I am getting around 35 volts overshoot then ringing on my class d output inductor.
Is there any way to reduce this ?
Is there any way to reduce this ?
An externally hosted image should be here but it was not working when we last tested it.
This is before the inductor.
I am using a 16 amp indcutor so it shouldnt be saturating.
The power supply is only a 2 amp supply.
This was without an input signal.
There is a 470nf capacitor after the inductor.
Deadtime is 125nS, I tried 25ns but it made no difference to the problem.
The mosfet gates have a 10R before them.
I put two inductors in series just to try it and there is a slight improvement.
I have ordered a 28 amp inductor to see if that helps.
The inductor gets very hot around 120 degree C.
It is a toroid 22uh and 16amps.
Odd thing is my supply is just a bench supply that gives out 2 amps so I cant see why any overcurrent is being triggered unless it is drawing very narrow pulses from the smoothing capacitors.
I am currently using this toroid.
BOURNS JW MILLER|2305-V-RC|TOROIDAL INDUCTOR | Farnell United Kingdom
correct, i drive 350w on output, yet i have 130w max supply, most of the power comes from caps
what material did you use
Hi,
You use this..currently at output of D class stage??
Use adeguate type AMIDON eg. T29 or Txx of this family. not only
inductance need.
Doesn't this waveform come from the class D module that you offer for sale in your web?
This looks like a class D amplifier with class AB PCB layout and decoupling. Class AB experience actually gives very limited electronics knowledge, there are a few more lessons to learn for class D. You are far from obtaining a good module that can be sold without giving a bad name to the whole class D community. Please don't sell anything with such an ugly switching waveform.
Supply rails are resonating badly, this is probably the result of paralleling capacitors in the way high-end guys do (without having a clue about the resulting RLC resonators and impedance peaks). You should model pcb tracks (L), deocupling caps (RLC) and mosfet parasitics (LC) and make a critically damped system by playing with values, high and low ESR electrolytics and snubbers. Learning to do this may take some time. It's easier to learn if you do it with a prototype and a simplified simulation model (ideal switching but including all LC parasitics of interest) and trying to correlate both. Modelling the circuit as separate parts with ideal stimulus applied (supply decoupling, switching node, output filter, modulator) can save a lot of time and computing power. Later you may be able to design directly for good waveforms, and already get them with the first PCB revision and very close to predicted component values.
The inductor has not much to do with the ringing, except that lower inductance will make ringing worse because more ripple current would be switched.
Conventional inductors you can buy from a catalog are not suitable for class D, the core material is optimized for low turns (high permeability) and low cost assuming low AC flux, not for low losses with high AC flux. Micrometals -2 material is a very popular one optimized for low loss with high AC flux above 50khz (at the expense of low permeability).
You should learn to calculate and measure inductance and saturation current for a given core and amount of turns. A long time ago I wrote a long post explaining how to do it.
This looks like a class D amplifier with class AB PCB layout and decoupling. Class AB experience actually gives very limited electronics knowledge, there are a few more lessons to learn for class D. You are far from obtaining a good module that can be sold without giving a bad name to the whole class D community. Please don't sell anything with such an ugly switching waveform.
Supply rails are resonating badly, this is probably the result of paralleling capacitors in the way high-end guys do (without having a clue about the resulting RLC resonators and impedance peaks). You should model pcb tracks (L), deocupling caps (RLC) and mosfet parasitics (LC) and make a critically damped system by playing with values, high and low ESR electrolytics and snubbers. Learning to do this may take some time. It's easier to learn if you do it with a prototype and a simplified simulation model (ideal switching but including all LC parasitics of interest) and trying to correlate both. Modelling the circuit as separate parts with ideal stimulus applied (supply decoupling, switching node, output filter, modulator) can save a lot of time and computing power. Later you may be able to design directly for good waveforms, and already get them with the first PCB revision and very close to predicted component values.
The inductor has not much to do with the ringing, except that lower inductance will make ringing worse because more ripple current would be switched.
Conventional inductors you can buy from a catalog are not suitable for class D, the core material is optimized for low turns (high permeability) and low cost assuming low AC flux, not for low losses with high AC flux. Micrometals -2 material is a very popular one optimized for low loss with high AC flux above 50khz (at the expense of low permeability).
You should learn to calculate and measure inductance and saturation current for a given core and amount of turns. A long time ago I wrote a long post explaining how to do it.
Last edited:
I am not currently selling any, if you look at my website they are up and coming modules.
I am having problems with a couple I sold as sold as seen prototypes which I sadly let go without checking on full power first.
I have some more research work to do before the modules are fit for public consumption. I have sold hundreds of class AB modules and never had any problems but they were fully tested before being let out.
Class D is a long and steep learning curve but I am getting there, I get a little wiser every day.
Last edited:
Yep . . .
You seem to be missing the core of what eva is telling you . . . that you need to stop thinking about the power stage as an audio circuit and start thinking about it as a high speed switching circuit. That probably means scrapping the present board/layout altogether and re-designing from the ground(s) up.
http://www.irf.com/technical-info/appnotes/an-1138.pdf
http://www.irf.com/technical-info/appnotes/an-1135.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.