I have built a few irs2092 based class d amplifiers.
I thought it was working fine.
It does on low volumes but on high volumes and low frequencies it keeps resetting. If I apply 1khz then it clips well and doesnt reset.
If I apply 50Hz it gets to about 50% of B+ and keeps on resetting.
I have tried extra decoupling.
Tried increasing most of the capacitor values around the chip used for decoupling with no success.
I get an improvement if I add another pair of MOSFETs but the resetting still occurs but at a higher output voltage. This suggested overcurrent protection kicking in but raising the threshold makes no difference to the resetting problem.
I thought it was working fine.
It does on low volumes but on high volumes and low frequencies it keeps resetting. If I apply 1khz then it clips well and doesnt reset.
If I apply 50Hz it gets to about 50% of B+ and keeps on resetting.
I have tried extra decoupling.
Tried increasing most of the capacitor values around the chip used for decoupling with no success.
I get an improvement if I add another pair of MOSFETs but the resetting still occurs but at a higher output voltage. This suggested overcurrent protection kicking in but raising the threshold makes no difference to the resetting problem.
how are voltages around ic?
sounds like overcurrent or suppy pumping
The voltages around the 2092 look fine.
Under resetting the vaa, vcc and vss are fine.
I have added extra capacitance to these lines and it still resets.
More like supply pumping to me...
The supply voltages dont go up during resetting.
look at pins related with current sense/set. if this is the same amp as in other thread you have, could be that peak current is higher then set...so look at those pins
I tried making the overcurrent sense less sensitive but it made no difference.
I can only guess the inductor is saturating and taking masses of current.
I have ordered a 28amp inductor and will try that when it comes.
More like supply pumping to me...
If music is playing loud I get supply pumping.
The B+ goes from the normal 45 upto 70 volts.
I put a resistor across B+ to gnd to get rid of it but the resetting problem still persists.
but it doesn't have to go through inductor, shorts through fets are more likely. in any case, look if switching stops, if overcurrent is triggerd
I have changed the deadtime from 25ns to 125ns and it makes no difference.
The mosfets are irfb4019 which are very fast.
yes, so overcurrent, now is it from wrong settings or wrong material used for inductor, that I don't knowTried some different mosfets in with a lower rdson.
The level at which it resets now is twice as high a voltage on the output.
Still not quite there tho it should not reset well into clipping.
Ran out of mosfets to try 4 in the amplifier so have ordered some more.
yes, so overcurrent, now is it from wrong settings or wrong material used for inductor, that I don't know
I have ordered a recommended inductor and some more irfb4227 mosfets.
I will try one at a time and see what the difference is.
The inductor does get very hot so that is suspicious of the inductor.
report back, i bet my money on wrong core material
Which core material should I be using ?
The PCB layout can't handle the EMI from the body diode reverse recovery without the circuit disturbing itself. The very first steps when you are testing a class D or SMPS prototype are to check proper switching timing, to obtain reasonably good waveforms (little ringing) and to ensure things are still under control at high currents, when body diode reverse recovery starts to show up and cause trouble (imagine: one MOSFET turns on while the other, although turned off, is still going to show low impedance until some reverse body diode current is drawn for some time, like 20-200ns).
Body diode conduction starts when (negative) Id*Rds-on exceeds one diode drop.
Body diode conduction starts when (negative) Id*Rds-on exceeds one diode drop.
The PCB layout can't handle the EMI from the body diode reverse recovery without the circuit disturbing itself. The very first steps when you are testing a class D or SMPS prototype are to check proper switching timing, to obtain reasonably good waveforms (little ringing) and to ensure things are still under control at high currents, when body diode reverse recovery starts to show up and cause trouble (imagine: one MOSFET turns on while the other, although turned off, is still going to show low impedance until some reverse body diode current is drawn for some time, like 20-200ns).
Body diode conduction starts when (negative) Id*Rds-on exceeds one diode drop.
I have 125ns deadtime. and have found deadtime makes no difference to the ringing. According to teh datasheet the 2092 waits 450nS before looking for overcurrent anyway and that is after ringing has stopped.
The mosfets stay cold so I doubt there is shoothrough or any other shorting of the power supplies.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Class D
- irs2092 resetting with low frequencies.