LM3886 amp output weirdness

[IO390]

Hi,

I've just built an LM3886 based audio amp, the circuit is pretty much the one from the datasheet minus the inductor on the output and with a couple of other things modified. The supply is unregulated +/-17V and for filtering I have two 40V 10,000uF caps and two 100nF caps.

The problem is tricky to explain but I'll do my best.

So the center point of the AC output should be at 0V, obviously. After the amp has been on for a few seconds (it works perfectly in this time) I can see, on the scope, the output slowly dropping down towards the -17V rail (DC coupling used on the scope). After about 30 seconds the output starts to hit the negative rail and clipping occurs. Eventually the output just goes down below the rail and disappears.

I've tried a few things. The only one that made a difference was disconnecting the filter cap between the 0V and -17V rails, but leaving the other one in. This prevented the output from dropping but was, of course, filled with the mains crap.

Removing the filter cap from the 0V and +17V rails made no difference.

I've tried two different types of capacitors on this and the change made no difference either. I'm a bit confused on this one so if anyone could give me some suggestions it would be greatly appreciated.

IO

[Simon]

sounds like a capacitor charging/discharging somewhere. How have you built it ? can you show a photo if it's not on a PCB ?

[IO390]

I'm sure you hate block diagrams as much as I do, but I think this gets the point across.

The relay switch is just a simple 12V floating supply which powers a transistor switch for the operation of the solenoid. I've removed the relay and just hardwired the thing but it makes no difference.

[grumpydoc]

Circuit diagram of the whole thing would be useful.

What are the PSU voltages (at the LM3886 pins) doing as the output voltage moves towards the -Ve rail?

What are the voltages at the LM3886 inputs (inverting and non-inverting) doing?

[IO390]

Here's the amp schematic.

It's basically the datasheet one but with a couple of things changed. The 10uF cap at the bottom of the image might be the issue. I was using a 10uF electrolytic (negative side to 0V) and realised that was likely the issue. I changed it for a 1uF film cap and the output drops much more slowly. But it still drops.

I did also try removing the cap altogether but it causes the output to just instantly drops to the negative rail. And also changing it for a 100uF electrolytic made no difference from the 10uF.

[macboy]

It is nothing to do with the power supply, that is probably just fine.

Your 1 uF film cap at the input is being charged by the bias (leakage) current of the + input pin. Put a 10 to 100 kOhm resistor from ground to the point between that cap and RB (the 1kOhm). Then all will be good.

[SeanB]

Remove the input capacitor, you need to have a DC path for input bias current of the amplifier. If you need AC coupling place the capacitor before the volume control.

[IO390]

It is nothing to do with the power supply, that is probably just fine.

Your 1 uF film cap at the input is being charged by the bias (leakage) current of the + input pin. Put a 10 to 100 kOhm resistor from ground to the point between that cap and RB (the 1kOhm). Then all will be good.

I tried that and it now works absolutely perfectly. Thank you.

Remove the input capacitor, you need to have a DC path for input bias current of the amplifier. If you need AC coupling place the capacitor before the volume control.

If I have any other issues I'll try this... Or would it perhaps be a good idea to do this anyway? Without the capacitor might I see improved sound quality?