The AM Forum

What would cause this condition?

This is the circuit:  http://smallwonderlabs.com/Retro-75_Instr.pdf    <-see page 10.

Scope image from the Ant terminal.

Screen shot of a plastic radio driving a linear? Plate modulated TX? Class Whatever FET??? Looks more like an FET TX and filtering gone wacko. Extreme overload of a receiver, 455 picked off from an R390A to the scope? R390A's can overload! BTW
Splain to us how this shot was taken my brotha
TNX
Fred

Fred,
Sorry, transmitter output.  1kHz audio to modulator.  Circuit is at the link provided in Post 1.  Scope is connected to Ant. terminal shown in schematic.  Trying to troubleshoot this thing.

By the way, I have a nice clean 1kHz sinewave at the junction of C116, L3 and the bottom of T3.
Bob

You sure the scope is OK?  What's going into Ch2 that you're triggering offa?

Scope is good. First thing I thought of.  Using channel 2 as the trigger source to the audio generator.

This shot is of the transmitter.  Ch1 is on the output of the transmitter.  And Ch2 is at the junction of C116, the top of L3 and the bottom of T3 in the schematic.  Ch2 is the trigger source rather than the signal generator.

It looks like there is no 50 Ohm load on the transmitter output besides the scope probe.

Went directly to a dummy load rather than my switching setup and it looks like I have a bad antenna switch.

Give that man a cigar.

Envelope looks good and I learned something new.  

So the question, what is actually causing this waveform to look the way it does?

and further investigation... looks like I've been tuning other rigs into nothing for some time unless I had a DL crapout recently.

Thanks Tom and everyone else.

Bob,

Posting the manual for the transmitter was a big help.  The schematic shows that the AM transmitter is a drain-modulated (plate-modulated, HI) FET PA.  The drain supply is modulated via a transformer.  These are the necessary important clues.

The scope’s envelope waveform shows a large residual amount of r.f. voltage below the negative modulation peak.  The modulation itself has good linearity.  It’s (almost) impossible to create the envelope you showed with transformer-coupled modulation, yet you have it!  I was thinking before I studied the schematic that the PA must be series-modulated because this is the only way you could produce that waveform with a normal 50 Ohm load.  However, since the PA is not series-modulated but transformer-modulated, there has to be another explanation.  

The FET has a high gate capacitance and with no load, the gate drive voltage would appear at the output with a high impedance load only.  I don’t think there a term for this with AM transmitters, but an old term of “backwave” was used with CW transmitters with this malady; significant output key-up.  Some transmitter schemes modulated the prior stage also, this would help the final drive at positive peak to ensure full PEP plus help to ensure no output at the envelope minimum.

It would be nice to see the proper modulation envelope now.

Hi Tom,
Thanks for your help on this.

Below is the waveform using a DL direct with the same scope connections as in the previous post.

It's not 100% modulation.  I can't get that, at least not yet.  I believe ALC is involved to prevent it from over modding.

OK This should be a weekly quiz to excersize the grey matter. I'm serious! It makes a thought process happen.

We would bug R390As at Ft Monmouth and fellow classmates would have to trouble shoot the problem.

fred

Bob

If you can't get higher modulation by figuring it out, I'd contact the outfit that made it.  I'm sure they want it to look good.  I bet there's a simple answer

GL  Al

Tom
Bob
et.

The first waveform that Bob posted looks like the following:

AM modulated r.f. + the audio modulating waveform (not r.f. added to the AM modulated r.f.).

Yes, without a 50 ohm load connected, the audio modulation would feed right through to the output. With the 50 ohm load connected, the 0.1uF capacitor C37 forms a voltage divider with the 50 ohm load... and C37 drops most of the audio frequency voltage.

If the real load (not the dummy load) does not present a low impedance to ground at audiio (not r.f.) frequencies, then the output at the r.f. connector will still look like Bob's original display (even though the modulated RF is getting through to the antenna).

Putting a "safety" choke to ground at the output would also fix the problem... since the safety choke will allow C37 to block the audio frequency signal.

Stu

Bob
Et al.

Regarding the modulation level:

C33 represents 1.5 kohms of impedance (at 1 kHz) in series with the output of the modulatior (point B in the schematic). C116 in parallel with C37 (when the 50 ohm load is present) are forming a voltage divider with C33. Thus, only 1/3 of the audio modulation appears at the drain of Q12.

I would suggest that you increase C33 from 0.1 uF to something closer 1uF

Stu

Hi Stu,
Thanks for the tips.  

One thing that still has me baffled.  When going through my coax connections and switching arrangement and determined to be OK that waveform anomoly still persists.  The only thing I can think of based on yours and Tom's assessment is the antenna tuner is the issue.  The tuner is a T type with input and output variable caps and an inductor forming the vertical part of the T. An ohmmeter shows an open as I would expect since the path is through the tuner (capacitors).  Yet when using an antenna analyzer it shows a good match at the transmitter to the 50 ohm load.  Why would the little transmitter see anything different or why would the scope show the waveform that it did?  Or should I assume this little transmitter is not seeing a load when keyed even though an analyzer says there is a good match? 

I found with the LED disconnected modulation percentage increases too but not sure if the little mod transformer would be up to the task.  I'll try the increase in capacitance and see how it responds.

Bob

The tuner probably does not present a low impedance to ground at audio frequencies. Thus, if you measure the waveform between the transmitter and the tuner, you will see the audio added to the r.f. (which produces the display you are seeing on your scope). You will probably not see the strange waveform at the output of the tuner... because the inductor in the tuner will provide a path to ground at audio frequencies.

To fix this problem, so that you don't see the strange looking waveform at the input of the tuner...you need to add a choke between the output of the Retro 75 and ground.

Any choke having an inductance of 20 uH or greater will do if the self-resonant frequency is above 4 MHz. If you use a 1mH or 2.5 mH choke, it is okay if the self resonant frequency is not too far below 1 MHz.

What I have done in some cases is to make a choke for this purpose by wrapping 8 turns of insulated wire (24 guage hookup wire is okay) around one of the ferrite cores that we use to build transformers for class E rigs. At 3.885 MHz, and this low power level, it will do the job. I put one of these across the output of my Johnson Ranger. It it works like a charm.

One way to do it (if you don't want to wire it in permanently), is to use a T connector to add the choke in parallel at the output of the transmitter. You just attach an appropriate connector to the choke. (e.g., a PL-259)

Stu

The problem with bugging stuff like that is the buggers many times aren't skilled enough to simulate a real world fault.

Wanna feel old? My kid is working at Ft. Monmouth this summer ;D

Now I wonder how many other Retro 75'ers are producing the same waveform but not having any real effects..?

73

Jack.

Thanks Stu.  

Since a dipole typically doesn't have a DC return, I wonder what the effect would be if one were to use a resonant dipole with this little TX?

what mix toroid do you use?

Bob

The toroid I use for things like this is: type 43 material, 1 inch OD, 0.5 inch ID, 1.125 inch long... but I don't think its critical. As long as the toroid doesn't get hot when you turn the transmitter on for a few minutes (at carrier level), you're good to go.

Any plate or drain modulated transmitter that doesn't have a "safety choke" at the output will have some of the audio appear at the r.f. output connector... if it is connected directly to a simple dipole (no r.f. transformer).

For example... in a typical plate modulated tube transmitter... if the DC blocking capactor is 1000 pF (much smaller than in the case of this solid state transmitter), and if the combination of the tuning capactance, the loading capacitance, and the capacitance of the feedline (~20pF per foot) is (for example) 4000 pF; then about 20% of the audio voltage (i.e. modulation of the B+) will be present across the r.f output connector of the transmitter.

The presence of most or part of the audio voltage at the output r.f. connector is not a problem in terms of whether the transmitter works as intended... but it is a problem for safety; and also for measurements (like the measurement you are making).

Note that in Steve's (WA1QIX) modulation monitor, he includes a choke to ground at the input of his r.f. pickup/envelope detector (sensor) to remove audio that might be present.

Stu

I don't know why I didn't notice the lack of a safety choke.  I would make sense to have it in a solid state TX too.  One never knows what's coming in from the outside too aside from what is coming from the inside (TX).  I would imagine that it would also help with static build up on a feedline or antenna  thus sparing the solid state device.

I learned something new, again. 

I just happened to have an FT50-61 already wound.  Must have used it for something similar some time ago.  Put it in place and it is doing the job.

Stu,
C33 is actually 100uF.  It's an electrolytic.  The PDF file shows a decimal point wherer as a paper print out doesn't show the dot. So Xc should be pretty low.

Bob

Yes... since the coupling capacitor has a value of 100uF, the reduction in the modulation will be negligible.

If you put in a keep-alive circuit to hold the drain voltage above (for example) 1 volt, you should be able to get bigger positive peaks by turning up the peak-to-peak amplitude of the modulated B+.

I wonder if the modulation transformer turns ratio was chosen to limit the amplitude of the modulation to 50% (or so) of the B+... in order to prevent the drain voltage from dropping too low on negative peaks. If so, you may need to add a 1:2 step up by using a different transformer.

Stu

Hi Stu,
You are correct in the selection of the transformer.  Modulation along with the limiting circuit is held at about 75%.  I've thought about trying a different transformer to experiment. 

I've removed the LED to turn off the limiting circuit and I can see what you're talking about regarding the drain.  The negative portion of the waveform begins to distort or breakup.

What would you suggest for a keep alive?

Bob

Try something like this. The diodes are all 1N4007. The purpose of using 3 diodes in series is to produce a keep-alive at starts at 5 volts, and levels off at 5 volts - 2.1 volts = 2.9 volts.

The heavy line that is labeled "FET Drain" feeds into the existing LC filter that blocks the RF on the drain from getting back into the modulated B+

Stu

OK, I'll give that a shot. Thanks.