To quote myself from a post on the previous page made some three days ago:
regarding turning up the pot, there's no way to break the ASIC unless other things are going seriously wrong. You can spin the pot screw in circles and you'll only ever get between 550 and 800mV, which is a completely safe range. The only problem would be if you took the voltage all the way up and then ran it without a fan might damage something, but you'll easily notice that the heatsink is getting very toasty.
One thing to note is this:
0.77V is about right for if you're planning on running up around 375MHz. Sticks come off my bench about 610mV and Novak runs 'em at 200MHz off that voltage. Errors are probably around a couple percent; it's hard to get an accurate read with the 512 vardiff we end up with but according to Bitmain's chip data 650mV is good for clean at that speed.
Sounds like your ASIC is damaged. Drawing excess current could be a bad FET on the buck or a near-shorted load downstream of it. If the heatsink is getting hot that means it's taking most of the power, which indicates the buck is probably working right but the ASIC isn't.
Another thing to consider is the buck isn't working right and is overvolting the ASIC, which is shunting the excess power and cooking that way. The real test is going to be measuring the Vcore and seeing what you find. If you can do that and let me know what it looks like would be a good first step. A different good first step would be to send it back for a replacement. I don't like selling things that break.
Regarding using cgminer to generate a config file, well that'd be something Novak'd have to look at. I've never used a cgminer config file, outside whatever's autogenerated by internal controllers on things. I've always run command line, or a batch file on rare occasions.
I read your answer to alh but hadn't remembered it when I posted. Thank you for the refresher and appreciate the further clarification. I understand what vcore is and that I did not need to have it set that high for running 200 or 250 but as I was playing and since I have played with many other devices which utilize a vcore I didn't see an issue with leaving it set higher than needed. I understand it burns more power and heats things up, but didn't think it would be an issue since I had a good fan and ac on the situation.
I actually hadn't given it much thought at all because what I'd read from the Compac threads I didn't see a red flag with using a maximum vcore adjustment with a low frequency.
I am assuming this is the root of my failure.
Too hot:
I don't know how I did it and will be using a couple of more fans when I actually perform real overclocking with my survivors. I examined my running pieces with a nice lighted magnifier and do not see anything on those two which resembles the damage in the picture above. Hopefully it was a fluke.
The vcore measures .86 to 1.1 and bounces between those values.
As a reference I measured one of the living sticks while hashing and it was at .663 rock solid.
If you do want it back confirm the ship to via PM and I will send it out.
I was working late the past few nights so am just installing these tonight.
I think the other 3 for the 5 looks very nice and they are running great.
That is what I call customer service man, and I respect and appreciate it when that happens consistently.
I am running them at 175 for a warm up run, have an arctic cooler fan blowing on them, and I am going to put a 120 mm fan on them before I go any higher.
I found a rhythm to installing these when using Zadig, I believe I've seen more than a couple of people mentioning hot plugging. Novac did as well.
I installed one at a time and let it be recognized by windows.
I reinstall Zadig Win USB driver on every Compac in the list (Have list all devices selected)
I start Cgminer batch file or however you select your settings
While CG miner is running take the compac out, wait a few seconds and plug it back in the same slot.
Never had that fail
Seriously, I appreciate good work and decisions. Thanks guys.
) 
