did you take this picture on your breast? Is this your shirt beyond?
Wishful thinking  That's my shorts; the bump to the left of the pic is my err... you know. |
|
|
|
With 1% per month there is definitely something wrong with the design of the system, especially for Bitcoin mining where power is constantly on full throttle, tanks don't actually need to be opened often and hardware goes in and out very seldom. Close to zero (as in 0.0x%) losses is what we achieve with our systems, but they are our third generation built on industrial grade and we have spent a considerable amount of engineering since early 2012. And only if you ignore the "spilling" when people handle the fluid, but that's down to human errors. What's also more important is the loss after a hardware refresh, for instance when you retire some 40nm hardware and move to 28nm. When everything is said and done, how much fluid did you "throw out" with the old hardware? We've been working hard to minimize that - and the spilling of course  Getting fluid loss under control is what consumed most of our resources. Large systems would mean large losses if that is not handled properly. (contrary to popular belief we have to pay 3M for the Novec so every little bit counts) Are you running 7100 or what? |
|
|
|
I am using the 16AWG cables that came with the AX1200i; the fault is clearly with the KnC "design" and not the cables.
What makes you think the AX1200i comes with 16AWG cables? I've never seen one with anything other than "type3" 18AWG cables. I've seen quite a few of them as well. I've also had them melt/smoke/burn spectacularly: http://imgur.com/a/UUpEGditto.... the 1200i's we have all have 18awg, while the regular ax1200's we have all have 16awg. The "i" must stand for "inferior" "Need to know" Basics... Plugging in the pci cable fully can be tricky, as most of the cubes have the pcb mounted about 1mm too low, and the plug tends to catch on the bottom of the hole in the cube. In order to get them to fully plug in FULLY, you must slightly LIFT UP and push in on the pci plug after it first gets a bite for it to fully slide into the plug socket. If you miss it, and it fails to fully latch, it may indeed melt; especially if your psu employs 16awg wires. Seeing a single wire on the pci plug begin the meltdown is a sure sign of a poor connection... and if you use a dab of conductive grease on each pci connection, it will also help make good contact, and avoid all this. power delivery is 95% on these imho Plugging and unplugging the pci plugs too much or not gentile can also quickly loosen the connections. I am always impressed with how civil and polite tolip is with regards to you guys. You say "especially if your psu employs 16awg wires" so are you running 14AWG? You are arrogant and wrong wrong (as usual)... I understand that reading is not your strong suit but here is a photo of one of my AX1200i PCI-e cables:  ...and now with zoom:  You can't argue with this photographic evidence but I look forward to seeing you try. u27 |
|
|
|
Yes, the warranty period for your order is 12 months.
I believe that you need to RMA this board, I will send this request over to our RMA team to issue you with an RMA number.
Med vänlig hälsning | Best regards
I know it's 12 months and have tried to return a cube that has a dead die on it only to be told that it's not covered because the machine as a whole is still slightly over 3TH. I have since had a PCI-e socket melt and KnC have publicly blamed power supplies and cables for this (which is not the case). I am using the 16AWG cables that came with the AX1200i; the fault is clearly with the KnC "design" and not the cables. Look at the ant S3 for instance; they have four PCI-e sockets per 350w machine to meet the 100w spec per PCI-e 6-way. KnC must have saved <$0.50 per cube by skimping on power sockets which makes no sense on a $10 to $13k machine. Not a happy customer. |
|
|
|
So they are... well I guess the hf hardware had to end up somewhere didn't it? I would still buy it if they were proved to actually exist and the price was right! |
|
|
|
If knc dont refund neptune buyers then surely they should be investigated by a company trading authority,they just dont seem remotely care from what i have been reading...
from what I understand they don't even acknowledge there IS a problem..they blame the power supplies (yeah corsair 1200i units are so 'not mainstream') usual bait/switch/stall .....must have paid for the BFL seminar ....I mean really ...like people with Neptunes can't read specs recommended on the power plug they used they were cheap but hey with the 90 day warranty now...they can probably skate on some of this till that 90 days is gone and your stuck with it miss the old KNC.....frigging evil KNC twin figging took over the kingdom me thinks Searing You say 90 day warranty, but didn't we all get the 12 months as stated when we ordered? I never agreed to a revised T&C knocking 9 months of the warranty! |
|
|
|
I get this for the site using safari:  You are a mac user so your opinion is invalid; but that said would this even be a wise investment based on free power? At this point you could get a shipping container full of 132nm for that price? |
|
|
|
Site works fine for me
same |
|
|
|
Does "rigorously tested and tuned at BitCrane’s testing labs" sound like "second hand" to anybody else? What's $1580 these days? About four S3 boxes which would be nearly double that speed and nearly half the power consumption per gh? Looks like they didn't skimp on VRM's (do I see 24 per chip?) but 500w per chip is a crap load of 0.7-0.8v to deliver:  you are way behind in efficiency. 1w/GH @ 28nm ?
no thanks.
Somebody got in there quick... 1w/gh is very poor for 28nm; why is performance so poor on a brand new line of hardware? How many chips per miner?
"hash rate of 550Gh/s per chip" this answer does not require complex mathematics. But tapeout who where when, temps (with water cooling) and vid would be good. |
|
|
|
Whats the life expectancy of this liquid?
It is stable so in theory there is no limit; however depending on your setup you will always see some loss from either opening the tank or from leaks, expansion, contraction of the gas in the process. All these things can be reduced with better seals, bellows to allow expansion without venting and so on but it is very difficult to eliminate all losses/ 1% per month is a number I have heard banded around but I have a buddy running this stuff and his losses are higher that that so far. u27 |
|
|
|
One thing is for sure... your Monarch numbers are no longer valid. Spondoolies is unlikely to meet those numbers at teh wall either; more like +10% power consumption. Same goes for the S3 that's more likely their 12v power consumption so +10% at the wall? u27 |
|
|
|
just got 3 wire cable Y style splitters from cablez on here 12 bucks each ...3 of them .with the 6 heavy duty leads...help spread out the load on startup on my corsair 1200i these should solve the problem.....others on here have had to use other then corsair 1200i units to get the neptunes up...will see how this works I guess for my use on a Titan should I choose to use this unit.. probably not a bad mod to have anyway to use on anything in the future...should get around this whole startup issue with load and the corsair 1200i power supplies if this was to apply to the Titan (will likely just be a coward and use whatever KNC recommends in the future and use this corsair 1200i for something else....coward that I am....)
so it goes good to know you had no problems
Searing
*yak yak, 1200i good, EVGA bad, yadda yadda, some comment i accidentally deleted and can't be bothered to retype... does anybody have it to copy paste or undo my edit?* Funny thing is, KfC can't now fix the PCI-e power problem without admitting that it exists and offering compensation. u27 is right...
I like this guy ^ he sounds like he knows what he is talking about. |
|
|
|
I can't be that "lucky"
You're not lucky, you're failing to measure it correctly. The chips run 300w or so each stock; how do you think you're getting from 220v AC to 0.7v DC with 220w/<15% total loss? Fairy dust and magic beans? u27 |
|
|
|
1450w stock settings 5 cubes or 1 whole neptune 3.45Th/s
Are you *always* stoned? Is that it? The above is exceptionally optimistic; if your cubes run stable on the stock 475Mhz then you're looking at 0.65TH per cube. Most people are able to get 0.7TH per cube by dialing in the clocks and voltages (very simple to do in stock web interface). The chips run at around 300w each but that's at 0.7v after VRM's which add around 20% inefficiency; so you're going to be pulling around 1800w from your 12v supply. With a half decent PSU you're looking at 1900w at the wall for 3.25TH = 0.58 w/gh or with minimal tweaking 2000w at wall for 3.5TH = 0.57 w/gh. That's if KfC don't screw you with dead dies that they refuse to RMA and your PCI-e plugs/cables don't melt. u27 |
|
|
|
Pretty much nailed it dogie, as usual, but I wouldn't have given KfC 7/10 communication from my personal experience; 9/10 when I ordered last year but now 4/10 at best. Correct they refused my refund (requested before any were dispatched) and also refuse to RMA my dead die. Not to mention the fire hazard. u27 |
|
|
|
For those in my boat with a corsair 1200i they may have considered for use on the Neptune (don't work see post below) and or the Titan (see post below)
I got this reply from Kurt at KNC Forums...so definitely don't get this for the Neptune and probably not a good idea for the Titan either
again for people in my boat who may have one of these lying about ..they 'thought' they could likely use with the Titan...er not looking likely now imho
Kurt's reply to my question on the corsair 1200i and its use with a Titan...1st line is in regards to the Neptune...2nd line is in regards to the Titan....
I quote:
"To answer your additional questions, the 1200i is a multi rail power supply that cannot handle the frequency set at 425MHz, there is no fix for this as the power supply is inadequate.
We do not know the exact specifications of the Titan miner yet, so it is best to wait until final specs are released."
end quote:
Anyway just a heads up to those having these corsair 1200i units they planned to use with Titan (or Neptune) ...and/or to those who were gonna buy such.
Probably that would be a bad call imho...
anyway for your consideration
Searing
I'm running 1200i supplies no problem what-so-ever getting enough power to ASIC with Y splitter cables; they will even run single cable if you set 40A per rail in the software. Also they run very efficiently with two cubes per PSU. It is not a multi rail PSU, it just has OCP (which is a good thing!). The problem is not with the PSU, as KfC are trying to suggest, it's with their 100w per cable "design". I spoke to Kurt before buying AX1200i, he told me they "should be fine" and I have a recording of the phone call so don't accept that answer! I was told in the same call that a single PCI-e would be fine and they hadn't had any problems using good quality cables because the socket is "fan cooled". KfC still refuse to do anything about the dead die on my Neptune because it's over 3TH as it is; not exactly what one would call "over deliver". On a side note I fried my first PCI-e plug at the board end today. Thankfully the OCP kicked in and prevented a possible fire but this was with 16AWG:  I'm glad my lawyer got me a refund.
Cost me around 1000€ in the beginning, but at the end, I got back a full refund of my order and KNC had to pay for my lawyer.
I would very much like to have a copy of his contact details; if he speaks English? |
|
|
|
I'm weeks into an RMA for two cubes and a controller card that didn't function. I finally received the controller card but absolutely no news on the cubes.
I'm getting the same crap from KfC; not as bad as you but I have a dead die on a cube so it's running at 75% speed. KfC admitted that it's not working properly but refuse to do anything because they say the malfunctioning machine is over 3TH anyway. I'm now planning to dump it in my buddies Novec and see if I can kick it into action with massive over voltage. So I'm not really going to agree with "Terrible early yields" just because our dreams of a 6th/s Neptune didn't come true...
Yeild, as in number of usable chips out of the number on the wafer... not the speed of the chips that do work. |
|
|
|
well sp30's hit a snag...so this will tell if a) Spondoolies really is aggressive on market share and future customers/satisfaction/fairness as they have stated and offer
more hash/refunds/or whatever to solve this problem they have with Sp30's not getting the hash they expected
or
they will go the other more common route and stall/make excuses/and allow no refunds or some other method to keep as much $$$ in their hands They already said that they are looking into compensation plans so I don't see how can they allow no refunds. Just a couple more days of wait. They have to offer full refunds; unless they offer 10% refund for 10% speed reduction I guess? More importantly their power spec is going to be screwed; I always said 0.34 W/GH/s was going to be ball crap even at chip... does this make Neptune the most efficient ASIC chip? Not to mention their 2x1200w PSU somehow delivering 2500w nominal power 24/7. u27 |
|
|
|
|
Show Posts
