Opinions? Electronic bias for linear amplifiers - improve idle efficiency

[K1JJ]

I came across this idea from an MFJ product. The board wires into a linear amplifer to bias the amplifier into cut-off between words. (CW and SSB) It is on two small circuit boards that have three wires and ground. Two wires go in series with the amplifier's cathode (fil CT) to provide shut-off bias and the other lead connects to the amplifer's RF input to sample RF in.

Here's the description:

"Add-on Electronic Bias System . . . $39 95 for AL-80A/82/1200/1500 Amps EBS-1, $39.95. EBS reduces heat by reducing quiescent current to low values during periods when exciter output falls below 100 mW. Features remote detector board. EBS-1H, $49.95, for other amplifiers."


This unit is used mainly on CW and SSB, though, on AM there's no reason why the bias cannot be increased between words to provide a lower class of operation for better carrier efficiency.

My question is: On ssb, do you think the time constant, however  short it can be, is fast enough to prevent turn-on splatter of the first leading edge syllable?  I know that some older ssb rigs have slow attack ALC's that cause splatter because of the instantaneous flat topping effect due to slow ALC action on the internal final.

I was thinking of using an 11N90 FET for the cathode device and use a fast-acting diode to sample some incoming RF. Maybe it could work fast enuff if the FET is turned on hard. Blower noise might be a problem to keep the signal below 100mw, so there will be a need for a threshold level.

I spend at least 50% of my time on ssb working DX, so this could save a bundle in heat between syllables. I like the idea, but don't see it employed on mainstream amplifiers, so I'm suspicious. It reminds me a little of the G4SEK? scheme that does a similar thang.

T

[AB2EZ]

Tom

Assuming that you are using VOX on SSB, you are already are doing a good job of cutting back on the power consumption between pauses. Other than an opportunity to try something different (or in MFJ's case, to sell something additional) it is not clear to me what the value of this really is.

Happy Holidays!

Stu

[WQ9E]

I built one into my homebrew amp (not the MFJ board) and it definitely reduces the heat when active although I have never run a measurement as to the exact amount.  I put a defeat switch on the circuit but I have never noticed any additional distortion on SSB when it is active and it isn't apparent on the scope nor has it been mentioned.  In the winter time the additional heat is welcome and the 4CX800 tubes are running well below their rated dissipation and temperature anyway. 

Unless your tubes and/or power supply are running near their thermal limit or you have suddenly got the green religion I wouldn't bother.  I put it in mostly as an experiment since it was easy to include during the early part of the design.  It has been about 7 years and I can't remember now where I "borrowed" the circuit idea.

[w4bfs]

hi Rodger ... how's it going?

[K1JJ]

Thanks for the info, Rodger and Stu -

Stu, I don't run VOX on SSB. There are so many loud relays clacking in the shack during TR that it would sound like a jackhammer on the air... :-)  In fact, I have a key-up delay in my audio of about 100 ms to mask the relays.  So it's always PTT and there would be a savings in heat if I could kill the amp idle during long pauses in speech, in addition to the fast syllable pauses, assuming the circuit is designed to be that fast. The attack could be VERY fast, while the decay slow, if desired, too.

Rodger - To run a test for splatter when using the board, tune yourself in using a receiver that has a 12" piece of wire for an antenna. Use a dummyload for the transmitter. Then tune off frequency by 3.2 kc and listen to your side ssb products.  If you have a 40db over 9 signal on freq, you may hear an S9 or whatever 3.2 kc away on ssb.

Then say "pop"  (or any other hard sound) to get the amplifier bias to turn on with a short attack. See what your S-meter says each "pop."  Then shut off the electronic bias board and try it again. If you see no difference, then the board is transparent.  If it adds some additional trash up the band on the short attack transistion to idle from class C to class AB, then you will know it is causing your band neighbors some heat...  

If there is a problem, it will sound sporadic and not like a continuous splatter when an amp has bad IMD all the time. You follow what I'm saying, I'm sure...


T

[AB2EZ]

Tom

Okay...

So this would be the equivalent of VOX, except you wouldn't open up all of the relays to go from transmit to receive. You would just bias the final(s) off instead.

Stu

[KM1H]

Ive heard it is possible but Ive yet to hear an EBS that doesnt add splatter. The old Alpha 77 was about the worst. Its pretty hard to control a RF signal at an audio rate withough a sequencer and thats just the beginning of the problem. A VOX or foot switch is rather foolproof.

Roger, publish that circuit!

Carl

[WQ9E]

Tom,
I will have to see what it sounds like in a critical listening test.  The FT-1000MP/amp only gets used in contests and I haven't had time to be serious about that in awhile.

John, things are going fine here.  I just spent an hour on the tractor plowing the driveway and part of the county road so we can get out to the main road. 

Carl,
I will test the results and then dig through my notes on the amp because I am sure I documented the circuit.  I mostly put it in there for CW and I remember at the time seeing some notes on how bad the Alpha circuit was about creating distortion.  I thought about at the time using a length of coiled coax between the exciter and the amp.  The "trigger" circuit for the EBS would be located at the exciter and the coax line would provide a delay similar to that used in the vertical delay line  of a scope so that the trigger event can be viewed.  I think this would be the best way to allow the EBS to stabilize operating conditions before the RF hits the grid.

[K1JJ]

Tom
Okay...
So this would be the equivalent of VOX, except you wouldn't open up all of the relays to go from transmit to receive. You would just bias the final(s) off instead.
Stu

Yes. In addition, the pause between syllables that is too short for VOX could be covered with a FAST acting bias circuit - but maybe that is fantasy due to the attack crud just discussed.  BTW, hope you're having a nice holiday season, Stu. I haven't forgotten the generous gestures you made with the digital RX filter and SS audio amp last year. Tnx again, OM!


Rodger, you made a good point. If there were a tiny delay enuff to let the amp come up and stabilize, there shud be no problem with getting clean operation.  Though I think the microseconds in a short piece of coax wud  not be long enuff unless the attack time was truly stellar. As Carl said, these audio to RF interfaces are hard to get right without generating crud. I'm pretty sure it's all a matter of fast attack times and providing a slight delay to cover up the attack key-up limbo time.

I wonder how fast an RF generated signal into a diode, driving a chain of FETS could key an 11N90? Then there is transcient times for the tube to come up due to internal circuit C effects. Maybe the between-syllables requirement is axing for too much and a decay-hold of sorts will be needed too..

T

[WQ9E]

Tom,

I think trying to use an EBS circuit is probably pushing it on SSB and it is better suited for CW.  It might be the perfect circuit for some of the old high power sweep tube rigs since a little more distortion will probably go unnoticed among the existing issues and it might stretch the tube life   I have a TS-511S (a rare sweep tube final Kenwood) awaiting attention, maybe it will become a candidate for EBS V. 2.0

I think a combination of a delay before excitation circuit, very fast switching, and slow bias decay would result in a usable circuit.

[WA1GFZ]

Tom,
I would drive the FET with an IXDD414 to make it react very fast. Then you just need a 5 volt control signal. This could come from the audio chain rectified into say a voltage window comparator (LM239) to drive the IXDD414. The LM239 is a quad so you could use a section or two the do the drop out delay time constant. This way you take advantage of the delay through the rig to have the FET saturated by the time the RF arrives at the final.
Sounds like a lot of work to save a little power. keep the heat in the house to keep you warm.

[Steve - K4HX]

Just go with the Taylor hybrid Tom. It never fails.

[K1JJ]

Yep, the Taylor hybrid can do anything. I may do a TV info-commercial and sell a bunch to the housewives for $19.95 plus S/H. You interested?.


Frank, that's an interesting approach - to use the low level audio as a triggering for the bias FET, driven by a IXDD414. I'll have to give it some thought. It will make a good project into the winter.  If anything, it will let me run the blower lower for less heat. The HUXman gots some big Taylor hybrid plans and when he starts bitting my tail into Eu I may do something.

T

[WD8BIL]

"Add-on Electronic Bias System ....

WHY?