Yes sorry should have been more clear, the log starts count on core #0 so if all 64 are good it will read 63 in log output.

Put "-D 2>log.txt" at the end of the bat command. Start it up and it will spit out the log.txt file in same directory. In there you will see startup sequence and it checks for number of active cores. Should say 64.

How did you manage that? These pots are pretty good with abuse, but either way all it takes is a very small screwdriver and some very fine and lite movements to change the voltage. Im assuming you went crazy on it and turned it way more than 180 degrees.

Pots are no where near 1.5 ohm...top one is around 5k and bottom is 100k if I remember correctly.

80C at the top of the small heatsink translates to about 90C chip core temp. I would not go above this for prolonged periods of time, but anything below 80C is actually quiet cool for an ASIC (especially for scrypt ASICs, since they don't have a very large compute die). I would not worry unless you reading temps past 80C.

The main issue with heat is actually more stress on the mosfets, since these chips scale very badly with heat (current draw reaches exponential curve as the chip temp goes up). Your most likely to fry the mosfets/cause a overheat buck shutdown than kill the asic. 

Yep, most of my distributors listed on the main page accept crypto.

Welcome to mining!

You should do some general research on the web for mining, each coin usually has its own pool (for example litecoin pool only mines litecoin obviously, and you can't mine any other coin on there). For example if you want to mine verge, you just edit the bat file and change the pool info to miningpoolhub.com (one of the verge mining pools).

Sounds like a placebo effect  lol

The only option my driver recognizes is "clock=" ...everything else is useless that you have. The only thing that could have made a difference is the d=2048, and those USB fans moving some heat away and causing them to run a bit cooler/more stable.

How long was it running with a detached heatsink? That should have completely fried it and probably the reason its not working how it should. These heatsinks are impossible to pull off as long as they are attached correctly (15 lbs force for 15 seconds). Sounds like the distributor you got them from didn't attached this one correct. Where did you buy it from?

Please check the support thread and read at leas the first few pages....all your questions will be answered there, as well as detailed instructions on how to get started on a Pi.

Yea to get under 1A and stable operation off a standard computer USB port your going to need to lower the voltage and run under 600 MHz.

yes, but you will need to manually restart it currently.

Yea 20 is near the max for that hub especially if you are running them past stock speeds. The most I have ever run stably is about 15 on it (at max speed).

Few sticks I have come across show this issue. Its rare, but the IO coms sometimes spit out garbage that is picked up as a hardware error, unfortunately no way to filter this out. As long as they hash fine you can ignore it (you just have to figure that 8.5% is 0% error for your stick when tuning).

Your hub is not providing enough power. Try lowering voltages/freq.

Thats normal, especially if its been running straight for a really long time. Some cores will destabilize and shutdown eventually. Driver has a fix for a complete core stall, but not drop in hashrate yet. Will have this fix in on the next release.

Cant really know ahead of time for issues like this, otherwise Im sure he would have made an announcement. Not his fault amazon is having issues in Europe. Either way previous two pre-orders were a 2 month wait, 10 days is not a pre-order :p

933 mhz x 2 (DDR) x 4 (quad channel) x 128 bit wide bus (8 x16 bit DDR3 chips) / 8bits = 120 GB/s

Remember stock ram for motherboard is 64bit wide no reason why you can't go higher, if they are really using something between 72-96GB per board and I'm assuming article is wrong and its 72GB which would make sense for 8 x 512mb chips per core for 128bit wide bus.

Thats incorrect, if they are using quad channel spread to 8 16bit 512MB DDR3 chips for each core thats ~14mh per core (assuming around a 1GHZ ram speed, some DDR3 can go up to 1600mhz. 14x 6 cores x 3 boards ~ 250 MH.

Whats a joke about that? Anyone commenting on the pure engineering/tech of using DDR3 for eth mining has no engineering background and no idea what your talking about if your laughing about it.

Give me 10 mill to develop an eth ASIC core with a 4 channel DDR3 mem controller and it will blow any GDDR5 based GPU out of the water in terms of perf/$ and perf/watt

Here you go :p

https://www.micron.com/parts/dram/ddr3-sdram/mt41k512m16ha-107?pc={00EED26F-83AE-4CE6-9A28-EB8B033361E8}

1GB x16 wide  933mhz chip super cheap DDR3 chip. Either way something is wrong with the article since the numbers don't add. If they are doing something like this would make sense to use 512MB x16 wide chips for lets say 4GB of 128 bit wide ram per core. Then your looking at like 7 or 14 MH per core depending on number on memory interfaces per core.

You don't need anything crazy on the core to do the hashes...going lower than 28nm for an eth specific asic is what would really be "stupid" as you put it.

Only reason eth asics have not come out is because GPUs have still been more economical, but with current eth prices that changes. An eth asic is essentially a GPU with a much smaller compute die and cheaper memory.

One of these cores would be about $10 bucks for the asic core/memory controller, and another 20 bucks for the 4GB of DDR3 ram. So your looking at around 30 bucks per 14 MH core (lets say it has 4 memory interfaces).

So bitmain is looking at the current market and are like hey 30 MH GPUs are selling for 400 bucks, we can make the equivalent with 2 ASIC cores for 60 bucks and sell it for 200 with a fraction of the power consumption = win.