I'm working on a Class-D design at this moment.
I think the design is almost ready, but I only have one major problem.
After the LP filter I have huge spikes.
see:
500ns_200mV
The square wave looks very clean in my opinion, so that can't be the problem.
50ns_10V
Those spikes also appear on the voltage rails.
Deadtime is already set short (roughly 35ns or so)
As a power supply I use 2x470uF than a bunch of 8A ferrite beats and than again 2 x 470uF.
All capacitors have a 100nF in parallel
Sometimes people use a small high power resistor in series, but I think these ferrite beats have a similar (even better?) function.
Unfortunately I can't give you a schematic, but maybe you can give some ideas, because at this moment I'm out of them 😡🙁
I think the design is almost ready, but I only have one major problem.
After the LP filter I have huge spikes.
see:
An externally hosted image should be here but it was not working when we last tested it.
500ns_200mV
The square wave looks very clean in my opinion, so that can't be the problem.
An externally hosted image should be here but it was not working when we last tested it.
50ns_10V
Those spikes also appear on the voltage rails.
Deadtime is already set short (roughly 35ns or so)
As a power supply I use 2x470uF than a bunch of 8A ferrite beats and than again 2 x 470uF.
All capacitors have a 100nF in parallel
Sometimes people use a small high power resistor in series, but I think these ferrite beats have a similar (even better?) function.
Unfortunately I can't give you a schematic, but maybe you can give some ideas, because at this moment I'm out of them 😡🙁
I had the same problem with a bigger dead time.
The MOSFETs are just running normal, they don't heat up.
The inductor moreover, is getting pretty warm (around 80 deg C)
If you are using a single ended scope probe, this may be a measurement artifact. and not real.
See if the noise is also on the supply common.
Those spikes also appear on the voltages rails.
I connected the "signal ground" slightly different, and the problem is much smaller, although still there.
I'm using a 4 layer board, with one big ground layer above the bottom layer (with all the SMD parts on it)
I connected the "signal ground" slightly different, and the problem is much smaller, although still there.
An externally hosted image should be here but it was not working when we last tested it.
I'm using a 4 layer board, with one big ground layer above the bottom layer (with all the SMD parts on it)
Last edited:
My guess is that the noise is a measurement artifact. You'd need a high performance differential probe to accurately measure the voltage
in the presence of high common mode noise.
Differential Probe - High Voltage | Tektronix
Ok, but if I measure other amplifier modules (from specific known brands 😉 ), I don't have these problems at all.
Yes, it's hard to be sure without all the high ticket test equipment. Layout can be tricky. Did you use low ESR capacitors?
I use four BC MAL203859331E3 330uF 100V on the voltage rails.
Parallel with 100nF X7R caps per each electrolytic cap.
I use them two on two and in between an bunch of 8A ferrite beats.
But I highly doubt if that will make such a huge difference.
Parallel with 100nF X7R caps per each electrolytic cap.
I use them two on two and in between an bunch of 8A ferrite beats.
But I highly doubt if that will make such a huge difference.
Tada!
The difference?
Remove the standard ground clip from the probe and put a wire around it with multiple turns and connect it as close as possible to the output.
I guess I've been trapped like all the others here with high frequency signals 😛 😀
(the difference in amplitude is because I haven't made all the improvements I made as the other PCB)
An externally hosted image should be here but it was not working when we last tested it.
The difference?
Remove the standard ground clip from the probe and put a wire around it with multiple turns and connect it as close as possible to the output.
I guess I've been trapped like all the others here with high frequency signals 😛 😀
(the difference in amplitude is because I haven't made all the improvements I made as the other PCB)
Excellent! You can also build probe test probe socket so that the probe (less tip and ground wire) can plug right in to the pcb.
Some probes come with the parts, otherwise you can make one yourself.
https://www.youtube.com/watch?v=-4q8geE5ef8
Last edited:
I was more thinking about this one:
1GHz Active Differential Probe – Daniel Kramnik's Project Log
Although, I think it's not (yet) usable for high voltage signals
1GHz Active Differential Probe – Daniel Kramnik's Project Log
Although, I think it's not (yet) usable for high voltage signals
- Status
- Not open for further replies.