HELP with LM3886 X3 in parallel.

[umut1001]

Transformer was för F5 amplifier 4x18v ac and ps caps are 4x10000uf.
It is also connected to stereo lm3875 gain clone.
Voltage is now +-20vdc
Maybe it is is a problem?It seems voltage limit of lm3886
And when I am adjusting The first two ones Thé third amp dc voltage changes! How it happens?

 

[umut1001]

Finally I could adjust all of them to around 19mv and connected output resistors all together. It plays now.I will play louder tomorrow.It is night here and my baby is sleeping

 

[umut1001]

The problem about low voltage(20v instead of 24v) was at CL60 NTC inrush current limiter.It is burned.It keeps 20v on itself.I removed it and connected CL101 and it is 24v symmetric now.And voltage increased to 24v again.Now it plays good music.

 

[sfthurber]

First I soldered all emitter 0r22 resistors...

in looking at your pic i see it is not one of the yj or jim's audio red boards so common on ebay. do you mind if i ask the vendor where you bought it?

 

[PB2]

The amps do function well as designed, but there is a rather detailed and involved process to match all the DC offsets. There is a good post about it and I will put it up when I find it. For now please look at this site.

This is the base design.

Since it has been mentioned twice on one page, it appears I need to revisit mine after buying some closely matched resistors. You can see in the picture I just used some generics purchased on eBay.

I'm not into chip amps at all but have an application where a pair in parallel might work well.
As far as DC offset goes I would build the amp, test it and make sure both amps work and the DC offset is say below 100mV to start. Then assuming the output short circuit protection actually works, I'd short the output and measure the voltage across the combining resistors as a measure of current, then trim the DC offset for minimum across those resistors. Then check it with the nominal load of 4 or 8 ohms, run it hard and check the voltage again hot to make sure it doesn't drift too much. Try this at your own risk of course.

I don't really like the high frequency compensation on these parallel amps, does anyone have the original Roland circuit, are people sure that this is the same circuit?

 

[AndrewT]

National show DC servos on the two channels that are to be paralleled.

I would start by ensuring that at switch on just as the mute unmutes that the unservoed offset is near zero mVdc. Then the servo takes over and tries to maintain that near zero mVdc offset as the chip warms up.

Doing that means the chips do not send current into each other, neither at start up nor as they warm up.