Another cause of big dc on output is not connecting zero volts on amp to zero volts on power supply. This leaves amp zero volts line floating around.
I agree. But the V+ and V- don't come even close to the input/feedback components. They are at the other side of the LM3886 and only go to the respective pins. So the more I think about it, the more I suspect the chips themselves. I'll see what happens when I replace them.
Agree and add: looking and seeing all connected is not enough.
Please turn amp off , wait for voltages to discharge and remeasure resistance from any of the red marked ground points to all others, twice.
We expect zero or a few milliohms of course.
Silly?
It may look silly, but trusting my sight has bit me in the ankle many times, so ....
Nothing beats actual measuring.
Added test: measure resistance from IC pin 10 to power supply ground.
I mean IC actual pin (no its PCB soldering pad) to capacitor legs joined together, you should read 21k.
Also: turn amp ON and measure voltage drop across Rin
Please post both results.
Do not put a cap between the inputs, creates instability
replace Rf2 with a short
see that all components on the +in and -in are soldered directly near the pins
replace Rf2 with a short
see that all components on the +in and -in are soldered directly near the pins
No, not silly. Have had enough of such things happening already. But each GND measured to the GND connection on the PSU board gives 0.3 Ohm, same as putting the leads of the meter together.
Pin 10 to GND gives 950 Ohm. So cut the pin 10 and measured from the stump to GND and have 20.9 kOhm as expected. So there is something going on in the LM3886.
With the amp on there is 0.46V over the 1k. Which is 40 µA, far too much to be normal.
Really? Attached shows the National/TI circuit. The figure is from the LM3886 data sheet.
The LM3886 tends to show a bit of spurious oscillation when it enters and exits clipping on the negative rail. That's where the VAS enters and exits saturation. The compensation components address this.
Tom
Attachments
Ok, mistery solved.
I replaced the LM3886 with another LM3886F that I got from another source. And everything is normal now. So those LM3886 I got from that place far away on the net are fake/duff/defective ones.
Offset is now a couple mV as expected. Just the same pcb and components, only replaced the LM3886.
Time to populate the rest of the pcb. And order more at a better source!
I replaced the LM3886 with another LM3886F that I got from another source. And everything is normal now. So those LM3886 I got from that place far away on the net are fake/duff/defective ones.
Offset is now a couple mV as expected. Just the same pcb and components, only replaced the LM3886.
Time to populate the rest of the pcb. And order more at a better source!
I learned the Chinese and ebay lesson a long time ago.
Especially power devices.
There are even some ebay Chinese sellers selling flaky components at inflated prices to lull you into a false sense of security.
Especially power devices.
There are even some ebay Chinese sellers selling flaky components at inflated prices to lull you into a false sense of security.
Dang! Thanks for following up. Glad you got it working.[/QUOTE]
A few years ago I saw pictures of electrolytic capacitors where the fraudsters had put a smaller capacitor inside a larger can. Not only did you not get the capacitance you thought you'd paid for, the capacitor was also a 250 V type instead of the 400 V marked on the outside can.
Tom