Tube AMTX Technical Assistance

Welcome to the board from another tube guy who dabbles in old tube radios. My Crosley tombstone from 1936 is working great after I replaced an open screen grid resistor and unbelievably not replacing the electrolytics. Still working on an old Philco...it has an open winding in an IF can and I just haven't had time to tear into it. RF and audio work, but the signal gets lost in the IF transformer and it ohms out open.

Now regarding your question. I think I saw this on youtube. Is that correct? Nice construction work!

The problem as I see it is that you are applying modulation to the control grid of the final amplifier and the oscillator signal to the screen grid. Thus the mixing (modulation) is happening in the final amplifier control elements. This is going to be nonlinear right at the start since the plate current is a nonlinear function of the audio voltage (almost a square law) due to tube characteristics. Modulation is of necessity not linear but this circuit is non linear in the wrong way. Crystal control won't help at this point but you could add it later, but this will not fix the problem. The FM you report is probably because the screen grid is acting as a tube plate and feeding the audio back to the oscillator.

I recommend some minor changes so you don't have to start from scratch. Connect the screen grid to B+ through a resistor and bypass with a cap. Feed the oscillator signal into the control grid. Put the RF bypassed secondary of the audio transformer in series with the plate. Make L2 a resonant circuit by adding a variable cap in parallel with it and tune for a dip in the cathode (or plate) current. Also, I would put a secondary winding on L2 to couple to the antenna. You will probably need to experiment with this to get the right turns ratio. The present cap coupling is no doubt radiating more harmonics than intended.

I recommend you get your hands on a old ARRL Handbook which has tube transmitter circuits so you can see a typical class C plate modulated circuit. I built a 25 watt AM unit from a schematic in the 1962 handbook which worked extremely well for a carrier current application on a college campus. Nearly perfect trapazoidal test up to 100% modulation. Do a search using key words such as "Class C AM" or "plate modulated AM" and so forth.

I don't think it will require a major rework since your oscillator and power supply are in place but I also don't think you will get satisfactory results with your final amplifier configured as it is unless the modulation percentage is very small, and you won't like that.

Neil

Awesome! Just the stuff I was looking for. One thing that I should talk about. What about power out (input power)? I have NO INTENTIONS of violating Part 15 power requirements with this TX. I know that a 6L6 can put out upwards of 25 - 35 watts and am concerned about this a bit. Much like the "speed limit" post found on this forum, I'm not real worried about keeping the input power "hard" at 100mW but keeping it close is a concern. Do you feel that these modifications would grossly exceed Part 15 rules? I will (and do) check to make sure that my signal does not reach beyond my property, which is the only goal here. I use the TX to broadcast my own selection of content to my antique radios in my home. I like being able to listen to my programming outside while working in my yard and such and I experiment with antennas and placement within my home to achieve these goals so again, keeping within the 100mW spec is not real important for my particular use. Still would be nice to know what you and others think.

THANK YOU AGAIN for your time and effort.

Hi there: Looking at the pictures of the tube transmitter
you built, I think you did a beautiful job. Tube electronics
really are wonderful to look at. I can only give you some
general comments, because I have bad eyes and have to
do most things with a magnifying glass.
Long ago, I modified a Ramsey AM-1 for crystal control.
The AM-1 transmitter had an LC oscillator. The result was
an FMing signal, just like with your transmitter. I had heard
that with the AM-1 transmiiter, it would be possible to
take the oscillator capacitor and coil out and have just put a
crystal across at that point and have a working crystal controlled
transmitter. Actually, I didn't do it that way, because I heard
that idea after I had already done something else that worked.
What I did was, I took the whole oscillator stage out of the
Ramsey transmitter. I built a colpitts crystal controlled oscillator
and put it in it's place. The FMing was then gone (at least as
far as I could tell). The audio sounded better.
The AM-1 (as far as I remember because it is retired from service
now) had an oscillator feeding an RF power amplifier.
I don't know if your transmitter is configured like that. I suppose
current through the crystal might be a consideration, but I don't
really know.
My Ramsey AM-1 might not have done 100 percent modulation,
but I don't think it was too far from that.
Again, I think you have done a beautiful job. I can't give you
any more info than the above, but I'm sure somebody out
there will be able to help you. The transmitter really really
looks great!
Best wishes and good luck!

This might be one of the best thread ever!

I feel like I should send both MICRO1700 and radio8z tuition checks - those posts were like a really interesting electronics class.

When I look at the blog - which is a mighty nice blog btw, and look at your photos there are some things that jump out at me. These observations, if they turn out to be relevant, might help with output power.

Some of the wire runs might be a bit long. Perhaps use a few more terminal strips and rearrange to keep connex short. I'm not sure if "ugly-bugging" so many components is best.

I wonder about winding coils on forms made of cardboard. Done all the time in receivers (though they're usually beeswaxed) but I'm not sure if that's best in the XMTR. They might absorb moisture and detune slightly, change dimension, or shrink and no longer hold the windings firmly, perhaps causing slight detuning or shorts.

I'm concerned about the mounting of the oscillator coil being laid against the rather substanitial chassis. There is so much mass there it might have a dampening or attenuating effect on the coil.

Running some of the connecting lead through the inside of the coil form isn't an approach I've seen used. Usually the coil leads terminate outside and close to the coil form and nothing passes inside the coil.

It might be worth trying to reconfigure the oscillator coil to have short leads outside the coil, and perhaps to mount it vertically on the topside of the chassis.

With both coils, should they be close-wound or would Q improve by opening that up a bit?

Also, if that magnet wire is from the Radio Shack three roll pack, the red is usually the lightest gauge. IIRC, it's natural, green then red going largest to smallest. Is the smaller diameter red wire introducing excess resistance?

Well, I misread the schematic and got a detail wrong. The oscillator output is not applied to the screen grid...I missed the bypass capacitor which blocks it. The screen grid is connected to the B+ as it should be.

The oscillator output is applied along with the audio to the control grid and likewise the audio is coupled back to the oscillator. This nicely explains the FM you mentioned.

This is not a good way to mix the audio with the RF and my suggestion about plate modulation is worth pursuing. The problem with this circuit is it is depending upon the non linearity of the tube to mix the signals. This generates strong harmonics which not only waste power but cause interference. By letting the tube amplify the RF and the modulation transformer modulate the plate voltage you gain efficiency. The tuned LC circuit in the plate circuit ("the tank") is very effective in removing the unwanted harmonics provided it is resonant and lightly loaded by the antenna.

The power input (which don't forget includes the screen grid power) can be controlled by the screen voltage to a certain extent but usually it is best to control it with the RF drive to the control grid as well as the control grid biasing. The more negative the grid with respect to the cathode, the less plate current is drawn. This circuit is biased by the resistor in the cathode and changing this will change the grid bias but it will also be necessary to set the grid drive (oscillator signal into the grid) to the proper level. My guess from the readings on your schematic is that the input power is about 3.7 watts. (cathode current = 7.6V/680 ohms = approximate plate current. This times 332 volts at the plate yields 3.7 watts and doesn't account for the screen power.)

There is not an unlimited range of screen voltage which is acceptable. If too high, it steals a lot of current from the plate and can overheat, and if too low, it inhibits the electrons on their way to the plate. The tube data sheet might have the recommended screen conditions.

Sorry about my mistake earlier and I hope this information helps.

Neil

I have reworked my AMTX a bit and posted new and current pictures on my server. I invite you all to come and check them out. The TX is working much better after applying some of your suggestions. There is one more thing that I would like to tackle to make this TX really "sing". I'm getting some oscillator whine or howl coming through the audio. It's not real bad but it is there and it gets annoying after listening for a while. Any suggestions on how I can either eliminate or reduce it somehow? I was thinking about a low-pass filter but I'm not sure this will work. I'm also not sure just where I would apply a low-pass filter in the circuit. Maybe you have a better suggestion? You can find the new pictures at: http://2xl.homelinux.com/riverty/amtx/index.html

Thank you all for your time.

rt...

Very interesting project. Great pictures.

I'm curious about the point at which the audio is applied.

As Neil pointed out, high level (plate) modulation would be a better way to go.

However, working with what you have I wonder if screen grid modulation would have less affect on the carrier frequency compared to applying it to the control grid along with the oscillator output.

Depending on the turns ratio of the audio transformer there may or may not be enough audio voltage to modulate the screen grid.

The Lafayette KT-195 I have was designed with screen grid modulation. The modulator plate feeds audio to the screen grid through a series R/C network. B+ is applied to the screen grid in a normal fashion. Simply moving the audio from the control grid to the screen grid might work with attention to the blocking capacitor.

The KT-195 dosen't have a separate oscillator and final output stage and as such there is some carrier shift with modulation. But where you have separate stages I would think there would be less FM'ing.

I modified the KT-195 for plate modulation and noticed an improvement. You can find the schematics at THIS LINK.

rt,

Glad you are making progress. I need a little help in understanding what you mean by "oscillator whine". Please give a description. I checked your site and your nice construction is pleasing to view. It appears that you have not substantially changed the circuit. Did I miss something?

Under what transmitter and receiver situations do you hear this?

Is it there with no modulation? Low modulation? Full modulation?

Is it happening when you are close enough to the transmitter with the receiver where the other stations on the frequency are swamped?

Anything else you might report might help us help you "bring this baby home".

MRAM beat me by four minutes. I have to learn to type faster but for now he's the champ!

Neil

First, I forgot to answer the question of "is this from YouTube?" The answer is yes, it's from Rick ?? (I forgot his last name) and you can find his page here: http://www.vacuumtuberadio.com/AM_Transmitter/.

As it is now the TX works better than it ever has. Audio is full and "punchy" and the FM'ing problem is much better if not gone. I used to have to tune on the fringes of the carrier signal to get audio. Now, I have a nice peak to tune on to. To visualize this, before the signal was more like a "V" where the audio was on the "ends" of the carrier signal. Now it's much more like an "A" where the audio peaks on the intended frequency of 1290 AM. I should also mention that I have experienced very little frequency drift as the TX warms up. Good Stuff!

You are correct in that the circuit has not been modified. Well, it has been modified numerous times before but I thought it would be best to get the circuit back to original for discussion and troubleshooting sake. What you see on the schematic is what it is now.

About the "whine" or "howl" I'm talking about. Honestly, I'm not hearing it today! Go figure. I guess everything is great right now. I'll have to watch this in the future.

Radio's that I use to test with vary. I have a 1948 Stewart Warner that I use mainly and the TX sounds great. I have also used a Sony Walkman, my car radio parked in the garage, an AM/FM car radio that I use in the garage, and a 1980's vintage Technics sterio amp and radio in the garage. I use my Grundig G5 to help fine tune the TX frequency, a cheap GPX AM/FM/CD under-counter radio in the kitchen, and an Emerson AM/FM radio in the bathroom. As I type this, everything sounds great. The Grundig G5 gives a little whine when I first power it up but the whine goes away after a few seconds as (I'm guessing) the radio filters this out.

riverty,

There is no arguing with success and I am pleased that you are getting your unit to work as you expect. No problem at all with reverting to the original circuit but as you know it would not be my preference.

OK on the youtube link and thanks. I thought your pictures looked familiar. My former boss used to comment about my photographic memory and I used to retort that he must mean my pornographic memory but that is another discussion for another board.

Your description of the audio tuning in in a V makes sense to me. Diagnosing from afar this indicates to me that the audio was severely distorted and by tuning off the carrier you were tuning to one sideband which probably accentuated the higher audio frequencies which are produce by distortion. It is almost by accident that this sounded "better". Control grid modulation and mixing as shown in the schematic is not a good way to do this. Now that you have the unit working better, which as I understand it, came about by you improving the lead dress you are probably not willing to change things. If the transmitter is working to your satisfaction then you should leave it be.

That being said, keep in mind that I think, based on the circuit schematic, that you are airing a very dirty signal. Listen up and down the band for artifacts. Proceed with caution.

Neil

First, before I forget. I have recorded a sample of the "howl" I'm getting in the audio. I guess it "comes and goes" depending on ?? I dunno yet. Either way, if you reload the page link above you will find I updated the page with a sample in wav or mp3 format. You can clearly hear it in the sample although it might not be apparent right away. Check it out if you like.

"Your description of the audio tuning in in a V makes sense to me. Diagnosing from afar this indicates to me that the audio was severely distorted and by tuning off the carrier you were tuning to one sideband which probably accentuated the higher audio frequencies which are produce by distortion."

I wouldn't say "severely distorted." I would say "squashed." At that time, the audio just sounded pinched in a more volume way rather than distortion. It's still distortion, and your right in what you say. I'm just commenting on the "sound" of it then. I was told that this is a symptom of FM'ing.

I did some walking around the house tonight with my Sony Walkman. Earbuds help in sampling audio and performance for me. Your right about a "dirty" signal. it's pretty crappy with a little distance. Within my house the signal is much better but after a little testing, it could be much better. "Scwis" got me this far. Better lead dress and more accurate wiring was needed.

"By letting the tube amplify the RF and the modulation transformer modulate the plate voltage you gain efficiency. The tuned LC circuit in the plate circuit ("the tank") is very effective in removing the unwanted harmonics provided it is resonant and lightly loaded by the antenna."

I'm reminded of a joke from "Friends" my lovely wife watches all the time. "I don't know what you just said!" Not really. This is where things get a little fuzzy for me. I understand the theory of letting the tube amplify RF, and modulating the plate voltage. Sooo, do I simply remove the audio transformer lead from the 0.0047mF cap and reconnect it to pin 3 on the 6L6? - or - Do I have to design an LC circuit for this mod? Honestly, this is the stuff I lack. Experience in tube modifications. I think if I had more knowledge here, I wouldn't have the problems I'm having - LOL!

Anyway, I would love to try your suggestions, and I believe what you say, I'm just not sure how to get there.