|
You can send those other S7s if you want. I got room for 'em.
|
|
|
|
|
For those of you who don't know me (which is expected, since I don't altcoins in general), I have built a small business around supporting the mining community and like finding ways to help home and small-scale miners get an edge against the takeover of industrial operations, including affordable power hardware and hosting. These days a lot of small-scale miners are running GPU rigs, so it's probably time I made good on some projects that have been rolling around for a while to help you guys out.
I work a lot with server PSUs and breakout boards. I also make cables. I have recently expanded my cable-making abilities to do things faster and with wider variety. I can now make 6+2 cables, PCIe splitters, and 4-Molex which I hear are useful for GPU powered risers. It is my assumption that, if someone wanted to run a GPU rig off a server PSU, it'd be handy to have cables with 6-pin on one end and 6+2 on the other to power cards directly without adapters. It'd probably be handy to have a 6-pin to EPS 8-pin native cable for motherboard 12V power.
I am also working on a small board which would draw power from 6-pin jacks and fan out into three 4-Molex cables for plugging into powered risers, integrating a 5V converter and with entirely modular cabling.
I might, if there is enough demand, develop a small picoPSU-style board with 5V and 3.3V converters (probably 12-15A each) that takes in power from 6-pins and attaches directly to ATX24 connector for main motherboard power.
If there's something y'all GPU guys think needs to be done or something else that would make things easier for you, let me know and I'll see what I can do. If any of this interests you, hit me up and we can talk numbers.
|
|
|
|
|
Bump for DPS1200/800 boards are in stock.
I'm also working on prototypes for some hardware to assist with building GPU rigs powered from server supplies (no ATX required), which includes custom cabling. I can do 6+2 cables, splitters and adapters, and 4-Molex connections for powering GPU risers.
I also will soon have prototypes for the new Dell 750W board, which are made to take PCIe 6-pin jacks for quick cabling and can be paired to load-balance and operate from a single on/off switch very easily. This provides you with probably the cheapest reliable/reputable option to run an S7 or S9 off 120V power.
|
|
|
|
|
I sell a board which works with both HP Common Slot and DPS-800 and has 12x PCIe 6-pin.
|
|
|
|
|
Share diff only affects what shares are sent to the pool and has no effect on the work being done. The chips idle when work is not sent from the server and cgminer has nothing to feed them with. Bitmain chips (at least up through BM1384) return all shares diff 1 and higher and cgminer filters out the results based on what the pool is asking for.
|
|
|
|
|
I could see it being an issue from mining starting and stopping instead of being fed continuous work, if the boards cool down and drop out before mining kicks back up. That shouldn't be affected by share diff but by availability of work.
|
|
|
|
|
So when you say "bitcoin spinoff" what exactly does that mean?
|
|
|
|
|
For reasons which have probably been explained extensively, it only works with 135-chip S7 and not 162-chip S7 because you're adjusting the output voltage of the main regulator on the board and with 162-chip miners there is no main regulator to adjust. The same applies to the S5.
|
|
|
|
|
It's probably worth noting that, in addition to hosting, I can hack S7s down to a lower voltage/hashrate for more efficient running (in the neighborhood of 4TH/950W) and I can provide PSUs for them.
|
|
|
|
|
23 days of real warranty is about 23 more days than most people were getting anyway.
|
|
|
|
|
Not sure what all those "#" are for, but it makes me read your reply in the voice of a teenager.
Buck converters do not jump up the voltage.
When you talk about the bucks frying, do you mean the bucks in the DPS2K or in the miner? Because I've load-tested a DPS2K at 196A total output for seven days in July heat with zero problems - not bad for a PSU rated at 165A. I've had them running 110% rated output for six months straight with no failures. They're friggin' tanks. If you mean the bucks in the miners, well, the output current is completely independent of the input voltage; the RMS inductor current can change a bit as you shift the input voltage closer to the output voltage but I think it actually decreases the peak inductor currents so the RMS actually gets more favorable and less heat is generated.
|
|
|
|
|
I don't know anything about xnsub, but I do know there's some very simple and straightforward build instructions for cgminer posted a couple times in this thread as well as in the README files bundled with cgminer source.
|
|
|
|
|
300MHz across three sticks should see 50GH. Ideally the heatsinks won't be too hot to touch. The real key to how fast you can push them is how cool you can keep them. There are people here who had sticks running at 400MHz for months, pulling up to 12W, because they had good cooling. The stick itself can probably handle more juice than your hub can - provided you keep it cool.
|
|
|
|
|
300MHz for me usually requires about 7W to run stable, which means 5V 1.4A in. 225MHz or so should be able to run on 0.9A but you'd have to have a pretty lean chip to get much higher. Every stick is tested for stability at 200MHz before shipping out, and pulls about 850mA at test voltage and that frequency.
|
|
|
|
|
Remind us again, what software are you using and what frequency are you trying to operate at?
|
|
|
|
|
It actually is kinda annoying to read the same [mostly incoherent and thoroughly profane] conversation in two different threads. You'll have a lot more people paying attention to your message if you're less of a mudslinging jerkass, especially to veterans.
|
|
|
|
|
3-phase is pretty nice if you know how to manage it. I had 110KW of miners and general shop circuits running off 208V 3-phase (120V line to neutral) with only 6A of neutral current last time the utility guys came out to check. But that's not a n00b installation either, like this guy would probably end up with.
|
|
|
|
|
The question is impossible to answer without knowing your current limits. 100A per phase? 200? 50?
If you don't know enough to know how to solve this question for yourself, I'd seriously reconsider spending money on... actually, pretty much anything.
|
|
|
|
|
It probably wouldn't be stable at stock voltage, but with the right tools stock voltage isn't really a limitation. S7LN boards should cool more effectively than stock S7 boards so, as far as power and heat goes, 700MHz should work better on the LN if you get around the stock voltage.
|
|
|
|
|
Show Posts
