AM Pulse Width Modulaiton

[W9LCE]

Admittedly, I have been most interested in tube equipment - but the Modulation Transformers are more costly than my budget.  I have considered using an old transmitter (a pair of 4-65As) and PWM.

It seems like the on-off square waves of PWM modulation (we built a Pulse Width generator years ago - in Grad Physics Lab) means that the Power Mosfet (IRC?) would have to stand the full plate voltage at turn-off, being a near short at turn-on.  It would have to pass the plate current as it shorts.

As I remember from the Lab, we used a sawtooth oscillator at about 100 KHz to set the pulse width of the audio signal.  I imagine there would need to be good filtering of multiples of that frequency - ie. an audio low pass filter?

I have been looking, and am not finding much information about using those Power MOSFETs as Pulse Width Modulators.   

I am assuming this is a form of Cathode Modulation - giving high efficiency -
While everything I have found to date, is solid state, I assume this can be a Cathode Modulator for a tube amplifier.

Tube Cathode Modulators divide the voltage between the Power Amp and the Modulator, so both levels would need peak voltages above the supply voltage, and efficiency would be moderately low, considering the power input.  A Power Mosfet as PWM would have a low shorting voltage - so the tube would be using almost the entire voltage power.  Am I right?

someone - recently (here?) - mentioned an IRC1750? or something like that - I assumed it was a high voltage Power Mosfet.  The question I ask here is - what would then be a peak tube voltage that the Mosfet could take? 
I have several power supply voltage levels available - from 400v to 600v to 1000v to 2000v - or even higher (an old Pole Transformer at 4500v).

Where can I get solid information about the design for PWM, in Ham Radio, and possibly schematics for such, to study?
advice - and/or information desired

[DMOD]

Some good info here:

http://classe.monkeypuppet.com/

Phil - AC0OB

[KA8WTK]

http://www.classeradio.com/pdm_article.html

Also here

[w1vtp]

Larry, NE1S & Steve  WA1QIX  are the tube PDM guys

[steve_qix]

Well, a lot of good questions.

I've built a lot of tube pulse width modulators and a lot of solid state ones, too.

The efficiency is super-high - the highest you can get - approaching 95% for solid state pulse width modulators and 90% for tubes.

Solid state PWMs are easier to design and build than tube PWMs, but don't let that dissuade you if you want to stay in the tube realm.  It can be done quite nicely - you just need a little bit of extra circuitry.

If you go hi-bred (tube RF, solid state PWM) don't try to go too high in voltage.  Keep it around 1200V total power supply voltage and you can use reasonably inexpensive 1600V IGBTs for the modulator.

A couple of 4D32s or something like that in the RF amplifier running at, say, 450V with a 1200V power supply would work out nicely.

If you with a go tube modulator and RF, you can go higher in voltage. A pair of 813 modulated by something would do nicely, and the total power supply voltage would be around 5000V - 2000V (at 400mA) across the RF amplifier and the remaining 3000V for positive peaks (150% positive in this case).

Lots of options, but read the article posted that's on the class E web site.  It's got a lot of good info.  http://www.classeradio.com/pdm_article.html

Regards,  Steve

[K1JJ]

For what it's worth, here's a journey from inception to testing of a quad 4D32 PDM'd by quad 6LF6 sweep tubes.

It will put out 500 watts of carrier at 180% positive.  It's my favorite AM rig next to Fabio II. Very compact.

As you will read, the 1600V IGBTs as PDM modulators did not work well. Even at 200 watts of carrier they developed thermal runaway and smoked. I tried a copper spreader, forced air and anything suggested.  In contrast, the 6LF6 sweep tube modulators work FB and have never failed.

The rig runs anywhere from 1500- 2200 volts. It's amazing how those 4D32s come to life when at 2KV peak voltage under modulation. Amazing tubes.  The 6LF6 sweep tube modulators have plenty of emission to support the finals.

I use a low level QIX PDM generator to drive the 11N90 MOSFET, which in turn drives the cathodes of the 6LF6 sweep tube PDM modulators.

The 4D32 finals use a broad-band toroidal input for no-tune simplicity.

The rig runs cool in the summertime at 80% RF and 90% PDM efficiency. I used big air coils for the filter left over from an old 4-1000A PDM project. The rig sweeps clean and is all I could axe for.


Initial design:

http://amfone.net/Amforum/index.php?topic=35443.0

Final Shakedown:

http://amfone.net/Amforum/index.php?topic=35596.0

[steve_qix]

I ran 1600V IGBTs as pulse width modulators, and did not have an thermal runaway problems at all.

But, my design was rather different than Tom's so that may have something to do with it.  In fact, the modulators ran quite cool (as they should) modulating 6  6DQ5s in parallel.

I'm not into low power, but if I were, I would probably build a pair of 4D32s modulated by 1600V IGBTs.  The best of both worlds 

[W9LCE]

I would be very interested in seeing a schematic of your transmitter - especially as it might be the low power 4D32 approach

[Opcom]

I realize you have done other stuff with this project but in an attempt to understand this better I redrew most of the diagram shown above using LTSpice.

Is the drawing schematically correct? I could not see everything properly on the pencil version.

[K1JJ]

I realize you have done other stuff with this project but in an attempt to understand this better I redrew most of the diagram shown above using LTSpice.

Is the drawing schematically correct? I could not see everything properly on the pencil version.

Very, bery cool, Patrick!  

Thanks for taking the time to lay it out.

I did a once-over and it looks OK so far.  


Notes:

Anyone building it should add current meters to the screens and grids of both the RF and modulators.

I use a Variac on the HV supply to go from 1200 - 2.2 KV.  1800V is a good level if it is fixed.

D1 and D2 are actually a series of six fast acting damper diodes. I used the same ones as used in the solid state class E rigs. They are mounted on Plexiglas in series.

I have air blowing on all of the tubes.  A slight breeze using a variac to slow it down. Makes a huge difference from running hot to where I can put my fingers on the tube glass after a transmission.  The fans are overhead and push air DOWN on the tubes, creating a nice, even flow.

The aluminum plate for the RF finals sits on six 500 pf door knobs. This plate floats for the PDM modulation.  

The RF final filament transformer needs its secondary to float at 6KV rating(or whatever) with the CT going to the metal plate.  (You might add that to the schematic.)    

The PDM modulator tubes use a standard low voltage fil transformer.  CT to the cathode.

The input RF toroid uses 5KV doorknobs for C16 / C17 to isolate it from the HV at the aluminum plate.  Adjust secondary turns for 50 ohms RF input.

Warning:  The RF aluminum plate is at full HV potential as well as other components associated with this aluminum plate. Consider anywhere in the rig to be dangerous for HV.

For testing, the PDM tube screen voltage is increased until the pulse wave saturates under full RF power under full modulation. HV probe on scope needed.

I'll think of some more notes later, Pat.

Again, thanks for the effort. Might as well archive this since it can be used as-is or as a guide to design another tube PDM rig.

Tom, K1JJ

[Opcom]

I updated it. Just threw in the power supplies without much detail, that part would depend on what is on hand or actually required. Did you regulate the bias or just use a stiff supply for it?

Maybe you can remove the directory details from the bottom of the page. The free PDF printer "PDF Creator" is annoying in that way, there seems no way to prevent that.

[K1JJ]

Very nice, indeed.

Another note:

The secondary of the bias transformer needs to be rated at 6KV isolation and the diodes/ capacitor afterwards need to float above ground. (preferably on a Plexiglas board.)  I wound my own transformer using an old Variac core. For the secondary I used about 50 turns of Teflon coated wire. I taped the Variac core real well using PVC electrical tape before winding. Worked FB.

No, I did not regulate the fixed grid bias supply for the 4D32s. I didn't see any signs of it charging up from RF drive rectification.  It is used for protective bias. The grid leak resistor provides the rest of the bias. The total is about -120V IIRC. This puts the finals hard into class C, almost D for high efficiency.

Again, very nice job, OM.

T

[W9LCE]

has the 3E29 pulse tube been tried as a PWM modulator?  (829B equivalent)

It is rated 85watt input - 15 watt dissipation - so it may be small for the job, but that is to 5000v and 10 Amp plate current

[steve_qix]

A 3E29 might work.

I have used the 6DQ5 as a pulse width modulator, and that tube works quite well.  It has a very high peak current capability which is necessary for a pulse width modulator.

[W3GMS]

I updated it. Just threw in the power supplies without much detail, that part would depend on what is on hand or actually required. Did you regulate the bias or just use a stiff supply for it?

Maybe you can remove the directory details from the bottom of the page. The free PDF printer "PDF Creator" is annoying in that way, there seems no way to prevent that.

Its nice to see a well done schematic! 

Joe GMS

[flintstone mop]

I updated it. Just threw in the power supplies without much detail, that part would depend on what is on hand or actually required. Did you regulate the bias or just use a stiff supply for it?

Maybe you can remove the directory details from the bottom of the page. The free PDF printer "PDF Creator" is annoying in that way, there seems no way to prevent that.

Its nice to see a well done schematic! 

Joe GMS

YES on the re-worked schematic......
I have that AM Ham operator need to make one HB 'high-power' AM rig in my lifetime; and it should be a beauty and have outrageous audio. I prefer the tube approach as, using solid state devices and alignment of class E or D transmitters, and the modulator involved there, is beyond my skills to set-up.
Thanks Tom for the scheezo and Patrick for taking the time to re-draw the original.
I'll take any extra notes Tom posts about some important parameters for the PS and isolation issues and building hints.

Fred

[W3GMS]

I updated it. Just threw in the power supplies without much detail, that part would depend on what is on hand or actually required. Did you regulate the bias or just use a stiff supply for it?

Maybe you can remove the directory details from the bottom of the page. The free PDF printer "PDF Creator" is annoying in that way, there seems no way to prevent that.

Its nice to see a well done schematic!  

Joe GMS

YES on the re-worked schematic......
I have that AM Ham operator need to make one HB 'high-power' AM rig in my lifetime; and it should be a beauty and have outrageous audio. I prefer the tube approach as, using solid state devices and alignment of class E or D transmitters, and the modulator involved there, is beyond my skills to set-up.
Thanks Tom for the scheezo and Patrick for taking the time to re-draw the original.
I'll take any extra notes Tom posts about some important parameters for the PS and isolation issues and building hints.

Fred

Another guy that has a legal limit tube PDM rig on the air is Larry-NE1S.  He built it in the 90's and had it on during the AWA AM net yesterday.  Of all Larry's transmitters I think it sounds the best.  
http://ne1s.rfburn.org/PDM.html

Joe GMS  

[W9LCE]

Steve
back on the 3E29 tube - you said "you think so"

did you look at the specs?
what I see is a tube, that for a very short period of time, can handle immense power.

specs
rectangular wave
for 7 microsecs - 5000v max  at a peak current of 1.5 AMPS
but then for 1.5 microsecs - 5000v max at a peak current of 10 AMPS

if the PDM pulse rate is 70 KHz - that is 1.43 microsecs - within the 1.5 limit

could this single tube handle the biggest Final Tube we would use?  Of course, there are other factors involved, these I do not know enough about to apply (like - what voltage drop will this tube require as a cathode modulator - or PDM?).  (I'm sure the information is on the tube specs - I just don't know how to find it.)

I have several old GE transmitters that use the 832-829 combination - thus I have several 832s - 829s - 5894s, so why not use a 3E29? (same base) - and  - so then use - like a 5894 - cathode modulated by an 829 (or a 3E29)

[steve_qix]

The pulse rating they are showing is not useful for pulse width modulators in our domain of AM transmitters.

The duty cycle of a pulse width modulator (for AM transmitters) varies from 1% to 100%, so the "on" time of the tube has to be figured for 100% (tube is on for a 10th of a second or more, repetitively).  So no matter what the switching frequency, the duty cycle will be 100% occasionally.  This is where the DC current rating of the tube comes in.

So, tubes such as the 6DQ5 and other high current capability tubes make very good pulse width modulators for 6146s, 4D32s, etc.