Many monoblock power amplifiers have two heatsinks, one to each side. This appeals to me: it could allow either more total heatsink area or two lower-profile heatsinks (instead of one tall one), or some combination of both.
I can see how this would be valuable for class A, or a sensible choice for bridged stereo AB. But for a straight mono AB amplifier, two opposing heatsinks would presumably mean quite some distance between push and pull output devices or output pairs... and I've read threads where having even short fly-leads between the amp's PCB and output devices was said to compromise the performance (or stability).
So of course I'm wondering: would this be a valid concern, do special measures need to be taken, or am I misunderstanding something?
Thanks, Kev.
I can see how this would be valuable for class A, or a sensible choice for bridged stereo AB. But for a straight mono AB amplifier, two opposing heatsinks would presumably mean quite some distance between push and pull output devices or output pairs... and I've read threads where having even short fly-leads between the amp's PCB and output devices was said to compromise the performance (or stability).
So of course I'm wondering: would this be a valid concern, do special measures need to be taken, or am I misunderstanding something?
Thanks, Kev.
The base collector & emitter lines are not too RF sensitive. The heat sense lines that feed back to the pre-drivers are high impedance and could cause oscillation at that high gain point. My obsolete ST-120 with the output transistors on a separate heat sink from the driver board, I put a 6 turn coil series the flying heat sense wire back to the VAS. On the VAS end, not out on the heat sink. Wasn't needed OEM because 1966 RCA homotaxial OT's were so slow. Ft 200 khz.
An audio amplifier with modern transistors needs to be designed and laid out as a RF circuit.
If you are building an existing design, I suggest sticking to the existing layout.
Ed
If you are building an existing design, I suggest sticking to the existing layout.
Ed
Class AB is more tolerant than Class D.
I sold a Class D amp a few years back.
Buyer complained it didnt work.
I asked for a picture of his setup and he had removed mosfets from PCB and put them on a heat sink 6 inches away !
I got him to put them back on the pcb and it worked fine.
Having said that keeping all wires as short as possible is the best way.
I sold a Class D amp a few years back.
Buyer complained it didnt work.
I asked for a picture of his setup and he had removed mosfets from PCB and put them on a heat sink 6 inches away !
I got him to put them back on the pcb and it worked fine.
Having said that keeping all wires as short as possible is the best way.
Thank you everyone! So it may or may not be a problem, but it sounds like I'm justified in at least being cautious, then. That is very useful to know.
I'll almost certainly be using a proven design of amp (maybe even PCB) and it probably isn't worth risking that by inexpertly adapting for two heatsinks. There are plenty of case design options for using just the one heatsink, after all.
Thanks again,
Kev
I'll almost certainly be using a proven design of amp (maybe even PCB) and it probably isn't worth risking that by inexpertly adapting for two heatsinks. There are plenty of case design options for using just the one heatsink, after all.
Thanks again,
Kev
They usually specify 1nH of inductance for 1mm of track/wire.
So with long wires your basically putting an inductor in series with the wires.
So 6 inches is 150nH of inductance per wire.
So with long wires your basically putting an inductor in series with the wires.
So 6 inches is 150nH of inductance per wire.
Gosh yes, that is quite... busy. It isn't quite the style I was intending, but I'm glad to hear it works for you. Though I suppose that doesn't 'necessarily' mean it would work so well for me; a different design of amp and different circumstances could presumably mean a different result.
Thank you, that makes sense. I can see how it might be possible to at least factor that in to the design now, though with my level of (in)expertise it would probably still be safer not to go there. I suppose there could be additional risks, such as interference picked up by loops, unless it was done quite competently, too.
Cheers, Kev
You have to define “perfectly”. Getting it to work, sure. Getting every last ounce of performance out of the components/topology, probably not.
I don’t try to get amps to work DC to 2 MHz either, even if it would be theoretically “possible”. I have no beef with 4MHz fT outputs, because I’ve found ”better” ways of minimizing crossover distortion. Some of which are considered sacrilege.