I'm trying to repair the Class D amp out of an Acoustic Research powered subwoofer which I'd managed to get running again but had a mishap with a scope probe that blew my isolation transformer's fuse and may have damaged some parts in the process. I have managed to identify one of the NPN and one of the PNP final transistors as having failed short (IOW, ~266VDC in communication with each other @_@):
When I first got this in to repair last year, one of those transistors had already gone and I sourced and installed a replacement, but this break/fix go-round the replacement was one of the shorted ones. It looks like reliable sources of the original models (2SC3281 and 2SA1302) can't be had anymore but the intertubes are telling me that I can substitute 2SC5200N and 2SA1943N and I've ordered three of each. I've seen posts about output transistor matching in Class D amps but I need to be clear on something: what is more important to match, Q13 to Q14 and Q17 to Q16 or all four to each other? I've got one of those inexpensive universal testers with the little color LCD screen that will kick out Hfe and a couple of other numbers; will that be sufficient to help me find the closest matches?
When I first got this in to repair last year, one of those transistors had already gone and I sourced and installed a replacement, but this break/fix go-round the replacement was one of the shorted ones. It looks like reliable sources of the original models (2SC3281 and 2SA1302) can't be had anymore but the intertubes are telling me that I can substitute 2SC5200N and 2SA1943N and I've ordered three of each. I've seen posts about output transistor matching in Class D amps but I need to be clear on something: what is more important to match, Q13 to Q14 and Q17 to Q16 or all four to each other? I've got one of those inexpensive universal testers with the little color LCD screen that will kick out Hfe and a couple of other numbers; will that be sufficient to help me find the closest matches?
Last edited:
That’s not a Class D amplifier. It’s a good old Class AB.
Q14-Q17 should all be as close as possible in terms of hFE. The same with Q13 and Q18. They don’t need to be matched between the two groups. Keeping the two output NPNs and PNPs equal to their peer is important, since you avoid having one work harder than the other. The degeneration resistors take care of some of that, but it’s still a good idea to keep them matched.
Q14-Q17 should all be as close as possible in terms of hFE. The same with Q13 and Q18. They don’t need to be matched between the two groups. Keeping the two output NPNs and PNPs equal to their peer is important, since you avoid having one work harder than the other. The degeneration resistors take care of some of that, but it’s still a good idea to keep them matched.
I'm no Class D expert but let me just say that it's throwing around 1MHz signals inside so let me provide the rest of the schematic and you tell me:
This is from an Acoustic Research AR1 loudspeaker; this is just the amplifier board. There is a second board that feeds this one; it takes speaker-level signal and appears to filter it down in analog (I don't have a schematic for that board but I was able to get it working again). One reason why I was probing around was because when I had the whole schmeer wired up with another power amp feeding it with low-frequency sinewave input, I was getting sinewave output with 1-MHz negative-going impulses superimposed on it. I think maybe I should check on the condition of R28 and C11.
An article I found from 2000 reviewing the AR1 stated the amp was Class D; I'm not well-versed in Class D topologies so I wasn't prepared to conclude the author was wrong.
This is from an Acoustic Research AR1 loudspeaker; this is just the amplifier board. There is a second board that feeds this one; it takes speaker-level signal and appears to filter it down in analog (I don't have a schematic for that board but I was able to get it working again). One reason why I was probing around was because when I had the whole schmeer wired up with another power amp feeding it with low-frequency sinewave input, I was getting sinewave output with 1-MHz negative-going impulses superimposed on it. I think maybe I should check on the condition of R28 and C11.
An article I found from 2000 reviewing the AR1 stated the amp was Class D; I'm not well-versed in Class D topologies so I wasn't prepared to conclude the author was wrong.
Ah. You didn’t show me the full schematics. The part in your first post is definitely a Class AB amplifier, but it seems to be fed through variable supply rails (which I believe makes it a Class H). Those rails seem to be supplied from a SMPS of sorts that closely resembles a Class D amplifier. So it’s a hybrid.
The section where you believe you have fried transistor is identical to a Class AB power amp.
The section where you believe you have fried transistor is identical to a Class AB power amp.
I have to run, but the outputs of U4A and U4B should have clean square waves that are pulse width-modulated with the input. You should see the same on the sources of the MOSFETs before the filter coil.
