The only miner I've paid up-front for (not counting a handful of USB BEs) was an ASICMiner Blade about August 2013. Everything since then was paid for with mining revenues or related handiwork - repairs and upgrades. I can't say all individual miners turned a profit, but mining as a whole was profitable for me until I started putting everything into buying museum stuff instead of viable gear. The only thing I've been openly burned on was a pair of BlackArrow X1 ordered from Minersource, traded up to a pair of Technobit HEX4M boards, which have yet to be delivered or refunded. Not like that's a perfect trifecta of ripoff artists or anything...

I never bought into BitFury gear back in the day, it was always way too expensive. Even before I had industrial power rates at the shop and was mining residential, my rate was below 9 cents so extreme efficiency wasn't really worth it to me - plus I kinda started with about $300. I have some boxes of Bitfury stuff on the museum shelf right now, bought a few HexFury USB miners to compare with my Compacs, got some OneString boards still new in packaging, but that's about it.

In case anyone's looking, I currently have about 30KW of power available. $65 per kilowatt per month (8.9 cents per kilowatt-hour, to be precise), paid monthly, no contract required. I can provide PSUs for about any miner.

That sounds about like it's supposed to behave. Glad you got it sorted.

Okay, now I understand what you were looking for on cgminer. Yeah I had the same problem a couple weeks ago while setting up test stations for production 2Pacs.

I've never used BFG so I honestly don't know what you're asking for. CGminer runs pretty much entirely on command-line flags with a basic menu as documented in the README files.

I recommend purchasing a USB power meter that'll display how much current the stick is eating while running. Sounds to me like it's drawing way too much power. For reference I have a stick here cooking at 4.2W and it's too hot to hold onto for very long but not too hot to touch. Appropriate stable voltage for 100MHz should not be uncomfortable. Given the stick is running hot but also throwing errors, I'd consider the voltage is stuck on too-high and it's been run hot for too long so the ASIC is just about dead.

600mV is about right for 100MHz up to maybe 150. Also I misspoke - they're tested to 200MHz at 640-660mV before shipping. A Vcore of 0.6 at 200MHz will return nothing but errors.

Regarding grounding, it depends what you mean by "external reference ground". As long as it's tied to local ground of the stick, it's ground. The Reset pad just need to be taken below about 1V and the chip will reset. This isn't tied into hashing at all though, so it's entirely possible (indeed likely) it won't reinitialize correctly and start hashing.

Heat means power. Power means voltage times current; voltage is the pot adjustment, current is proportional to operating frequency. If it's too hot to touch at 100MHz the voltage is set way too high, because at optimum settings 100MHz should be about 2 watts.

I can't speak for used hardware, but when sticks ship from me they're tested to stable operation at 150MHz, usually voltage set between 640 and 660 millivolts. Over 660 and the ASIC gets replaced because it's running too hot. Nothing ships if it can't hash at <30 errors per hour and <660mV. If your stick can't do that there's probably something wrong with it.

Also, there is no "seller GekkoScience" on Amazon. GekkoScience (me) is the manufacturer but has never directly sold them on Amazon. Far as I know, about the only Amazon seller I've directly supplied was BitMiner Joe, aka CrazyGuy here on the forums.

Not really sure what you mean by "bottom end". I believe VH has programmed the driver to take down to 6.25MHz - about 340MH. By that chart 15GH would be about the 5W point.

Bitfury has reassured me several times that I'll have access, and they've actually shipped me sample chips twice. Punin mentioned on the forum last year that I'd get sample chips and I hadn't even asked for them yet.

Regarding a full run, if I can solve some of my manufacturing problems I'll be a lot closer. And you know I won't be self-mining up the butt because I only have 100KW of datacenter available and most of that is farmed out to hosting customers. Y'all have always been my priority.

The first handful of sticks are shipping out today. I'd have more done and tested, but my minion left after 3 hours during which time he only got about 40 minutes worth of work done - 40 minutes at my pace, anyways. He's still learning the benefits of batch processes and power tools for making short work of large modularizable tasks.

I'll be queueing on an 80/20 rate, which means 80% of manufactured sticks will go to big orders and 20% to small orders. Right now that's pretty much how the breakdown is. So that means I'll ship the first 40 sticks to small orders and save the next 200 for a big order, and so on. If I can get the manufacturing pace picked up to near where I'd like it there'd be about 240 moving every week, but that's going to mean tweaking processes to get rid of the metric buttload of manual rework currently required.

I'll be working a fair bit over the weekend to get more sticks out, and ironing out some kinks with the new testing setup that's causing more issues with output. The next few orders on the list will be shipping out Tuesday.

The Bitshopper 2Pac PCBs are in, so I'll get started manufacturing those before long. The same order also had Biggie pod prototype PCBs, so maybe I'll get to building one of those soon and seeing how it works.

The biggest problem I'm having with 2Pac assembly is paste application and wetting on the pads of ASICs and other fine-pitch QFN chips, which thoroughly dominate the complexity of the Biggie. So I'll have to work on the assembly process to solve that problem and make sure the minion currently in charge of running the pick-and-place does it right.

First orders should be going out tomorrow. These things are taking a lot longer to iron out than I expected, mostly because of issues with parts placement. I've got a minion running the pick-and-place and I think I need to give him another run-through on how to do things right. Paste isn't wetting properly so parts aren't sticking well - part of that is it's been cold lately, and part is the paste isn't being stirred right before application so it's too stiff, and part of that is I think he's just not laying it down right. Yesterday he somehow gouged the stencil so badly it took a chunk out of the PCB.

I did some figuring this morning and changed the program a bit to help out with one issue I was seeing on the board, but about two dozen sets (one "board" is 10x sticks in a panel) had already been run through so those will have to be reworked manually. And then a lot of ASICs are not stuck down right, which if I can figure up a solution for that or just do it all myself I'll end up re-doing about 20 panels but there's already 30-odd already finished I'll just have to deal with. Long story short, daily total output is half what it should be until I can solve the problems. That's a bad thing for everyone.

Good news is, the first orders should be going out tomorrow. I'll probably queue things 80/20 since that's about how the orders are lining up, which means in every 100 sticks 20 will ship for small orders and 80 for big orders.

I don't think he was recommending you start your own pool. There's a bunch already operating.

Right, I've got a whole box of 470uF 250V pulls from old supplies. I got a box of retired (working) Dell server supplies I could scrounge some 330uF 450V Nichicons out of if you're interested.

I see the point of reducing losses, but as noted earlier, the nonlinearity in your rectifier is gonna give you a pretty terrible power factor. Course, for just a test, that's probably not too bad. I'm looking forward to your results with an Active PFC load. I actually suggested something similar to a guy a couple years ago who was trying to figure out how to chain S5s onto a 277V line.

Not sure, been too busy with other stuff and haven't fired them up yet. I was hoping to get working prototypes built in January but had a lot of stuff piled up and no help. As soon as I have results, believe me I'll make it known.

Yeah, S7 and S9 are stupid density. The only way I'll build a >1KW miner is if someone takes a bunch of my 250W-max boards and puts them in a rack case.

No, because the S7 and S9 have no hardware in common with the older versions and run in a power density I'd rather not work with.

I've got assurances from Bitfury to have access to new chips and have been given a bunch of samples already.

I'm gonna go ahead and say no. I mean in general yes that'd probably be the case, but not this time. The S1, S3 and S5 all use the same mechanical parts but different controllers, so if you built it compatible with S5 only you'd have to buy a new controller for the S1 and S3. That's what the other guys are doing, I think - or are they making a whole new custom controller for all of them? I forget. If you used the existing controller then you'd have to reverse-engineer the hardware of it, which includes an FPGA for port multiplexing and a specific IO interface that might not be signal-compatible with your chips, so you'd have to put something onboard to translate your data anyway. I'm trying to build around a generic USB interface and open-source cgminer so you can use S1, S3 and S5 chassis interchangeably and almost literally any controller you want, minimizing the total amount of proprietary stuff that causes warranty concerns later on. My design principle is "simple, durable, reliable" - which means I like interchangeability but not complexity.

That's just the way I'm going, anyway. The basic features of my design have been codified since December 2014 or so, I just haven't had the resources to make it real until this year.

Ever think about just stealing the passive PFC rectifier/cap portion out of an old ATX supply? Seems like that's most of what you're trying to build anyway, and around a 400W supply should be input-rated for your draw.

If by "the sticks are out" you mean "some sticks not advertised in this group buy are out" then yes you are correct. If you mean BF16 sticks are out, not so much.

So you're proving vapourminer's idea can work. It'd be better with a much shorter cable.