Bitcoin Forum

As it stands right now, I'm looking to get my hands on a couple of erupters to prototype some aluminium water blocks for, sounds like overkill, which I'm sure it is. However I've been entertaining the idea for awhile and I have all the tools necessary (and fabrication experience). Ideally one would want a stand alone system (radiator, 1 water block per erupter, tubing, connections, pump, etc..), made in such a fashion it would suit most USB hubs. I was initially going to start off with aluminium (since I have blocks of it sitting around) then go to copper if it was a successful gain.

The real question is, would it be worth it, if so, being cooled to X degree would yield better performance.

Anyone else have some input on the idea? Trying to decide if it's worth attempting at all, what everyone's over all opinion is on the idea.



(I would have posted this else where, but since I've decided to lurk the forums for awhile and not post, I still have newbie status... so hi, I'm a newbie, not really...)

Most people use a cheap usb fan to cool 5 to 6 eruptors, which works just fine. Why would anybody want to water cool them?

i would stick with a copper water block

also i would be willing to test a water block for you if i get a block erupter

I would say it is next to useless. Watercooling isn't a magical fix.  

For watercooling to make sense it should be economical, improve cooling, and be used on a device which can significantly expand its power consumption.
What is the computing power per waterblock?  You want those lowest waterblock cost in terms of $/GH and that means larger hash power chips will be better choice.
Is air cooling acting as a limitation?  Chip temp, size of heatsink, noise of fans.  If the device is only warm to the touch with no air flow then water cooling isn't going to improve anything.
Do you have extra power to overclock?  Can you supply more power to raise the clock to benefit from the increased cooling. Increased hashing power out of the device offsets the increased costs.

USB miners fail hard on all three.


Better choices for exploring watercooling (from most benefit to least benefit):
Hashfast (uses watercooling. I wonder why?  See above)
KNC (100GH/s 250W ASIC and board can supply up to 320W)
BFL (not jalpeno)

No significant benefit IMHO:
ASICMiner Block Eruptor blades (will need very large heatsinks)
Avalon (will need very large heatsinks)
Bitfury (stock board has no spare current)
USB Miners (see above)

As far as the cost of water blocks themselves, yes, could be costly if you were buying them from a manufacturer. I have a few blocks of aluminium laying  around (oddly enough) and the fabrication experience to make the blocks, in theory, they could be made pretty cheap after a mold was made. I'm just merely tossing out the idea of water cooling to see if it would have a sustainable difference. It's more of a "what-if" project from boredom, just because I have a few blocks of aluminium laying around. Granted copper would be ideal, but when you had materials just sitting around... :D

I personally have made a custom rig from a old gaming case, which has (2x) 200mm, (1x) 230mm, (3x) 120mm, (1x) 140mm which all fans are ran from an separate power source from everything else. It's quite efficient for a case that was laying around for 3 years collecting dust and it's completely silent as well..

If you want to stick with block erupters I would suggest using a long water block that maybe 20 block erupters can fit on  ( 10 on each side 5 facing each way )

I have considered this, though, it would have to be design to the hub they are being used on (for spacing etc).

not if they used usb extention cables or you made a small board with the usb ports ether side of the water block

hmm, yea, I see where you're going with that now... gives me a few new ideas, thanks.

Check this thread (https://bitcointalk.org/index.php?topic=241652.msg2562583#msg2562583) for overclocking/overvolting (TL;DR: requires oldschool crystal swap, resistor swap @ volt. reg makes it glitch out into overcurrent mode).  The things only put out ~5W at stock clock though.  Have fun.

nice thread, thanks for the input...

5w x 20 is 100 w and that's worth water cooling it wouldn't take to much to make a water block for 40 with a built in usb ports or you could make it for the k1 nanos or the k16 ( they should be easy to overclock )

Please be realistic. You will never see positive ROI with USB Erupters. Unless bought for 0.05 BTC or less. I encourage to use calculators with difficulty increase factor taken into account.

I did to calculation with double the current diff and a 50% increase in a year and I said just over 11 months to pay for itself they will last longer than that so they will make some btc in there lifetime

Here: http://mining.thegenesisblock.com/

I am positive if I paid for them in GBP that they would eventually make me some money buying BTC directly would make more money but that not an option for everyone

I'm not sure.  How much are you paying for them?  I'm not saying it's not fun -- if it is, it is.

I am into alt coins but I have been offered a fair few at less than  £30 each

Again, if you're doing this for fun, go for it.  Keep in mind that these things only do SHA256 (won't work for most alts).  Click on the link i offered above (the page is not very intuitive, click on the blue buttons to get drop-down menus), stick in your price & you'll see that your calculations are way off.  At least i think they are. 

It might be fun because you will be likely 1st doing this. Post pics when done  ;)

I am sure there not off yours might include the running of the comp but I already have them on 24/7 mining alts so the com power is less than 2w same if you use a pi

Ok.  I understand.  Click on this link: http://mining.thegenesisblock.com/a/3acaa699bf
I've entered 0 cost of electricity, zero shipping and zero associated costs (like usb hubs, whatever).  At this rate, you're still short $13 at the end of the year & will continue to be $13 short.  As long as you expect this, you won't be disappointed.

that is assuming it goes up crazy amounts every 30 days and the btc doesn't go up

if you buy 1 block erupter at the price i said and it doesn't make back its cost in 365days i will cover the difference :)

if you would like to take me up on the offer we can make it formal and even put some cash in escrow if you want no loss to you :)

I'm not going to take you up on it.  And i'm not going to talk you out of mining with the thing -- it's fun & cheaper than hookers & blow.  You have my *sincere, sarcasm-free* best wishes.
Have fun & make some coin.

Thanks

but i don't have any intentions of mining bitcoin i am just interested in the water cooling side of things :)


( alt coins is what i mine and trade way more money and more fun :) )