I have aquired a few commercial pa speakers that use the 2092s chip and IRFB4227 FETS to drive the woofers. Looks like a mix of iraud7 and 9. Of the 5 speakers 3 would not turn on and 2 where extremely distorted. Of the 3 I have looked at so far 2 have had shorted fets and one that would turn on would not start oscillation. After removing the fets and the 2092 the power supplies all started and looked fine. after replacing the fets and 2092 on one of the modules it was found that the high side was not working and the low side would oscillate a few times then shut down at about a 350hz interval. These 350hz pulses where enough to pump up the the high side suplly well over 100v. I think this is what ultimately cuased the failures. Not sure if the 2092 went first or the fets? I looked at Voltage on VB and noticed the voltage collapsing. Turns out the bootstrap cap was bad. No suprise as it sits 1mm from the output inductor. So i replaced the 4.7uf electrolytic and the amp seemed to work fine. I thought, well i should replace the elecrolytic with a ceramic cap on the undeside of the board. (The previous owner has sent all the speakers back at least 2 times each for warantee repair .) I wanted a better solution than just replacing the cap to have it fail in 3 months. I only had a 10uf XR5 cap so I put it in. I figured the bigger cap should not cuase a problem exept maybe interfere with the pop supression a bit. I guess I was wrong it fried the 2092 chip immediately. This is the second chip i fried trying to fix these. The first was my fualt when i shorted Vss to Vref with a prob. It also took out the 4227's too🙁
Was I wrong in thinking a ceramic cap would be a good replacement? What mechanism woud cuase the 2092 chip to die from the bigger value cap? I built an amp with the dip version about 12 years ago and dont remeber them being so touchy.
TIA for any help.
Was I wrong in thinking a ceramic cap would be a good replacement? What mechanism woud cuase the 2092 chip to die from the bigger value cap? I built an amp with the dip version about 12 years ago and dont remeber them being so touchy.
TIA for any help.
Since the bootstrap capacitor charges through the low-side MOSFET, only modest values should be used. A small resistor (<10 ohm) in series with the bootstrap diode restricts this charging current, in case of high value low-ESR caps.
The bootstrap diode and capacitor are external to the driver chip and are therefore not capable of directly damaging it. A secondary failure looks more likely.
The bootstrap diode and capacitor are external to the driver chip and are therefore not capable of directly damaging it. A secondary failure looks more likely.
In my case mostly irs2092s survived after shorting out mosfets on either sides. This was the case for genuine irs2092s. As for the distortion firstly change the diodes. Use ultra fast one like mur120.
Disconnect the input from pre amps and see if problem persists. Also check for feedback circuit failure.
Try to reconstruct the amp schematic (just ouput section) so you will ve able to check section wise
Disconnect the input from pre amps and see if problem persists. Also check for feedback circuit failure.
Try to reconstruct the amp schematic (just ouput section) so you will ve able to check section wise
Thanks for the replys.
Here is what I've sketched out so far.
Pretty similar to the IR designs. I can't seem to find the bootstrap precharge resistor that is shown on the ref designs. I measure 120k between B+ and D7 cathode. AN1138 says the bootstrap cap needs to be precharged for "proper start-up". i guess I can check with the fets removed to see if it gets charged without the low side able to turn.
Everthing seemed to be working fine after replacing the bootstrap cap with another 4.7 electrolytic. I changed nothing but the cap to a 10uf ceramic, did't move scope probes or even really move the board. Double checked for no short on the cap and continuity from pins to new cap. The 2092 shorted out as soon as it was powered up. The 2092's are supposed to be genuine. Not sure how I would ever know though.
Here is what I've sketched out so far.
Pretty similar to the IR designs. I can't seem to find the bootstrap precharge resistor that is shown on the ref designs. I measure 120k between B+ and D7 cathode. AN1138 says the bootstrap cap needs to be precharged for "proper start-up". i guess I can check with the fets removed to see if it gets charged without the low side able to turn.
Everthing seemed to be working fine after replacing the bootstrap cap with another 4.7 electrolytic. I changed nothing but the cap to a 10uf ceramic, did't move scope probes or even really move the board. Double checked for no short on the cap and continuity from pins to new cap. The 2092 shorted out as soon as it was powered up. The 2092's are supposed to be genuine. Not sure how I would ever know though.
IRFB4227 are the original FETS beheringer used on this amp.
What changes are you thinking? My thought was better caps, heatsink on the 2092, and possible reducing the carrier from 400k to 200K.