I haven't done the calculations, but - the chips inside an S9, with all that airflow, see 120C junction temperatures while dissipating roughly 6W per chip. At 300MHz your 2Pac will also be dissipating about 6W per chip, so without airflow it should be getting even hotter.

A good rule of thumb for anything electronic is avoid >125C junction temperatures.

It shouldn't catch on fire if you run it too hot, but it'll definitely stop working pretty soon. And before then it'll be a safety hazard to anyone who touches it without welding gloves. And if it fails it won't be covered by the warranty.
I don't recommend exceeding 100MHz for any length of time without some kind of fan.

No, they had too many reliability issues so I shut 'er down to work out whatever kinks were causing problems. I'll have a new and vastly superior model in a few months though.

Thanks for the nondescript acknowledgement that definitely wasn't spam, Mr High Schooler. Glad you figured out whatever the problem apparently was and got something back up and running.

I really don't like pre-orders. I've relaxed the policy when necessary in the past but only for designs that had already been proven and I just needed to raise funds for an initial batch. This guy hasn't been proven in full yet - we've got 2 chips hashing, but not in a stick shape and not with controls. Definitely won't take money on pods since that scale of device hasn't even started testing yet.

It'll be tight but I should be able to at least get started on batching sticks. But the more 2Pacs I can sell for at least the cost of materials, the less I have to take a huge loss on selling BTC to do it.

No kidding, if someone asked for ONE THOUSAND 2PACS right now I could ship by Friday.

Oh hey guys, here's a bit of an update.

I have today sent off for prototype PCBs for what I hope to be the final version of a new 16nm stickminer. It's going to feature not just improved hashrate and efficiency, but a fully integrated control package that, once we finalize the firmware and driver support, will mean we have provision for temperature monitoring (of both PCB and ASIC die) and software voltage control. So no more tiny screwdrivers. We could have rolled a stick out two months ago if I'd been okay with tiny screwdrivers, but there was room for improvement so I designed the whole control package.
Following on the stick's heels will be a much-improved pod miner, building on knowledge gained from the (ultimately disappointing, from my standpoint) Terminus R808 project and integrating an expanded version of the stickminer's controls. Improved housing, improved cooling, power use monitoring, fan speed control, it's gonna be nice.

However.

In the last two months, bitcoin has taken some pretty massive dumps. That may be nice for long-term investors who are using this opportunity to buy, but it sucks for everyone who actually uses the currency as a currency - that is to say, as a means of exchange for goods and services. You know, guys like me. Not so friendly.

Anyway, the year's downturn in general has killed the resale market for my bread-and-butter 2Pac USB Stick Miner. So operations are consuming most of my cash savings, while my BTC savings are shriveling on the vine, and 2Pacs are stacking up on the shelf.

Which I see as heavily unfortunate, because I need the revenue from the final batch (or heck, even the parts cost of the final batch) to help sponsor the initial batches of the new better gear. I could have stopped production batch before last, but I knew the new stuff's production was still a ways out and I have employees who need paychecks, and "take care of your people" runs deep in my family.

Long story short - I need to get rid of my 2Pacs in order to afford to manufacture the new stuff that's better than 2Pacs. Do any of you guys want to toss in on some cut-rate 2Pacs to help fund batching of exciting new stuff?

Hardware issue, needs repair or replacement. Contact the seller.

There's also the thought that anyone paying attention would know a fivefold increase in efficiency from one model to the next didn't really even happen back when ASIC mining was new and still working down from 130nm. When someone suddenly comes out with a machine that's far and away better than what everyone else has, it is ALWAYS a scam. Especially if it's in anywhere near the same price range. Obvious ploy is obvious (and oft repeated).

Double especially when the company claiming to be releasing it has been hacked, and sent you an email saying they were hacked, and then sent you an email saying the miner rumors say they're selling in fact does not exist and you should not buy it.

And if you're on 208V and derate to 80%, a 20A circuit gets you 1 machine (and 48% utilization); 30A circuit gets you 2 (and 64% utilization).
Though a 60A PDU and 80% derate would run exactly 5 without a lot of overhead.

208's pretty common for industrial-scale electric, even in datacenters.

And the point still stands that the power density and cooling requirements mean these'll only run where your intake temperatures are regulated to probably 60F and lower. In my farm, that means I could run one for at most five months out of the year. If someone bought one and sent it here for hosting I'd have to turn it down. Got enough trouble keeping S9s cool May through August.

BM1385 in the S7 is a 28nm chip, and has a bottom-clock efficiency around 0.18J/GH with decent hashrates still around 0.22J/GH; the 700MHz operating point setpoint of 45-chip S7 was past the knee of the efficiency curve and we also lost ~10% of power, board-level, to the 45A single-phase main regulator plus around 50W machine-level just for fans. Chip-level, the BM1385 was not much worse than some others' early 14/16nm designs.

(if by 2 you mean 4)

So what, they're using air with a higher specific heat? Epoxy with zero thermal resistance? 300CFM fans that don't spin at 10KRPM and draw 100W to do it? Have they discovered a heatsink profile with no solidity and perfect heat transfer, and made sure to place every single one to avoid vortices?

Quality or not, the physics of air cooling is a big limitation. Look at all the issues we already have with heat unreliability on 16nm and 10nm and think how much worse it's going to be on 7nm, which has basically never been tested before this point.

I've also been saying that full control of core clock and voltage are necessary features but nobody does that anymore. Probably for idiot-proofing. It is handy some have at least partial options.

I've been saying exactly that for three years, since S7 rolled out, but it looks like that kind of high-dense chipsinked package is becoming everyone's de-facto standard (except Avalon, thankfully).

There's no good reason not to build KW/sub-KW miners. Site density stopped being an issue back when people started using shelves instead of racks.

1.4KW is already difficult in that case size, without adding half again to it. Especially since all evidence points to smaller nodes having more issues with heat density in the dies, so the chips probably can't be allowed to run continuously at 120C die temperatures without a high risk of roasting before the machine's expected end of viability.

And assuming it doesn't fall apart or catch on fire a long time before it pays itself off.

10% monthly increase on BTC sure would be nice, instead of the 20% overnight dumps we've been getting lately.

Forum user "Optimizer" makes everything here: http://www.price-technology.com/bitcoin-mining.html

I'd start by talking to him.

Not necessarily. Some people just might not want to (or be able to) operate a single 2KW miner.

Unless you have children, any children getting hurt would be trespassing anyway so sucks to them. Same reason I don't understand requiring extensive fencing around a backyard pool to keep other people's kids from drowning themselves - discipline is supposed to do that for me.

Anyway, I think the idea is to have a small enclosure (a doghouse was mentioned?) probably right at the panel, to reduce the need for inverters or long wire runs. Grid tie negates some of that, but if it's not required all you'd need is internet access and nothing else travels more than a few feet. Depending on how panels are rigged, voltages could stay relatively low/safe.

Just don't forget that the support-component costs scale per miner, not per chip, so tacking on controls and power regulation for "a few chips" costs roughly the same as for a few dozen chips which gives the "few chips" miner a much higher initial cost per hash.

You know I'm not a fan of super high power miners, but there are practical limits to how small a miner can be and still be cost-effective.

If it's still mining, it hasn't failed. It's just an issue with the voltage level talking to the USB chip being low enough that it trips the threshold on the LED flash circuit.

"Seconds" means they failed regular testing at stock settings but passed at a higher voltage. Right now I have none available, but prices have come down so much that a regular stock stick costs about $3 more than a Seconds stick last time I had any.

The new sticks are still on the workbench. Getting them hashing was the easy part, so I wanted to integrate a few extra features and that's adding to the R&D time. I'm working on firmware right now that'll communicate with cgminer for temperature monitoring and software voltage setting, and once that's tested we'll finalize the board layout and run some prototypes. A production batch is, absolute best-case, at least six weeks out. Also just between you and me, 20-60GH looks to be a very conservative rating.

Here soon I'll probably be running a deal on 2Pacs to clear the stock shelves and bring in some funds to help pay for the first batch of the new sticks. Need to get rid of those things.