Hey Folks,
I have been a happy user of my custom made TPA3255 boards for years now and still am, despite a recent incident at high volumes
I have a MCU in the front end that detects clipping/error conditions and lowers the volume when clipping or any kind of overload occurs. I am using them with two 20V 170W Lenovo power supplies that are stacked up for 40V DC output. The amp is configured to run in stereo BTL mode, which is meant for loads from 4-8ohm. PVDD capacitors are two 1000uF 50V Panasonic FMs.
When testing the amp at a friend's place on his B&W 703 S3 I finally pushed it a bit harder and it did overload at some point - clipping was detected and volume was lowered as expected by MCU. However, on few occasions and before clipping was detected a loud cracking sound occurred, like there was a hard cutoff at the output - not a very pleasant thing to hear as you can imagine. I was not expecting this, as I was hoping that any error conditions would be caught by internal circuitry and errors raised to be handled by MCU.
Now, I know that according to specs, 8ohm 703S3 can go as low as 3.1ohm at some frequencies, so I am wondering if this was maybe the culprit? Or maybe the PVDD capacitor bank is not big enough, seeing that eval board is using 2x4700uF? According to the datasheet, if OC protection kicks in, the device will go into error state which needs to be cleared by _RESET.
Any ideas what could be causing this kind of behavior?
I have been a happy user of my custom made TPA3255 boards for years now and still am, despite a recent incident at high volumes
I have a MCU in the front end that detects clipping/error conditions and lowers the volume when clipping or any kind of overload occurs. I am using them with two 20V 170W Lenovo power supplies that are stacked up for 40V DC output. The amp is configured to run in stereo BTL mode, which is meant for loads from 4-8ohm. PVDD capacitors are two 1000uF 50V Panasonic FMs.
When testing the amp at a friend's place on his B&W 703 S3 I finally pushed it a bit harder and it did overload at some point - clipping was detected and volume was lowered as expected by MCU. However, on few occasions and before clipping was detected a loud cracking sound occurred, like there was a hard cutoff at the output - not a very pleasant thing to hear as you can imagine. I was not expecting this, as I was hoping that any error conditions would be caught by internal circuitry and errors raised to be handled by MCU.
Now, I know that according to specs, 8ohm 703S3 can go as low as 3.1ohm at some frequencies, so I am wondering if this was maybe the culprit? Or maybe the PVDD capacitor bank is not big enough, seeing that eval board is using 2x4700uF? According to the datasheet, if OC protection kicks in, the device will go into error state which needs to be cleared by _RESET.
Any ideas what could be causing this kind of behavior?
Last edited:
how are you lowering volume? maybe some kind of biasing issue on tpa inputs caused by that volume lowering thing?
edit: i think it is better to use clasic compressor/limiter circuits on signal inputs set in a way so amp never clips as anyway tpa thd rises very steeply when clipping occur so better to avoid it alltogether.
i have made some pcbs for pa duty tpa that have preamps with compressor for that exact purpose but havent got around to manufacture and test those yet, but tested tpa with external compressor on active boxes i made some time ago and it does help mask that steeply rising distortion.
edit: i think it is better to use clasic compressor/limiter circuits on signal inputs set in a way so amp never clips as anyway tpa thd rises very steeply when clipping occur so better to avoid it alltogether.
i have made some pcbs for pa duty tpa that have preamps with compressor for that exact purpose but havent got around to manufacture and test those yet, but tested tpa with external compressor on active boxes i made some time ago and it does help mask that steeply rising distortion.
Last edited:
