2KW solid state linear amp using two MRFE6VP61K25H devices

[ka1tdq]

QST did an article sometime last year about a 1kw 2 meter amplifier using a single MRFE6VP61K25H device.  It was designed be used in plasma generation for the semiconductor industry for ion implantation, I guess.  A complete kit is available from Communication Concepts for $350 (minus the power suppy and heat sink materials of course).

I was thinking about building an amp that used two of these modules for 2kw out on 75 meters.  Driving power is very low and would match up great for my QRP solid state transmitter.  By adding a 2-port slitter on the RF input to the two modules and a 2-port combiner on the output should work.  I already have a massive 50 volt (or whatever I need) transformer for the power supply. 

The only design changes would be the impedance matching on the front of each module and the output of each for 75 meters instead of 2.  My .43 watt carrier transmitter should be able to hit the legal power limit at 200% positive peaks. At least that's what the numbers say on the MRFE6VP61K25H data sheet, unless I'm missing something.

Hey, I'm excited...

Jon
KA1TDQ

[W1ITT]

Was there any indication in the QST article as to what the 3rd order IMD figure is at various power levels?  Industrial RF sources need not be squeaky clean, but we don't need any more crud on the HF bands.  It's interesting that current amateur practice regards -30db IMD as acceptable, and these numbers haven't changed much since the advent of linear amplifiers when SSB  became popular. (Although some of the transceivers with TV sweep tubes in the final did well to make -25db with a tailwind.)  The HPSDR exciters run -45 db or better IMD.  One would think that 40db might be a nice target for new construction amplifiers.
The more strapping one is, the more important it is to have a clean signal, if only to keep the adjacent frequency lads from coming on frequency and complaining.

[K1JJ]

It's interesting that current amateur practice regards -30db IMD as acceptable, and these numbers haven't changed much since the advent of linear amplifiers when SSB  became popular. (Although some of the transceivers with TV sweep tubes in the final did well to make -25db with a tailwind.)  The HPSDR exciters run -45 db or better IMD.  One would think that 40db might be a nice target for new construction amplifiers.
The more strapping one is, the more important it is to have a clean signal, if only to keep the adjacent frequency lads from coming on frequency and complaining.

You said:
"The more strapping one is, the more important it is to have a clean signal, if only to keep the adjacent frequency lads from coming on frequency and complaining."


I agree 100% - well said.

Another way to look at it...  Let's say  the "average" AM or SSB station on the air is running about 375 watts pep. If we average all the guys running barefoot with the ones running a linear, then this might be a pretty close estimate.

Now, if we are running a linear at 1500 watts pep, then we are going to be 6dB louder than the average signal on the air. This means to blend into the crowd we need our amplifier to be 6dB cleaner than the average signal. If the average signal is -30dB 3rd IMD, then we need ours to be -36dB 3rd or better... to blend in.

For example, if we run  -30dB 3rd IMD amplifiers, instead of having S9 crud up the band at 375 w, we will have S9+6dB crud (at 1500 w) in the same comparison.

Bottom line is we have an increasing responsibility to improve our amplifiers as we increase power. The alternative is to develop a reputation for splatter and have guys beating us up for adjacent channel crud. I've been there and it's no fun.

But, there is nothing better than watching our own spectrum analyser and seeing -45db 3rd (and better) at 1500 watts out. I'm been there too - and it's pure joy.  My current class A / class AB1 linear system exhibits -50 to -55dB at 1500w out. I still can't believe it, but it took months of building and optimizing to achieve.


To respond to your question, Jon...   since you are a EE and work at a company, I would give the manufacturers of the MRFE6VP61K25H devices a call and insist on an actual IMD spec using a real HF amplifier. Maybe it's good enuff, I dunno.

I do know, however, that the 600 watt CCI kit using four MRF-150's cannot do much better than -22dB 3rd in linear service. That kit is a crud generator. I have one here and use it only as a class C carrier source.  I know Frank/ GFZ  agrees with these findings.  It takes a special effort using over-size ferrite cores, NFB and other techniques that are not in the Motorola manual to see > -30 dB 3rd in an external (>500 watts)  solid state linear amplifier.  The average ones on the HAM market today do about -25dB 3rd. At 1.5 KW output, -25dB 3rd borders on abominable compared to many (properly tuned and conservatively driven) tube amplifers.

T

[KF1Z]

current amateur practice may regard -30db IMD as acceptable.....

But of course FCC rules say any spurious emmisions must be -43db or better below 30mhz.
If that emmision lies "..outside the necessary bandwidth of a transmission.."

[N4LTA]

The best alternative is to Modulate a Class D or E stage with PWM or similar and not try to linear amplify an AM signal with FETs. Lots less expensive and less power hungry also.

Pat
N4LTA

[WD5JKO]

Industrial RF sources need not be squeaky clean, but we don't need any more crud on the HF bands.

This is only partly true since 13.56 mhz is allocated for industrial use, and the 2nd harmonic is CB channel 22A...who would notice!  

What is a concern with the folks using Industrial RF generators at 13.56 Mhz is that there is usually a RF matching network with very high Q...So any significant 2nd or third harmonics would impair a complete null of the SWR to 1:1.

I service some of these amplifiers, and the harmonics need to be at least -55 dbc.

Now Inter modulation products are another story since the amplifiers are not modulated. Any switching power supply ripple though creates sidebands to make an AM signal. These amplifiers need to have the sidebands at least 50db down from the carrier.

So as a result, these kinds of amplifiers can actually be cleaner in some ways then amateur transmitters.

Jim
WD5JKO

[KF1Z]

This is only partly true since 13.56 mhz is allocated for industrial use, and the 2nd harmonic is CB channel 22A...who would notice!  

27.12mhz is also an ISM frequency.

[Steve - K4HX]

Clearly the FCC does not consider emissions from IMD to be spurious. Otherwise, all the commercially available transceivers and amps with IMD figures in the -30 range (some of the newer ones even less) would not achieve FCC type acceptance.

current amateur practice may regard -30db IMD as acceptable.....

But of course FCC rules say any spurious emmisions must be -43db or better below 30mhz.
If that emmision lies "..outside the necessary bandwidth of a transmission.."

[ka1tdq]

To respond to your question, Jon...   since you are an EE and work at a company
 

Thanks for the pay raise!  Actually, no.  No paper and I'm a lowly government Comm Tech worker.  

The gain for the amplifier I described does seem high and I guess the tradeoff is distortion.  The design is for 2 meters, which has more room than HF and doesn't propagate far.  

I did build that small push-pull amp using IRF-510's which will boost a normal amp.  I haven't tested it since I removed the 1 ohm resistors from the source leads, but it should be around 15 watts carrier.  That's what I was doing with my old tube station back in Mass.  (driving a pair of 3-500's with 15 watts... the shack glowed nicely with the lights off at night... very romantic mood setting).  

I have basically a 1:1 power transformer for either 120 or 240 in.  The secondary has a center tap and additional taps for 102 volts out when I put 120 in.  So, I can configure this transformer for anything between 51 and 240 volts output.  I'm looking for a good solid state design which can take advantage of these power levels with 15 watts carrier drive.

Jon
KA1TDQ

[W1ITT]

To be clear, 3rd order IMD and harmonics are two entirely different concerns.  We can filter out harmonics easily at the output connector, but cleaning up IMD must be done at the source.  The HPSDR gurus are working towards EER, Envelope Elimination and Restoration ..a complex distortion feedback system which predistorts the drive then tinkers with the characteristics of an amplifier to create a very clean output signal.  But that's still in the "later" category.
I am working on a trade to get hold of some 4cx350J tubes.  They were made by Eimc, for a Collins government project, with the express purpose of making power cleanly.  The data sheets speak of -45db 3rd order numbers, and this without negative feedback or other magic.  They require a special socket...even though a first glance makes it look like a 4cx250 socket should work.... and the sockets are rarer than the tubes.  I'm thinking that a pair would make a lovely amplifier, but they'd have to be well matched so as to track well in parallel or some of the nice IMD performance could be eaten up.
I guess the concern for all of us is that "Bird Watts" are not the only measure of amplifier performance.  The manufacturers aren't concerning themselves much with IMD...at least not for the Amateur market...cheapskates that we are.  So it's good fun to try to advance the state of the art at home.  Tom -JJ has made some good progress in that arena.  If anyone else is working on squeaky clean amplification, I hope they will share their success story with the rest of us.

[kb3ouk]

But it doesn't matter what your 3rd order IMD performance is if your driver has higher IMD than the amplifier does. Your signal is only going to be as clean as the dirtiest source in the chain. So if you have an amp that has -45db 3rd order IMD specs, but you drive it with a transmitter that only makes -30db, your overall performance will be -30db. Tom did a real nice writeup on here a few months back about his low IMD setup, which is pretty impressive to read about.

[ka1tdq]

Well, there's an idea. I can make tube amps in my sleep. I didn't bring with me my 3000 volt / 1 amp HV transformer when I moved here to Arizona because it wouldn't fit in the car. I am willing to do an even trade of my 100 amp plus power transformer for a strapping 3000+ plus volt transformer. I don't have money for shipping though.

I still have the original homebrew plate choke from my last amp and a friend of mine back in Westfield recently gave me a 3-500 socket.

Jon
KA1TDQ

[steve_qix]

It might work just fine... Didn't look up the specs on the devices yet.

A few years back, I built a very simple but effective linear amplifier for 75 meters using 6 FQA11N90s ($4.00 each!) in a class B setup.  The amp used a tuned circuit, etc.  in the output, very much akin to a tube class B linear.  With only minimal drive, I was easily able to achieve 1500 watts  PEP.  There was a .33 ohm resistor in series with each source lead to aid in balancing between the very high gain MOSFETs.

[KF1Z]

Clearly the FCC does not consider emissions from IMD to be spurious. Otherwise, all the commercially available transceivers and amps with IMD figures in the -30 range (some of the newer ones even less) would not achieve FCC type acceptance.

current amateur practice may regard -30db IMD as acceptable.....

But of course FCC rules say any spurious emmisions must be -43db or better below 30mhz.
If that emmision lies "..outside the necessary bandwidth of a transmission.."

Right Steve, of course.
But the statement about the necessary bandwidth, COULD be a stickler... IF they wanted it to be.
Just another example of the vague language of a gov agencies rules.  :-)

[ka1tdq]

Steve, where are you hiding all these schematics?  If you have a copy of that one, I definitely want it.  I was getting all geared up for another tube amp and was ready to go with the parts in the picture, but 6 $4 fets are much better!  

[steve_qix]

Steve, where are you hiding all these schematics?  If you have a copy of that one, I definitely want it.  I was getting all geared up for another tube amp and was ready to go with the parts in the picture, but 6 $4 fets are much better!  

df

Hmm.... well, the only schematics that ever existed are in my head.  I never wrote it down - just built it, but the design is very straight-forward and simple.  The input was a 4:1 (turns ratio) transformer wound on a FB-43-1020 core, with the secondary connected to the 6 gates all in parallel.  There was a .33 ohm, 2W resistor (non wirewound) in series with each MOSFET source.   I made up a matching network to go from the driver transmitter (in this case, it was a DX-60) to the primary of the input transformer.  Something like an L network would work.

The "cold" side of the input transformer secondary went to an RF bypass capacitor (like a .1uF), and I applied a little forward bias at this point to slightly turn on the MOSFETs and avoid "crossover" distortion (where the devices are biased to cutoff for more than half of an RF cycle).

The output looked like a class E output with respect to the output RF transformers and the tuned circuit.

The power supply was a 70V supply, on a Variac.

I had it on the air for about a month on 75 meters.  Then went back to standard class E because the efficiency is SO much better!

Sorry I don't have any diagrams, but it's not so complex a circuit the one was really required at the time.

Regards,

Steve

[ka1tdq]

Yeah, I can put all that together.  I'll just use 50 volts DC unregulated and use a low-to-high impedance matching L network for the input.  The rest is cookbook.  So, yeah, awesome. Thanks.

Jon
KA1TDQ

[W1ITT]

In response to KB3OUK's comment regarding drivers with less than stellar IMD numbers, I agree that a super clean amplifier won't clean up a dirty input signal.  But the point is, we now have, if we care to look for them, exciters that are nice and clean.  If you follow the progress of the HPSDR  group  (  see openhpsdr.org , and the tapr.org pages  )  you will find that their Penelope transmitter  board puts out a -45db or better IMD signal, as does their Hermes transceiver  board.  Apache Labs, operated by one of the Hermes design team sells a Hermes and Angelia board.  The HPSDR boards only put out a half watt, but their PennyLane amplifier which is specified at 15 watts out will run at -45db IMD if you throttle things back to four or five watts out.  From there, I'm looking at using the aforementioned 4CX350J tubes to make some strapping power.
The multithousand dollar big three transceivers are still stuck in the 30db IMD range except for the Yaesu rig with the low power class-A mode.  I'm not sure what the Flex Radio series radios do.  It may be clean at low levels, but then they run it through a common solid state amp and throw the good IMD out the door.
To summarize, there are clean exciters available commercially, for not a lot of money (relatively), but you have to look for them.  Our next step is clean QRO.
73 de Norm W1ITT

[K1JJ]

In response to KB3OUK's comment regarding drivers with less than stellar IMD numbers, I agree that a super clean amplifier won't clean up a dirty input signal.  But the point is, we now have, if we care to look for them, exciters that are nice and clean.  If you follow the progress of the HPSDR  group  (  see openhpsdr.org , and the tapr.org pages  )  you will find that their Penelope transmitter  board puts out a -45db or better IMD signal, as does their Hermes transceiver  board.  Apache Labs, operated by one of the Hermes design team sells a Hermes and Angelia board.  The HPSDR boards only put out a half watt, but their PennyLane amplifier which is specified at 15 watts out will run at -45db IMD if you throttle things back to four or five watts out.  From there, I'm looking at using the aforementioned 4CX350J tubes to make some strapping power.
The multithousand dollar big three transceivers are still stuck in the 30db IMD range except for the Yaesu rig with the low power class-A mode.  I'm not sure what the Flex Radio series radios do.  It may be clean at low levels, but then they run it through a common solid state amp and throw the good IMD out the door.
To summarize, there are clean exciters available commercially, for not a lot of money (relatively), but you have to look for them.  Our next step is clean QRO.
73 de Norm W1ITT

Again, your comments are right on, Norm.  

I also have an HPSDR and marvel at how clean its output is at 300mW.  It seems close to -60db 3rd at that level.

It's good to hear that their PennyLane amplifier is -45dB at a reduced 5 watts. Outstanding.    

I use the 4CX-350J as a class A driver, as you know.  Possibly the best linear tube ever made. I use a lab-based 1 watt amplifier that does -60 dB  3rd to drive the 4CX-350J.   I ask only 50 watts output from the 350J to achieve about -55dB 3rd (at 2KV)  to drive the dual AB1 8877 amplifier conservatively.    At more than 50 watts, the 350J IMD degrades down to its normal speced -45dB.

My point is that we need the driver stage to be at least 10 dB cleaner than the final amplifier IF the driver is not to effect the final too much.  The final IMD will always be something less than the driver. Even a super-clean  -90db 3rd order driver feeding a -30dB final will always generate less than -30db out of the final, probably -29.5dB or whatever.


It pays to set our IMD design standards high, cuz in the real whirl results will usually be poorer due to many factors.  I've found that there are sweet spots in tuning the overall stages of my linear chain that can swing the IMD as much as -10 up or down.  If I quickly peak my linear chain, I will see about -45dB 3rd.  If I put a 2-tone through and play around with all the various parameters, I can tweak it up to -55dB 3rd.  With presets, the -55dB can be achieved on select frequencies very quickly.

I was emailing with one of the guys doing the pre-distortion tests. It has promise, though at that point the software needed some work to make it more stable across the band as well as human engr needed. But it will be mainstream one day and used with ham amps I hope. That will finally break the -30dB 3rd ceiling that has plagued us for too long. I foresee a day when -30dB amplifers will be regarded like spark gap rigs...  Most of the friction on the band is created by splatter, no?


The time to have good IMD is when the band is very quiet, like late afternoon and everyone's S-meter is low, like S5.  If we are running QRO on 75M with good conditions and S9+50 over 100 miles away, then with a -30dB IMD signal our side crud will be S9+20 over outside our normal audio bandpass.  The point is that an extra 10 to 20dB of IMD suppression can have a tremendous effect on side splatter at these times.  Even -10dB of less crud can mean the difference between being transparent or overloading someone's receiver and taking out a weak station up the band.  

Good luck with your HPSDR / Penny experiments and please post your 4CX-350J progress!!

Tom, K1JJ

[Steve - K4HX]

Part 97 definition of spurious.

(42) Spurious emission. An emission, or frequencies outside the necessary band- width of a transmission, the level of which may be reduced without affecting the information being transmitted.

Could include IMD related products. It would have to be reckoned with these sections of Part 97:

§ 97.101
General standards.
(a) In all respects not specifically covered by FCC Rules each amateur station must be operated in accordance with good engineering and good ama- teur practice.

§ 97.307 Emission standards.
(a) No amateur station transmission shall occupy more bandwidth than nec- essary for the information rate and emission type being transmitted, in ac- cordance with good amateur practice.

(c) All spurious emissions from a station transmitter must be reduced to the greatest extent practicable. If any spurious emission, including chassis or power line radiation, causes harmful interference to the reception of an- other radio station, the licensee of the interfering amateur station is required to take steps to eliminate the inter- ference, in accordance with good engi- neering practice.


I've never seen it written, but it would appear -30 dB IMD (+/-) is considered good amateur practice. That said, amateur radio has gone backwards to a certain extent in it's "practice" with respect to IMD. Rigs from 30-40 years ago like the Collins S-Line and KWM-2 and the Kenwood TS-820/30 were specified with IMDs in the high -30s. Most new transceivers are at -30 and some even at 28 and 29!. It would appear as though the bar has been lowered almost 10 dB over the years.

Given that a receiver specifications have generally improved over the years and it seems at though more HF ops are running linear amps than 30-40 years ago, the IMD bar should be set higher, not lower.

[n4wc]

Just a clarification question.  Is the 30db IMD y'all referenced, the ARRL method of 30db below pep output? or is it the "old EIMAC method" of 30db below one of two tones?
And does anybody do the three tone IMD testing where one of the tones is moving in relation to the other fixed tones?  I think that's how it is done.
And thanks for this discussion.
  73 Bill

[W1ITT]

Bill...
In talking about IMD I always speak in the Eimac language, as does most of the rest of industry.  Perhaps I'm looking for conspiracies where there are none, but I suspect that the ARRL method which gives a 6 db credit to transmitters was a gift to the ham equipment manufacturers to make the crummy sweep tube finals look acceptable back in the sixties.
I haven't explored the three-tone method much, but it appears that it has merits.  Also, because power supplies are not uniformly stiff as a board, a dynamic method which tests IMD under various rapidly changing amplitude conditions has value.  A voice signal dynamically works the power supply differently from a two-tone test, which is essentially a steady state signal.
At this point, I'd be happy if we all could target -40db IMD (Eimac method) and then go from there.  It will be literally an order of magnitude better than the world we now live in.  We've wallowed in -30db for too many decades.  And aren't we supposed to be trying to advance the state of the radio art?

[Steve - K4HX]

Nah, it's all about producing ever higher scores in contests. Don't you read QST?   

  And aren't we supposed to be trying to advance the state of the radio art?

[ka1tdq]

Actually, before I venture into Class E stuff I should probably start with something small like a single fet rig.  To get on the air now though, I will stick with what I know. 

I'll build a single 3-500 amp with the original plan of 2 microwave transformers sitting in a mineral oil bath for heat dissipation (in the cooer in th picture). I will use the big transformer's 100 volt taps so that I can use a center tapped 6.3 volt filament transformer. 

All I need to get is a tube and electrolytics for about $200 (same as a class E rf deck, but I don't need to experiment).

[WBear2GCR]

Minor detail here. Using a cooler for the oil bath for the transformers will only cool them to the point where a thermal equilibrium is reached between the xfmrs and the oil. I hope ur just storing them in the cooler? You'd want a metal box, or a metal box with fins or pipes that go to fins for cooling. Look at how a commercial dummy load is constructed or even a pole pig.

                      _-_-

PS. why not find 3 or 4 microwave xfmrs, run them in parallel so that the current draw is dropped and so is the operating temperature - IF this is a problem? Small microwave ovens are rated at 700w, x2 = 1400watts, which is quite a bit more than the input power on a hard running 3-500 tube I think... of course microwaves are intermittent duty. But to test you can drop a big resistor of appropriate value to draw equivalent current and see how hot it gets (I guess you do want the rectifiers and caps for the most accuracy, but just the resistor is good enough to see what they do thermally at "X" watts...