Sorry, I don't want to come off as lackadaisical. But because I'm also a audio engineer even just pulsing the amplifier with my finger I can tell if the sound is distorted or not.
And there's also the issue of my second phone dying that I use to test the amplifier because it has a built in dac. I use it to test and my main phone to film, just to show proof of the amplifier being fixed.
I apologize if this method is offensive.
And there's also the issue of my second phone dying that I use to test the amplifier because it has a built in dac. I use it to test and my main phone to film, just to show proof of the amplifier being fixed.
I apologize if this method is offensive.
You have to make the decisions on what's best for you but testing in that way doesn't find intermittents with switches, with intermittent inductors or bad connections. It also doesn't allow components to reach operating temperature. You can't check for the stability of the various power supply voltages or to look for ripple that is more likely to occur under load.
In my opinion, you need to test with a reasonably heavy load (I use dummy loads that are 2x the lowest rated load, 2 ohms for a 1 ohm amp) with the output regularly reaching clipping during the test and for at least 30 minutes at that level.
In my opinion, you need to test with a reasonably heavy load (I use dummy loads that are 2x the lowest rated load, 2 ohms for a 1 ohm amp) with the output regularly reaching clipping during the test and for at least 30 minutes at that level.
I know, I'm getting another audio source to play music and also test tones.
As you know I don't have dummy loads but I use 2, 4 ohms speakers.
And depending on workload I would play an amplifier for a few hrs.
As you know I don't have dummy loads but I use 2, 4 ohms speakers.
And depending on workload I would play an amplifier for a few hrs.
Set up a second/test bench to use for final testing. On a second bench, you can let the amp run while doing other things. Do final testing with music. I used the following:
As a side note, the replacement of the resistors essentially removed all over-current protection. The originals were borderline (too quick to protect) but probably gave good protection. I'm assuming that the resistors you replaced were connected to the CSH terminals.
As a side note, the replacement of the resistors essentially removed all over-current protection. The originals were borderline (too quick to protect) but probably gave good protection. I'm assuming that the resistors you replaced were connected to the CSH terminals.
Last edited:
A second bench makes sense.
The resistors I replaced are the same value in the other 2 amplifiers, over current protection for the output devices or the IC?
Also the low side FETs are switching at 400+Khz but the high side is switching at over 2 Mhz, is that normal?
The resistors I replaced are the same value in the other 2 amplifiers, over current protection for the output devices or the IC?
Also the low side FETs are switching at 400+Khz but the high side is switching at over 2 Mhz, is that normal?
Without knowing if the amp ever worked properly and not being able to test the amp to full power (do you have a large power supply?), I can't really advise. The originals gave the amp a better chance at surviving if the speaker wires shorted to ground but I don't know if the amp would go into protect prematurely with them.
A couple of things...
The protection circuits in the 2092, 2093 and 20957 amps are rarely ever properly designed. The only ones that I've seen that were properly designed were the reference designs from the IC manufacturers.
The protection circuit has to be tailored to the FETs being used as the output transistors. A significantly different RDS(on) requires different protection circuit resistor values.
A couple of things...
The protection circuits in the 2092, 2093 and 20957 amps are rarely ever properly designed. The only ones that I've seen that were properly designed were the reference designs from the IC manufacturers.
The protection circuit has to be tailored to the FETs being used as the output transistors. A significantly different RDS(on) requires different protection circuit resistor values.
I have a large PS 13.8 volts 42 amps. Are you talking about the 4 resistors I changed around the 2093?
I see, so those resistors are for the protection? I was thinking that those resistors had something to do with the output FETs getting hot.
These amplifiers do have a high failure rate from what I've seen. The technician who gave me these to fix had over a dozen of this particular one and several dozens of other types.
So this issue of the transistors getting prematurely hot, by the way, it only seems like channels 1 or 2 or both are getting hot. Would it be because the high side is switching at over 2 Mhz?
But also before I removed all the FETs only channels 3 high side FET was getting hot. Maybe there is a leaky FET? but then again when I removed both FETs from channel 3, then channel 4 FETs started to get hot. Then I removed all FETs from each channel except channel 2 and it's high side FET was getting hot.
Maybe the dead time for the high side drive needs adjusting?
I see, so those resistors are for the protection? I was thinking that those resistors had something to do with the output FETs getting hot.
These amplifiers do have a high failure rate from what I've seen. The technician who gave me these to fix had over a dozen of this particular one and several dozens of other types.
So this issue of the transistors getting prematurely hot, by the way, it only seems like channels 1 or 2 or both are getting hot. Would it be because the high side is switching at over 2 Mhz?
But also before I removed all the FETs only channels 3 high side FET was getting hot. Maybe there is a leaky FET? but then again when I removed both FETs from channel 3, then channel 4 FETs started to get hot. Then I removed all FETs from each channel except channel 2 and it's high side FET was getting hot.
Maybe the dead time for the high side drive needs adjusting?
Higher switching frequencies can make the FETs run hotter. As with the other amp where frequency was possibly an issue, you not only have to look at the frequency at idle but also when driven hard. Can you drive one of the 2MHz channels into a load at high power to see how low the oscillation frequency drops to?
What's failing on these amplifiers (if you have many failures)?
I have never had anyone who was willing to reprogram the protection circuit and test to confirm that the protection is OK.
42 amps is only good enough to test amplifiers (up to full power) up to about 400 watts. People who have proper supplies to test the larger amplifiers have up to 1000 amps of power supply.
Deadtime may be an option but longer deadtime settings will likely result in higher distortion (possibly inaudible).
What's failing on these amplifiers (if you have many failures)?
I have never had anyone who was willing to reprogram the protection circuit and test to confirm that the protection is OK.
42 amps is only good enough to test amplifiers (up to full power) up to about 400 watts. People who have proper supplies to test the larger amplifiers have up to 1000 amps of power supply.
Deadtime may be an option but longer deadtime settings will likely result in higher distortion (possibly inaudible).
I can do that, I have a 10 amp bench supply is that good to check for it?
Mostly because when people install them they don't install a fuse so anything that happens to the amplifier for whatever reason, because there's no fuse it blows up the power supply section.
I had no idea I was messing with the protection circuit, should I install the the original resistors?
I know 42 amps is weak.
I know that also, about the dead time. But why would the high side be switching at a high rate than the low side?
Mostly because when people install them they don't install a fuse so anything that happens to the amplifier for whatever reason, because there's no fuse it blows up the power supply section.
I had no idea I was messing with the protection circuit, should I install the the original resistors?
I know 42 amps is weak.
I know that also, about the dead time. But why would the high side be switching at a high rate than the low side?
The self-oscillating frequency is determined by the components between the input, comp and ground... and also, to some extent, the IC. Refer to section 2.2.2 (only one point in the note that refers to frequency).
Running only one channel, 10 amps may be OK. You may see a difference without a load, so you can try that, as well.
Without proper protection programming, you have nothing to shut down the amp when it's loaded too heavily or a speaker wire shorts to ground. With proper protection, the output stage shuts down and will protect the amp, even with no fuse. A properly rated fuse should be used to add further protection for other failure modes.
Running only one channel, 10 amps may be OK. You may see a difference without a load, so you can try that, as well.
Without proper protection programming, you have nothing to shut down the amp when it's loaded too heavily or a speaker wire shorts to ground. With proper protection, the output stage shuts down and will protect the amp, even with no fuse. A properly rated fuse should be used to add further protection for other failure modes.
Since the original 680 ohm resistors will be a safer option, yes. I can't guarantee that the amp won't go into protect prematurely. If it does, you may have to adjust slightly. I think I'd suggest driving one channel (with the 680 ohm resistor) to full power into your 2 ohm speaker load to see if it goes into protect or if it plays up to intermittent clipping with bass music.
Okay, I'm going to try that. Also I saw something else, on the schematic for this amplifier there's a 12K resistor in the OCSET and DT but on this amplifier it's 8.2K resistor. Is that an issue?
I'd leave DT alone at this point.
The low-side OCSET is important as well but they have it connected to the VCC which it odd. It's typically connected to the Vref. To make any decisions there, I'd have to have more definitive information. For now, let's just work with the high-side programming on one channel.
The low-side OCSET is important as well but they have it connected to the VCC which it odd. It's typically connected to the Vref. To make any decisions there, I'd have to have more definitive information. For now, let's just work with the high-side programming on one channel.