Gonna be watching this thread, in case I can magically afford it in the next two days. Gonna depend on what my pool does to be able to fund it, but turning two X1 into 1.2TH would be pretty cool. Should be able to scrounge up some cheap waterblocks and tie into a big watercooling loop too! Exciting!



If they're looking at pulling no more than 400W per board, you could get away with two boards per PSU but three per two PSU load-balanced would be both safer and more efficient (as each supply would be operating closer to its peak point). Hate to be a shameless plug but we (the guys that make the Dell PSU boards minersource sells) should be started moving DPS-2000BB boards in about three weeks, which can source about 2500W reliably but require 200-240VAC input. There's a preorder group buy https://bitcointalk.org/index.php?topic=637595.0 in case anyone's interested.

Oh hey, this. Yeah we'll start moving DPS2000 interface boards within 3 weeks. We do still have a bunch of the 750W kits still around, which should run a pair of S3 without issue.

We don't have any sample boards left to ship out, but I'll see about putting some mechanical specs up on the product page on our website.

Indeed.

Additionally, anyone emailing about preorders, if you don't receive a response in the next day or so it's probably because for some stupid friggin' reason our server's been blacklisted (I think by Microshaft this time) and your host isn't taking our messages. Used to be an issue every time DNS updated, but we're on a static IP now so I don't know what they're complaining about this time.
If you're worried about a pending order status or anything, send me a PM and I'll let you know the score.

That's why we're gonna start stocking the 36" cables, to help power distribution setups just like that.

PCBs should be here on or before the 23rd, and we've got our sources pretty well arranged for the rest of the parts so there should be no delays getting the first batch preorders out. As always, preorders are appreciated because it'll help us get materials in larger quantities, which means faster more reliable turnaround for manufacturing and also building up standing stock for new orders. Right now I can only guarantee that preorders will be fulfilled; we haven't met the threshold yet to be able to complete the entire first batch. I don't like having to do it that way, but parts aren't free and we're a small outfit with not a lot of capital.

They'd probably do 5 or 6 without issue. We had four S1 and a pair of custom-hot Cubes on one, pulling just about 2500W for a solid week without even blinking. How many depends on what they look like overclocked, but if the stock numbers on Bitmain's website are to be believed, this thing could probably handle 7 at the top end.

The overcurrent protection in the supply trips out typically around 76 to 80A. I've taken them up that far using a dummy load through one of our boards. Not sure how well they'll sustain that power, but we've had customers reporting running upward of 800W sustained without any problems.

I don't sell anything under 16AWG. I've pulled 27A per cable before with no issues - certainly don't recommend that for extended periods but they can do it. The $55 kit comes with 4 16AWG 18" 6-pin cables.

https://www.bitmaintech.com/productDetail.htm?pid=00020140630025130637RV8OhOwt06BC

Looks like these PSU kits will handle a pair of S3 units with room to overclock. $55 (plus shipping) to source power for 1TH I think sounds pretty good.

Having stock is sorta dependent on being able to get enough of the rest of the parts to build them. Right now I can only guarantee that we'll be able to build enough for the current preorders.

PCBs have been ordered. If we don't get a couple more orders (or some other magic happens) in the next two weeks it's gonna be tight to get the batch done up, but we always deliver on our word. We figure at this point it's better to trust the community that's been telling us for months it wants these boards, than to sit around and wait for someone to come along and just hand us a bunch of money.

So we ordered the PCBs for the first big batch. About 35 are already spoken for, which is a really good start, but more orders wouldn't hurt. We expect to start delivering on July 25th.


We're also prototyping a little board which'll pull from a 6-pin cable and convert to an ATX 24-pin (with 2-3A capacity on 5V and 3V3), even with the ability to wire it up so the ATX_ON signal will kick on our PSU. That'll allow folks to use these DPS-2000BB natively in things which previously required bigarse ATX supplies for blades and controllers. One DPS-2000BB and two of those little boards could replace a pair of 1200W ATX supplies in miners. We'll have some prototypes probably early next week, they're likely to cost around $15 each and would work just as well on the D750 boards.

Are you asking if you can modify one of the 2950 boards for a 2850? It's technically possible but it'd actually be easier to start over from zero because even the pin spacing on the connectors is different.

Pretty sure they use different connectors and pinouts.

I'm also pretty encouraged about the S3, looks like it could be a fun piece of hardware. I don't recall the specs exactly but it was something around 400W?

We still have plenty of boards and full kits available and the sale pricing is continued. These guys should be idea for powering stuff like the RK-Box. Hard to beat 750-800W at 90+% efficient for $55.

I'm an American manufacturer but please don't be scared to buy our stuff. I like to think that how BFL does business is not representative of American business in general (though, realistically...) I look to my uncle's die casting company for inspiration - he and half a dozen other guys left a crappy outfit because the management weren't allowing them to make their products better, so they started their own business and within a few years were doing more work (and better) because they actually put focus on doing the job right and taking care of the customer. That's what American business is supposed to do. Thanks for crapping the rest of us, BFL.

So, update. Large Potential Customers are dragging feet so we got fed up making everyone in the world wait and we're gonna do what we can to start without them. We're looking at assembling a batch of a couple hundred boards to be delivered in about 4 weeks. We're going to be doing Non-Current-Sense boards only, because we haven't had enough interest to merit current-sense boards (economy of scale and whatnot, you understand). We'll be gathering up funds for the next couple days and starting to place orders probably on Friday for PCBs. This'll probably be a 3-week turnaround so we should start shipping boards within 4 weeks to preorder customers.

Anyone that has requested information on preorders but haven't yet been able to give us money, it's probably because I still sorta hate preorders and wasn't terribly motivated to take money from people with no real guarantee of a delivery timeline (because of LPCs and whatnot). But we now have a fixed plan and we're gonna start rolling with it. We won't be offering CS boards at this time, but we will be gearing up for NCS boards and both lengths of cables in the very short term. I'll do my best today to catch up with everyone interested in preorders, and anyone else I may have overlooked or who was worried about delivery timelines, we Will Be (barring an act of God) shipping DPS-2000BB boards before the end of July.

Ah, groovy. You guys thought of everything.

Does the Habanero have input voltage monitoring? 12.8V is less than 7% out of spec for 12.0V, which most things should handle at least 10% tolerance.
I'd consider it might be an undervolt issue, either from bad or insufficient output capacitors in the PSU or from a poor connection in the edge connector. I prototyped a DPS-800 board but always had trouble keeping the edge connector in solid contact with the PSU output contacts. Could be when the VRMs fire up they're draining their input caps and putting a huge sudden current load on the DPS-800, which is undervolting and tripping from one of those reasons. It could be the PSU is providing ample current but the cables are either too thin or too long, or poor connector contact, such that line impedance (either from high resistance or inductance) is not allowing sufficient sudden current changes to make up the instantaneous load. I'd suggest tucking an at-least-1000uF capacitor on each 6-pin connector (slide it in to the contacts on the back of the cable-side connector) which will help buffer out brief current spikes like that at the device instead of at the PSU. If that's the issue it ought to help, otherwise I don't really know.

Not to be a shameless plug, but that's why we design our PSU boards with additional output capacitance and low-impedance screw terminals, and supply 16-AWG cables for good current handling.

I built a 300A-capable watercooled adjustable active dummy load in my basement in a couple weeks for stress-testing our 200A-capable PSU interface boards. Mind you it's designed to operate at 12V nominal but would probably work below 1V (I think the main limitation is low-side current sense resistors limiting my minimum voltage to .625V at a full 300A draw) but if one guy can do that, a multi-million-dollar business can figure out how to do sink 300A. Bolting 20 of http://www.digikey.com/product-detail/en/WSBS8518L1000JK/WSBSA-.0001-ND/2421966 together in series for .002ohm 720W-capable would be a good start for under $200 if you wanted a fixed load. Stick 'em in an oil bucket or something.