Homebrew transmitter progress updates

[KA2DZT]

Ron,

Most of those multimatch mod xfmrs are not exactly the same on each side.  There is a step-up from the total one side to the total other side.  Or, in reverse there would be a step-down.  Also, the taps on each of the four windings are not in the center.  You would want to use the entire winding on the mod tubes.  Then try to use the entire secondary, but you may use something less on the secondary to reach the 1-1 ratio.

I think, because the voltage on the modulator is only 2/3 the voltage on the final you will need some step-up to reach 100%  So, with a 8K load, the load on the 811s will be less than 8K.

My mod xfmr is 9K to 5K plus a 50hy reactor,  That's about 1.34-1 step-down but, I run 1500 volts from a single plate supply to both the 811s and the single 813.  My xmtr will easily do 150% positive.

Fred

[W8ACR]

OK Fred thanks. I'll start at about 1:1 and see how it goes. Perhaps some of the other veterans would like to weigh in on this subject.

Ron

[W8ACR]

Update#14 - 4/20/13 - The speech amp is finished. One more deck to go. The 300V B+ will be coming from a separate regulated power supply. (Just in case someone might notice there is no power supply on this deck)

[Todd, KA1KAQ]

As I understand it, this improves frequency response. I think this is feasible as long as the modulator has power to spare.

That's the case with the stock iron in the KW-1, Ron. It has taps for 18K or 12K, with 12K yielding better fidelity, though not as efficient to drive. Having some drive in reserve is a big plus.

It's coming along great!

[W8ACR]

Here is the speech amp mounted inside the cabinet just below its power supply. That's quite a power supply for a 10 watt speech amp. Dontchya think?   It's been sitting collecting dust for two years. It's such a nice PS I had to come up with a use for it.

Modulator construction to commence shortly.

[KA2DZT]

Ron,

Looks like a Lambda regulated PS.  You could run half the xmtr from it.

Everything is looking great.  Took me over 10 years to build my xmtr.

What exactly is the Ohmite rheostat for??  I'm hoping that it's not the volume control.

Fred

[W8ACR]

Fred,

Yes, it is a Lambda model 0-324 power supply. Will supply 200-325VDC@150mA. I actually have another one also. I got them for basically nothing at the W0ZUS estate auction. Very nice high quality piece of equipment. Two 5U4's and four 6L6G's. Heck, it would probably cost over $100 just to buy new tubes for it. I don't have a manual though, and I can't seem to find one on the internet.

The rheostat is 1000 ohms/100 watts, and its purpose is to adjust the cathode bias for the 6A3 drivers. I thought it would be more convenient than a WW resistor with a slider. Adjust for 80mA idling current. See the previously posted schematic. It is used in place of the 750 ohm/10 watt resistor.

Ron

[W7TFO]

Soft warm fuzzies throughout!

73DG

[K1JJ]

Tom, I have seen several types of NFB circuits, so I'm generally familiar with them. I have seen them come off the plate of the driver, and I have seen them come off the plate of the modulator. I'd be interested in seeing the NFB circuit you have in mind. How do I know which 811 to use to be 180 out of phase? There will probably be about 1000-1100VDC on the plates of the 811's. I should have plenty of spare audio unless I run the final at full bore. Even then I should have some spare audio.

Ron

Ron,


http://amfone.net/Amforum/index.php?action=dlattach;topic=33816.0;attach=37534;image

The NFB you axed about:   Take a close look at this picture (lower left) and you will see the layout.  I use about 2 megs of resistors in a ladder and tap the low level NFB signal off a 9K ohm resistor at the bottom of the 2 meg ladder to ground.  I have a 180 pf HV capacitor across the whole string.   I tap the HV audio signal off one side of the mod transformer secondary and connect to the top of the 2 meg ladder.   The 9K will be adjusted depending on how much signal you need to work with into the low level NFB pot to the preamp.

*** BTW, I just saw an error in the wiring.  The black HV lead from the mod transformer scecondary needs to go to the TOP of the ladder, not to the yellow cap. Glad I caught this before firing it up!  **


I use a pot to dial in the amount of NFB into the low level audio input.  

To determine the phase (180 degrees) we can simply look at the schematic and know that each stage gives a 180 degree shift and connect accordingly to end up with a net 180 degree shift. OR, simply set the low level signal pot to minimum signal and then slowly turn the pot up to add in some NFB.  If the audio modulator audio starts to go down, the phase is correct. If it starts to go up (or even take off) it needs to go back or forward  another stage to find 180 degrees.

T

[W1RKW]

Ron,
FWIW, I'm using a CVM-4 on 2 813s to modulate 2 813s.  All windings are used.  WA1HLR suggested configuring it as an autoformer.  The rig does about 125% or so.

[WD5JKO]

NFB Comment:

When applying NFB to audio stages in a modulator, consider the power stage headroom. If the occasional audio peak puts the power tubes in a momentarily overload, then only apply NFB (from mod transformer secondary back) if the audio power stage has transformer coupling at the grid side, and then limit the overall NFB to 3db.

Why?

When in NFB we have a close loop gain less then the open loop gain. When we overload we go to the open loop gain. If the audio stages are R-C coupled, then when going open loop, we get a momentary pulse such that the power tube overloaded has a bias shift. This bias shift might have a 1/4 to 1/2 second delay depending on the R-C time constant, and the result is dreadful sound. The R-C stuff applies more to those with class Ab1 modulators like the 6L6's in a Johnson rig where we pull the interstage transformer and replace it with a phase splitter tube.

With my Gonset G-76, NFB around the audio driver tube helps a bunch. I then apply only 3 db NFB around the mod transformer back. Anymore then that, the audio suffers since I operate close to overload on peaks. The driver stage should never be overloaded, so more NFB is used there.

With the case here, 6A3's driving a pair of 811's, some NFB from 811 grid upstream is useful. With my 1980's AM KW, I used a pair of 6B4's to drive a quad of 808's. The 6B4's were driven by a 12BH7 P-P, and preceded by a hot cathode phase inverter. I used push pull NFB from the modulator grids back to each 12BH7 cathode, phased correct for NFB. Cleaning up the grid signal to the 808's was all I needed...those tubes were so clean that no overall loop NFB was required.

Jim
WD5JKO

[K1JJ]

Hi Jim,

Good info, tnx.

Possibly to avoid these problems, the modulator and driver system should be designed so that they has plenty of headroom. So much that it will never go into overload in the first place.  If it does, then replace with a bigger one...

I plan to run some exhaustive THD tests on my 4X1 modulators soon. Now that I have the test equipment set up, I want to test with various levels of NFB, different bias levels as well as seeing what effect some grid current does to degrade things.  Using a well-regulated screen supply and a solid state MOSFET audio driver, I'm hoping to shoot for at least  -30dB 2nd harmonic THD or better as a goal.

It's amazing how much we can improve our AM rig's overall cleanliness by playing around with settings and values as we watch for sweet spots in THD on a spectrum analyzer.

In the past, all I did was fire up a new rig, run some tones through and look at the scope sine waves. I didn't know what else I could do.   To reach the next level of analysis and optimization  took the spec analyzer setup.

T

[W8ACR]

Thanks for the comments guys. Jim, I like the idea of pushpull feedback from the 811 grids, but in the case of a 6N7, the cathodes have a common connection to pin 8, so it seems that PP feedback to those cathodes is not feasible. Would PP feedback to the 6N7 plates also work? or would it be better to do single ended feedback to the 6SJ7 cathode? Also, since I don't have a spectrum analyzer, how could I make any meaningful conclusions whether the feedback is improving the audio. My speech amp schematic is posted earlier in the thread.

Thanks, Ron

[WD5JKO]

Tom,

  Yes going for more headroom is more of a modern approach, and therefore more applicable to your 4-1 thread, whereas this thread is more of a glorious retrospective of the earlier time when we sized the modulator RMS power at about 1/2 the RF Final DC plate input power. In a rig such as this one, maintaining a high level of modulation without the occasional overload is near impossible unless some form of audio processing is used.

   Another reason to limit the NFB (mod transformer secondary back to low level stage) to about 3 db is the phase shift from the two transformers in the signal path can be high at the audio extremes such that instability is likely unless we do some form of phase compensation.

  Ron,

 I think I'd omit NFB at this point in the design. This can be added later if need be. Since the 6N7 is the phase splitter, adding NFB will disturb the AC balance. It's a little late, but following the 6N7 with a 6SN7 P-P with separate cathodes would do. You should be able to run about 10db NFB this way, so maybe the surplus of gain gets used up with the lower closed loop gain. If there is still too much gain, maybe triode connect that 6SJ7.

Oh, the 6N7 resistive divider with a 470K and an 18K is not likely to be exact. Maybe replace that 18K with a 50K pot. Then later with sine wave tone testing, adjust the pot while looking at the modulator plate meter. There should be a slight dip in plate current when things are balanced.

If you have a selective receiver, you can look at the audio harmonics. I can do this with my ricebox with a 250hz CW filter set to the audio second harmonic. Balancing the audio should null the second harmonic in a P-P modulator. The main consideration is to keep the rcvr away from overload. Maybe set the carrier level to S9+20 by controlling the RF sampling level to the rcvr. Do this with a 1Khz tone and about 80% modulation.

If you calibrate your S-Meter in DB, when the 2nd harmonic is down 30db from the fundimental 1 Khz tone, that is 1% distortion.

Jim
WD5JKO

[K1JJ]

Tom,

     Another reason to limit the NFB (mod transformer secondary back to low level stage) to about 3 db is the phase shift from the two transformers in the signal path can be high at the audio extremes such that instability is likely unless we do some form of phase compensation.

  Jim
WD5JKO

Yes, exactly.  Trying to get more than a few dB NFB with 2 transformers or more is extremely difficult due to the transformers' cumulative phase shift.  When I used a backwards driver transformer, I could not include the mod xfmr without the NFB causing the modulator to take off.  But with the solid stae MOSFET audio driver, the only xfmr is the mod xfmr so, FB

I'll bet Ron can tap off the modulator plate cap and deal with just that driver xfmr and be OK.

T

[WD5JKO]

Possibly to avoid these problems, the modulator and driver system should be designed so that it has plenty of headroom. So much that it will never go into overload in the first place.  If it does, then replace with a bigger one...

   Yes, but overload on plate modulated AM goes both ways. Having the additional headroom to modulate beyond 100% upward brings about the increased tendency to "white-line" at 100% downward modulation. This causes another kind of distortion, that is far more serious then compressed positive peaks, and also puts the modulation transformer at risk.

   So sure there are ways to deal with all of this, and both you, I , and many others have. That said, there will never be a consensus on which way is better. So maybe an old style traditional modulator like this one is best without the extra overhead. The OM's back in the day knew a few things about this stuff.

Jim
WD5JKO

[Steve - K4HX]

Good points Jim. Get one thing, give up another.

The one thing that has changed since the old days is the plentiful and low cost supply of compressor/limiters to prevent over modulation and destructive peaks. Scope are more common now too. There's really no good reason to over modulate or flat-top now days.

[K1JJ]

   Yes, but overload on plate modulated AM goes both ways. Having the additional headroom to modulate beyond 100% upward brings about the increased tendency to "white-line" at 100% downward modulation. This causes another kind of distortion, that is far more serious then compressed positive peaks, and also puts the modulation transformer at risk.
  

As already mentioned by you and Steve/ HX, I like two ways to limit negative peaks:  One is a light compressor action and the other is a negative peak limiter.  

Steve/ QIX has an excellent low level NPL that works with simple diodes, but here's the key: He has an 8-pole low pass filter after the diodes that removes the garbage that clipping alone would produce.  

This NPL is incorporated within the PDM generator board he sells.  I plan to use just the NPL / filter circuit for my HPSDR rig, the FT-1000D and the new 4X1 rig.  It worked tremendously well with my old class E rig. The circuit is posted somewhere on this BB.

 In contrast, using the "3-diode" high level circuit worked, but generated some crud since it did not have a LP filter following it.

It can all be done low level. This may be the cleanest method to do it. We'll see.


I realize this is preaching to the choir here, but there are probably some others not familiar with what we are discussing.  Limiting negative peaks, however we do it, whether old school or new school,  is important.  

T

[K1JJ]

  So sure there are ways to deal with all of this, and both you, I , and many others have. That said, there will never be a consensus on which way is better. So maybe an old style traditional modulator like this one is best without the extra overhead. The OM's back in the day knew a few things about this stuff.

I think they knew a few things, but limited themselves intentionally. The old school hams were more interested in "communications" and doing it as cheaply as possible.  High fidelity AM was heard but not common in the 30's, 40's and 50's.   The commercial ham gear, well, we still hear the boat anchors today, were not very hi-fi.  But, the state of the art for hi-fi was there if desired. AM broadcash rigs from those days sounded very FB and had plenty of headroom.

I think it was simply the evolution of the ham trend to that point.  Years later, we have a trend for more hi-fi AM signals, whether modified or home built.

Still, the old school stuff is cool and worth pursuing as Ron is doing. Heck, if I wanted to stay totally modern, I wouldn't have torn down my class E rig and started building a 1950's style tube rig myself.

Just like tube audio amps vs: solid state.  We all know which has the best specs, so why do so many gravitate to the tube stuff?   Some like the nostalgia - and some like performance.  I like a combination of the two.

T

[W8ACR]

Update#15 - 4/21/13 - Mod deck is finished! Buzzard Baby is now a complete AM transmitter a la 1940's technology. Shown below are some pics of the individual modulator deck as well as the entire transmitter assembled and lit up. The only testing done so far was to apply power including the plate voltages, and everything looked good. I have to wire up the mic cable and put a relay on the speech amp deck and then we'll be ready for some actual transmitting testing. I'll give you all 24hr notice when I'm ready for the maiden voyage. It'll have to be on 40 meters and will probably be in the evening between 8 and 10 PM. I haven't been on 40 AM for a while. What frequency do you suggest?

[W8ACR]

More pics

[W8ACR]

Here she is all lit up

[Steve - K4HX]

Sweet! Congrats on your build. Hope to hear it on the air soon.

[W3GMS]

I'll give you all 24hr notice when I'm ready for the maiden voyage. It'll have to be on 40 meters and will probably be in the evening between 8 and 10 PM. I haven't been on 40 AM for a while. What frequency do you suggest?

Ron,

Any frequency that the broadcast stations are not on!  During the day here on the East Coast 7.290 is a popular gathering for AM op's.  I normally don't get on in the evenings so not sure about the frequency.  I will get on when your ready for the on the air smoke test! 

Congrats on a fine build vintage style.

Joe, W3GMS

[K1JJ]

Ron,

Congrats on completing the rig, OM!


You may now play this video and dance around Gangnam style as you call your first CQ:


http://www.youtube.com/watch?v=mIQToVqDMb8

T