Series Modulating two 4D32s

[N4LTA]

I am cleaning up after the fire. It was worse than I thought. Structure and cleaning is well above 200K and the contents losses are over 120K.
S1003 amp was upside down on the bench and gt filled with black fire hose water. It may be done for. The tubes survived and probably the transformer but the power supply and regulator board  may end up trashed.

The class D amp was in testing and the boards are ruined. the chassis and meters also got soaked with black  fire hose water. That stuff is corrosive as heck. a day later things were badly corroded.

My beloved R75 got a mild drink and where the water hit it the paint changed color. It didn't get hurt too bad and it still works. It need a good take apart and cleaning though. It was totalled so I got a pretty good looking later model today. Had been looking since the fire. Anyone want the old one for shipping let me know.

Anyway - I had been toying with the idea of a dual 4D32 cathode modulated transmitter for some time. I orderd most of the parts and a chassis and hope to build it in the apartment.

I'll uses a 900 volt Mosfet to modulate it. Believe it or not I found five 11N90 FET in the driveway by the dumpster. They look fine but I have some more coming.

I'll use DMODs circuit he did for me a while back on the current thread. Any idea how large the heatsink needs to be for the FET?  Looks like the current will be 500 mA? Also - any idea of the power out?

I'M here for probably at least 4 months so I need a project.


Pat
N4LTA

P

[KD6VXI]

R75.....  Icom?

I'll take it, lemme know shipping to 93421.

--Shane
KD6VXI

[K1JJ]

Hi Pat,

I think a pair of 4D32s is an excellent choice for a medium size class C rig.  You probably remember my four 4D32s that were PDM cathode modulated.  Except for the dangerous floating HV supplies and the difficulty to get a tube PDM pulse perfect I was impressed with the tubes.  I could get out about 150 watts per tube when pushed very hard with 2KV peak modulating voltage.

I would be interested in joining in a thread about your build. I followed the thread you mentioned about cathode modulation in the past. To be honest, I was not totally convinced that the end result was as good as you might expect for a cathode modulated stage.  There were a lot of attempts and failures along the way.  But it's worth another try.

The bottom line is that I am getting close to finishing up my new 813s X 813s plate modulated rig and want to stay on a building roll. I have 4D32s and sockets and would like to try a MOSFET cathode modulator. Maybe 250 watts carrier out would be a good goal at 1500V on the whole thing or whatever.

Maybe we could start the thread by posting the best suggested MOSFET cathode modulator circuit as a starting point and then come up with a classic 4D32 class C stage.  Post a link to that old thread as a stepping stone.

As a fall back, if the cathode modulation is not perfect, I can always convert it over to a plate modulated rig using my 6LF6 sweep tube modulators.

Within a month I should be ready to start building the project and join you and see.    I know there is a lot of pent up demand for a good MOSFET cathode modulator design based on the last lengthy thread.

To start: Let's say we have a 250 watt carrier. What amount of heat will the MOSFET modulator have to dissipate during dead carrier and during 100% modulation?  How many 11N90s and heat sinking will it mean?   Does the modulator run in class A?  

Yes, 500 mA sounds right for a pair in class C  - I was doing about 1A with four 4D32s, though that was pushing it.

Tom, K1JJ

[w9jsw]

I've still got your large chassis punch. Please advise where you want it sent by email. I expect you may want it for some meter holes on a front panel.

I also want to build one of these. I have the PS parts available. I can also do a simple PC board if you guys settle on a starting circuit, although from what I recall, point to point on the modulator would be pretty straightforward.

John

[km6sn]

I notice the Raytheon spec sheet for the 4d32 states
1. the filament must be heated for at least 2 minutes before application of screen or plate voltages, and
2. The heater supply should be connected to the cathode, BUT IN NO CASE SHOULD THE VOLTAGE BETWEEN
     THE HEATER AND THE CATHODE EXCEED 100 VOLTS!
That means the heater transformer must survive the modulating voltages.

[K1JJ]

I notice the Raytheon spec sheet for the 4d32 states
1. the filament must be heated for at least 2 minutes before application of screen or plate voltages, and
2. The heater supply should be connected to the cathode, BUT IN NO CASE SHOULD THE VOLTAGE BETWEEN
     THE HEATER AND THE CATHODE EXCEED 100 VOLTS!
That means the heater transformer must survive the modulating voltages.

Yes, the filament transformer must be rated for full HV and the CT tied to the final's cathode to keep that fil-cathode voltage potential at zero.  I have wound my own fil xfmr on a Variac core using heavy HV wire as the secondary. The original Variax winding becomes the 120VAC primary. Works FB.

Question: I never could understand the difference between a wasteful, heat producing class A series stage in the cathode of a class C final... compared to "cathode modulation."  Why would the 11N90 MOSFET circuit make this 20% series efficiency any different?   

My 6AQ5 PW class C rig used a 6AQ5 class A series modulator in the cathode. While the class C tube ran cool, I kept smoking the series modulators until I went to two, and even that was not enuff.


T

[w8khk]

Question: I never could understand the difference between a wasteful, heat producing class A series stage in the cathode of a class C final... compared to "cathode modulation."  

There is a big difference between "Series Modulation" and "Cathode Modulation", but a very small, subtle change in the circuit that makes them operate in a very different manner.  But the two terms are often used interchangeably, which adds to the confusion.  

Cathode Modulation is achieved by placing a variable resistance (modulator tube) between the cathode and ground (Plate Supply B- supply return), while the grid bias supply is also returned to the B- supply return.  It does not matter whether the grid bias is an actual bias supply, grid leak bias, or a combination of the two.

In Cathode Modulation, the variable resistance of the modulator tube, while in series between just the cathode and the negative return of the plate supply, causes a rather small change in voltage between the cathode and the grid, altering the conduction of the final amplifier tube.  This could also be viewed as grid modulation in the cathode circuit.  When the modulator resistance increases, a small increase in the cathode voltage results in an increase in voltage between the cathode and the grid of the final, causing a large decrease in final plate current.  Cathode modulation is the result of plugging a series-tube modulator into the key jack of a CW transmitter which uses cathode keying of the final amplifier.

So, in summary, with true cathode modulation, the result is "efficiency modulation" of the final, with very little dissipation in the modulation circuit.  While the series-connected cathode modulator tube sees the entire final amplifier plate current, the voltage is relatively low, resulting in minimal dissipation in the modulator section, but a great deal of dissipation in the final tube, similar to that encountered with grid modulation.

True "Series Modulation" involves one critical difference in the way the circuit is wired.   Instead of returning the grid bias circuit to the plate supply B- terminal, the grid bias must be returned to the cathode of the final amplifier tube.  Now the entire final amplifier "floats" above the plate supply B- and chassis ground.  More than half of the plate supply power is dissipated in the voltage across the series modulator tube.  This is necessary to allow for greater than 100 percent modulation peaks, as well as the additional voltage drop across the modulator tube at maximum plate voltage.  With series modulation it is necessary to float the entire grid and cathode circuit of the final above ground by the peak voltage across the final amplifier.  Series Modulation functions the same way whether the series modulator is above the final amplifier (between the plate supply B+ and the plate of the final amplifier) or below the final amplifier (between cathode and Plate Supply B-)  The former is analogous to the modulation transformer in traditional plate modulation, with one very significant difference.  The modulation transformer in plate modulation adds to, or subtracts from, the plate supply voltages, where the algebraic sum of these two voltages results in the modulated plate voltage.  However, with series modulation in the plate supply, it is necessary to bias the series modulator tube such that more than half of the plate supply is dissipated across the modulator tube, while providing less than half of the plate voltage to the modulated RF final amplifier.  Then as the grid of the Class-A modulator tube swings negative, the resulting voltage across the entire final amplifier is decreased, and the negative modulation peaks are produced by the final.  As the Class-A modulator tube grid swings positive, the voltage across the entire final amplifier rises to double the value of the resting (unmodulated) plate voltage, at which time the final RF amplifier produces an RF output with 100 percent modulation.  The additional supply voltage allows the final plate voltage to rise even higher, producing greater than 100 percent peaks, while there is STILL an additional voltage drop across the modulator tube, thus requiring a significantly larger plate supply voltage.  Remember that with traditional high-level plate modulation, the power to the final is provided by the plate supply AND the output of the class-B modulator, while in series modulation, ALL the power is provided by the plate supply, and "controlled" by the high dissipation of the series modulator tube(s).

So whether we vary the supply voltage of "THE ENTIRE FINAL AMPLIFIER" by placing a series resistance in the B+ side between the supply and final plate circuit, or a series resistance in the B- circuit between the negative supply and the final amplifier cathode/grid circuit, the results are the same.   With the former, it is necessary to float the filament transformer (and grid circuit) of the modulator tube, as was done with the tube PDM implementation.  With the latter, the cathode (and the grid) circuits of the final amplifier must float at a high voltage above chassis ground.  Both of these methods attain a final amplifier efficiency comparable to traditional high-level plate modulation, while the series modulator TUBE efficiency (Class A) is extremely low (but on the other hand extremely efficient (100%) as a shack heater in January).

The 6AQ5 rig mentioned is true series modulation, looking at the grid bias return on the final.

I am doing this on a somewhat larger scale, now that I have finally acquired an Eimac 3CX3000 tube for the series modulator.  The 5000 volt 2000 mA CCS power supply from a Collins 5KW FM broadcast transmitter, will provide 2000 volts at 500 mA to the final and handle positive peaks with ease.  It will not matter whether the series tube is above or below the final, so long as the final amplifier grid bias is referenced to the cathode of the final tube, not B- ground reference.

[N4LTA]

Shane

here is a photo of it sitting beside me with a 6 foot wire antenna. I don't think a lot of water got into it as there are no holes in the top. The speaker sounds a bit funky on loud signals and there is some mild staining on the LCD display You can see the color  change at the top.

Good luck with it. I think it might end up OK but I am not about to sell it to someone. I got a nice one coming  so I am good.

How much do you want to spend shipping - UPS  ground?  The box got burned and all my packing got burned so I will gave to scrounge something up to ship it in.

Send me your address via a message  on this forum.

Pat
N4LTA

[N4LTA]

Tom

I am ready to join you. I think it can be done and I would love to team up. We need to get DMOD on board. I can do a a couple of PC boards one we get it all hammered out.

My UV exposure frame and all the PC Board stuff got burned up , but I have a UV frame coming soon. It is made for T shirt printing but should work fine for PC boards.

Pat
N4LTA

[DMOD]

Tom

I am ready to join you. I think it can be done and I would love to team up. We need to get DMOD on board. I can do a a couple of PC boards one we get it all hammered out.

My UV exposure frame and all the PC Board stuff got burned up , but I have a UV frame coming soon. It is made for T shirt printing but should work fine for PC boards.

Pat
N4LTA

Who, me?  

I remember that circuit.

I have made some minor modifications to it so here is the latest schematic:

Phil

[K1JJ]

Rick:  Thanks for the well-thought-out explanation.  So bottom line is if the grid is referenced to B-, the cathode placed modulator gets more efficient while the class C final gets less efficient.  Or if we reference the grid to the cathode, the modulator gets very inefficient, but the final stays at full class C efficiency.  Sounds almost like a wash?

I'm curious why you are considering series modulation rather than cathode modulation for such a big rig?

Rick, could you take a look at Phil's two circuits and give us your opinion? Is it referenced as a cathode modulator as you described? What would you estimate the carrier efficiency to be for both the modulator or class C final - and during 100% modulation?


Phil: FB on the circuit. That sure makes it easier.   Has anyone built up that rig yet and run some tone tests with it? I followed the last long thread(s) about cathode modulation and my head was spinning after a while. Some guys went into screen modulation...     I am looking for a reasonably efficient circuit that will pass a square wave.  150-200 watts, 60% overall eff would suit me. Maybe plate modulation is the only way to get 250-300 watts out of two tubes.  BTW, I see you are using only 600V, which is the maximum peak voltage under modulation... shouldn't it be more like 1500V peak to make those tubes sing?

I found my pair of 4D32s, sockets and plate caps and decided to start from scratch with a new self-contained chasis.  IE, whether I end up with cathode, screen or plate modulation, I want everything including audio, power supplys, etc., all on one chassis.  I finished wiring my new 813 rig today but still have to build a new HV supply and wire up the mod iron, all in a separate cabinet.  And then test it all. So almost ready for a new project.


John: Did I send you a 3" meter punch or is that someone else you are referring to? I don't remember where my punch ended up so it wouldn't surprise me if I did... :-)  Cool on wanting to build a 4D32 rig too. Between you and Pat, we will have a good PC board to populate the modulator.

Pat: Yes, let's work together and try to finalize a circuit that works in simulation and also makes logical sense by visual analysis. I'm not in the mode to do a lot of experimental trial and error right now and would rather do it one time and do a good layout. I'm sure you feel the same way there.


*** Update:  Thinking back, most "AM efficiency" amplifiers (screen modulated, linear, cathode modulated) I've heard had to be loaded VERY, very heavily (more wasted heat) or they simply sounded mushy and distorted. I think that is the biggest problem guys have with them. They tune for maximum output and the operating point on the load line is way off for cleanliness. In contrast, a plate modulated final does not really care about loading and runs well at reduced or higher power regardless of reasonable loading settings.  So I think the trade off here is going to be OVERALL HEAT, no matter what method is used except for PDM or plate modulation. (even plate has modulator 60% eff issues)  I'm thinking that maybe even a linear amp run in class B/C AM efficiency mode might even be better. But then the exciter has to be pristine. No free lunches.

So back to where we started, that said, a pair of 4D32s in class C,  MOSFET cathode modulated with careful load tuning and somewhat reduced efficiency still sounds sexy... :-)

T

[w8khk]

Rick:  Thanks for the well-thought-out explanation.  So bottom line is if the grid is referenced to B-, the cathode placed modulator gets more efficient while the class C final gets less efficient.  Or if we reference the grid to the cathode, the modulator gets very inefficient, but the final stays at full class C efficiency.  Sounds almost like a wash?

I'm curious why you are considering series modulation rather than cathode modulation for such a big rig?

Tom, I do not look at it as a wash for several reasons.  Grid modulation and Cathode modulation are very similar, both considered "efficiency modulation" whereby the efficiency of the final is varied in order to obtain the desired modulation, using a fixed plate voltage.  It has tuning and loading requirements that are much more critical than plate modulation.  On the other hand, series modulation, whether above or below the final, is virtually identical to high-level plate modulation with respect to the operation of the class-C final, tuning, efficiency, etc.  What we now have is a class-A modulator instead of a push-pull class-B modulator.  Thus the modulator efficiency is low, but we still have a plate modulated rig with regard to the RF circuitry. 

Why would I ever want a class-A modulator, you ask?  I ask what are the weak links in a class-B plate modulated rig???  Maybe the modulation transformer, modulation reactor, and of course the driver transformer for the class-b triodes running grid current on positive peaks.  I throw away all these obstacles, weight, expense, and risk of arcing the transformer windings for a bit of inefficiency to heat the shack in the wintertime.  I would much rather spend money to spin the power meter for a transmitter than a space heater!

My goal is to be able to provide clean modulated high voltage to my P-P 304-TL rig, or any other RF amplifier that needs around 2KV modulated DC, with DC to daylight immaculate frequency response, and no risk of expensive iron crap-out.  Let the audio processor limit my bandwidth.  We know we can do that!

I am currently testing with a couple RCA 810 triodes, working out the details of the driver circuitry.  With class-A series modulation, no grid current is required, but a large voltage swing is paramount.  Perhaps a 4-125 tetrode could swing the grid of the modulator tube, direct coupled using a resistive load.  I am testing with a bank of 60 watt 120 volt incandescent light bulbs.  No RF amplifier needed to test the modulator.  Then scale it up to the 3CX3000 and fine tune it for linear operation.  I may succeed, I may not. PDM may be the only way to get the desired results, but I am not ready to give up.  And, as far as efficiency is concerned, so long as I do not make regular old-buzzard transmissions, my kilowatt hour meter on the back of the house will be none the wiser.  I am doing this to see if it can be done with good results, primarily with regard to the class-A modulator.  Wouldn't it be neat to sweep the rig with square or triangle waves, at full input power, 100 percent modulation, and look at the envelope and spectrum displays with no fear of blowing out expensive iron while we find the sweet spot?  That is my goal.  This might seem like audio-phoolery, the same way those folks with oxygen-free everything brag about their single-ended triode power amplifiers.  They do have an advantage in avoiding many of the issues that are present in push-pull audio, such as cross-over distortion.  They run with no feedback and have superb audio, just inefficiency and very little power.  By the way, I cranked up that 4-1000 deck I finished years ago, due to lack of a suitable modulator.  I did not want to go with grid or screen modulation.  Yes, I have all the iron needed for plate modulation, but I have a hankering to do something different before I settle down and finish my sand-state Class-E PDM rig ala'QIX.

Well, hopefully that gives you a better idea of what I have in mind.  You asked me to look at AC0OB circuits, but I am not sure which ones you are referencing.  Point me to them and I will look, but Pheel has the ability to provide much more insight, not only as the designer, but also with his wealth of experience in practical application and simulation. 

[w9jsw]

The punch I have is Pat's...

I saved some pdfs from the prior thread. Here is the PS supply that Phil proposed.

What spec must I look for with respect to the filament transformer? 600V or something else? I have a Triad 6.3V one that would likely fit the bill.

John

[N4LTA]

If anyone needs the punch - pass it around. I wont need it for a while. I'll used some smaller meters on this rig.


I just bought a filament transformer 6.3 volts at 8 amps, tested to 2000 volts, a 500mA  PS choke and  and a 1200 v ct power transformer rated at 329 VA.

Also got a chassis  - 13x17x3    - so that is a start.


Pat
N4LTA

[K1JJ]

OK Rick -

What I wondered is why did you pick series class A modulation for the big rig rather than cathode modulation?  I get the impression that cathode modulation is more efficient with just the change to the B- reference...

How does  the overall efficiency % of the complete rig compare from class A to cathode modulation?  That is what I am not clear on. It seems like we are getting a free lunch simply by changing the grid reference, but you are still choosing to go with class A ?  :-)


FB on the punches, Pat.  I have been using hole saws for my meters these days. My Greenlees started to get dull and I was told they cannot be sharpened without ruining the precise fit between them. It got to the point where I needed a 3' torque lever to cut thru a panel.

I have out the 4D32s on the table standing in sockets and just thinking about the right modulation method.   I'm still working on the 813 rig. She will be a beauty when I'm done. Pictures coming soon.

Cool on scoring the chassis, fil, choke and plate xfmr.  Youse is surely committed to a project now!

T  

[K1JJ]

I'm still keeping an open mind in the thinking stage and comparing all possibilities...

I was looking at the success some of the guys like Billy/W2BTC, Stu/AB2EZ   and the Tron have talked about using the Antek power transformers as pri/sec OR as auto-transformers for plate modulation. They strap all the windings together in phase.  They sound excellent on the air, sweep very well and cost about $150-$195 or so.  Do they require a coupling cap or modulation reactor?  Not exactly sure what they are doing yet.  

I was thinking of a pair of 4D32s plate modulated by a quad of 6LB6 40 watt sweep tubes, triode-connected, maybe 1,000V on the whole thing for a 250-300 watt carrier out if pushed hard -  200 watts all day.    I'm still not sure how much a pair can put out cathode modulated, maybe 75 watts carrier with a lot of heat?

The 1500VA  950 V version 36 pounds - (at the bottom of the page) or somethong like it is what the guys are using, I think.
https://www.antekinc.com/1500va/       Stu had a thread here about 10 years ago that shud be loooked up.

I'll bet a smaller and less expensive xfmr wud work in this application. 

Anyone have info or thoughts on this?

Just thinking out loud.

T


** TRANSFORMER SPECS AND WINDINGS BELOW:

[W2NBC]

http://amfone.net/Amforum/index.php?topic=13307.0   

[K1JJ]

http://amfone.net/Amforum/index.php?topic=13307.0    

Wow, you hit the motherlode, Jeff!  :-)

I see talk of using an 8 Ohm solid state amplifier to plate modulate thru one of the low impedance windings... got my attention.

Let me absorb all their FB work and come up for air later tonight.

http://amfone.net/Amforum/index.php?topic=13307.0

T

[K1JJ]

To cut to the chase, here's a 2.04:1 impedance ratio if we use all the primary windings tied together  and separately tie the secondaries together.  Like a standard mod transformer.  1.43 : 1 turns ratio.  All tubes. A pair modulated by a pair.

If used with a modulation choke reactor and coupling cap, this might work FB.   Rated at 3500 V.  

There are other smaller transformers available with similar ratios too.


T

[W2NBC]

More info:

http://amfone.net/Amforum/index.php?topic=16683.0

[K1JJ]

OK, now back to the cathode modulation thread....

I think a pair of 4D32s cathode modulated could potentially make a great driver for a linear amplifier. It could probably pass a squarewave and be super clean. So a little power wasted is no big deal, especially if the linear is biased into class B/C.

I notice that I need about 30 watts carrier to drive my 4-1000A linear amp with the FT-1000D, which turns on the fan and shows the strain under dead carrier.  A 50 - 75 watt driver would be a better solution, especially since cathode modulated it will pass a squarewave like my modified FT-1000D will.

The 4D32s are still sitting on my desk in their tube sockets awaiting assignment....  

Who will be the first to build Phil's circuits?

T


* A friend just did this up. Different.  Cool car actually.

[WD5JKO]

Here is a version using a PDM Cathode side modulator, 2 X 807 RF and 6DQ6 modulator:

http://www.lu8jb.com/PWM_807.htm

The PDM board is shown on other rigs listed.

Jim
Wd5JKO

[w8khk]

OK Rick -

What I wondered is why did you pick series class A modulation for the big rig rather than cathode modulation?  I get the impression that cathode modulation is more efficient with just the change to the B- reference...

How does  the overall efficiency % of the complete rig compare from class A to cathode modulation?  That is what I am not clear on. It seems like we are getting a free lunch simply by changing the grid reference, but you are still choosing to go with class A ?  :-)

Tom, it was more of a personal preference, not at all related to overall efficiency.  I will probably only use it in colder weather anyway.  

I did not want to do cathode or grid modulation on my existing P-P 250-TH or P-P 304-TL rigs, for several reasons.  First and foremost, tuning adjustments for maximum linearity are much more critical for efficiency modulation vs high-level plate modulation with an efficient modulated class-C final.  Second, running efficiency modulation runs the final at a much higher level of dissipation for a reasonable output power, and I do not have sufficient spares for these older tubes, thus I wish to baby them while they are still good.  Third, I have just wanted to try scaling up the class-A series modulator approach to legal limit, but did not want to use a bunch of tubes in parallel to try it. And fourth, a pair of 3CX3000F7 tubes and required hardware were obtained in a monetary-free swap, and fifth, all the components required for the 5 KV Collins power supply were obtained through another swap.  It is just something I have wanted to try for the fun of it.  I have done cathode and grid modulation in the past, so I really don't have any interest in playing in that direction.

I still have all the parts to finish the QIX Class-E rig with PDM, and need to find time to work on that project.  If the class A modulator does not prove to be as clean and linear as I expect, I have the option of running PDM just by revising the drive for the series tube and adding a PDM filter network.

It all boils down to a personal preference, and the fun or pleasure of making it work.

[DMOD]

The advantage of Cathode Class A modulation with a mosfet or mosfets is that it is more efficient than a tube because a tube has quite a bit of plate resistance.

Since the Drain-to-Source resistance is very low, on the order of 0.8 to 500 ohms (depending on gate voltage), verses say 6000 ohms or more for a tube, this means the cathode voltage can pulled down to almost 2Volts on modulation peaks.

Additionally, the amount of peak-to-peak voltage required for gate drive is on the order of 5 volts.

Another item is that the only inductors needed are low value RF inductors to keep the RF out of the modulator circuit and does not affect audio, so the audio is very flat.

Lastly, there is no filament power needed for the modulator and the power dissipation of each mosfet is on the order of 25 Watts per transistor for a dual 4D32.

The disadvantage is that the final's cathode filament power needs to be linked to the modulator's Drain voltage to insure the speced voltage differential between the final's filament and cathode isn't exceeded.

I think PDM, while more complex, is great for say 500 Watts and above transmitter systems, as in Broadcast transmitters for efficiency purposes.  

I am not trying to convince anyone that cathode modulation with mosfets is a superior way to modulate but I have good success with converting three different tube transmitters to the Mosfet cathode modulated system.

So many ways to modulate and so little time.  

BTW, sorry to hear about the fire.


Phil - AC0OB

[K1JJ]

The advantage of Cathode Class A modulation with a mosfet or mosfets is that it is more efficient than a tube because a tube has quite a bit of plate resistance.

Lastly, there is no filament power needed for the modulator and the power dissipation of each mosfet is on the order of 25 Watts per transistor for a dual 4D32.

Phil - AC0OB

Hi Phil,

Thanks much for the info.

25 watts MOSFET dissipation for two 4D32s is amazing for a class A modulator!  Plus no filament power. I will have to look into this more closely. What is the carrier power out when running 25 watts dissipation?

So the heatsink is not that big. But what if we use like, 1500V instead of 600V?   Obviously the 11N90s will not handle it. But how about using those IGBT HV devices? I'm told their prices have come down. I played around with them when building the 4D32 PDM rig using them as PDM cathode modulators. They did get hot though at 2000V on the rig and popped, even with a copper spreader/ heatsink.  1500-2000V will certainly ramp it up.  It all depends if the rig is to be used as a low power driver or as a stand-alone rig that can use the power to get thru.

Have you tried passing a squarewave thru and how does a 50 Hz triangle look?

If 600V is the limit, then I can still use a low power 50-75 watt rig that can pass squarewaves to drive my 4X1 linear amp.  It will also double as a low power daytime summertime rig. What's not to like?

Rick:  I see where you're coming from with the class A modulator plans. The critical loading adjustent would be my first reason not to go with efficiency modulation. It's like putting a test tone thru every time you want to tune it up on a different freq.   TNX, OM.


T