In the last few day, BFL has been making some further progress on their development of BFL ASIC miner. Josh, CEO of BFL posted a video contains a quick power consumption test on BFL’s very first PCB board with ASIC chip. The power consumption is higher than expected but still less than BFL’s competitor, Avalon ASIC. Butterfly labs plans to ship its first batch of ASIC miner first week of April.
Luke, a software programmer who will create the software for the chip and board touched his hand on the first ASIC.
BFL Josh Zerlan recently provided updates on the state of Butterfly Labs’ production in an IRC channel. The core of the conversation is this:
Given this is the largest set of answers we’ve had from Josh in a while, and it’s quite hard to parse the Shoutbox archive – here it is in its full glory: (edits for formatting)
BFL_Josh: Well guys, I had planned on updating everyone with a video of a board hashing here in KC tonight, but I haven’t been able to get that together yet, so I’m probably going to have to push it off until tomorrow. We are targeting a start of shipment next week, but I’m not quite ready to commit to that at the moment, given our past estimates. It’s imminent, though.
Lab_Rat: It hashes?
BFL_Josh: Yes, it hashes
BFL_Josh: I just don’t have the firmware here in KC yet.
enkidu99: nice josh. does this mean you’re able to get the next batch of wafers inbound as well?
BFL_Josh: I have to check on specifically where they are in the process, I’ve been primarily focused on the current batch.
Chip Geek: We need a few in the hands of customers before April 1 for the bitbet wager…
BFL_Josh: I know Chip Geek, I’m going to try to ship out a unit ,at least one, before APril first if it’s at all possible.
BFL_Josh: Imma pull an Avalon all the way, it’s a proven method!
Brian Harris: since the existing batch has been pushed back does that mean the subsequent batches will be moved forward?
Barrett: will there be any increase in the hash/rate as an bonus…
BFL_Josh: Probably not
Chip Geek: Still within the +/- 10% of the advertised spec though right?
BFL_Josh: I’m not sure. We may miss our power targets, that’s been part of hte hold up… we think there’s a problem with the power consumption and we’re trying to figure out where it’s having an issue. However, in the interests of time, we are going to be shipping what we have and going back and fixing while we are shipping, just so people can have their units. People are welcome to wait for the revised boards that should help out with the power issue if they want to delay their shipments… (heh, riiiight).
BFL_Josh: The power is still far less than any other unit, so it’s not like it’s something crazy or anything, but it’s not 1w/GH and we’re trying to locate the source of the power drain. It may require a revision of the substrate and/or the PCB.
BFL_Josh: Haha it’s WAYYY less than 620 watts even at it’s worst.
BFL_Josh: What’s causing even more consternation is the fact that the wafer we burned for tests runs at far less power than a second wafer we mounted on the BGA package… so it may be a wafer by wafer thing, and sicne we only have two datapoints, it’s hard to nail down the issue. We’re rolling the other 4 wafers off the line shortly (they may be done already I think) – and that wil at least give us an idea of the power consumption between wafers with 6 data points.
Buck’O: if we don’t care about the power suck can we just go ahead and get ours ahead of those who do care?
enkidu99: josh would you consider swapping these units for “power optimized units” if we took them as is for now, at a later date (even for some fee)?
BFL_Josh: I don’t know for sure, but probably.
Bogart: You guys didn;t already begin packaging the rest of the wafers?
BFL_Josh: the rest of the 6 wafers, we have been holding off on the last 5 layers for the rest of the chips to be sure we don’t need to make a tweak in the metal layer due to the power issue. I think we’ve pretty much settled that hte power issue is NOT in the chip, but somewhere in the substrate or PCB (or if it were in the chip, it’s not something that’s easily and quickly fixed).
MrTeal: LOL… ANy chance you can ship out the first 5 units before the 1st, just for fun?
BFL_Josh: I’m going to try, not guarantees though. I would love to see that bet won, believe me.
Brian Harris: If you’re just riffing about dates, to get orders through the middle of februrary filled, is it going to be like a 1 month gap from the initial batch?
BFL_Josh: it’ll probably be more like two weeks.
Barrett: WILL the single still go 60gh// but with more power and heat..?
BFL_Josh: Depends on the power draw, it may be that the power draw is too high for the board and we need to scale the hashrate back. If that’s the case, we’ll send multiple units to cover the hashrate difference for orders.
Lyke: Do all the devices have the power issues or just specific ones?
BFL_Josh: All devices, Lyke… they all use the same board and substrate.
0x3d: So how overclockable will these devices be then?
BFL_Josh: Well, that’s a good question 0x3d: if we have to scale the hashrate back due to power, the chips themselves may end up more overclockable, but I honestly don’t know at this point.
Steve Honaker: @Josh if shipping could start next week when is our advance notice to send in our FPGA’s?
BFL_Josh: Steve: I should know more on that tomorrow [29 Mar]
SgtSpike: What’s the current power usage?
BFL_Josh: We are trying to nail it down. We have some boards at 1.76w, some at 2.5 and some at 4 and one at 6 <- although I think the ones at 4 and 6 have either a broken power regulator and/or a bad power brick, because they power brick is only 120w and it’s drawing like 195 watts which is ridiculous. When I get hte firmware here in KC, I’ll be running it off of an ATX PSU to eliminate the power brick as the culprit. So you see, this is why the power issue is causing some difficulties.
BFL_Josh: Yes, I’m aware that Bitcointalk trolls will have a field day with everything I’ve just said. Heh. Even a broken power regulator and a bad power brick is still better than an Avalon, but in all honesty, at this point with bitcoin prices at 90 bucks a coin, I’m thinking no one is going to care at hte moment they just want to hash. Which is why we’ll be shipping and then improving as we go.
b1nary1sfun: well that’s all we can ask for….but how many will ship, still the 1/3 plan?
BFL_Josh: Yeah, probably… although the 1/3 plan is just dumb in this current climate, but ti’s what we’ve committed to.
Bogart: The cooling and the power supply stuff and the traces and everything else will need to handle the extra power.
BFL_Josh: If the cooling becomes an issue, like I said, we’d scale it back and ship multiple units.
polrpaul: summarize the update for us, please
BFL_Josh: Dude seriously, chill out. I’m giving you what I can as far as information goes right now. We are not hashing with anyones units and we will ship as soon as we have product.
smracer: So you will get 2 30Gh machines for each 60?
BFL_Josh: If need be, yes.
X: josh honestly when did you think we will start seeing some units leave the factory?
BFL_Josh: I’m really thinking next week, but I will not commit to that until I have some more information.
polrpaul: is hte test rig sorted?
BFL_Josh: Yes, the test rig is sorted.
smith88: I hope you guys get the bug ironed out and do not have to ship multiples. You’ll get it; a year from now everybody will look back and see history in the making.
BFL_Josh: We will ship whatever we have to to get people hashing at their expected GH.
Lab_Rat: well then worst case scenario for a single would be 360W???
BFL_Josh: No, worst case is 195w, but the power systems on the board aren’t rated to handle that really I don’t think, so we’d back off the hashrate before we’d let it get that high.
Lab_Rat: ok thank you. they can handle up to what 120W?
BFL_Josh: Yeah I think 120w is a nice comfortable number. It can probably got a bit higher than that, but 195w is probably not where we want to be.
onryo: Wont a x4 unit blow the power unit?
BFL_Josh: Yeah, so we’d have to back off the hashrate on a per unit basis if that were the case.
BFL_Josh: We’d do like 2x 30GH/s singles in stead of 1 60 GH/s one
BFL_Josh: We will ship the purchased hashrate regardless of what it ends up costing us.
MrTeal: @BFL_Josh, do the LIttle Singles have all of the power supply circuitry populated, and could they run at 30GH/s even if the Big Single can’t run at 60GH/s?
HTL2001: I think he means, if you ordered a big single they might do 2x little singles since the power limit is per board, not per chip
BFL_Josh: [HTL2001 is] correct
SysRun: Asking again, what are the remaining steps?
BFL_Josh: Just mounting, assembly and shipping.
smracer: So people that ordered a minirig will get 2 minirigs at half the hashrate?
Crystallas: Are you able to share how many orders are currently committed?
BFL_Josh: No, sorry.
kaega: Did you say and estimate on the production line output? 100/day? etc?
BFL_Josh: I estimate around 400 a day on average in the beginning just to clear back orders.
X: josh do we have any idea of if we will reach sept or aus orders in this batch
BFL_Josh: In the first 5 wafers? No, I doubt it.
jjiimm64: Josh, will the paired down units still have all the chips in them?
BFL_Josh: Yes they will
kaega: thanks for the reply. How long to get caught up (estimated)? By June? I placed additional order 3 weeks ago or so.
BFL_Josh: I should have a better idea on that in the next few days.
Barrett Griner IV: Will BFL work through Easter weekend…..to get units out…
BFL_Josh: Yes, if need be
onryo:if you guys could make a rig in the 10k USD range a lot more people might be intested….
BFL_Josh: I agree 100%. I have been floating that for awhile, but we’ve been so focused on otehr thigns, we haven’t gotten down that road yet.
oldbushie: so… 300 whats a day? boards? chips? shipments?
BFL_Josh: 400 completed units a day
Barrett Griner IV: So at this point the 8 Chip single will now be 2x 4 chip singles….will that continue….or til the bug is fixed…
Ivan Frimmel: BUG?
BFL_Josh: There is no bug
BFL_Josh: Sorry I don’t have the video guys, I really wanted to float that tonight.
Lyke: would have been great to see video – Guess we’ll see tomorrow?
BFL_Josh: Yeah, tomorrow should be good I hope!
SysRun: Josh, every time you use the word ‘should’ you are now required to put a bitcent in the ‘should’ jar. BFL_Josh: Not a bad idea