Recently I made LM4780(obsolete) chipamp on a hifidiy,net PCB (no further info on this site about schematics, etc.) as per standard BOM values printed on the PCB itself later changed after running into oscillations issues. LM4780 pin 11 was broken (-V) but "patched". As it was lying in my drawer for the past 10yrs, I ordered PCB from Aliexpress to finish these projects. This PCB lacked any place for spk gnd and input capacitors. So I connected spk gnd to supply gnd and somehow connected input DC blocking caps . I have also recv. XY LM3886 PCBs for another pair of LM3886. Any suggestions for improvement will be incorporated in the final build of both chipamps.
During initial testing phase itself the IC would get crazy hot and put out 1.6V DC at both outputs ,with very faint audio and heavy static and weird noises. So I slapped in 18K resistor +47pF ceramic series in feedback, as I had only those available in my drawer. The DC got reduced to 0.75V but still unbearable noise levels. Then I soldered 47pF on both + and -ve inputs each. LM datasheets recommends 200pF between + and -ve inputs and 20K +50pF The amp is now behaving normally with one ch. 18mV DC and other 110mV DC heatsinks runs cool to touch. No zobel was used , I just installed 2.2R resistor just to protect from accidental short. I will be making changes as per datasheets BOM incl Zobel in the final build
I was wondering how could other diyer's iteration of such PCBs would have turned out, as no ceramic pF for reducing oscillations are used in even Audio Sector PCBs too.
My question is -
1. Is 47pF ok on + and -ve inputs? Should I bother to change it to 200pF?
2. Does the large heatsink, directly mounted to metal tab which is connected internally to -V, plays any role in noise or oscillations. Should I isolate it using Mica washer?
I won't be using any boutique parts, this project is just to make a budget biamp, based on single LM4780 configured as stereo / dual for highs, & pair of LM3886 for mids and lows with lower carbon footprint in mind.( +/- 20V ). I will be extensively testing them first on elcheapo speakers and dummy loads for oscillations to avoid damaging tweeters.
Thanks
During initial testing phase itself the IC would get crazy hot and put out 1.6V DC at both outputs ,with very faint audio and heavy static and weird noises. So I slapped in 18K resistor +47pF ceramic series in feedback, as I had only those available in my drawer. The DC got reduced to 0.75V but still unbearable noise levels. Then I soldered 47pF on both + and -ve inputs each. LM datasheets recommends 200pF between + and -ve inputs and 20K +50pF The amp is now behaving normally with one ch. 18mV DC and other 110mV DC heatsinks runs cool to touch. No zobel was used , I just installed 2.2R resistor just to protect from accidental short. I will be making changes as per datasheets BOM incl Zobel in the final build
I was wondering how could other diyer's iteration of such PCBs would have turned out, as no ceramic pF for reducing oscillations are used in even Audio Sector PCBs too.
My question is -
1. Is 47pF ok on + and -ve inputs? Should I bother to change it to 200pF?
2. Does the large heatsink, directly mounted to metal tab which is connected internally to -V, plays any role in noise or oscillations. Should I isolate it using Mica washer?
I won't be using any boutique parts, this project is just to make a budget biamp, based on single LM4780 configured as stereo / dual for highs, & pair of LM3886 for mids and lows with lower carbon footprint in mind.( +/- 20V ). I will be extensively testing them first on elcheapo speakers and dummy loads for oscillations to avoid damaging tweeters.
Thanks
Attachments
Last edited:
Apologies to forum members! I have sinned!! , Horrible soldering. I will clean it up in my final build. I finished it in hurry, just to check if the IC is working or not
Attachments
Last edited: