Hi
OK, this is a longshot, but maybe one of you have experience with this.
Background: I have some odd/random failures in a FB (full-bridge) classD amp based on IR2184.
I have build HB (half-bridge) version, with very similar setup, but without this kind of error, so I'm trying to figure out why the FB is failing.
The only direct connection from one half-bridge to the other is the 2 SD ( shut-down) pins of the 2x IR2184. In my amp they are tied together (and can be pulled down to Vss with a transistor). - Could that be problem? I have a decent layout, acceptable decoupling and very short critical tracks.
Im pretty sure that the error only happens on turn-on (maybe turn-off, but only detected at the subsequent turn on).
I use SD to disable the IR2184 until supply voltage is high enough, and to disable if current is too high.
I remembered seeing a similar FB in an Yorkville subwoofer amp, but in that product the 2 SD-pins are separated although they are activated simultaneously .
Maybe not clear in the picture, but the 2 SD-pins are not tied together.
The circuit that pulls the 2 SD-pins high:
Why have 2 branches doing exactly the same?
Implementation on PCB ( With through-hole components...)
The only reason, I can see, for splitting the SD-signals, is that there is some distance from one IR2184 to the other IR2184 and maybe Yorkville designers tried to avoid tracks from directly from one HB to the other HB?
Any ideas?
Note: YES, the IR2184 have looong deadtime and distorts. I get below 1% in my prototype and thats plenty fine for a 500W sub-amp.
Kind regards TroelsM
OK, this is a longshot, but maybe one of you have experience with this.
Background: I have some odd/random failures in a FB (full-bridge) classD amp based on IR2184.
I have build HB (half-bridge) version, with very similar setup, but without this kind of error, so I'm trying to figure out why the FB is failing.
The only direct connection from one half-bridge to the other is the 2 SD ( shut-down) pins of the 2x IR2184. In my amp they are tied together (and can be pulled down to Vss with a transistor). - Could that be problem? I have a decent layout, acceptable decoupling and very short critical tracks.
Im pretty sure that the error only happens on turn-on (maybe turn-off, but only detected at the subsequent turn on).
I use SD to disable the IR2184 until supply voltage is high enough, and to disable if current is too high.
I remembered seeing a similar FB in an Yorkville subwoofer amp, but in that product the 2 SD-pins are separated although they are activated simultaneously .
Maybe not clear in the picture, but the 2 SD-pins are not tied together.
The circuit that pulls the 2 SD-pins high:
Why have 2 branches doing exactly the same?
Implementation on PCB ( With through-hole components...)
The only reason, I can see, for splitting the SD-signals, is that there is some distance from one IR2184 to the other IR2184 and maybe Yorkville designers tried to avoid tracks from directly from one HB to the other HB?
Any ideas?
Note: YES, the IR2184 have looong deadtime and distorts. I get below 1% in my prototype and thats plenty fine for a 500W sub-amp.
Kind regards TroelsM
Hi.
Thanks for the reply. Sorry for not following up.
The failure is complex to pin-point as both half-bridges share power supply and supply to drivers. I'm pretty sure that the drivers fail first and depending on the specifics they may take output-fets down with them.
In at least one failure-scenario I had a 10uF cap directly on the SD-pins because that gav the best start-up. I'm wondering if the same cap destroyed the SD in put on shutdown because the SD-pin would be held high some mS after the power was lost?
Should probably have a fast diode from SD til driver-supply.
Kind regards TroelsM
Thanks for the reply. Sorry for not following up.
The failure is complex to pin-point as both half-bridges share power supply and supply to drivers. I'm pretty sure that the drivers fail first and depending on the specifics they may take output-fets down with them.
In at least one failure-scenario I had a 10uF cap directly on the SD-pins because that gav the best start-up. I'm wondering if the same cap destroyed the SD in put on shutdown because the SD-pin would be held high some mS after the power was lost?
Should probably have a fast diode from SD til driver-supply.
Kind regards TroelsM
