a 600 V B+ high power AM transmitter

[w4bfs]

There are plenty of high voltage designs ( say 1200 Volt or higher ) and plenty of lower voltage designs ( say 160 Volt or lower ).
Something Carl mentioned in another thread got me started thinking about using multiple tubes in a 600 Volt design and just seeing how close this could get to a 375 Watt rig.  I've taken plenty of pot-shots at other hams ideas, now its your turn.

Suitable 600 Volt tubes are usually a 6L6 family derivative ( 6L6, 807, 6146, etc. ) or getting a bit larger  4D32 or 829b.

I have always wanted to do something with 829b tubes.  There are still a lot of them out there.  They are mostly NOS.  According to the RCA transmitting tube manual, 90 Watts out plate modulated, ICAS, 75% efficiency.

Enuff beating round the bush ... 4 829b tubes in parallel at 600 Volts B+ would need up to 800 mA plate current.  This yields a 750 Ohm modulated impedance.  An Antek toroid of 300VA rating could handle this load at these voltage swings without having to worry so much about insulation breakdown because we only need 600 Volt peak swing for 100% mod.  800mA of dc current is another matter.   Finding a modified Heising reactor for this current level may be a show stopper. Let's see .... at 750 Ohm mod z we need only 3 Henrys of inductance in the reactor to be good down to about 50 Hz.  This opens a lot of possibilities  for using inductors or other purpose transformers.  One of the first ones that come to mind is a mot   ( microwave oven transformer) with some experimenting called for here .

600 Volts at 800mA calls for unusual supply transformer specs.   I have grown fond of capacitor input pi filter designs.  This may not be the best approach for this design due to the higher load current.  

Tell me what you are thinking ....

[KA2DZT]

Problem with the idea is that the plate load would be very low.  Probably be able to drive the antenna directly from the plates.

Fred

[w4bfs]

Problem with the idea is that the plate load would be very low.  Probably be able to drive the antenna directly from the plates.

Fred

actually not far from the truth, if using open wire feeders ... a push pull amp would do that better ...

since class C the rf load impedance would be approximately 1/2 the modulating impedance or 375 Ohms ... a pi net would work with appropriate components .... might do better with a tapping arrangement  ... 73   John

[WQ9E]

There are a lot of 1625 tubes out there which would also be a good low cost choice.  Going low voltage/high current creates some issues but in tradeoff it does solve some other issues.

Take a look at what WRL did with the output network for the Galaxy 2000 amp they brought out back in 1967 and it may give you some ideas.  It used 10 6HF5 tubes in parallel.

[kb3ouk]

I was going to suggest third harmonic resonators in the plate and cathode circuit (or just in the plate circuit) to boost the efficiency, which might work. Using 4 4D32s at 600v 880ma 75% efficiency, the power output is close to 400 watts. Using 3 4D32s at 600v 660ma with the 3rd harmonic circuits in both the cathode and plate circuits would give about 360 watts at 90% efficiency.

[WD5JKO]

In the Jul/Aug 2006 issue of QEX there is a 4 sweep tube push pull parallel amplifier using resonant input and output circuits along with broad band toroidal impedance transformers. The article name was, "An Innovative 2-KW Linear Tube Amplifier". There was a thread about this here a few months back, and a scanned copy was linked to an external website. This was a class F amplifier that was ~ 83% efficient at 1KW out. The innovative part was the power supply was modulated, and this allowed SSB linear amplification (or AM) when the RF tubes stayed in switchmode class F.

There might be some circuit details in that article worth looking at.

Jim
WD5JKO

[KM1H]

Other possibilities are a pair of 4CX250B's or clones at about 750V, tubes and sockets are everywhere.

The 8072 is often very cheap, even NIB since the Motorola/GE UHF FM rigs and repeaters are no longer in use. Make an anode cooler from a scrapped 3-500Z, 4-400A, etc and use a muffin fan, increase the screen voltage to get 400W out of a pair. Get the complete rigs and you get the BeO block and socket and then use a bigger heat sink and muffin fan. Moto ran them at 100W+ out on 460 MHz.

I ran a test of a pair in a NCL-2000 with 3-500Z coolers and they were happy at 600W CW/SSB out in the CW position which runs 1800V plate and 300 screen. Cut the coolers off 8122's (I believe they are resistance welded) is another possibility as that is the only difference between the two. Then use a muffin fan down below for a minimal and quiet air flow.

The transverse cooled 8121 is another in the family that has been showing up cheap, they were used in lots of low power TV use and demand is nil. Their cooler is an aluminum press fit, a bit of heat should loosen it up if its a scrap tube and put it on a 8072. The creative will find a pair an interesting possibility for a MLA-2500 swap.

Another cooler with potential is off the 3CX100/7289 family, its held on with a screw and looks like an easy swap to a 8072, but I havent done the measurements.

Carl

[Steve - K4HX]

W1AJW used to run a rig with five or six 6146s in the final. It can be done.

[N0WEK]

There was a junk box transmitter in one of the magazines back in the 1960s that used a voltage doubler or tripler for a power supply and a whole gaggle of 1625s.

They used an old TV chassis and didn't worry about the extra holes. It's upper frequency limit was about 40 meters because of the paralleled inter-electrode capacitance. Apparently it worked pretty well but it wasn't pretty.

Anybody remember it? I read Popular Electronics and QST back in the day.

[WU2D]

Back in 1980 I saw a beautiful full legal limit plus homebrew amplifier with 3 4CX250's on their sides in a 2U rack built by a ham that worked at Motorola. At the time this was an impressive amplifier for someone to homebrew. 

I have used 829B's for RF and even audio amplifiers. They do take some drive but are roughly equivalent to two 6L6's.

Mike WU2D

[VE3AJM]

I recall that Nick W8VUR :victory under Roosevelt:,used a HB AM tx using a whole bunch of 1625s in parallel in the final. Worked FB on 40m.

Al

[Opcom]

you may find something like a 3H 1A inductor in an old 3-phase equipment where low inductances are more common.

[w5omr]

Brett/N2DTS built a transmitter that runs (3) 4D32's.  300 or so watts output, modulated by (if I recall correctly) a pair of 811's in push-pull Class B.  I saw the thing, one of the last times I was in Joisey.

[steve_qix]

I have done this.  I used 6   6DQ5 tubes in parallel.  The rig ran a lot of power, although I can't remember exactly how much a this moment - I seem to recall something like 600V at 1 amp.  It used a PI section output, appropriately designed for the impedances in question.

Anyway, the modulator was the interesting part.  I used a solid state pulse width modulator made with high voltage IGBTs.  This was quite a while ago.  There are even higher voltage IGBTs available now, so the design will be even easier.

Here are some pictures:

[steve_qix]

Oh, the power supply voltage for the above mentioned transmitter was around 1800VDC at around 340mA.  You could get away with 1500V (at around 410mA) and still modulate up to 150% positive.  With 1800V and 600V at carrier on the RF amplifier, 200% positive modulation would be possible.

[kb3ouk]

PWM sounds like it would be the most practical solution for a low Z amplifier like that, since you would definitely have issues finding mod iron to match such a low impedence load.

[Opcom]

A thought to your foundations for a TX. -Find one of those old tube-regulated power supplies. Or a pair of them. They have what you may want.

One rated for 0-300V @ 500mA  would maybe have eight 6CD6's, or the like. One for 0-500V/300mA may have six 807's. Either way, there's a chassis and a hefty CCS rated transformer with more voltage than the unit makes, sometimes 200V more, depends. A second unit would make a foundation for a heck of a modulator.

or even a series modulator (done that for QRP, easy to mod one for a series unit by removing a few caps). Two such supplies also may be stacked in series, done it lots of times, they are probably rated for some decent hipot, so elevating one by the HV of the other should be OK.  link  http://www.montagar.com/~patj/amxmtr_dc_classamod.html


Each of these type things usually has a couple of nice meters too. I hate to say chop up old power supplies, but unless I'm wrong, not many people use them any more. I do, anyone else??

Just a suggestion.

[w4bfs]

Thanks to all who have responded ... it would take a long time to properly answer each .... maybe it is better to acknowledge that everyone is coming from a different place, stock in the junk box, etc.

All suggestions have their merits and every design is a series of trade-offs.

keep bringing it up .... 73 ... John