Learning To Walk
When we learn to walk we make many mistakes and fall over many times before we are able to walk properly. Nigel has not built a clone but has designed his amplifiers almost from scratch. He has learnt many valuable lessons about class d design.
If you knew all these things then you have more knowledge than I do. Nigel' s experience have added to my knowledge.
When we learn to walk we make many mistakes and fall over many times before we are able to walk properly. Nigel has not built a clone but has designed his amplifiers almost from scratch. He has learnt many valuable lessons about class d design.
If you knew all these things then you have more knowledge than I do. Nigel' s experience have added to my knowledge.
If the amplifier just worked then I would have learned nothing.
I have been all through the design looking for problems and blown a few mosfets and irs2092's in the process.
I even connected the power the wrong way round after one change but luckily the smoothing caps shorted the supply and did no damage just made my 2amp test transformer hum. I use a low current transformer for testing so that I don't blow so many mosfets if a fault occurs. If I hear the transformer hum I turn off immediately.
There is a lot that has to be right before it works.
VAA, VSS, VCC and VB have to be correct voltage and noise free.
The main problem was the rushed pcb.
It wasn't until after I designed the pcb that I found the iraudamp9 application note.
From this it was obvious where some of my problems were.
The amp is playing fine now and sounds very good despite there still being pcb short comings.
It has certainly put in a few hours for me where normally I would be bored wondering what to do.
I have been all through the design looking for problems and blown a few mosfets and irs2092's in the process.
I even connected the power the wrong way round after one change but luckily the smoothing caps shorted the supply and did no damage just made my 2amp test transformer hum. I use a low current transformer for testing so that I don't blow so many mosfets if a fault occurs. If I hear the transformer hum I turn off immediately.
There is a lot that has to be right before it works.
VAA, VSS, VCC and VB have to be correct voltage and noise free.
The main problem was the rushed pcb.
It wasn't until after I designed the pcb that I found the iraudamp9 application note.
From this it was obvious where some of my problems were.
The amp is playing fine now and sounds very good despite there still being pcb short comings.
It has certainly put in a few hours for me where normally I would be bored wondering what to do.
If you have a decent sound card you can use a free software solution like RMAA.
A voltage divider and filter is needed form amplifier output to soundcard input.
Download. Audio Rightmark
A voltage divider and filter is needed form amplifier output to soundcard input.
Download. Audio Rightmark
Last edited:
D
Deleted member 148505
There are a few problems with pcb.
The 2092 should be closer to the mosfet driver transistors.
The mosfet drivers need decoupling.
Decoupling should be closer to 2092.
The regulator should be better decoupled.
Star grounding needs to used for both amplifiers.
Decoupling should be tight up to the mosfets as possible
I have an older single channel that works so I might revisit that instead.
The layout is much tighter on that.
I think trying to get two amplifiers on the same pcb made things more difficult.
The 2092 should be closer to the mosfet driver transistors.
The mosfet drivers need decoupling.
Decoupling should be closer to 2092.
The regulator should be better decoupled.
Star grounding needs to used for both amplifiers.
Decoupling should be tight up to the mosfets as possible
I have an older single channel that works so I might revisit that instead.
The layout is much tighter on that.
I think trying to get two amplifiers on the same pcb made things more difficult.
Ideally, the size should be equal or close to the area occupied by the components
You are right but I put the resistors 200 thou apart so I can get in with croc clips so I can scope the circuit.
Its also not always possible as I need to get tracks out between 2 components.
Last edited:
Reworked the pcb again tonight.
Managed to get the 2092's closer to the mosfet gate drivers by about an inch.
Not a lot but every bit helps.
The pcb was noisiest on channel 1. So I had a good look around channel 1 2092. I spotted I had the audio input going very close to Vs which wasn't good. I fixed that.
Managed to get the 2092's closer to the mosfet gate drivers by about an inch.
Not a lot but every bit helps.
The pcb was noisiest on channel 1. So I had a good look around channel 1 2092. I spotted I had the audio input going very close to Vs which wasn't good. I fixed that.
The pcb is much tighter now with the components grouped together around the irs2092's.
I have also shortened greatly the -60v rail so I don't get ground bounce.
The smoothing capacitors were moved closer to the mosfets.
I have also shortened greatly the -60v rail so I don't get ground bounce.
The smoothing capacitors were moved closer to the mosfets.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.