemble into a side-manual to go with the transmitter.
Keeping an eye on the power generated by the transmitter will be of interest to many Part 15ers, and toward that end I eventually will start discussing metering that can be in-built to the Big Talker for ease of monitoring.
interesting
This is the most interesting thing I have read in a long time, I didn’t even realize i had read the whole thing till just now!
Keep the info coming!
OUTPUT RESULTS
The two versions of Big Talker have now been evaluated for OUTPUT POWER, measured with a 50-ohm dummy load.
Version 1.0.3 – The Radio8Z
Range of power input to the Q3 Final Stage using the 1.5 to 12VDC Power Control — 13.65 to 698.40mW
Version 1.0.4 – The PhilB
Range of power input to the Q3 Final Stage using the 1.5 to 12VDC Power Control — 6.9 to 242.4mW
HOLD ON! Before quickly opting for the higher power range, you should read ANTENNAS AND FILTER TALK by PhilB posted 6/21/11 here in this thread. PhilB ran simulations with three different antenna types, and determined that the Part 15 field strength for 13.560mHz of 15,848 microVolts/meter at 30 meters would, in each case, require ONLY 5mW input to the Final Q3 Stage!
The higher power levels are intended for less efficient antennas, indoor antennas, and experimentation.
Updates to the Big Talker page are being made this evening.
Hi Guys!
OK. I’ll be working on the Big Talker.
Thank you to you, Carl!
Also, thank you to PhilB and
Radio8Z for all of the great input.
This is a pretty big deal for the Part15.us
group and Part 15 broadcasting in
general.
I’m going to have to get more parts,
but I do have a lot of them now.
Looking forward to this!
Bruce, Dog Radio Studio 2
(Formerly MICRO1690/1700. If anyone
wonders why I changed the name, the
story is on the “13.560 MHz Again” thread.)
BFO For 13.560 MHz Test Receiver
Hi Carl!
Somewhere around this thread you mentioned
that you needed a BFO for your receivers, so
you could listen to the stability of the Big Talker,
against a beat note.
There are some fun ways to do this, although,
I’m not sure how stable they are – but they are
fun. You can use the local oscillator of one
receiver and beat it against the signal that another
radio is picking up. When I was a kid, I used
a Radio Shack FlavorRadio (just AM, remember
those) and I was able to listen to 160 and 80 meter
cw and ssb on a fairly crummy multiband portable.
I could use a harmonic of the local oscillator from
the Flavorradio for 80 meters. I think the fundamental
took care of 160 meters. At night, I used to fall asleep listening
to cw marine stations in the 4 MHz band. It was a
little tricky, but it worked and was fun.
Also, the Ten Tech (Sp?) ham radio company has
some good kits. They have a Universal BFO kit
which you can use for a radio that has no BFO.
I was thinking of getting one for my Grundig S-350.
If you can get the thing to work on 455 kHz, you
might just be able to get RF into the IF strip, and
do it that way.
Anyway, it might be a fun thing to try. The Universal
BFO kit is only about ten dollars, I think. And Ten Tec (Sp?)
is a great company.
Best Wishes,
Bruce, Dog Radio Studio 2
HF and Wireworund Resistors
Of course 13.560mHz is right in the heart of HF territory (High Frequency) and the RF engineers know that wirewound resistors can be a problem.
A few of the negative results I had while building the Big Talker may have been due to a wirewound resistor.
At one juncture I put a 600-ohm resistor in series with a 200-ohm for a total 800-ohms, but the transmitter gave very atypical results.
The 600-ohm resistor had been salvaged from an audio patch-panel where it served to terminate an audio line, and is now believed to be wirewound.
The Radio Handbook has this to say…
Inductive reactance becomes a problem when wirewound resistors are used in high frequency applications.
Wirewound resistors act as low-Q inductors at radio frequencies.
The Big Talker became very stable when the 600-ohm resistor was replaced.
Big Talker Transistor Question
I notice on the parts list for the Big Talker,
that the transistors are specified as PN2222As.
My project for 13 MHz started with an actual
Pixie 2 kit, that was bought from HSC Electronics
in Sacramento, CA. The transistors that came
with the kit are specified as 2N2222s or
equivalent, on the parts list. I have not been
able to magnify the actual transistors yet, to
see what they really are.
Do you guys have any comments on this? I am
figuring that the PN2222A and the 2N2222 are
pretty much the same, because the original
Pixie transmitter circuit wasn’t changed too much
for the final outcome.
However, I could be wrong.
Best Wishes,
Bruce, Dog Radio Studio 2
Big Talker Transistor Question
Bruce,
The 2N2222, 2N2222A, PN2222 and PN2222A are all equivalent in all the important parameters. The 2N versions have a metal TO-18 package which might be important for applications requiring a heat sink. The PN versions have a plastic TO-92 case. The PN versions are much cheaper. Plastic TO-92 transistors are almost universally used in modern design.
My favorite source is http://mouser.com Fairchild PN2222As (Mouser part 512-PN2222ATF) are priced at 6 cents each and there is no minimum order.
Some General Observations on the Design
Hi all,
Here are some of my observations to date.
CLASS C MYTH
Not that it really matters, but many tend to classify this type of output transistor configuration as Class C. Class C means the transistor operates with an on/off duty cycle of less than 50% and traditionally about 30 to 40%. This design operates with a duty cycle of 60-70%. Class C operation would require very careful adjustment of the Q3 base input amplitude to cause Q3 to conduct for only 30-40% of the time.
CARL’S INPUT POWER MEASUREMENTS
Carl’s measurements show a higher input power for Radio8z’s buffer vs. PhilB’s buffer. This is due to higher Q3 base drive level in Radio8z’s version of the Q2 buffer circuit.
Q3 BASE DRIVE VS. MODULATION LEVEL
When 100% modulation is applied, the output waveform is pretty crappy looking for both versions of the Q2 buffer. The crappy output waveform is attributable to the constant Q3 base drive level. At negative modulation peaks, the drive is too high. PhilB’s version of the Q2 buffer doesn’t supply enough drive to swing the Q3 output to 100% modulation. Radio8z’s buffer does supply adequate drive for the peaks, but the waveform is still crappy because of overdrive of Q3 at low modulation levels. If you want hi-fi audio, you won’t get it with this circuit because the buffer supplies a constant Q3 base drive level.
Q2 BUFFER CIRCUIT OBSERVATIONS
Using Radio8z’s crystal model In simulations, I found the oscillator frequency to vary between the 0% and 100% modulation levels. The variation was most pronounced when no Q2 buffer circuit was present. With either the Radio8z or PhilB buffer circuits are present, the variation is reduced significantly. An insignificant edge goes to the PhilB circuit, but it’s not important when compared to the advantage of the Radio8z buffer in providing a higher base drive level to Q3.
RESONANT ANTENNA IS ABSOLUTELY REQUIRED
There is a low pass filter in the output, but a resonant antenna is still required. The filter only operates properly when a resistive load of approximately 50 ohms is connected. To achieve this, a 13.56MHz resonant antenna is required in the form of a properly dimensioned dipole or a series resonant base-loaded vertical. Just connecting a random length wire to the output of the filter will give totally unpredictable results. Measured input power and antenna range will vary all over the place.
CRYSTAL
I have used Radio8z’s crystal model for simulation with a change to the C value to 57pf to more accurately produce a 13.56 MHz frequency. Actual circuit operation depends on the type of crystal being used: series vs. parallel resonant and the associated circuit parameters. From, what I see, the circuit doesn’t match the required parameters of mainstream commodity crystals available from mouser.com or digi-key.com. The actual result may be off-frequency operation or overstressing the crystal due to too much power dissipation. I don’t have a recommended solution other than a radical change to the design of the oscillator circuit. To Carl: what is the part number and manufacturer of the crystal you are using?
CRYSTAL DESCRIPTION
The crystal used in both versions of Big Talker are from ICM, International Crystal Mfg. Co., Inc., 10 N. Lee Ave., Oklahoma City, OK 73102…… Phone: 405-236-3741, 800-725-1426
In Oct. 2009 during the Talking Pixie2 project my general request for a 13.560 crystal brought a communique from Mark Hadley at ICM, who wanted more details about the intended use of the crystal…… I sent him the Pixie2 schematic, and based on that he wrote…
“Dear Carl Blare: The schematic helped a lot. I would recommend our catalog number 43468 at 13.56MH for this application.
“This is an HC49U package with the proper load of 68pF and a calibration tolerance of +/- 30ppm from -10 to +60C.”
Incidental Details
Bruce, you will be interested that I used the circuit board built for the Talking Pixie2 when I built the Radio 8Z Big Talker, therefore it has..
Q1 = 2n2222; Q2 = 2n2222A with Top Hat Heat Sink; Q3 = pn2222A…
In other words, it uses all three types.
The PhilB Big Talker contains all three pn2222As.
Now, on the subject of the higher punch given to the base of Q3 by the Radio8Z, that is done by the addition of the diode D1.
When I tried the diode with the PhilB I had poor results, but then I found I may have been using a wirewound resistor somewhere in the mix, which I have written about a few posts back.
THEREFORE, this weekend, I will try another experiment by adding the diode to the PhilB, and I will report the results.
ALSO, when I took the output power readings both transmitters were being modulated with an audio stream, and I noticed the milli-Amp readings were swimming up and down, then I read the reminder that these readings need to be made with no modulation. That is also on the weekend agenda.
One last comment…
the modulated quality of these transmitters, with these 70.7-V type transformers wired to the lowest impedance winding, sure sound like high-fidelity. But that remains to be proven.
More on the Simulation
Phil’s comments in his previous post adds good information to the technical discussion of the design. He is correct regarding “Class C” and the simulation agrees with his conclusion re this. With the driver angle greater than 180 degrees there could be a decrease in efficiency but efficiency was not a design goal.
There are usually compromises when modifying an existing design rather than starting from scratch where the stages can be integrated more carefully and more complexity would probably be needed to optimize the circuit. This design is a simple circuit which should (and apparently does) operate for the intended purpose.
Regarding the simulated crystal used in the simulation, from the oscillator schematic it appears that the crystal is operating in parallel resonant mode so a parallel resonant LC circuit was used. The operating frequency was not trimmed to exactly 13.56 MHz but it is close enough for the simulation. Also, since this is an approximation to a crystal it is not valid to assume that the simulation will predict exactly the performance of the actual crystal in the real circuit, especially in regard to frequency pulling with modulation. This will have to be assessed by measurement using the real circuit.
Phil made a very important point regarding the antenna load seen by the transmitter. The simulation showed the second and higher harmonics were down 48 dBc which is very good but this was done only with a 50 ohm resistive load. As Phil stated, results with a random length antenna are unpredictable though my experience has been that the filter used here doesn’t require an exact match to perform well. Just be aware of a potential problem with harmonics with non-resonant antennae.
If you use a random wire it would be a good idea to have a 50 ohm resistor in the circuit in parallel with the antenna. This will consume some power but it will also keep the Z seen by the transmitter nearer to 50 ohms with a random length wire.
Neil
TEST ANTENNA
The test antenna used for the building and testing is a common dipole based on descriptions from two antenna books — Radio Handbook 20th Edition, Wm. I. Orr, and 73 Dipole and Long Wire Antennas, Ed M. Noll.
Mr. Noll gives a simple method to design a dipole from common lamp-cord, with the lead-in section of the cord left intact and the antenna portion split in two opposing directions.
The length of the antenna is just slightly less than a half-wave because Mr. Noll says “The physical length of a dipole is shorter than a calculated half-wavelength, thus the physical length of a dipole must be made about 5% shorter than the calculated free-space wavelength….”
“Feed-point impedance will be about 72-ohms, varying with proximity to other conducting materials.”
One of the books says the feed-line should be an odd-fraction of the wavelength, the other book simply says the feed-line should be a multiple of the wavelength.
In any case, with the test set-up, the mA readings I get with antenna vs. 50-ohm dummy load are nearly the same, which I guess suggests the antenna is performing well.
Example:
V = 12 mA = 58.2 mW = 698.40 WITH 50-ohm dummy load
V = 12 mA = 58.7 mW = 704.4 using dipole
Keeping in mind that I was AM modulating at the time, these readings could be nearly identical.
Perhaps I should do a test using a poor random antenna……
Good idea. Will do.
PhilB Diode Test
With all resistors being of the right kind….
Diode D1 in place between Q3 base and ground, black band toward base….
Works only up to 4-Volts applied to Q3, draws 15.6mA, producing 62.4mW.
Up to that point modulation is clean and distinct.
Starting at 4.5V the transmitter starts “puttering,” a noise similar to a motorboat. Kind of like too much feedback.
This was a comparison test to see if the general range of power output levels could receive a boost.
BETTER RESULTS!
Dear Big Talker Friends:
Tried it (the diode) again only this time with 50-ohm dummy load instead of dipole.
IT’S ALL GOOD!
The PhilB powers all the way up to 12V 33.28mA 399.36mW
Crisp, clear AM modulation (with radio real close)
The “motorboating” experienced with the dipole must actually have been RF feedback.
THEREFORE THERE IS PLENTY HERE FOR THE EXPERIMENTER TO TRY.
QUESTION TO THE DESIGN TEAM:
Would it make sense to have the diode switched in or out for higher and lower power ranges?
Thank you from Bruce
Thanks you guys very much
for the transistor info.
Yesterday I ordered the
recommended audio transformer.
You guys are doing great!
I’ll be back.
Bruce, Dog Radio Studio 2
10 DB Attenuator
I’ve been reading all of
your reports.
You could set the transmitter
to run at 50 milliwatts out, and
put a 10 dB attenuator pad between
the transmitter and the antenna.
The attenuator would get 90 percent
of the power, I think. Then the
output filter and the final RF stage
would perhaps operate in a more stable
and desired fashion, even if there were
small variations in the output load.
Is this correct? I’m no RF engineer.
I suppose that there are variations on
this idea.
Best Wishes,
Bruce, Dog Radio Studio 2
I AGREE
I agree with your idea, Bruce, Dog Radio Studio 2, because it seems just about the same as something an FM engineer once told me about audio attenuators.
In the old days, you know this, professional audio lines needed to be terminated in 600-ohms, and it could get pretty complicated with patch bays in the mess. What he did was add a 1 dB loss pad to every device, so the device was always terminated with 600-ohms, and whatever was at the other end was also always terminated in 600-ohms. I think your attenuator for RF is based on the same principle.
May I request other input on the idea?
Audio Transformer For Big Talker
I just wanted you guys to know that
the audio transformer that you specified
has come in the mail.
Now that I have that, I’m going to start
figuring out where the rest of the parts
need to go on the board, what is left
to get – and –
I’ll probably have some more questions
for you guys.
Nest Wishes,
Bruce, Dog Radio Studio 2
To The Dog Net
To Bruce and the Dog Studio Network….
Somebody will be here on duty to take your call.
I’m Glad You’re There Carl Because I Do Have A Question
After a busy couple of days, I finally got
a chance to look at this audio transformer
that I received from MCM Electronics.
I haven’t had a chance to look back at your
website Carl, so I’ll just ask you this question.
The transformer is an Atlas Sound LT-72. I
thought it was supposed to be an LT-70, but
I could be wrong.
My other question is this: On the 8 ohm secondary
side, there are four insulated wires and they are all
different colors. No problem there. But in addition,
there are two thin copper wires coming out of the
secondary side. There is no plastic insulation on
these wires. They are a couple of inches long, and
I suspect that one would have to scrape them with
some sand paper to get to the actual conductors.
Does your transformer have wires like that?
And there is no wiring diagram and the markings
on the transformer are very poor, so if you know what
wires need to be used in the circuit – that would be
good, too. That’s assuming that the color arrangement
on your transformer is the same as mine.
Thank you very much!
Bruce, Dog Radio Studio 2
Transformer
Hi Bruce,
Click on the data sheet panel in this link:
http://www.atlassound.com/pn/LT72
The color coding is given at the bottom of the PDF datasheet.
Hope this helps.
Neil
About the Transformer
Hi Bruce:
Yep. I’ve been sitting right here at the switchboard for 6-days straight, but I knew you’d call.
Those transformers are interesting, and I think I’ll link to a sight that goes into the history of them. They are designed for an industrial type of loudspeaker public address system known in the U.S. as a “constant-voltage,” or maybe “constant-impedance” system.
The side normally viewed as the secondary does have two bare wires which in a P.A. system would solder to the voice coil input of an 8-ohm speaker and those specialized speakers have mounting holes so the transformer gets bolted on. But our application is the reverse, and the bare wires are the primary side for Big Talker. That’s where you drive with 8-ohms of audio input from your audio amp.
The side with all the wires has a “C” marking, meaning “Common,” and it should be red or bright pink. It will be your B+ voltage input.
Directly next to the red wire will probably be a violet or blue wire, and this will be your RF side that goes to L1.
From the “C” up the row each wire is slightly more winding on our “secondary,” and you want the shortest amount of winding, which might be marked “5” or whatever the highest number you see. In P.A. use that would represent the amount of power transferred to the speaker, i.e., 5-watts.
It’s interesting to experiment with other windings when you test the transmitter, but I found the lowest impedance winding gave the best results, which is to say, louder audio, and maybe higher power levels in the RF output.
I think almost any of these type transformer will do the job for modulation, since we are not transferring 4 or 5 watts, and they seem to be a good off-the-shelf impedance match, since Neil advised a 1:1 low-impedance relationship rather than the step-up type we had before.
Standing by for your report.
The Story of 70.7V Sound Transformers
This Wiki – Link, (NOT wiki-leaks) , tells about the 70.7V speaker systems within which these transformers play a crucial role. Just keep in mind, they are talking about loudspeakers and we are talking about RF audio modulation….. so some of what they say may not apply to our situation.
http://en.wikipedia.org/wiki/70.7V_speaker_system
Thank you Neil and Carl!
I sure appreciate the information! It’s a relief to
know that those bare wires belong there and that
it’s not some kind of mistake in manufacturing.
I’m putting the components on a large perf board.
This will help me if I make a mistake. On the first
transmitter, I built it too small and then I couldn’t
trouble shoot problems easily.
I have most of the parts. The Idea is too get everything
onto the board and then see what is left. I still don’t
have the parts for the buffer amp, but I hope to pick
them up this week. That should be most of it. The final
amp needs a few extra parts and I am going to make
some kind of output attenuator.
I guess you guys said that either buffer amp will work,
so I’ll just go with whichever one is easier to get parts
for.
I have a couple of different power supplies to run the
transmitter. I have a variable supply that will run the
final amp, or I can run the whole thing on batteries,
just to get it going. I also have a good audio amp that
will drive the transformer, so I don’t have to build one.
If I can just the pieces working, then I can worry about
making it look nice later.
I turns out that this project is the easiest to do right now.
I guess next we’ll go back to the Gates board. Besides putting
the AM back on the air, there are other projects I would
still like to try. I still have all the solar cells for the FM
BCB repeater idea we were talking about last year.
We my have to move to another location. After being in the
same house for 20 years, it’s getting hard to make ends meet.
This is largely because of my eye condition.
I don’t want to get into this too far because it’s out of the
scope of our nice Part 15 board here – but the priority is
to get my son into college and then my daughter into college
three years from now. We have been fortunate to have two
really great kids and we want to launch them into life as
best as we can. That’s why I haven’t been on the board as
much lately. We have a lot of major decisions to make.
On the other hand, being here on Part15.us has really helped
me when times have been difficult. There is a really nice bunch
of people here and I am very grateful for that. Not only has the
Part 15 aspect been really nice, but I have enjoyed the radio
stories so much and have learned more about electronics.
So we’ll take it one day at a time. It would be really nice to
have this 13 MHz transmitter going in a week or two. That’s
the goal.
Thanks so much again.
Bruce, Dog Radio Studio 2
To a Good Father
Bruce, you are a good father.
I looked up the Atlas LT-70, in as much as you expected to get a LT-72, and found that you got a better, heavier duty transformer.
To compare (and the numbers are the wattages these transformers are designed to handle:
LT-72_ .5, 1, 2, 4 Watts
LT-70_ .5, 1, 2, 5 Watts
Everything is perfect.
Carl:
Thank you for the kind words, Carl.
I’ll be back as soon as I can!
Bruce, Dog Radio Studio 2
Dog Radio 13 MHz Transmitter Report
Hi Guys:
Right know, (and it usually isn’t like
this) I lack the funds to make a
buffer stage.
So the plan is to build the 2 transistor
version (like you did, Carl) and add the
buffer when things calm down around here.
I plan to test run it at a very low voltage. I
believe I have a working schematic for this
and I am putting the parts on the board
right now. I actually can run the 2 stages
at different voltages if I want to. The
oscillator can run at 9 volts, and the
final can run at 3 or 4 volts, or whatever.
It’s funny – there is so much going on here,
that this is probably the worst time to be
working on this thing. Then, on the other
hand, it may be the best, because it’s getting
me away from the other stuff for a while
so I can think.
I will keep you guys posted.
Best Wishes,
Bruce, Dog Radio Studio 2
Crud!
I just looked at the transistors from
this other Pixie kit I am using and
they are 2N4401s! The schematic
says they are 2N2222s. Crud!
Well, I’ll try them. I know they are
close.
It’s OK to say crud here, right? Carl,
what do you think?
Bruce, Dog Radio Studio 2
Why Good Humor is Good
Bruce, you are one of the many people that spend time on this website with a keen sense of humor.
Only an opinion, but NOT a humble opinion, humor is just like part 15, it keeps us from madness and the absurdity of life.
If anyone criticizes you for your humor, just say these words:
You FOOL! You don’t get it do you. Life is funny, I’m telling you! It is a RIOT! If you don’t get that, I have no time for you. So laugh, and stop being such a crud!
Laughing Very Hard And Oscillator Circuit
You’re right Carl! I laughed really hard
when I read your message!
By the way, the new oscillator circuit is
on the board. Physically, it is much better
arranged and easier to work on than the
one from the original 13 MHz transmitter.
I am arranging things so that if anything
blows up or has to be changed, the part
will be easy to find and change out.
In the original 13 MHz rig, that wasn’t the case.
It became a rats nest very quickly.
I don’t have a voltage source, or I would power it
up now. Hopefully, I can do that tomorrow night
and then we’ll take it from there.
Best Wishes,
Bruce, Dog Radio Studio 2
A Few More Comments
I think I have the audio transformer
figured out, thanks to the info from
Neil and Carl. Thanks, you guys.
For the output filter, I’m going to have
to make the inductors myself. Considering
their low values, I don’t think that will be
too hard.
Best Wishes,
Bruce, Dog Radio Studio 2
Tidying Up
Lazy Carl Blare is still edging toward getting the Big Talker into its metal enclosure. Tomorrow for sure.
It would be nice to have meters just like big transmitters, but probably only one meter is really all that’s needed, that being a DC Volt meter on the B+ for the final stage. Once a chart is made from milliAmp readings taken with a multi-meter, it is easy to deduce the output in milliWatts using simple math.
Still, a milliAmp meter could be included, but so far as I know there is no way to directly meter the output in milliWatts. Make that a question: is there?
Bruce, if possible, when you hand-make the output inductors post your instructions. That sounds like a nice touch for a self-made transmitter.
Hi Carl!
Yup, I can provide the info on the
output coils that I make. It may
take me a while to get the parameters
to be correct, though.
I am building this new Part 15 13.560 MHz
transmitter very very carefully. (So it
probably won’t work, HA!) I am putting
the components farther apart on this
version so I can tell where everything is
in case I have to swap something out. But
I guess I said that already.
We’ll keep you posted.
Best Wishes,
Bruce, Dog Radio Studio 2
13.560 Osc. Running
We’ll, I wouldn’t have believed it,
but the new oscillator works.
I am listening to it right now on
a Kaito 1103 receiver.
I thought it wouldn’t work because
usually when you construct a
circuit carefully and slowly Murphy’s
law says it won’t operate. The thing
you slap together in five minutes
that looks like garbage will run for
100 years.
Now onto the final amp, the modulator,
and a place to put the buffer circuit later
on, when I expand the transmitter.
Bruce, Dog Radio Studio 2
One More Note From Dog Radio 2
I didn’t put the .01uf cap across
the power leads of the oscillator.
The thing still runs, though. I’ll
have to add that.
Maybe I can scrounge around my stuff
and see if I can come up with some other
parts to build one of the buffer designs.
It will take me a few days to figure this
out.
Rebuilding this thing was definitely the
way to go. The components are much farther
apart now, and I can see what’s going on.
Aside from the unknown of making the inductors
myself, I think this thing is really going to
get built now. It’s just a matter of how long
it will take.
Thanks to you guys.
Bruce, Dog Radio Studio 2
P.S. Carl, I know that you want to figure
out a way to mark your power control so
it will be marked as power out, instead of
final supply voltage. The only thing that
I can think of is to run the transmitter into
a 50 ohm dummy load and measure the RF voltage out.
Then you can calculate power out. But i don’t
know how easy it will be to do this. You
will have to make an RF probe, I think, and
then measure the voltage and do the math. Because
this is 13 MHz, and not the AM BCB, I think your
measuring set-up (the RF probe) will have to be
made a certain way, but I don’t know what that
is.
I have a QRP watt meter. If I didn’t mention it
already – QRP is the ham radio Q signal for “low
power.” My QRP meter goes down to about 5 mW.
So when I get the transmitter running, I will
be able to measure the power out.
This power meter was quite useful when I built
a very famous ham radio transmitter design,
known as, the Tuna Tin 2. Designed in 1976,
by a guy named Doug Demaw, W1FB, it ran on
the ham radio 40 meter band, on 7.040 kHz.
The idea was that you built it in a tuns fish
can, although my version is just inside a box.
This transmitter uses 2 2N2222s, and works really
well. It is a QRP transmitter, and most people
who built it got 300 – 400 mW out with a 12 volt
supply. This is morse code of course – on/off carrier
keying. My version of the rig runs about 390 mW out.
I found that during the day, into a low 40 meter dipole,
the TT-2 was heard very very well in Rhode Island,
Mass., and New York. I had absolutely no problem
having QSOs with other hams in nearby states with this
little transmitter. But, of course, conditions had to
be good.
Right now, the little TT-2 doesn’t work. I think there
is a broken lead somewhere where power goes into the
transmitter. I have to fix that. Also, most QRP dudes
on 40 meters have moved down to 7030 kHz now, because
7040 is being used more by hams that are running digital modes.
This little gem is sitting in my ham shack just waiting to
get fixed. Many hams built this thing, in fact, hundreds
were made.
Well, I guess that’s enough for mow. Best wishes again.
Bruce
Science At Its Best
Making the board with ample spacing between components is the best science for an experimental circuit, where you can see everything that’s going on.
The .01uF from DC power buss to ground is based on the Talking Pixie2 diagram, but later on, the Radio8Z circuit simulation shows a 1uF from DC power buss to ground. I have done it both ways, and both seemed to work the same… I think this part is meant to keep RF from riding the DC power.
The output power controls for my present Big Talkers are basic card mounted screw-driver adjustable variable resistors, but ultimately this should be a large chassis mount control with a knob. Those little circuit-mount types are not meant for constant adjustment, which no doubt the part 15 operator will always twiddle with.
The 13mHz broadcast band, 21-meter band, here in the U.S. is like an abandoned shopping mall.. nothing going on. But part 15ers are moving in and bringing the old neighborhood back to life.
Some Construction Hints
Bruce,
It is great that you are keeping us up to date with the progress of your transmitter build since sharing information such as this allows (and I hope encourages) others. I have a few tips to offer for your project.
Though I understand your reason for spacing the transmitter circuit as you describe, be mindful that high frequency RF circuits may require close parts spacing with short leads. If you experience spurious signals or self oscillation of the circuit it may be due to lead dress. No harm in trying but keep this in mind.
The capacitor from power to ground is to prevent signals from one stage from getting into another stage via the power connections, especially from the final to the previous stages. If not bypassed the circuit may oscillate. It is recommended to use at least two capacitors for this, one in the 1 to 10 uF range and another in the .01 to .1 uF range in parallel. The reason for this is the higher value caps bypass low frequencies well but not so much high frequencies. The smaller caps bypass high frequencies but not low frequencies. By using both types in parallel the idea is to bypass a wide range of frequencies. In some VHF (146 mHz) amplifiers I built I found it necessary to use three caps in parallel (1 uF, .1 uF, and .01 uF). You don’t necessarily have to do this but if you have spurs this may be a way to fix it.
Neil
Circuit Upgrade
Thanks Neil for writing about the cap subject and the circuit will be upgraded to match your description.
Part numbers and a revised circuit drawing will be posted this week.
Thank you Carl and Neil
Thank you Carl for your ongoing encouragement
on this thing.
And Neil, I really appreciate that info on the caps.
NOBODY has ever explained it that way, and it
completely makes sense! Also, I understand the
need to not have leads too long for 13 MHz. If
the thing gets unstable, I’ll tell you guys about it,
and probably change things. I guess we’ll all
learn something. So I appreciate that, too.
So we’ll see where this thing goes.
I’ll be back soon.
Bruce, Dog Radio Studio 2
Circuit Update
The Schematic Diagram has been updated to show the addition of capacitor C14.
http://kdxradio.com/bigtalker.html
Big Talker Buffer Circuit
Hi Guys:
I think I have enough junk around
to build Phil’s version of the
buffer circuit.
Not all of the parts are exact, but
I think they will be close enough.
So it will end up being the 3 transistor
transmitter instead of the smaller design.
More details on that to follow.
Thanks very much!
Bruce, Dog Radio Studio 2
Phil’s Buffer
Great Bruce DRS2!
I think there’s a link up toward the top of my circuit page that shows PhilB’s exact part suggestions…I have the paper note in my hand, and it says…
R6, R7 = 51k
R9 = 500
R3 = 800
Just to encourage you, I tried several different close values while I was searching for parts, and they all worked. You should be fine substituting values
Pin Diagram for LM317
The LM317 is only 3-pins, but it is a very tricky circuit to wire once all the parts are included, and the pin arrangement is sort of mixed up.
I added a pin-diagram link on the circuit in the parts list with IC2, but here is the same link now…
http://www.national.com/images/pf/LM317/00906366.pdf
Preparing To Make Buffer Stage
I bought a Radio Shack 1/4 watt
resistor assortment.
500 resistors.
Was this a mistake?
It will take me longer to sort them
out than to build the buffer stage.
I’m sort of joking. There are ten
of some resistors and five of some.
The ohm meter should make it go by
pretty fast (I hope.)
Actually, when I do sort parts, I put
them in business envelopes, and label
the envelopes.
Well, here we go. By the way, Carl,
I liked your last reply on the hurricane
thread. More on that later.
Bruce, DRS2
Phil’s Buffer
Carl,
In the interest of simplifying your whole project, I suggest eliminating my buffer option. Just go with Neil’s version. The performance difference is negligible and not worth the complexity of perpetuating both versions.
Are We Ready To Simplify?
Is the project finalized?
Answer To Self
The Big Talker project is not finalized, at least not for me, because of the following…
And the question of “Are We Finalized” actually comes in response to PhilB’s offer to simplify the results by dropping his design version.
I vote for keeping both versions for the time being and for keeping this project alive, because of the following….
The Following
1. The PhilB Version allows lowering the input power to Q3 closest to his recommended 5mW for all three of the antenna simulations that he published so to meet the FCC permissible radiant power;
2. Previous posts have opened a number of design concerns that remain unresolved, including questions about the crystal, oscillator isolation from the output circuit, and other matters;
3. Future students who pursue this project will likely wish to know the entire result of this early work, just as those of us who explored the Talking Pixie2 tried to unearth whatever earlier variants might exist.
Of course, PhilB reserves the right to withdraw his intellectual property, and as the keeper of the manuscript I will honor any such request, but I hope we can preserve the status quo for awhile.
Therefore the answer to my part of the question is… this is an ongoing project, not final.
Speaking Of Ongoing Projects
I’m still sorting out resistors from that
pack of 500 I got at Radio Shack.
After that is done in the year 2030,
I’ll start building the buffer.
Best Wishes,
Bruce, Dog Radio Studio 2
P.S. Then I’ll get back into my time
machine, and come back to September,
2011.
MID MONTH BIG TALKER REPORT
This morning I re-measured the indoor dipole which has been running since we built Big Talker.
Through re-calculation, I deducted some length from the transmission line and added a few feet to the indoor dipole. The first time I put it in I must have calculated slightly wrong.
The result is entirely different than before. Now the indoor strength of Big Talker is overloading the front ends of three shortwave radios, to where I get scratchy static from the station up and down a large bandwidth around 21-meters. To get clear signal I need to go outdoors away from the building. This brings to mind what PhilB wrote back on August 5.
SOME GENERAL OBSERVATIONS ON THE DESIGN – PhilB – 8-5-11
“Connecting a random length of wire to the output of the filter will give totally unpredictable results. Measured input power and antenna range will vary all over the place.”
Antenna design is very critical for Big Talker.
Whole House
In the previous blog I described the re-design of the indoor dipole, and now it is later in the day.
Walking the floor from front to back examining the overhead dipole I notice that now it goes tip-to-tip, from the front wall of the house to the back wall. That’s interesting. Doesn’t that mean the house is a 13.560mHz resonant cavity?
Maybe we should call a dentist.
Don’t You Hate It When
Well I do. I mean, I hate it when I think I’m going to do something interesting, but unexpected annoyances get in the way. You’ll see….
It all started when I was testing Big Talker, the shortwave transmitter, now that it has a newly cut dipole. All was going well, but the date was repeating in my mind: September 15. September 15. It seemed familiar, as if it meant something. Bing! Well ya, it’s the deadline for filing corporate tax. Ah geez. So all attention got switched over to the momentary break with freedom and went through the slow process of preparing an extension to file, then I’d need to rush to the post office……but, there was someone knocking on the door without an appointment. “Go to…” I thought. “Go to…” I refused to answer, even though I could be seen inside the window at my desk. The pesk knocked a few more times, so I didn’t answer the door a few more times. He went away. I drove to the post office. When I returned the guy was back so I drove past my own home and pretended I was a stranger driving past.
At last I was back to the Big Talker and found that, much as Radio8Z first predicted, an output voltage setting of about 3-volts was optimum. At this setting the signal is solid for hundreds of feet outside and the modulation is as wholesome as FM with clear musical quality and strong voice presence.
Whereas before, with an antenna cut too short, I was able to span the whole range of power settings and get good modulation, but now I think the overall tune of the antenna is closer to being correct.
More testing scheduled tomorrow.
General Antenna
I started a separate thread titled “General Antenna” which talks about the antenna being modified here in our Big Talker test laboratory. In that blog I asked about the fact that one of my antenna books reports that dipole antennas should be cut about 5% shorter than their calculated wavelength, and the response is so interesting I will reprint it here
5% SHORTER
Submitted by radio8z on September 14, 2011 – 21:32.
The reason a wire antenna is cut 5% shorter than the calculated length is because the velocity of the signal going down the wire is 95% the speed of light. This makes the physical wavelength on the wire shorter by 5% compared to the velocity in free space. The 95% figure is known as the velocity factor.
A transmission line which is an integer multiple of a half wavelength presents the same impedance to the source as is present at the load end of the line (antenna). This is not needed if the antenna Z is matched to the transmission line Z in which case the line length doesn’t matter except for loss.
Neil
Transmitter Readings
With dipole cut very close to 1/2 wavelength at 13.560mHz the present input to the final stage is 44.7mW.
Volts to Final= 3.0
mA dissipated by final = 14.9
Scheduled tonight: Will remove .88-foot from each side of dipole to comply with the 5% short method described by Radio8Z.
Night Tests
As planned, .88′ was removed from each wing of the dipole, representing 5% of the resonant length of each side, 17.6′.
Small as it was, the transmitter stopped being slightly unstable at power settings above 3V and I tweaked the power as follows:
4.5V 20mA = 112.50mW
6V 31mA = 186mW
9V 39mA = 351mA
A mystery hum continues to occupy certain points in space indoors (as received while listening to 13.560) and around the house outdoors, but out in free field the hums disappear.
While cutting one length (which I did with the RF turned on) I noticed that touching the end of the antenna with the small cutters canceled the hum. I was physically not in contact, as the cutters have an insulated handle, but of course I was present through capacitive coupling. Also, by simply moving to a different location, if I could do that while still holding the cutters on the antenna, would the hum have been cured? Still don’t understand what is causing it.
Also tried a choke on different stages, such as AC power and audio line, but that did nothing the received hum. I already am very confident the hum is not being generated by the transmitter or its audio system. It seems to be entirely an antenna induced problem either because of proximity to indoor systems or because it is resonant to some unknown RF.
Something About Dipoles
By themselves dipoles are perfectly symmetrical and in that sense are balanced.
However, the output of our transmitter is far from “balanced,” the RF side has all the RF filter components which makes it quite different than the negative side, which is circuit ground. The RF chokes alone must radically change the wavelength of one side of the dipole.
That big imbalance doesn’t seem to get mentioned in the literature.
Balancing Act
Perhaps a BALUN at the antenna would help.
Still Working On the 13 MHz Project
Hey guys!
I had mentioned that I probably have most
of the parts. But I bought a bag of resistors
(500 resistors, in fact) which I had to sort out
with a multimeter. They are sorted out now,
it took me about 3 weeks.
Now I can pull all of the parts together and see
if I can build this whole thing.
You know, this project seems to be coming out
really well. There are though, other ham radio
low power (QRP) circuits that also can be tried for this
application. (GLAAARRRRRG)
Sorry – that’s the sound of my brain exploding.
Best Wishes,
Bruce, Dog Radio Studio 2 (Named in honor of my
wonderful dog Betty, who is in Doggy Heaven now.)
New 13.560mhz station getting close!
Hi to all those following this blog!
After 2 years I’m FINALLY about ready to roll out my Pt. 15 SW. I’m not a do-it-yourselfer, so I’m using the oscillator/buffer stages of an old ham xmtr to do this.
Still to do: borrow a good field strength meter to make sure I’m within limits, and…
Get a mod transformer and blocking cap & install them to modulate my carrier.
A complete update is in the Pt. 15 Forum under a new heading,
“New 13.560mhz station almost ready”. (or something like that….Can’t remember how I titled it!)
Anyway, check it out if interested, and let me know if any of you are getting close also, or are already on the air. Thanks!
Ron Kocher, Palm Coast, FL
Welcome Again
Good, Ron, you found our project thread for the 13.560 transmitter we built.
Of all the transformers I tried for modulation the very best is a 70.7V type audio sound system type, which is essentially a 1:1 transformer at 8-ohms. They are available from MCM electronics. Look for part # here
http://www.kdxradio.com/bigtalker.html
Announcing Big Talker 2
For those who followed this thread, Big Talker was the 3-transistor shortwave transmitter designed by a committee of members for operation on 13.560mHz, and was a good success.
Because we ended up with two designs for the driver and output stages, I am going to build a Big Talker 2 based on this part 15 rule:
15.227 Operation within the band 26.96-27.28 MHz.
(a) The field strength of any emission within this band shall not
exceed 10,000 microvolts/meter at 3 meters. The emission limit in this
paragraph is based on measurement instrumentation employing an average
detector. The provisions in § 15.35 for limiting peak emissions apply.
When things get started I will open a fresh string.
Brief Signal Collision
On Apr. 26, 2012, at about 8:38 AM CDT, while running Glenn Hauser’s World of Radio during a test period on the Big Talker shortwave transmitter at 13.560mHz, a strong carrier suddenly turned on on top of 13.560mHz, knocking Mr. Hauser’s voice into the background but I heard no voice or other audio to identify the invading carrier.
Very quickly I switched Big Talker off and scanned in the vicinity of 13.560mHz, but could find no other signal. Therefore the interruption was very brief.
Our guess is that a nearby utility vehicle used a mobile radio at very close proximity, probably on some other frequency, but the RF was so strong that it temporarily swamped our Sangeon receiver.
What guess might you have, given this limited description?
Other Users
I’ve monitored that freq occasionally and usually receive a constant, unmodulated signal. Not very strong but obvious.
I always figured maybe some type of RFID system nearby.
A Watt Meter
Over on another thread Seanwk40 linked a very interesting Watt meter that measure all the way down to 5mW, except that it was pointed out by several members that part 15 AM transmitters will not match with this meter because of special output loading characteristics.
In fact the meter requires a standard 50-ohm RF input, such as provided by the Big Talker part 15 shortwave transmitter. Therefore I think this meter and the Big Talker are a match! Tell me if you agree (or disagree).
http://www.ohr.com/wattmeter.htm
Shortwave Activity
The Big Talker Shortwave Transmitter, to remind, has two versions. The significant difference between the two is the buffer stage, one designed by PhilB and the other by Radio8Z. Up until recently I planned to build two distinct transmitters, one for each version, and of course you are free to build either version yourself. But I now plan only one version and I’ll explain why.
The PhilB version most easily matches the estimated legal operating point, calculated by simulation which is part of this thread, and amounts to an output power of 5mW to a properly matched antenna, therefore that will become the one and only Big Talker in operation here at KDX.
BUT, the Radio8Z Version has a new and exciting purpose, as part of the Deep Voice Long Wave Transmitter now under development. Mr. Radio8Z has already presented a simulated output circuit which, when combined with his buffer stage, will produce the 1Watt to the final stage permitted in the long wave experimental band.
The extra set of inductor parts have been offered to Doggradio S2 for his 13.560 build, now focusing on the final stage.
Other Big Talkers
The ears perked up this morning when Ian and Mark of Free Talk Live mentioned their new affiliate is “The Big Talker” in Wilmington.
What good are slogans, I tried to find “The Big Talker” name infringer, but there are 13 Wilmingtons in the U.S.
Luckily Ian came back to say that their Wilmington is the one in North Carolina.
O.K., so there’s me at radio-locator.com scrolling through stations with slogans like “The Dove”, “Family Radio”, “Coast 97.3 Jams”, “Sunny 104.5” and “God’s Country 1340”, but no “Big Talker.”
So it’s probably licensed to a town on the outskirts of Wilmington.
This is KDX “The Popular Station”.
Big Talker Will Talk Again
Possibly the longest thread in part15.us history, Big Talker is the project in which several members designed a legal Part 15 shortwave transmitter for 13.560mHz per 15.225. The schematics are posted at
http://kdxradio.com/bigtalker.html
Carl Blare and Dog Radio Studio 2 each built one, and Carl’s Big Talker was strung out on a work table with an indoor dipole antenna, in operation for awhile. Then the parts were stored in a box and procrastination set in.
The other day the circuit cards were organized on the workbench, awaiting mounting in a permanent metal enclosure.
Following that, an outdoor End-Fed Dipole, per MRAM 1500’s recommendation, will be built and a hole punched in a wall for the transmission line.
I will probably open a special thread on the subject “How To Put a Professional Hole In the Wall,” since that’s a good general intrerest topic.
By the way, a 2nd Big Talker project has a thread…
http://www.part15.us/forum/part15-forums/general-discussion/big-talker-2
That is a shortwave transmitter for 26.96-27.28mHz under 15.227.
Way in the future will be Big Talker3, 49.82-49.90 under 15.235, but that thread doesn’t exist yet.
There is yet another thread regarding Deep Voice, a 1-Watt long wave transmitter under 15.217, at this link
http://www.part15.us/forum/part15-forums/general-discussion/longwave-broadcast-dx-171-khz
The schematic for a proposed design has not yet been posted, but the KDX page for that is up
http://kdxradio.com/deepvoice.html
Get ready for project action.