LM3886 bridge mode... calling all gurus of Gainclone

[peteb2]

I have a DynAudio active speaker with a simple internal amp pcb consisting of wired two LM3886s in bridge that feeds the main speaker & a third single LM3886 for tweeter. The fault is random. Powering on randomly results in full on motor-boating oscillation for the main driver speaker. It’s extremely high volume and has seriously given a few folks a fright more than once. Checking data sheets and reading design references it looks as though there are a few suggested parts not included!

The unit is not old, maybe 18months. The problem has always happened & yes warranty would have been a better route.

So far I’ve replaced the entire set of ‘one-hung-low’ no name brand electrolytics but the very random oscillation issue  continues to suddenly happen at power on in a very random fashion.

At the bench I pulled the bridge rectifier and have a substituting external laboratory bench supply set at the usual 40VDC which is normally supplied by the onboard supply.  I’ve been ‘scoping the + & - terminals of the LM3886s wired in bridge. I’ve a 1kHz ref tone at the input and a 4ohm dummy load (driver disconnected).

For multiple power-ons via the bench psu,  the amplifier comes up perfectly with a clean sine wave 1KHz... and draws just over 1A but if I give it a large grouping of on - offs say 30 seconds apart there’s a sudden single failure as there’s obvious low freq motorboating oscillating then slowly the 1KHz appears both + & - feeds seriously cropped on the upper half of the tone signal cycle. It takes a good minute to finally settle back into a proper clean sine wave with no clipping on the upper peaks & multiple power cycles things are fine until oops a random bad one happens.

I’ve never dabbled with the Gainclone chipset (so hence my call to any gurus) but what’s happening? Is it possible like the lousy no-name brand electrolytics hints that the LM3886s are fakes that are not working correctly? I have some genuine spare devices but to swap them is a really huge process on account of the layout. Any experiences of similar trouble or ideas of what to do appreciated.. TIA

[peteb2]

Solved the problem... replaced the two LM3886's set up in bridge mode, no more random cases of motor-boating oscillating upon switch-on even after 50 attempts so i hope its ok now. With 400Hz signal applied the scope traces for output + and - come up perfectly together every time even during a very quick power up, off and on...

[SilverSolder]

Interesting failure mode...  I guess there aren't that many parts to swap, one advantage of a chip amp!