Well, this is interesting - BFL also offers not mounted chips now. Assuming BFL chip yeld is reasonably good, then it might get interesting. It all depends on real chip availability - we have not seen Avalon chips available en masse yet but considering appearance of 1...3Thash clients, there probably is significant number of ASICs available already. I would speculate for high volume chip buys before availability to general public orders and fast rising network hashrate at the moment might be direct result.

https://products.butterflylabs.com/65nm-asic-bitcoin-mining-chip.html

Tho most interesting part - it goes away for a day, but growth continues. Almost linear growth, 3.5Thash reached.

 You can get rack mountable fire suppressors, but they are not cheap. Easiest would be to use canister of CO2 with temperature sensitive release - either temperature to pressure converter acting valve or simply melting tube (not that good on long term).

 Might it be, that indirect approach can be considered - typical ephedrine use has quite strict temperature, pH and other environment limits, that human body dictates. Lets say the molecular structure is unchanged but there are added substances, that for example are inert at body temperature but react unfavorably over 60c and reduce drastically possible methamphetamine production. I am sure purification by different ways is possible but if main goal is to rise production cost beyond usability for "chemist", what could be possible approaches?

Few points that may or may not be relevant:

* Avalon 10x slices allow sandwiching between 2 radiators. PSU components are separated, reasonable design feature and this might be favorable layout considering overclocking and extending.

* There might be chain member number limit dictated by serial speed and driving capability, so it might be reason for 10 chip per chain.
Reasonable would be to calculate theorethical max datarate needed to be in the ballpark.

Well, the (pseudo)ephedrine question seems popular... Lets reverse it for fun and profit: what could be effective means to significantly hinder or stop  easy removal of hydroxyl from any ephedrine stereoisomer and still keep reasonable bioavailability and vasoconstrictive properties for typical use?

Do not forget to connect PSU grounds and check earthing.

It is probably easiest to take 8 pin PIC AVR or (insert your favorite MCU with internal oscillator here) nowdays... but:

For fun:
http://www.falstad.com/circuit/e-vco.html

Use suitable RC filter to smooth the input enough... Probably 555 is not that good - bandwidth change is not enough to cover whole RPM range. Still - 99% motherboards do not care about fan speed change range as long there is variable signal present.

rKxEXdc4MEFcfV6ngFVZvV2vtsmp2bhJeo

So, depending on Avalon order number sequentiality - do we have 1 PetaHash/s mark in the air ?

Bump up to that question.

So,

lets bet how how low it can go ?

Moo! 

Hi,

 as I cannot yet post in more appropriate sections, let me start here.

 Recent relatively high volume rushed Avalon bare chip sales have left me wondering what is the real reason behind inventory dumping.
At first glance one might suggest board manufacturing capacity limitations, chip manufacturing limitations (large minimum amount of chips
in one batch), the wish to calm 51% attack fears or other fairly benign reasons. However, at the current moment Avalon has almost
monopolistic hold of the market - why should one rush to inventory dumping in this kind of situation?

 I can see but only one actual reason - to further leverage monopoly, they chose cheapest way to kill arising competition - flood market with
promise of reasonably cheap ASIC rigs. There is no better way to do it than dump already known working chip design to multiple vendors. Anyone
interested notices either chips now or in later phase, the readymade boards just because the publicity. The use of distributed parallel manufacturing of rigs still containing Avalon chips is brilliant in at least three ways:

*Avalon can have reasonably high price per chip now, with the water written promise of high ROI of rigs made from the chips later. Buyer's fallacy is to expect ROI on current projections of difficulty rise that are !mostly linear!.

*Avalon pretty much stops new orders made for competitive products that cannot hope to compete with numerous board manufacturers as a single entity. Assume BFL preorder sell or cancels appearing en masse. I would suggest, that selling the first generation Avalon chips this way is actually the most significant event since the announcement of ASIC chips.

*Avalon has a extremely fast way to raise capital for second generation shrinked die faster and bigger chips. One might be surprised how soon the second generation can arrive.

 Any thoughts are welcome.