If it picks up in cgminer-gekko but doesn't mine (or spits only errors), start turning the voltage knob clockwise to increase the voltage. At some point it should level out. Take a moment to go over the troubleshooting tips in the first post.

No, the wire gauge allows you to run the higher current required for a 1600W PSU. You'll be looking at 7-8A, which you'd want minimum 16AWG (and 14AWG is better) to keep it from being a fire hazard.

Unlikely that any given rail is at least 400W? That's what you'd need for a single board to avoid stuff exploding.

Yep, sub-500W miner would be nice.

It really does depend on what you're plugging into. There are a dozen different outlets for 240V.

I usually look at Iron Box Electric's eBay store for cords. I know that doesn't help the desire for Amazon Prime but they got a wide selection; about a third of my hosting (on 208V) is running through miscellaneous cords I got from them.

Given how late they were saying they weren't working on 14/16nm, Christmas might not be a stretch.

I would guess they keep hashing on whatever work was passed in last and then fall idle whenever that runs out. But they're doing it with no fans, and whatever power draw they pull at idle, all that heat would still be inside the box. Sounds like Tupsu has been on the receiving end of a problem like that with destructive results.

Have you read through the first page of this thread?

BFG doesn't require WinUSB driver, and for the hundredth time 5.4.2 doesn't use hex for the frequency.

The cgminer-gekko Windows download is missing about an 8MB DLL file that can be copied over from a stock cgminer download or elsewhere. That info's in here too.

If your stick is set at 0.5V that means it's broken. It should probably be set more like 0.65V, and the entire adjustment range is from 0.55 to 0.8

NFW - I was rocking a similar roll-it-forward from about September 2013 through summer 2014 - started with a $400 AM Blade purchase and turned it into upward of 3TH by about June without any additional outlay. But that's when I shifted from pure mining to collecting interesting hardware. I traded a bunch of S1 off for an S2 even though I actually lost hashrate because I didn't already have one. I think the S3 in September 2014 was the last new miner I fetched until just the other day. Had I not switched gears and kept rocking maximizing hashrate I'd probably be upward of 100TH by now.

NFW, how much of your farm is on free power?

Biodom - Actually I was just winging it and using numbers that would backtrack to S9 purchase price for comparison. But using dmwardjr's mining return numbers...

So an S9 would cost me a quarter coin in power if coin is at $1500, leaving me 1.95BTC from the 2.2BTC mined if I roll the entire power cost forward each month for a recompense in December. Though unrealistic, that makes the comparison easier (and also the most optimistic).

So I should buy 1.95BTC right now for about $1150 and cash it out for $2900 at the end of the year for a net profit of $1750?
Or should I buy 3.6BTC right now for about $2100, buy an S9, pay $360 in electric over the next 6 months, mine 2.2BTC and cash it out for about $3300 at the end of the year for a net profit of about $800?
Or should I buy 3.6BTC right now for about $2100 and cash it out at the end of the year for around $5400 for a net profit of around $3300 and no monthly expenditure?

If I'm paying monthly for power on an S9 by selling the required amount of coin monthly, I only actually keep around 1.86BTC by the end of it, so I'd cash out for about $2800 and a net profit of $700. Or I could have bought 1.86BTC for about $1100 and cash it at the end of the year for a $1700 profit.

Your numbers aren't very favorable for buying an S9 fresh. If an S7 runs 1/3 the hashrate it should net 1/3 the coinage, so about 0.73BTC which is currently valued at around $430. I can get an S7 aftermarket for around 0.5BTC and spend 0.25BTC on power (if I roll all my costs forward to recompense in December) for an ultimate breakeven. Or I can keep the BTC and cash out at year's end for a $450 profit. Dang, as much as I like the idea of $1500 coin those numbers really aren't encouraging at all for miners.

I've got about half a dozen S7 running in hosting on dual Dell 750s. The difference with a new batch will be the option for 6-pin cabling, which people seem to like because, though far less versatile, it's quick and clean. I'd probably be selling these boards out the butt if I had any for the last six months, but things have been tight around the shop between Compacs and DPS8/12 boards and people waiting a month and a half to pay their bills. Money definitely does not grow on trees around here.

A 208V-fed DPS1200 will run a 135-chip S7 on stock settings. Runs hot, but it runs.

Good schematics would be nice. Heck of a lot easier than reverse-engineering, and cheaper. Tomorrow's task will be hacking away at Avalon stuff since they didn't provide an A3218 datasheet. After that will be hacking away at an S7LN board. Making things work and making things work better are always fun.

So if coin increases to $1500, should I buy 1.4BTC right now for about $800 and cash it out for $2100 at the end of the year, or buy 3.6BTC right now for about $2100 and mine around 1.4 BTC by the end of the year to cash out for $2100?

I've had interest in something I was already thinking of doing, adapting the 750W board to take screw terminals or 6-pin jacks like the new DPS8/12 board. The all-in kit I would provide would be two boards with 6-pin jacks and 11 cables - one about 6 inches long. There'd be a jumper wire in there to tie the SHR pins together to load-balance the two PSUs, and you'd use the short cable to jack the two boards together and establish a common rail. That leaves five jacks per board just right for an S7 or S9.

I could build a single dual-PSU board, but since the same functionality could be provided with two regular boards and a $2 cable I'd rather not have to split my resources between two different board runs.

Some of them did. Anyone replacing gear has to make sure and only count the resale value of the S7 once; it can't both contribute to S7 breakeven and S9 initial purchase (in S9 breakeven calculations).
I assume you're replacing 3xS7 with a single S9 to get that reduction?

Hm... that is a good point.  It'd take a bit of rigging but you could use the 500W to kick on the 850W, but then you violate Bitmain's recommendation to have the controller power up last. With a bit more rigging you could probably cross-latch them to avoid that problem, but if you know enough to do that you probably also know how to wire in a 240V circuit.

I'm working on a redesign of my Dell 750W board and hopefully a new batch, which with that you could load-balance a pair on a common rail and get 1500W with 10 cables off 120V for about a hundred bucks. Too bad a full production batch is more than a month off.

Or get about an 850W PSU for two boards and a 500W PSU for the third board and controller.

Assumed?

That's only 20K chips, actually not too bad. Probably under $200K, hopefully closer to $100K.

Remind me to continue not using that service...

Spondoolies had simulation results at 0.15J/GH and BitFury was ordering engineering samples for a 0.06-0.15J/GH chip. Given that the BM1384 could run 0.25J/GH but the S5 was clocked at 0.45J/GH and the BM1385 could do 0.18J/GH but the S7 was clocked at about 0.26J/GH; if the S9 chip is clocked at 0.09J/GH we can assume it's probably got about 30% more efficiency hiding in the basement for a bottom clock around 0.06J/GH.

I'd bet that Bitmain and BitFury used all the tricks they'd already learned optimizing designs for 28nm on this 16nm chip. Is there probably room for more improvement? Seems likely. But I doubt we'll see another 30-50% drop in power use like we had with optimizations in previous generations.

I don't really have a timeline, unfortunately, given how full the plate is with manufacture, manufacture, manufacture and dev. But I'll do what I can.