actually, pool circulation generally runs 24/7. its the heating functions that can be turned on or off.

fact is that if the water isnt circulating, or the pool is warm enough, you will have to turn off the unit or dump the heat elsewhere when mining.

^theres too much focus on home mining units. As mentioned above, there are three types of buyer:

1) NOVELTY. Anything that produces <100W will never be cost-effective ROI. There is a market for 5000 units, each worth <$150, or <$750,000 total market.

2) Small Investors. The guys who want to quietly operate a miner at home or in the office that stands a good chance of producing a positive return on investment. They will buy 800-3,200W of power consumption (ideally each being around 800W). At ~$400/800W, theres probably a market for 20,000 units, or around $8,000,000.

3) Big Investors. The Farms. The Early Miners. The Centralized. Those with the ability and desire to build more complex cooling/power in order to use anywhere from 10,000W to 10MEGAWatts, These units are efficient, loud, and should be plug-and-play in a typical server rack (likely with a built in PSU).
   Each unit would likely draw >2500W, and might be built with a L6-30 (30amp, 250V) plug. This level of investor (Assume an average 1000x 20kW for the entire group) would probably be spending $10,000 to purchase 20kW - or a market of almost $10,000,000.

Note that the above all being true would mean that ~40MW of equipment would come online from this manufacturer, spread over a span of 3-4months. That would contribute ~80PH (+25% of the current network hashrate, or about +3% to 7 simultaneous difficulty adjustments.) _{which i think sounds about right, maybe a bit on the high side.}


tl;dr 40% of sales is in midsized units, 55% is in industry-standard hardware for scale deployment, and 5% comes from selling novelty miners like USB devices.


thus, there are two good options for SFARDS:
1) build a single device, that uses 400-800W, and can be easily stacked and deployed in large quantities as well as individual sales.
2) build two devices: one that uses 400-800W for smaller miners, and another that is >2500W and rackmounting for the serious investor.

that would actually be a fun thing to try with some usb miners. only thing is it needs about 25-40W to create the lava, and requires cool underlighting in the form of blinking leds

a smart gui will have 'n+1' or 'n+2' available pool slots, with 'n' being the number of pools already filled in. most users will only have 2 pools setup, but some advanced users might like the option of 4+ pools, as well as the ability to do load-sharing between them

The 'ideal' unit would be something small enough for home use, but large enough that it could be used in larger farms profitably. This is best seen as:


1) 1500-5000GH/unit
2) $400-1500/unit
3) external (user-provided) power supply, hopefully drawing less than 250W/PCIe connector
4) rack-mount dimensions (2U-4U) and stackable. Half-depth would be a smart way to keep the unit small but remain rackmountable.
5) fans that are less than 70dB. That should make it reasonable for home use (at least in a spare bedroom or basement). This means ideally 120mm or 140mm fans only.

6) string design for chips - this is >10% more efficienct
7) design the webUI similarly to what spondoolies-tech built (very detailed settings, stats, logging, scheduling, etc)
design the ability to modify BOTH the voltage AND frequency, as well as the fan settings if possible. this allows wider range of users to run it at peak or in quieter modes.

9) basic outer case. put the power and ethernet connections on the same side, use a basic sheet metal enclosure. 2-4 leds is sufficienct, anything more becomes 'too blinky'. something like the SP3x design is ideal.
10) consider a small screen that can readout hashrate and/or IP address.


MY 'DO-NOT' LIST:
1) fans <80mm are loud
2) miners <1500GH mean buyers end up with a huge stack of units. dont build anything with <$400 MSRP
3) miners >40 pounds (~16kg) can be a bit heavy to move around
4) miners that act as nightlights. all i want is a singlular LED that is lit solid in normal operation and isnt blinding.


-Klondike

yup. you couldve mined back 20-40% of the cost by now, and been able to resell the units for >100% of what you paid

AFAIK bitfury is at the limits of funding - have heard little or nothing about their franchising over the last few months. I think their 28nm chip is not ready (by delay or failure?)

whats the difference between an SP30 and SP31? I presviously thought it was related to the PSU limitations

not entirely sure i understand your description - but i only sell the three cables types listed in the OP, and dont really do small custom designs

+1 thanks.

I have a few pictures, but uploading from my phone sucks - and they simply wont do justice. The Kialara peices are gorgeous and extremely durable (glass and stainless steel).

Putting this up because I have a few extra Kialara Cold Storage 'Artworks' available in Toronto, ON - including:
2014:              $140 USD
2015 Silver:     $135 USD
2015 Ceramic:  $125 USD

If you're interested in a Kialara but can't manage to buy the MOQ for international shipping, this could be ideal for you. PM me if interested. Shipping cost extra, local pickup might be possible

lots of PCIe-PCIe back in stock!

for what it is - its at least an available industrial racking unit. noone wants to install 100 S5 units when they can just do less than half as many of these, and not worry about extrenal PSUs (which requires careful stacking or shelves)

The price isnt ideal, but theres no other offerings that make sense for a larger buyer. The best value comparison would be the SP30 (can run between 3.5TH/1.8kW & 4.5TH/2.6kW). Its 2U racksize, with a pair or NAMEBRAND swappable 1200W PSUs. Pretty fair comparison, if not a better product based on stats.

They go for $1350-1600 USD *shipped* used right now.

i love these altcoin runs - i make 10-50% more renting out, and it helps keep down the bitcoin difficulty.

though im not sure what altcoins are able to have such an effect beyond that paycoin garbage. is it possible that a nicehash renter could direct the hashrate at pools for a withholding attack - would that be profitable or serve any purpose beyond stagnating the difficulty?

actually, you can:

0 23 * * 1-5 /usr/local/bin/change_voltage  685      #(This is the crontab entry to change voltage monday-friday)

1) ssh into unit root/root
2) "vi /usr/local/bin/change_voltage"   to enter the code related to scheduling.
3) press "i" to modify.
4) change the number from '5' to '50' below in the 3rd line
5) press escape, type ":wq" to close

NOTES:
a) i suggest no more than a 20mv spread, or ramping up will take a very long time
b) you can also use the crontab to change fanspeed too, by changing it to "0 23 * * 1-5 /usr/local/bin/change_voltage  685 60"  #(60 being fan speed. crontab will implement the fan and voltage changes simultaneously

Im so glad i sold my 3x S2 over a month ago. I was on the fence, waiting for the upgrade - but got a good price for them

pretty sure its supposed to be 5mv.

for example, i set my machine in crontab to 640 for daytime (635mv start, 640mv peak) and 700 for night (695mv start, 700mv peak). I have zero problems with this.

I dont know why you would want it to be a 50mv range (0.05v) between start and peak - thats a massive range, and could take 1-2 hours just to slowly ramp up to the peak.
ps: 0.73-0.75v is about the highest you can go before pushing stability of either the chips or your psu. lower voltage is more efficient anyways

tldr: set it based on the maximum you want. it starts slightly below that and ramps up over 10-30min timespan to ideal conditions.

^that sounds correct. Thats a 30A receptacle, meaning you can load 5.5-6kW on it (208 and 240V respectively, achieving 80% load). It would connect to your breaker panel using a 2-pole breaker (plus GND)

Now if you have a 30A 3-phase, thats a bit different and is quite unlikely to be seen in a residential building. (90% of homes are 2-phase, each at 120V, to allow 240V when running both phases). A 3-phase breaker will only work properly on a 3-phase panel - on a 2-phase panel you would have 2 rows of 240V sockets, and one row of 0V (since its 2 hot lines of the same busbar)

3-phase is generally for industrial, where power is stepped down from a 600V service, giving 3 hot lines, each at 120V or with a difference of 208V between any two lines. You should balance the load across the three phases

it wont be. guarenteed thats at the lowest voltage/frequency, and probably for just the chip (doesnt include regulator loss, PSU loss, other power draws like a controller/fan)

realisticlaly in a final product, this would probably range 0.28W-0.50W depending on settings (ie: similar to the Bitmain chip)

Its not that GPU mining 'cant' sustain the network - its that ASICs are more profitable, and as a result pushed up difficulty.

dfficulty =/= capability  (and is driven almost entirety by price