Designing a QRO Series Modulated AM Transmitter

[w8khk]

Chapter 1 - Conception to Birth

For several years I have been contemplating the construction of a QRO series-modulated AM transmitter.  With the tubes at hand, it seemed fruitless, because the series modulator tube is so inefficient.  Might as well stick to 6AQ5s or 6L6s, as most others have done.  Then one day, I had the opportunity to swap some old relic RCA 851 tubes for a few Eimac 3CX2500F3 triodes, with a dissipation rating of almost 4000 watts.  Now it just might be possible to make a series-modulated QRO rig, that would double as a 100-percent efficient "Shack Heater" in January.  A few months later, I acquired the requisite power supply components, consisting of the plate transformer, bridge rectifier stack, filter inductors and oil capacitors from a Collins FM broadcast transmitter, everything required to make 5 KW CCS at 5000 volts.  Now it was finally possible to start looking at options for a legal-limit rig, designed conservatively enough for reasonable reliability.

Rather than jump off a cliff with an untested design, I decided to start a bit more conservatively, but still somewhat greater than the PW power of a little seven-pin receiving tube.  One design constraint is to utilize the bottles I have in the junk-box.  No eBay or RF Parts purchases allowed.  One option might be to use several RCA 810s, or maybe an RCA 833A, to series modulate an RCA 8000.  I performed tests of the tubes in my inventory, to determine that the filaments were intact, no leaking seals, no gassy tubes.  A pine-board Frankenstein "widow-maker" test setup was assembled in the lab.  Variacs to control everything from a distance were set up across the room.  Being that my xyl is now my ex. my winter lab is the main living room in the front of the house, making it very comfortable for design and testing activities.

I attached a few photos showing the testing in progress for the 810, the 833, and finally, the Eimac 304-TL.  The long-range plan is to series-modulate a pair of 304-TLs with a single 3CX2500F3 in the cathode follower configuration.  This is expected to provide a more linear transfer characteristic, with a driver that needs to provide only a voltage swing, no grid current on the series modulator.  I hoped to use my father's (W2DU, SK) push-pull 304-TL deck that he built in our apartment in 1949 when I was only two.  But after further research, I decided to construct a new final deck, floating the input circuit and filaments, such that modulator driving circuits may be simplified.  The current plan is to run a pair of 304-TLs in parallel, pi-network output, neutralized with a balanced grid drive circuit.  This approach allows input RF coupling isolated via a link circuit, and a grid leak eliminates the need for a floating bias supply.  

Initial testing indicates a moderately powered rig may be assembled with either two or three of the 810s as series modulators, with a single 8000 as the final.  Further dynamic testing will be performed with a resistive dummy load representing the final amplifier, proving the driver and modulator circuits before building the RF section.  Using this same validation method, the mid-range rig might be skipped, moving on to testing the QRO modulator without an RF deck.

The reason I selected the relatively rare 304-TL for the final is the awesome reserve of available emission, making it possible to provide positive peaks in excess of 100 percent, with relatively low plate voltage.  The modulator design will include a "power control" allowing the adjustment of the plate voltage on the final amplifier.  A single 304-TL is able to reach one kilowatt input with as little as 1500 volts on the plate.  A pair can do it with 1000 volts, although the combined plate current is significantly higher.  One hundred percent positive peaks may be achieved with the series modulator when running around 2000 volts on the final, at reduced plate current, maintaining legal limit output power at much higher efficiency in both the final and modulator.  Other tubes that require a much higher plate voltage to reach legal limit output would not provide the positive peak headroom without excessive plate voltage on the modulator tube.

[w8khk]

The attached Excel graphs illustrate some of the static testing performed on 833 (grid voltage v output voltage with a 10 kohm load, and on the 810 showing cutoff bias and zero bias plate current at various voltages. 

It can be seen that the transfer characteristic is relatively linear.  All design will be limited to dissipation levels well within the tube specifications.

An interesting point regarding the series modulation method of efficiency modulation is the fact that maximum dissipation of the series modulator tube is observed at carrier with no modulation.  As the modulation level is increased, the series modulator tube dissipation decreases, and that power now appears in the sidebands of the output waveform.  In contrast to class-B plate modulation, the current demand on the power supply is static.  That is to say that between no modulation and maximum modulation conditions, the average current demand upon the plate supply does not deviate.

[w8khk]

My original plan was to place the series modulator between the positive terminal of the plate supply and the final amplifier, allowing the use of the classic grounded cathode RF deck.  This approach would complicate the drive to the cathode follower series modulator tube.  Another alternative would place the series modulator between the ground return of the final RF deck and the power supply negative return.  This approach would place extremely high voltages on the plate side of the final amplifier.  I wanted to minimize the exposure to extremely high voltages as much as possible, so I will experiment with a compromise, that entails grounding the plate circuit of the final amplifier (as was done in the Collins/Continental one KW power pebble) but I will float the power supply such that the cathode return of the modulator may be grounded.  This will simplify the audio drive, but the disadvantage is the possibility of audio high frequency attenuation caused by distributed capacity in the power supply.  I expect this issue may be addressed by coupling both the negative and positive sides of the power supply to the load through filter reactors.  This assumption must be validated via resistive dummy load testing.

Attached are initial prototype drawings of the proposed circuit configuration.  The driver circuit has yet to be finalized, and it is expected that the 3CX2500F3 may be driven by a much smaller tube than an Eimac 4-250A, and the voltage swing is not expected to be extreme, even though the tube is operated as a cathode follower.  This improvement results from referencing the cathode of the series modulator to ground, and floating the power supply.

[w8khk]

The two circuits in the previous post look somewhat abstract and unusual.  I am attaching here a couple alternative views illustrating the circuit is somewhat conventional, with the cathode follower series modulator above the RF final amplifier.

The unusual variation is the fact that the entire circuit is grounded at the plate circuit of the final, and the cathode circuit of the series modulator.  If leakage capacitance in the power supply may be addressed with two filter reactors, one in the positive leg, and the other in the negative leg, then I will go ahead and construct the transmitter as designed.  Initially, I may test the overall design with the mid-size 810/8000 bottles, but if the driver design proves to provide the desired linearity without undue high-frequency attenuation with the floating supply, I may then turn my attention to muffling the loud noise resulting from the 3CX2500F3 cooling air blast.

I have considered the possibility of changing to a class-H modulator design, with a carrier and a peak tube, but at the present time I would prefer to keep it simple and attempt to create an extremely linear modulator with a single series tube, ignoring the dissipation and associated inefficiency.  Come spring, my thoughts may change toward a more summer-compatible design.

Please note that the audio driver circuitry is a conceptual design, and has not yet been thoroughly planned or tested.  Stay tuned for more developments, as the conceived design hopefully becomes a practical reality.

[K8DI]

It took me a minute of drawing the two sections on one paper to follow what’s going on. One thing unsaid but important is realizing that the center tapped filament transformers are in fact RF chokes.  It also seems to me the modulator filament transformer secondary will be passing audio/modulation, which may effect high frequency audio response.

If you really think about it, any circuit is a loop. The loop may be the same or different for DC, audio, and/or RF. There’s no requirement that any particular point in the loop be grounded. Wherever you do RF ground it, the other side of the RF loop will be a point you can pull off ground-referenced RF.

Ed

[K1JJ]

That's a lot of good info, Rick!

I wish I had that much for my 813s series modulated rig, but I lucked out and it worked FB despite.

If you build this big rig and later decide the modulator is just too much heat except in that narrow band between November and March, then it would be an easy task to change it over to a high efficiency PWM transmitter.  All the floating stuff is there.  Just add another 3CX-2500F3 to the final and you're Mr. BIG!   

But until then, good luck with the experiments.

The 304TL using lower HV is a nice feature. With 5KV you will have plenty of headroom.


T

[w8khk]

It took me a minute of drawing the two sections on one paper to follow what’s going on. One thing unsaid but important is realizing that the center tapped filament transformers are in fact RF chokes.  It also seems to me the modulator filament transformer secondary will be passing audio/modulation, which may effect high frequency audio response.

If you really think about it, any circuit is a loop. The loop may be the same or different for DC, audio, and/or RF. There’s no requirement that any particular point in the loop be grounded. Wherever you do RF ground it, the other side of the RF loop will be a point you can pull off ground-referenced RF.

Ed

Thank you for your comments and suggestions, Ed.  Yes, I thought the separate pages of the RF and modulator stages might be a bit difficult to digest separately, hence I provided a couple simplified sketches in the last post.  I could not fit the detailed version on one page with enough resolution to be readable after imaging it.

I have considered the series reactance of the filament transformers, and I believe I have addressed this issue in the design.  This will be verified during breadboard testing.  The filament transformer for the series modulator is bypassed to ground with capacitors with sufficiently low reactance at audio to negate any audio losses here.  The RF final filament transformer is bypassed for RF, but not for audio.  Considering the large wire size and small number of turns here, the reactance at audio here is extremely low compared to the source impedance of the modulator driving the final.  I was more concerned with shunt capacity, which would likely attenuate the higher audio frequencies.  But testing the transformer for capacity between the secondary and the frame, as well as secondary to primary, revealed just over 100 pF, a negligible attenuator at audio frequencies.

The main concern is the capacity of the power transformer and associated devices, floating at the audio frequencies.  I expect the reactance of the filter reactors in both the positive and negative legs will alleviate attenuation here, but that needs to be verified both with measurement of the components involved, and measurement of actual performance in a breadboard configuration.   Prior to that test, a suitable drive circuit is needed, and that is undergoing test as we speak.  (The circuit diagrams show a filter reactor in only the positive leg, but I have an identical reactor to place in the negative supply leg as well.)

I agree that any point in the loop may be grounded.  The two primary goals are to minimize risk of danger by minimizing the areas where high voltage is exposed, and provide a simple and reliable method of driving the modulator at audio frequencies with a stable method of controlling the bias and operating point.  Having the cathode of the modulator at ground potential means the grid voltage may be controlled independently, and does not have to deal with the simultaneous variations of the cathode voltage, as in a traditional cathode follower.  If the plate voltage was fixed (with a grounded negative power supply leg) then the grid drive for the modulator would need to be designed considering the cathode potential would be varying as will the full modulator audio output, thus a grid swing of greater than the output variations would be required.  With the modulator cathode grounded, a small variation in grid voltage provides full modulation, no grid current is required, and the modulator still performs as a cathode follower.   Instead of powering the driver stage from the main plate supply, a lower voltage split supply, providing positive and negative voltages with respect to chassis ground, will be used to power the driver triode or tetrode.

[W3SLK]

W8HKH said:

Being that my xyl is now my ex. my winter lab is the main living room in the front of the house, making it very comfortable for design and testing activities.

I wanted to comment on that statement! It harkened me back to the visit I had at  Phil, K2PG's QTH in Wolf Creek, NJ. Where the living room was 'occupied' by 2 Collins 20V's and a Gates BC-1T

[w8khk]

That's a lot of good info, Rick!

I wish I had that much for my 813s series modulated rig, but I lucked out and it worked FB despite.

If you build this big rig and later decide the modulator is just too much heat except in that narrow band between November and March, then it would be an easy task to change it over to a high efficiency PWM transmitter.  All the floating stuff is there.  Just add another 3CX-2500F3 to the final and you're Mr. BIG!   

But until then, good luck with the experiments.

The 304TL using lower HV is a nice feature. With 5KV you will have plenty of headroom.

T

Tom, I am glad you have been successful in developing your 813 series mod rig.  That is a rather large step above the original idea of a couple 4D32s modded by sweep tubes!  

We have made considerable progress testing with the 810, 833A, and 250-TH, and it appears we may have a circuit configuration that will fly with a very simple drive circuit for the modulator.  Since no grid current is required, and only a small voltage swing is needed, a single 6L6 may suffice.  If we decide to run grid current in the modulator, then the GFZ board can be plugged in.  

Another benefit of all this testing is that my entire stash of the above-listed tube types have been verified operational.  Many of the tubes are 80 or so years old, yet NONE exhibited any issues related to leaking seals, NONE were gassy, and ALL had intact filaments with sufficient emission capacity!  A good number of them are WW-II surplus.  

However, one NOS 4-65A filament went up in a cloud of white precip on the inside of the glass, but I believe this one was the subject of careless impact damage on the plate seal, as it was the only one opened for inspection before I received them.  I opened two factory-sealed boxes, and those tubes tested perfectly; the other six bottles have not seen the light of day since leaving the Eimac factory in 1957.

I really think the pair of 304-TLs in the final will be sufficient munky swing for my desires.  As for class-E, I need to take the time to finish up my 8-pill QIX sand-state rig, as I now have all the components - heat sinks purchased at the Montgomery AL fest last month completed the inventory for that fun project.

If I am successful in safely squelching the air blast QRM from the 3CX2500F3 bottle, then I may turn my thoughts to a linear based upon a 4CX3000 that is currently weighting on my surplus tube shelf.  That would make an extremely clean linear which would run very conservatively.  But only if it can be done very quietly with regard to the cooling airflow.

[w8khk]

A bit more development work has been done on the Frankenstein Widowmaker series modulator design.  As the test voltages go up, the test personnel are religiously practicing social distancing from the mass of clip leads stolen, er, actually borrowed from the QIX shack.

Tests of the driver circuit have been performed with a 6V6, 6Y6, and 5881  (6L6GC) in both triode and tetrode configurations.  The best linearity and overall gain was experienced with the 5881 in tetrode mode.  

Early tests included an 810, 813, 833A, and 304-TL.  A 5881 driver with a single 810 modulator was evaluated, then a pair in parallel, and finally a quad of 810s, progressively reducing the load resistance from 20K, to 10K, then to 5K, while increasing the power supply voltage from 1000 volts up to just over 3000 volts.  The target is to modulate a single, or pair of 8000 triodes to about 200 watts DC input, 1000 volts at carrier, 200 milliamperes, for a modulating impedance of 5000 ohms as in the dummy load resistor tests above.  While creating the graph data, the bias was adjusted to increase the measured plate current in increments of 40 milliamperes, then the drive grid and 810 modulator grid voltages were recorded.  No component failures were experienced, even though most of the power tubes are almost 80 years old!

The 5881 driver plate is directly connected to the 810 grids, pulled up to positive 300 volts via 57,000 ohm 10 watt resistor.  The 5881 cathode is tied to negative 300 volts via a 3,500 ohm resistor, and the 5881 screen is grounded to the supply return.  (The test supply for the 5881 driver is a Heathkit HP-23 HV section, powered by a variac, such that the 800 volt supply is reduced to 600 volts, and the center between the two HV caps is at ground reference, as is the driver screen.  (Chassis is hot at negative 300 volts.)

Bias to the control grid is provided by a separate 100 volt supply via a 250,000 ohm pot, with the positive side of the supply referenced to system ground.  Plate supply for the 810s is a UTC S-48, powered by a 120 volt Variac, with a Far-Circuits PCB consisting of 20 each 3 amp 1000 PIV diodes in a bridge, with twelve equalized  500 uF 400 volt electrolytic capacitors.

Attached is a photo of the test setup with the pair of 810 tubes.  Also attached is an Excel graph, showing plate current across the top X axis, and negative grid voltage on the left Y axis.  Both the 810 grid bias (red) and the 5881 driver grid bias (blue) are included in the graph.  Many variations and adjustments were made to arrive at the current configuration, which appears to be rather linear.  All tests thus far are static.  The next step is to set the bias around -12 to -13 volts, and inject sine and triangle waves into the driver stage, and review the output waveform across the 5000 ohm load resistor, producing a modulated HV supply at simulated 200 watts DC input.  It appears that ten volts peak-to-peak (3.5 volts RMS) input will provide 100% modulation.  If this is coupled via a 600 ohm matching transformer, only 20 milliwatts are required to drive the 5881 grid.  If a high-impedance source is used, no power is required.   Assuming the linearity test results are acceptable, it will be time to start punching some chassis holes.  

The current plan is to run a quad of 810s at around 2000 volts at 50 milliamperes each, dissipating 100 watts each steady state, albeit lower dissipation with modulation.  This should provide around 1000 volts at 200 milliamperes to either one or two 8000 triodes in the final amplifier.  With 3000 to 3500 volts from the power supply, positive peaks over 100 percent should be possible, but not really required with reasonable audio processing.

The bias adjustment pot can be used to reduce plate voltage for tuning, and to adjust the operating point, along with output loading, to arrive at the desired RF power output.  At first glance, it may seem overkill to use a quad of 810s to modulate a 200 watt input AM rig.  But spreading the dissipation across four tubes will allow them all to run very conservatively, and potentially eliminate the need for a cooling fan.  With sufficient ventilation, both the final and modulator should be happy with convection cooling.  The only iron in the rig will be the plate transformer and a pair of filter reactors for audio isolation.  With 3000 volts from the supply, the 810s will normally see only 2000 volts plate to cathode, and this is within the ratings for the tube as it is used in plate modulated service.  Under full modulation, the demand on each modulator tube is significantly reduced.  Dynamic performance results should be available in the near future.  Assuming this rig is successful, the next step will be the 3CX2500F3 modulating a pair of Eimac 304-TLs.

[w9jsw]

Can you provide an exact schematic at your current test point? I think the 810/8000 setup may be a nice intermediate transmitter by itself as you move on to loftier goals.

BTW, on the Arduino front, I have started on the band control software and have arrived at a nice bezel for the display.

John

[fg5fc]

Hello W8KHK
i'am a newbie, but i readed a paper written by some french ham radio guy, it seem this serie modulation is
much more easier to do, what do you think about the diagram joint?
John

[w8khk]

Hello W8KHK
i'am a newbie, but i readed a paper written by some french ham radio guy, it seem this serie modulation is
much more easier to do, what do you think about the diagram joint?
John

Hello John,

That seems like a very straightforward design, including microphone and line preamplifiers, but of course no processing or equalization.  Looking at the specifications for the IRF840 FET as the series modulator, I estimate it could handle an RF carrier input power of around 50 watts.  I am not able to translate much of the French text, so I leave further comment to those more fluent in your native language.

73, Rick

[w8khk]

Can you provide an exact schematic at your current test point? I think the 810/8000 setup may be a nice intermediate transmitter by itself as you move on to loftier goals.

BTW, on the Arduino front, I have started on the band control software and have arrived at a nice bezel for the display.

John

John, after reading what you and Brett (N2DTS) have posted about the Hermes Lite 2, I am seriously considering adding one to my toy box.  The Arduino is a great swiss army knife for many applications, ham-related and otherwise!

I am attaching a rough sketch (scan to .PDF) of the planned transmitter.  Please keep in mind that only static testing of the modulator section has been completed, and there is a good bit of work to be done before the circuit may be duplicated with confidence.  Many of the components in the test setup are what I had on-hand, and are not necessarily appropriate for the final build.  

I have been simulating the RF deck with a resistive dummy load of 5000 ohms, 400 watts dissipation.  After dynamic tests, including distortion measurement are completed, I may commit this rig to a rack and panel assembly.  It will be a less challenging build than the big rig, since there is no need for forced air circulation and the associated noise abatement challenges.  I did notice one error in the drawing; the instructions to set the bias pot for -2000 volts at the cathode of the RF tube should instead read -1000 volts.

[fg5fc]

Hello Rick, thanks alot for your quick answer i'm very interested by the project of serie modulation, i have been very active on ten meters 15 years ago with a rice box, ft101E, i made alot of DX with only 20 watts out, i rembered also a lot of QSO on 20meters with Bill W8VYZ and also Doug VE4BX.(NOW both SK)
I have more or les all parts to built medium power transmitter , but the modulation transformer is the big probem. i'am waitting for your futher test with serie modulation.

73'S
John

[w9jsw]

Thanks, Rick.

Perhaps we should have a QSO soon so you can hear the HL2 as well. Would be very happy to explain more of its features. I am very happy with mine.

On the schematic - my intent is to start looking thru my junk box to start assembling a pile of parts in the hope that your experiments yield a solid design. I don't plan to proceed to a build just yet.

Will it run with a pair of 810s? Those see to be a bit pricey on Ebay at least. What about 572's?

I usually take hand drawn designs and input them into Kicad so I can have a nice rendition of the design for future builds. I always have a schematic that exactly mirrors my homebuild systems as my memory of things fades with time. No telling when you will have to go back to repair a unit and I don't want to scratch my head and try to decide how I built it.

John

[KD6VXI]

I'll third the hermes lite 2.

You won't find better for the money.  And PreDistortion!

--Shane
KD6VXI

[w9jsw]

Rick,

Parts sourcing - HP23 trans. It has a 282Vac HV side and a 95V on the bottom of the LV side. If you set the input to 75Vac, that equates to secondary's of 170Vac and 57Vac, right? How can you get +/-300V from 170Vac?

Thinking I can use a small antek with 2 230V secondaries and a couple of 6.3 V secondaries and not need a variac.

https://www.antekinc.com/as-1t230-100va-230v-transformer/

For the HV, are you using the 2500v taps and cutting it back to 3kv with the variac? I have the S-49 trans so I guess I can use the 3000V taps and a full wave with CT instead of a bridge? Are the dual chokes necessary in that case?



John

[w8khk]

Hello John,

It is a bit early to begin sourcing components.  Static testing seems to reveal a relatively linear transfer between the input voltage and output power, but until dynamic testing is done, with a fixed resistive load, and the varying load of a final amplifier, I think it might be counter-productive to attempt to collect all the pieces to duplicate the initial test configuration. 

I made a quick attempt to start the dynamic testing yesterday, but tabled that until a more reasonable layout is produced.  As expected, with all the clip-lead haywire, unintended feedback produced instabilities, making it impossible to perform any meaningful testing with sine, square, or triangle waveforms.  The testing accomplished thus far indicate that it may be possible to create a reasonable modulator with the tubes at hand, and they seem to validate the tubes appear to be in usable condition.

The plus and minus 300 volt supply was chosen because it was convenient, but the voltage and current capacity is way overkill for the driver stage.  Variacs were used in the initial tests such that it was possible to start with a lower voltage, and raise it incrementally to see the effects of linearity and available headroom with the circuit.  If this driver stage is used in the final build, the current requirements will be miniscule, and the voltage requirements are as yet unknown.  The split power supply was used to allow the driver to produce the negative swing required by the 810 grid in the cathode follower configuration.  Screen connected to the ground point simply provided a fixed 300 volt screen potential without the need for a separate screen supply.  This may not even be close to what is needed for a low-distortion driver.

To answer your question about voltages available from the Heath HP-23 transformer, consider the fact that there are only a couple milliamperes load on the supply, allowing the filter capacitors to charge to peak value.  Looking at the variac again, the input rating was 115 volts, connected to a 126 volt line input.  The output knob pointing to 75 volts to the Heath transformer primary was obviously higher than the 75 volts annotated in the schematic.  As I said, this is a rough representation of the overall circuit, written quickly from memory.  Since the secondary is not under load, the output voltage would probably be higher than spec even with the expected primary voltage, therefore the 300 volts on each side of the doubler is not unexpected.  When testing, I metered the voltages applied to the various tube circuits, and did not concern myself with the voltages provided by the variacs. 

When I get everything mounted on a chassis with proper wiring techniques, I will probably include electronic voltage regulation for the driver circuits, thus enabling all the parameters to be varied while observing performance.  I will also make the design open to trying various other driver tubes.   Once the dust settles, it will be rather easy to define what components are needed to duplicate the performance, but that is not possible at the present time.

I expect to table this project for a few short weeks until after the AM Rally in February.   I have quite a bit of work ahead of me to organize the shack for that event, as a lot has changed since I last had an antenna in the air free of the trees.  I also want to take some time to look at the Hermes Lite 2 while it is available for orders.  Thanks for the pointer to the Hermes forum!

[w9jsw]

Sure thing. Just got my curiosity going. Will wait to see how this develops. I also need to get ready for the Rally and the inauguration of the 813 rig.

John

[KF7JAF]

Rick,

What are you using for a filament transformer on the finals? It should be insulated to several KV, if I recall correctly.

Dave
KF7JAF

[w8khk]

Rick,

What are you using for a filament transformer on the finals? It should be insulated to several KV, if I recall correctly.

Dave
KF7JAF

The initial tests of the RF stage used a pair of filament transformers from a Temco broadcast transmitter, originally powering a pair of 872 mercury vapor rectifiers.  These tubes required 5 volts at 10 amperes.   With the secondaries of two in series, 10 volts at 10 amperes was sufficient to light a pair of 8000 tubes, which required 10 volts at 4.5 amperes each.  The 872 filament transformers are insulated to withstand more than 10,000 volts.  

For the final rig, I plan to wind a secondary of heavily-insulated wire around a 120 volt variac core, thus providing not only the required high-voltage insulation, but also the desirable low capacity between primary and secondary, thus reducing the attenuation of the higher audio modulating frequency energy.  I have used this technique in the past for much larger tubes, requiring over 25 amperes filament current.  It is much easier to wind a secondary around a variac core, with sufficient insulation, than it is to make a transformer with E-I laminations for the same high voltage isolation.  Using multiple secondary windings for multiple tubes enables metering of individual plate currents.

Some interesting benefits of powering the final stage with a negative power supply are the elimination of the high voltage shunt feed plate choke, as series feed through the plate tank is provided through the typical "safety" RF choke at the RF output, and also the option of placing the plate current meters between this choke and ground, thus eliminating the high voltage danger at the meters.  The need for the high voltage, high RF current plate blocking capacitor is also eliminated.

[w8khk]

Sure thing. Just got my curiosity going. Will wait to see how this develops. I also need to get ready for the Rally and the inauguration of the 813 rig.

John

I CAN'T WAIT to hear that 813 rig.  Only four weeks until the rally!