The machine may have been powered on or rebooted just before taking that reading, or maybe its located in a other room or rack.
Whatever the cause, 10 out of 11 (or 11 out of 12 if we include your unit) is not an isolated incident.

Seems unlikely. In the thread you posted on KnC forum, it wasnt like you were the only one. To quote someone else:


Either its a software problem somehow, or they are having serious problems cooling their datacenter and its doing double duty as a sauna right now.

10 blocks is fairly meaningless, Ive seen it go anywhere from x0.3 to x5.

If you look at 120 blocks, there is still quite some variance, but at least the numbers mean something:
Last 120   09/10/2013 22:04   262631-262751   260 578 393   x1.38

If you add to that the upward trend, well 300 might get difficult, but 270-280M might still be possible.

220M would mean 1.57PH. Were are already at ~1.7PH and growing. Maybe if all those KnCs catch fire and we have a terrible streak of bad luck,  we could get down to 220.  But Im betting anything you want against that.

BFL outsourcing facility:

You think BFL stopped shipping all of their ~3000TH backlog ? Bitfury stopped shipping and expanding (~+100TH or so in their private mine alone this month so far)?
You sure asicminer will not deploy zero of the 500TH this month as friedcat claimed? Then there is the slim chance Hashfast would actually deliver on their promise.

We are already looking at 240-260M next week without any of the above. Good luck betting on 220 by the end of the month

cf the breaking USB connectors; this is something common apparently, I saw the same is happening with a completely unrelated product (multirotor flight controller) where most of these microusb connectors fell off.

To prevent it from happening: just put some hot glue or even epoxy glue under and around the connector.

If it already happened, anyone with some basic soldering skills should be able to fix it. If you are not comfortable with a soldering iron, find a friend or shop to do it for you.

Organ, may I suggest an analysis that I personally would find more interesting?
Not sure how to put it, but giving us an idea of how accurate live updating difficulty predictions and hashrate estimates are.

For instance, right now, when I visit
http://dot-bit.org/tools/nextDifficulty.php

I see a projection of 235M, but its based on ~600 blocks. What is the spread on that ?

Or put another way, I would like to see something similar as dot bit, but guessing *current* network hashrate and next difficulty adjustment with 50 and 95% confidence intervals.

This would have to be live updating to be really interesting tough.

No one will tell you without signing an NDA. This may give an indication though:



Thats for an older nextreme process, but the same principles apply.  NOw you have to realize the variable production cost for a traditional asic is extremely low. Think $30 for hashfasts chip, so even if nextreme3 is say 10x more expensive per chip at relatively low volumes, thats still a good deal for now. But it wont be in a year.


As above, without signing an NDA, I doubt easic will tell you. Im guessing on the order of a factor 2x or so.

So you have chips on the shelves doing nothing for months while they will depreciate something like 80% during that time and while they could be earning you many thousands of bitcoins ?
Right.

Why dont you point them to a pool and publish the link so we can see them in action?

How can you have a working prototype now, yet not be able to ship before January? Working prototype means you have done at least one production run; you should have several thousand chips.
Looks and sounds fishy to me.

Just a hint: why dont you reshoot some photos so that they are not blurry and out of focus? Takes like 2 minutes and doesnt look so amateurish.

10% is probably way too optimistic. ActiveMining is using a structured asic, The upside of this approach is low NRE, short time to market and relatively low risk of it not working. But the downside is that its way more expensive to produce as traditional asics per unit. Depending how much Ken invested (which I think was dependent on presales which dont appear to be stellar), 10% or even more might be attainable shortly after launch, but it will be impossible to maintain a year from here.

This is slowly turning in to a trainwreck.

September came and went, we are almost halfway october, most customers still seem to be waiting. Hosted miners were reportedly running at only a tiny fraction of their hashrate, and are not running at all now. Alarming levels of DOA. One customer had 5 out of his 6 jupiters catch fire (!).  Software and firmware clearly not ready for primetime.

I really enjoyed following orama's "behind the scene's" reporting (seriously, they should have made a full length documentary, I would have paid for watching it),  I wont fault KnC for lack of trying, and of course, we have seen much worse execution, but I think we can safely say KnC's schedule was once again way too optimistic and they started shipping too soon. These problems could have been avoided.  

Not sure how or when this one will end, but Im already looking forward to seeing the next trainwreck unfold with hashfast.

You can not measure the hashrate. All you can measure is how long it takes to find blocks, and from that estimate the hashrate (and future difficulty). Blocks finding is heavily subjected to luck, so you need a lot of blocks to get a reasonable estimate. Just have a look at these charts to get an idea:
http://bitcoin.sipa.be/

Im pretty sure that apparent angle is a result of the fire/explosion, no?

4? Ouch! Out of how many?
Either way, guess we now know why they are only tested for 30 seconds. Probably with a fire extinguisher at hand.

Seriously though, thats not good, at all.

IOW, even in the best case scenario,  "october" shipments almost certainly wont happen before november.
But everything is on track .

Neither. You may get listed in the top 10 bitcoin scam list though

Sure, If the machine is able to run for a few hundred years or so.

Electricity cost is still negligible today when compared to writing off the (28nm) hardware. One day, electricity will become the sole real limit on network hashrate, but by that time your miner will be earning at most 1% of what it would earn today, and possibly as little as 0.1% if there are many miners with "free" electricity.  That makes break even calculations pretty pointless.