This idea has been around for a long time, but I haven't seen it mentioned much. This technique is mainly used for increasing efficiency on ssb, but could be used on AM in a more limited way..
Here's a simple VOX board from China for $16.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=221121494057&ssPageName=ADME:L:OC:US:3160Here's how it works... Trigger the VOX board with some audio tapped off your existing audio preamplifier output. Set the amplifier bias so that the final tube idles JUST above class B, so that it has a small amount of idle current, enough to pass fan noise. When our voice triggers the VOX board on, a relay sets the amplifier bias so that it idles at a normal current. Set the delay for about one second so that during voice pauses, the amplifer goes to sleep until another spoken phrase. So, word spoken, vox on, regular idle bias --- voice pause, one second later increased bias to put amp to sleep.
The problem with using these in the past is that most configs put the amplifier completely into cut-off (class C) during a pause and the first word would cause a splatter peak for a moment. (slow attack time) But if we let the amplifer idle just above cutoff, this will not occur.
I have been running my class A/ AB1 linear amp system this way for about a year now and find it works very FB. The difference in heat is about 1/2, considering the amplidier is asleep about 50% of the time in normal speech. The SDR anlayzer shows no sign of first word splatter since it is already idling at a "reasonable" level above class B before getting kicked up to AB1 or wherever..
I am using a homebrew board, but plan to replace it with this slick Chinese board.
As for AM use.... it COULD be used the same way, but I would idle the amp a little higher during sleep time so that the carrier does not drop off and sound like controlled carrier. Maybe 1-3dB would not be noticed but could save a lot of heat. Most amplifers need more drive when using more bias, thus the carrier drop off problem.
BIAS: The easy way to change the bias with the board relay is to use a string of standard diodes in the cathode of the final tube. Put in enough diodes to let the amplifier sleep just above cutoff and short some diodes out with the VOX relay until the amp idles at its normal current for regular operation.
Combine the VOX bias idea with the AM linear "class C carrier technique," (as discussed in another thread) and we are talking some reasonable savings in power.
T