True it provides some latitude but you can't fix stupid.   BFL just "should" have known better (especially when FPGA were nearly 300% over power spec).  Still it would have allowed them to launch reduced hashing power boards in 2012 while working on a new design. 

The diagram in your post shows two DC supplies per quad for a total of 8 just like the PCB.  Still KNC indicated again they don't expect more than 250W nominal.  Using DC PSU capable of 320W is smart though as it gives 70W extra of tolerance.  A large portion of BFL's delay was their chips used more power than expected and thus they had to redesign the boards, buy new DC PSU, upgrade the cooling, etc.  

While nobody wants the chips to come out over power spec, worst case scenario say the KNC chips come in "hot" using 270W instead of 250W.  Not a big deal.  The board and heatsinks can already handle up to 320W without breaking a sweat.  No components need to be changed. Sure 250W is better than 270W but no delays is more important than delays.   The over engineering reduces the need for costly delays.   Had BFL done that even with all their other problems/lies customers would have gotten their rigs months earlier.

Well it does matter if you want to make sure your PSU will be able to supply that power without waiting till the last minute.


No.  That wasn't the point.

I never said the wires indicate the amperage the connector can handle.  Not sure where you got that from.  It is already established the connector can handle more current than is necessary, 324W (27A @ 12VDC).  However just because the connector can handle 324W (27A @ 12VDC) doesn't mean you will be able to pull 324W (27A @ 12VDC) unless the PSU can deliver that much power TO the connector.  The layout of the PSU connectors matters, it shows what current the PSU can safely deliver on that set of wires.  Multiple connectors on a single wire in series indicate that the PSU can deliver at least the combined current required by the ATX/PCIe standard for all those connectors.  For two 2 pin connectors that is 300W (25A @ 12VDC). 

"Sorry" (not really) about using only using watts without specifying voltage but I was pretty sure everyone knows we are talking about 12VDC here.  So so nobody remains "confused" I updated that post and this one, making them longer by repeating the voltage over and over.

Hmm interesting.  Well doing a pixel count of devices with known dimensions (power connector, GE DC modules, Cyclone IV) and comparing them to the package area it looks like the package will still be ~55mm square (+/- 2mm because pixel counts aren't exact).  So the prior 55mm/43mm spec is probably still being used.  Maybe they just lowered the pin count.

That is correct.  3 12VDC pins @ 9A ea = 324W max.

It is the same for the 8 pin PCIe connector as the extra two pins just contain extra grounds.  So 6 pin connector = 3 12VDC & 3 ground, 8 pin connector = 3 12VDC & 5 grounds. They extra two grounds are used as a "ground sense" by PCIe devices that is how a GPU "knows" not to power on if you only connect a 6 pin connector to an 8 pin slot.

Yes you can estimate it from any component with a known size (Altera Cyclon IV, GE DC power modules, 6 pin Minifit Jr connector, etc).
However it isn't necessary they have already indicated the package will be 55mm x 55mm (the lid is 43mm x 43mm).

Nope.  The only "premined" block is the genesis block and it is unspendable.

Here is a fun experiment.   Put 3x 7990s in a sealed 3U or 4U rackmount chassis and try to keep the cards from burning up.  There is a reason that people use open rigs.  Trying to remove that kind of heat is next to impossible.  Even if BFL delivers the form factor is next to useless because you are never going to be able to fill a rackmount server with 3 to 5 of them.

Chip wattage vs system wattage.  Remember you have two voltage conversions plus control boards, and other gear.  So the estimate is the entire system will use 350W however for the purpose of discussing the power and cooling of the chip it is only 250W.

Well design your own chip.    You didn't seriously think they were going to scrap a design and make another one because you don't like it do you.

Also you missed the point.  There is no fixed amount of surface area needed to cool X watts.   It is a curve.  The more surface area the less airflow you need all the way down to 0 CFM (passively cooled).   A radiator is generally much larger than an air cooled heatsink.  It is simple logistics.  Trying to mount a 12" x 6" heatsink directly to a chip is difficult (and dangerous as you could damage the die).  Using water to move the heat from the chip to a larger surface area is the whole point of water cooling.  Simply logistics mean you can't use as large of a heat sink directly air cool a chip, and thus need more airflow = more noise and more wattage.  Water cooling also has the advantage of lower delta T and cooler chips use less power. 

The company believes they have the expertise to design a large efficient processor.  If they can there is no reason to use a larger number of less powerful chips.  The less integration at the chip level the more integration that is needed at the board level.  More wafer cutting, more chip packaging, more complex boards, more supporting components, more fans, more heatsinks, more assembly, etc.  A larger number of smaller chips gives you more flexibility and makes cooling less of a challenge but the system isn't going to be cheaper or more importantly assembled faster.   Waterblocks are a significant portion of any water loop so if you are going to watercool less blocks is always better (i.e. water cooling two 5970s is going to be cheaper than 4x5870s).

There is more than one way to build a miner.  The nice thing about competition is you can pick the design you think provides the best chance of success.  I think it is great that you have different companies with different high level designs.  Bitfury uses almost 200 chips over 16 boards maybe you should look there?

With a large radiator you can use larger, slower, lower wattage fans.  Heat transfer to air is directly related to surface area.  A radiator is going to have much more surface area and thus need less airflow.   With a large enough radiator you could get by with no fans.  Another thing to consider is that the power consumption of an ASIC (any ASIC) rises with temperature.  Water cooling tends to have a lower delta T (difference in chip temp and ambient air temp).  

Who buys watercooling gear from newegg?  Sounds like someone who has never watercooled anything.

Each of KNC chips is rated at 250W.  So either 250W can't be cooled with air (in which case KNC is screwed) or HF could just as easily cool their 250W chip with air.  So why choose water.  Simple water has 16x the thermal conductivity of air.  You can cool a large wattage device more effectively and quieter using water.

By "commodity" I assumed they meant "off the shelf".  Since it has the package size roughly that of a CPU it will work with any waterblock designed for a CPU.  Case, fans, radiator, power supply it is all commodity gear.   The only parts which are unique to HF are the ASIC boards and controller board.  Now compare that to BFL cooling and power solution.  Think there is even a single vendor selling an air or water cooler for a BFL single?

My hope is in their next batch they allow DIY kits.  Sell me three ASIC boards and a controller board and let me build my own (with flourinert).

It will be my last post because honestly I don't care enough.  Maybe an investor who doesn't want their funds seized could ask someone like MSantori in legal forum.

However A BANK isn't an MSB because it is a BANK (there is an exclusion in MSB regs specifically for banks).   However it means banks, not partnering with a bank, not licensed by a bank but actually a bank.  If the Panamanian bank was offering this service they likely would be good but I don't really see how this is complex but .... CFIG isn't a bank so the bank exclusion doesn't apply.

You won't find any swedish banks on the MSB list but you also won't find any US banks either. 

Also merely accepting funds isn't money transmission otherwise all companies in the world would be money transmitters.  So every company in the world which accepts funds from a client isn't a MT and won't be on the MSB list.  The action of money transmission has a specific definition.  Paraphrased it involves a third party (CFIG) accepting funds from one person (client) and sending those funds to a second person (exchange).  Note exact definition is included above.  CFIG offering a service where customers can send funds to an account and at their direction CFIG will transfer them to an exchange so the exchange can credit the client's exchange account is the very definition of money transmission.

Since the company intends to accept US customers and they are NOT a bank and FinCEN has indicated foreign entities are subject to MSB regs I know this is crazy but I would kinda expect that in due diligence the company would have retained US counsel and got an opinion letter or if they are so sure FinCEN doesn't classify their activity as an MSB file for an administrative ruling.  If FinCEN says no and later says yes the administrative ruling is a "get our of jail and keep your money" card.  If nothing else it is cheap insurance.

I expected some realistic answers from CFIG not "we aren't a money transmitter despite doing the TEXTBOOK definition of money transmission involving US clients".  If they are that inept and investors victims are willing to accept anything and everything at face value I don't really see the point of going further.

If CFIG isn't a money transmitter then essentially nobody in the world is a money transmitter including PayPal and WU.

PSU are rated on DC load.  So a 910W PC P&C means it has a peak load of 910W DC.  If 90% efficient at 100% load that would mean 910W DC = 1,011W AC (910/0.9).

950W AC = 855W DC (950W*0.9)

I assumed 90% but you should look up efficiency curve for your PSU.  If you cant find one and it is an "80Plus" rated unit you can check the 80Plus website.  They have test which show efficiency at 100% load.

Please don't get him started.

Hence the ~ 

Read the linked bitmit description.  You send coins and AFTER three years he sends back 3x as much.  There are no payments over time.   It is stupid, and "investors" are almost certainly going to lose everything but it isn't a ponzi.

Not everything is a ponzi scheme.  I have a feeling if someone mugged you on the street you would be telling the policeman you were a victim of a ponzi scheme.

The oppertunity cost is ~$0.  It can be merged mined right along side Bitcoin.  The hardware, labor and electricity would still be spent if one was only mining Bitcoin.  Nobody is mining this at a loss it is just a free bonus on top of the BTC.

Yeah PSU have gotten much better in the last decade.  Finally.  The "80Plus" program has actually done some good.  It sucked that at one time ATX PSU had a sweet spot of maybe 80% efficiency in a narrow band of 50% to 60% of peak power.  No reason that it took this long to finally get decent PSU.  Thank the rising wattage requirements and electrical rates.