Troubleshooting the LM3886 MyRef

[Groundloops]

Dear congregation,

as I had to find I´m not alone and knowing that there are many others who more or less succesfully built their MyRefs in whatever Rev, I think it´s about time to consider the trouble that may occur. Maybe it would be a good idea to start a wiki, but I don´t know how.

The short story: Built the RevA, got it running, was very pleased. Built the RevC, got it running, was pleased even more. Sold the RevA, shortcut the RevC, no more pleasure anymore. Managed to fix at least one channel, killed it again, black despair. Buffoon-dimwit-muppet-twerp begs for help.

So I pass the first entry myself (thanks to Andypairo):

Symptom: No more function after shortcutting the outputs while amp was running. Relay doesn´t switch when powering up.
Relief: Replaced fried R11 (1 Ohm resistor between signal and power GND)

[DcibeL]

The MyRef board can be troubleshooted in the same manner that you would troubleshoot any amplifier.

-Check power and ground connections to both the LM3886 and the Lm318.

-Check resistance from output to ground when the relay is open, and when the relay is closed. Check various known resistances.

-If all looks good at this point, it is time to get out the signal generator and oscilloscope. Trace the signal path until the signal is no longer present. search for a fault in the vicinity.

Without an oscilloscope it is harder to determine a fault, but for the most part can still be done with a known constant signal input and a multimeter. A computer can be used as a function generator for this purpose.

[Panelhead]

The active componets are fairly robust. The LM318 is easier to kill than the LM3886. You have to work hard to kill it.
After checking all voltages, the offset is usually a good indicator of where to go. High offset and no relay go together. Offset is either a problem with rails or the feedback loop.
An open in the feedback loop will give high offsets and prevent the realy from closing.
Aftter that look and smell for toasted resistors. It also is possible to smoke the zeners, and the 1K dropping resistors, small diodes feeding the relay/led circuit, and other necessary parts.
But always start at the rails for the LM3886, AND LM318. Then output dc offset. This should ID 75% of any issues.
I am running in a new one using the mono boards now. Hope to post pics tonight. Love all three listened to so far.

George

[es44]

I must admit i find it quite easy to kill a 3886
You just have to reverse polarity of the psu. When you are able to hear and see again, ( ears ringing from condenser blast and then the smoke), the first thing i saw was a hole in the ceiling where the can from the condenser hit, and the lid of the 3886 was gone.