can anyone offer help understanding the flow of this circuit

I can tell you what I would do. IIUC you had the amplifier powered up before and the only problem was that the output was near the + rail, is that correct?

If so, I would put it back together again and power it up. Just don't connect a speaker to it. I would measure some voltages to figure out about where the problem is on the board. Its not hard to do that way. Once you know you are very close to where the problem is, its usually a lot easier from that point to get to the final answer.
 
The schematic shows no source resistors so it can do without TO3 insulators. However the two heat-sink's should not connect to each other.
Have you checked that?
The heat sinks are separate. There's one for each channel each holds six mosfets 3 n channel and 3 p channels
 

Attachments

  • image.jpg
    image.jpg
    781.1 KB · Views: 44
  • image.jpg
    image.jpg
    593.5 KB · Views: 44
If there is any question about heatsinks just get some insulators and install them. You're wasting time fretting over it when you could just take care of the concern once and for all.

Look, I've been doing this stuff for more than half of a century. Do you want to get to the bottom of this or not?
 
Is there any way to check the driver bored now that the mosfets are out with the power on?
Maybe if you can find some other way to close the feedback loop without creating an oscillator. Do you know how to do that?

There might be some other way too, but it would likely be very touchy with the feedback loop open. That's because there is a lot of open loop gain. Of course one could change that by modifying the circuit, but is that likely to be the best approach at this point?

EDIT: Why the hesitancy? You seem to have no idea of what to do at this point. There is another tact you could take. Is it that you don't want to do the work to reassemble the thing?

Its like this, sometimes troubleshooting requires localizing the problem area. Sometimes there is not a one-step, direct route to easily finding the problem. Imagine a big system that is spread out over multiple rooms full of equipment. Do you think you open everything up, take everything apart and try to find the problem with only a DVM and your eyeballs? Probably not, unless maybe you already have a pretty good idea from experience on exactly where you need to look and for what. Sometimes you have to use other tactics, is all.
 
Last edited:
I have spent years with these Mosfet amps. I always repair and check the whole cirquit without the Mosfets installed. This is no problem and you only can blow cheap parts. When you see a clean signal at all gate resistors, you should be done and solder in the Mosfets as a last step. So leave them out by now.
I'm quite sure you have a very simple fault. The way to find it is to have both channels installed and powered (except Mosfets) and compare voltages left to right, left to right and so on... Print out the schematic and mark every point where voltages differ at the defective side. You can not miss the fault this way, simply impossible. Don't think and speculate, check!

I had a look at your PCB pictures, there is a region where things get ugly, because traces are too narrow to tell them apart. Exposed copper can grow out over time and make a tiny connections. A brass wire brush is a great tool!
Also some solder points look bad. I put it on your little experience, but it is just 5 seconds to touch them up when you look at the PCB. Don't leave that to "I do it later".

PS use something to magnify the traces, a lot of light and a high resolution camera help. You will see how relaxed x10-20 is compared to staring at the area 1:1.
 
Last edited:
Maybe if you can find some other way to close the feedback loop without creating an oscillator. Do you know how to do that?

There might be some other way too, but it would likely be very touchy with the feedback loop open. That's because there is a lot of open loop gain. Of course one could change that by modifying the circuit, but is that likely to be the best approach at this point?

EDIT: Why the hesitancy? You seem to have no idea of what to do at this point. There is another tact you could take. Is it that you don't want to do the work to reassemble the thing?

Its like this, sometimes troubleshooting requires localizing the problem area. Sometimes there is not a one-step, direct route to easily finding the problem. Imagine a big system that is spread out over multiple rooms full of equipment. Do you think you open everything up, take everything apart and try to find the problem with only a DVM and your eyeballs? Probably not, unless maybe you already have a pretty good idea from experience on exactly where you need to look and for what. Sometimes you have to use other tactics, is all.
The reason I sent that message to Indra was to know if the test required mosfets in or mosfets out. Not because im too lazy to dig in. I've been working on this thing every free minute I got.
 
I have spent years with these Mosfet amps. I always repair and check the whole cirquit without the Mosfets installed. This is no problem and you only can blow cheap parts. When you see a clean signal at all gate resistors, you should be done and solder in the Mosfets as a last step. So leave them out by now.
I'm quite sure you have a very simple fault. The way to find it is to have both channels installed and powered (except Mosfets) and compare voltages left to right, left to right and so on... Print out the schematic and mark every point where voltages differ at the defective side. You can not miss the fault this way, simply impossible. Don't think and speculate, check!

I had a look at your PCB pictures, there is a region where things get ugly, because traces are too narrow to tell them apart. Exposed copper can grow out over time and make a tiny connections.
Also some solder points look bad. I put it on your little experience, but it is just 5 seconds to touch them up when you look at the PCB. Don't leave that to "I do it later".

PS use something to magnify the traces, a lot of light and a high resolution camera help. You will see how relaxed x10-20 is compared to staring at the area