Topic: Heating a Space via Mining (Read 3352 times)

Yep, it is the quantity: https://www.beemster.nl/en/expertise/three-phased-connections/

Sorry, 1 phase, yes.

I still highly doubt that the average house will use anything 3 phase, the stove you describing is probably business grade, now if it is a matter of "quantity" then it makes perfect sense, here if the house has more than a single floor, the utility company would run 3 phase to it, each floor gets 1 phase and neutral, say 6 flats each two share a single phase, but still, inside that house you can not just plug and play 3-phase appliances, you would need to run the other two phases directly from the main panel.

I understand that things are different across the globe, so ya, to each country its own design.


Correct, the naming is rather confusing to many people, it is hard to explain that 2 wires are a single phase, it is also hard to explain why would a clamp meter read zero when you get the two wires together, electricity isn't all that simple after all.

Just for technical clarity, 2-phase power does not exist outside of some *very* specialized applications.

Utility power is either single phase (which includes split-phase) or 3-phase.

Sorry, slow response. Not sure about the exact technical reasons, but as it was explained to me by an electrician, is that there is a lot more electric usage these days, so for over a decade they do newly build houses with 3 phase to be "prepared". By itself 2 phase would be working for many induction stoves, but (if I'm not mistaken) it's 5 or more hobs and/or downward extraction of air, it can go over the specs for 2 phase.
As he explained it would in case of washer+dryer+dishwasher+induction potentially be too much, and that's not taking into account car chargers.

the 75kva at your link was 4800 used and install is at least two grand

So say 7000 to do 14 miners that is 500 a 335th hydro plus coolers I have seen them at 600 for one unit

so say 7000+7000= 14000 for 14 335th hydros.

makes me buy the s19xps.

Oh I certainly agree. My point is that because it is not a major issue to them, BM just does not care. A 75kva autotransformer is only $1,1700 so smaller farms do have a viable solution to it. If companies like RIOT and Marathon who run sites that pull 10's to 100's of MW don't care then BM won't care either.

True, but that does not change the fact that potential clients with regular U.S 460v would find it hard to justify the extra cost and time needed to set up new transformers just for the hydro, I mean I see MicroBT tweaked thier gears to run on 277v just for U.S clients, I expected Bitmain would do the same given that the U.S is the largest mining hub today.

Getting way OT here but price for that isolation xmfr is too high. https://store.maddoxtransformer.com/products/64d50fed-3483-587e-9848-1f3005736804 has way better but again - applies only to 'small' farms that must use incoming 480vac from their existing utility supplied distribution transformer. Actually - that site has several excellent articles dealing with power distribution https://maddoxtransformer.com/resources/articles/

The mega farms pulling 10's to 100's of MW per-site have their own substations tied to The Grid's main transmission lines and can go a whole different route. Typically it is incoming xmfrs taking in 33kvac or higher from the utility and dropping that to a distribution voltage of 13kv or so which is sent to the mining warehouses to then be dropped to whatever voltage the miners need - in this case 400V - by several final transformers located at each warehouse or pod.

When starting from scratch with a blank slate makes it very easy to accommodate different needs just by having different final xmfrs to provide 240 or 400 or 480v and segregating the loads accordingly.

Or they have connections with companies selling this unit.
this is an 80kva If I go with copper over aluminum it will cost 11k plus shipping and install costs say 14000.  This would allow 14 new 335th hydro's and I need to buy liquid cooling. It is a lot to convert over.

If their major markets were in Europe or Asia then building gear to use 400v would make sense but since that is not the case, I go with 'stupid'.

Since W=V*A, using a higher line voltage of course means that for any given power less amps are pulled which means that one can use components that are rated for lower current and are easier to cool since is the current that drives how much heat they waste. Fun fact, when designing power systems the component current rating has a much larger impact on price than the component voltage rating does.

My guess is that BM just does not care because multi-MW transformers are *not* what one would call a 'standard in-stock item'. They are made-to-order and that being the case the mega farms just spec theirs for 400v.

One would assume that since the largest mining operations exist in the U.S now, Bitmain would make something that works there right out of the box without all the complication and the added cost, it's probably one of the reasons why even the 19 hydros are selling for pretty cheap now since all these resellers are stuck with them and nobody wants to buy them, heck even their hydro container ANTSPACE HK3 V6 that sells for 110k needs an input voltage of 400v +-5.

I am no way near an expert in the electrical field, but are the folks in Bitmain too stupid to make something that is hard to run in their NO.1 market country, or do they know something we don't?

In the US industrial power is 460-480v. For new installations it's not hard to get transformers that deliver 380-415v but ja those are strictly for dedicated uses and even if you buy the service transformer the power companies will really push for a different solution unless you need more than say 1MW service. For existing services we would just use auto-transformers to knock 460/480v down to the desired range, they are fairly small and relatively low cost.

Why would you need 3 phase for induction cooktop? You just need 220v for that, which is L+L (split phase) for U.S and N+L for the majority of the world, probably for the car charger?

So most homes around the globe will have two wires, N and L, where the voltage between the two is 220v to 240v.

When we talk 3 phase there are two different setups

1- Three-Wire

Just 3 L wires no N, the voltage between each two phases is 380-400v ( not sure what that is in the U.S) but google tells me it is 460v which won't work on the hydro miners if the number is correct, such setup only exists in industrial areas.

2- Four-wire Wye

3 L + 1N , the voltage between L and L is 380-400v , between L and N is 220-240v, these 4 wires will run along the street and then house one gets L1+N , house two gets L2+N and so on, so despite having 4 wires running under or above your house, you get only 2 since as a normal house all you need is that, when it is a business or a factory they give them the 4 wires because while most of their machines would require only L1+L2+L3 , they would still the N to get 220v for the lights, TV and what not.


So in the average country, residential buildings won't get 3 phase wiring, electric car chargers might change that concept, might have already changed that in some places.

Back to weird U.S numbers, Phill's numbers match the result I got from Google, 460v is probably too much for the hydro miner given that the specs clearly state 380v-415v, I am sure you can find transformers in the U.S that can get you 400v instead of 460v, just not sure how popular those are.
 

It reads that the psu takes  380-415ac

that's 3 phase.

which means 3 hot wires not 2.

most homes in usa have 2 separate 120volt hot wires which will give 240 volts single phase.

my mine is 3 phase  460 volts there are 3 hots which our transformer drops to 232 volts 2 hots

If you are buying this miner it is commercial if you are in USA as home do not have 3 phase.

thanks to mikeywith for finding that.


"https://shop.bitmain.com/product/detail?pid=000202309271329420372NbWET50062F

Specifications
Purchasing Guidelines
Warranty
| Specifications

Product Glance   
Model   S21 Hyd.
Algorithm | Cryptocurrency   SHA256 | BTC/BCH/BSV
Hashrate, TH/s(1-1)   335
Power on wall @35°C(1-2), Watt(1-1)   5360
Power efficiency on wall @35°C(1-2), J/TH(1-1)   16.0
Detailed Characteristics       
Power Supply       
Power supply AC Input voltage, Volt(2-1)   380~415V AC   
Power supply AC Input Frequency Range, Hz   50~60   
Power supply AC Input current, Amp(2-2)   12   
Hardware Configuration       
Networking connection mode   RJ45 Ethernet 10/100M   
Miner Size (Length*Width*Height, w/o package),mm   339*163*207   
Miner Size (Length*Width*Height, with package),mm   570*316*430   
Net weight, kg   12.3   
Gross weight, kg   13.6   
Environment Requirements       
Inlet water temperature, °C   
NEM(2-3)

HEM1(2-3)

HEM2(2-3)

35

45

50

Water flow, L/min   8.0~10.0
Water pressure bar   ≤3.5
Working fluid(2-4)   Deionized water/Pure water
Liquid PH   8.5~9.5
Diameter of water pipe connector, mm   DN10
Storage temperature, °C   -20~70
Operation humidity(non-condensing), RH   10~90%
NOTE:

(1-1) The Hashrate value, Power on wall, and Power efficiency on wall are all typical values, The actual Hashrate value fluctuates by 3%, and the actual Power on wall and Power efficiency on wall fluctuate by 5%.
(1-2) Inlet water temperature.
(2-1) Caution: Wrong input voltage may probably cause server damaged.
(2-2) Three-phase AC input, 10 A per wire.
(2-3) NEM: Normal Energy Mode; HEM: High Energy Mode.
(2-4) If the water conductivity is ≥100 μs/cm, the fluid must be replaced. The water conductivity is less than 20 μs/cm when the system is running at the first time."

Never mind, totally forgot what the hydro specs are, was still thinking in the 200~240 area of the S19 I'm used to.
Thanks Phil!

Hmm, I don't get the part about the v here.
Speaking just for my area but in Europe, we've changed from 380v to 400v transmission before consumer so the phase is up from 220v to 230v but we are on 30A here so my 4000W oven has no trouble with it.
Isn't this more a thing of maximum kvA rather than voltage?

Yeah, with a move from gas for cooking to induction (or similar) a lot of people have to adapt their fuse box and get a 3 phase installed. Right now in some regions there's a several month waiting list though

240v won't cut it for the hydro miners, you need 3 phases 380-415v (10A on each phase), depending on your country specs, for anyone who doesn't know what 3-phase wiring is, you can't just use three wires coming from the same phase and assume the miner will work, most houses will likely run on a single phase setup, so it won't be easy to supply 3 phase appliances, it's easy to figure that out just by looking at the wires coming into your house if they are just 2 wires -- you get a single phase wiring, if they are 4 wires then you got 3 phase which again, is very unlikely.

Oh that is some serious stuff, let me look into that, thanks!

Bitmain has a newer version

https://shop.bitmain.com/product/detail?pid=000202309271329420372NbWET50062F&locale=en®ion=en


and there are a few aliexpress companies that will provide a single unit cooler for it.

you will use 5400 watts so you need a 30 amp 240 volt circuit

5400 watts is about 18000 Btu of heat.

Getting a notification email for this topic, thought to chime in. Last year had to get a room warmed up so the new plaster would dry. After two days of settling, it had to have both air flow and light heat for the water to evaporate, so I dusted off the old L3+ miner and set it up to run air around the room while running an air dryer at the lowest setting. Didn't bring much in in coin, but was a fun to show to the workers that dropped by.

Another thing I've been thinking of is linking liquid cooling from for example a bitmain machine (https://www.coindesk.com/tech/2022/03/14/bitmain-says-new-liquid-cooling-miner-is-its-most-power-efficient-model-to-date/) to a central heating system. Mostly it's using gas to heat water to run through the radiators, but with an abundance of solar panels, even in winter there could be overproduction. Instead of pushing that back into the grid, could be used to heat and mine. unfortunately bitmain isn't giving them away exactly very cheap.

I remember those Rockminers. They were nice quiet units.

Long overdue update. 

I ended up heating the office trailer for 2 winters (2014 and 2015) using miners.  First winter was Rock Boxes second winter was Spoondoolies Jacksons. 

With the power of HODL it was well worth doing compared to space heaters.

I would love to see what you've done. 

I really think that future of decentralized home mining depends on also using the heat.

Here's a simple back-of-the-napkin calculation - I would be curious to hear if people think this is wildly unreasonable or not.

Let's say the whole thing costs you $1500 to build a 1000W system and you spend about $1700 in electricity for the season (Nov - Feb @ 24h x 30 days x 4 months x 6 cents per kwh) to run it.  As long as your mining efficiency is better than break even for the power you'll make $1700+ in bitcoins.  Assuming you were going to  have to burn that electricity anyway for heat, your net cost is $1500 instead of $1700 for the season. 

Next season you only swap the mining boards so your cost is $1000 for 1000W, but you can reuse the $500 radiator unit.  Assume you again break even, now your net cost for the season is $1000.  Or, if instead you keep your old boards which are now mining at a 50% loss, then you spend $1700 in electricity for 1000W of heat and make $850 of bitcoins.  Either way, you cut your power bill down from $1700 to $850 or $1000.

If the bitcoin price goes up, your mining efficiency is > break even on power or you can use the heat all year (eg. to augment a water heater) then the economics look even better.  If the bitcoin price crashes or mining difficulty increases faster than expected, you'll lose - but nothing ventured nothing gained. 

This isn't a get rich quick scheme, this is a get rich slow scheme.    And hopefully fun; people spend much more to overclock their home PCs with no financial gain.

The three financial caveats you need to accept are:

1) you were going to spend the electricity for heating anyway (24x7 during the coldest months)
2) your mining revenue is greater than, or equal to your electricity costs to run the mining boards (during the season)
3) the up-front cost of the mining boards is less than the cost of a season of electric heating

Notice that the only real unknown is the mining revenue.  However, it is safe to say that the expected return for any mining boards should be greater than the cost of the electricity or else no one would even consider buying it.  From a strict business perspective, the mining operation still runs at a loss of the invested capital ($1000) even if you break even on electricity.

A clever accountant might also recommend you treat the whole thing as a home business.   Then you can deduct the cost of the mining equipment and electricity against the bitcoin revenue and deduct the loss on your taxes for a few years.

Please let me know where my math is wrong; I know I am taking a big risk by proposing a profitable scenario in this forum. :-)

S-3's would work if you find the correct fan's  I am going to do a thread on s-3's with different fans vs the r-boxes


3 r-boxes and a rpi = 400 usd  and  390gh  power =  400 watts

1 s-3 plus the newer fans = 365 usd and 450gh power = 365watts


the s-3 has wifi with an antenna added  if may be better but and here is your but  I do not have a good enough fan as of today.  so far every set of fans that works good enough to cool the s-3 is too loud.

I have this set coming  to me on thurs:

http://www.amazon.com/Silverstone-Tek-Computer-Cooling-FHP-141/dp/B00A460TK6/ref=sr_1_1?ie=UTF8&qid=undefined&sr=8-1&keywords=silverstone+141

As feature sizes get smaller and chip density increases air cooling is going to be a losing proposition.  Some sort of passive liquid cooling is the ideal way to do it, especially if you want something that your wife will let you put in the living room.   

The ideal heat exchanger / radiator would allow you to easily swap out obsoleted boards for new ones.  That way you could amortize the heater over multiple generations of mining hardware.    Even better is enough room to keep older generation mining and just add new next generation boards so you can squeeze the last coins (and heat) out of old hardware.

All mining rigs are going to need the same thing - power, ethernet and cooling; the rest is just mounting.  So in theory the radiator part can be fairly universal.

Make sure you have a way to transfer the heat if you are heating the space. You don't want to inhale heat produced from ASIC..

Why not just make some kind of fire-proof housing for 2x S1's?  Creating a solution from scratch for a small market geared towards home heating would defeat the purpose of using older, less efficient miners that otherwise would have been tossed or sold at a fraction of the purchase price.  Otherwise it would just make sense to buy something like Spondoolies or S3's which would come in at the same price point (or less).

I'm currently working on something like this heater for me.
standalone with RPi or TPLink router inside.
Might have a few pics within 1 week or so.

I am sure if we could put a group together interested in home heating from bitcoin mining that we could get an affordable and elegant device made specifically for this purpose.

What I am thinking of is something clean and quiet that could sit on the wall in the living room or den and not have to be hidden in a closet.  Another option would be a device that hooks into a hydronic water loop heating system.

I'd like to start an interest list and make this happen, but we need a large enough group willing to invest in such a system to make the economics work out.  If we could get one of the mining rig firms interested that would also be helpful.  I've made some back-of-the-napkin calculations and I think the goal of nearly free heat is very attainable, even accounting for the initial investment in the mining hardware and heat exchanger.

I'm looking down the barrel of a nordic winter so I could use this kind of system today.  Who's with me? 

Ideal candidates for this system have the following characteristics:
 1) already using electricity for home heating
 2) long and cold winters
 3) own their home

The folks at Nerdalize are developing computational heat exchanger modules specifically for home heating that would be perfect for mining also.  They use general purpose computers for the computation/heat source, but  I have talked with them about adapting their technology for Bitcoin mining.   They are interested, but they would need a large enough order to justify adapting their modules for bitcoin mining.  They have already run evaluations of a KnC Jupiter mining board and their systems could easily handle the high power/heat densities of bitcoin mining chips.  They just need to see customer demand to justify more work in this direction.
 
One thing to keep in mind also is that mining is about to start running into Moore's law.   That means we could see 20 nm mining chips stay competitive longer than 28 nm chips, and future feature sizes even longer.

This could also be a way to decentralize Bitcoin mining which now seems to be dangerously headed towards data center centralization.

yeah I am in NJ and my winters are never as bad as yours but last year's winter was a bad one.  The miners liked it and I had a very warm home last year.

My heating (natural gas) bill for 2013 winter was reduced about 2/3 overall with last winter mining, while I don't have exact KW stats on hand, I'd estimate last winter I consumed about 4 to 5 KW running 24x7.

This winter I will consume ~6.5KW/hr which comes out to ~22K BTU's/hr.  Still not enough to heat my home, but I guarantee it makes a good dent in my gas bill this winter as it's a 24x7 source of heat.  Currently this week it's been in the lower 40's to lower 60's, a few windows open, and home is comfortable in the lower 70's.  Will have to close all the windows once it starts to stay in the 30s-40's.

Michigan winter gas bill:
2012 mild winter      ~$3,500 (No Miners)
2013 Horrible Winter ~$1,500 (miners)
2014 Winter TBD        TBD (miners)

you will like them they are sealed, quiet and 4 will use about  550-575 watts..

see if crazy guy will pre-program a rasp pi for you. I am happy to use it.

Hi Phillip, it may be funny but it is also true!

The new rboxes is what I am thinking, as you suggested.  Perhaps 4 of them. 

No the cow is.
[/quote]
You made me laugh out loud.
[/quote]

Good I thought it was funny when I wrote it.  Every once in a while I can be funny.
[/quote]

Phillip is correct, the cow was a lesbian but we ate her last year.
[/quote]

that is also pretty good one.  

Beef it is whats for dinner.  Back to the poster's question.

 I am running the newer r-boxes from crazy guy.  

Just the 2. They are pretty quiet units.  You could run a few of them.  2 of them put out 250gh hash power and they use 275 watts

Are you really Lesbian? 
[/quote]

No the cow is.
[/quote]
You made me laugh out loud.
[/quote]

Good I thought it was funny when I wrote it.  Every once in a while I can be funny.
[/quote]

Phillip is correct, the cow was a lesbian but we ate her last year.

Thanks, I am looking at this option as well.

Good I thought it was funny when I wrote it.  Every once in a while I can be funny.

Antminer S1's are going to be the best bang for your buck GHs wise. There were a bunch that just sold for $50 each.  I would setup 4 S1's with 2x of the DPS-800 GBA 850W (1000W on 240V) power supplies and gigampz breakout boards.  It is the absolute cheapest hashing power, bar none.  That's what I started out with a few months ago now,  it's really not much of a learning curve. Configuring routers and setting up wallets is basically the extent of it.  I'd be happy to help you with any roadblocks by PM if you go that route, the Antminer web interface is (nearly) dummy-proof.  Being in Canada, I plan on taking my leftover S1's and spreading them out through my business, home and cottage for some heat+hash.

www.gigampz.com

You made me laugh out loud.

No the cow is.

Are you really Lesbian? 

I have set up my S3s in vertical airflow configuration primarily to improve the units' air cooling efficiency.  Coming winter, I intend to place the setup on the floor where it would draw in the cooler air at floor-level and blow the hot exhaust towards the ceiling.  Hopefully, this will help you come up with some ideas to suit your purpose:

https://bitcointalksearch.org/topic/m.8227268

https://bitcointalksearch.org/topic/m.8471571

Construction takes place inside the buildings while my office trailer is outside the buildings so physical dust should not be a problem.  Air coming out of the buildings is filtered through several chambers.

Hi Phillipma1957,

Thank you very much for all your valuable input.  This looks very promising.

Cheers,

Lesbian Cow

I found a new miner for you.

https://bitcointalksearch.org/topic/m.8731623


the seller is good and these may be okay. I had the older version  the older version was quiet

http://www.amazon.com/Rockminer-R-BOX-32-37GH-Gen3-TSMC-Miner/dp/B00KXL0QWS/ref=sr_1_3?ie=UTF8&qid=1410208149&sr=8-3&keywords=r-box

price was as low as 50 bucks for 37gh.  the new version looks hopeful  for your needs.

Since you blow all that air through the miners, you only "collect" the dust there, you don't generate it per se .. Construction areas usually have dust to collect every time when it's hot outside, so I wish you lots of rainy days!

I have no idea who this guy is


http://www.ebay.com/itm/2-X-AntMiner-S1-180-Gh-s-Bitcoin-ASIC-miner-PLUS-Corsair-RM1000-PSU-/331311092897?pt=US_Virtual_Currency&hash=item4d23ae14a1


but 2 and a psu decent deal.  no adjustments this will use 700 watts but it will be loud.


if you downclock and undervolt it will use  350- 400 watts  and be quiet. it will do 300gh

 I am looking for the thread on how to do it.
found the thread

https://bitcointalksearch.org/topic/guide-undervolt-antminer-s1-119wgh-at-the-wall-526060

here is more info

https://bitcointalksearch.org/topic/guide-undervolt-antminer-s1-119wgh-at-the-wall-526060


quick math  2 units  cost say 230 as the ebayer wants an offer.
$0 for power
they go in the black in feb-march. if you slow them down to 150 gh each
if you can stand the noise and run them at 180gh
they profit in dec-jan

two under clock/volts = 360-400 watts

two stock clock/volts = 720 watt

I advise these as they are cheaper then the s-3's but do have wires .  if you want better looking get the s-3's

This is where we need someone to make a case that actually looks like a space heater to house the S1s.

S1s really look like crap with their exposed circuit boards, and you also have to deal with an exposed power supply and PSU wires.

I certainly wouldn't want to use them to heat a normal room in my house unless I could hide them.

Its better to heat the house with those wood pellets they are dirt cheap.

you want s-1's and you want to underclock them and undervolt them.  they are very cheap used and the fans run very quietly.

I am in the north east . NJ

I'm have hashfast chip boards (VMC) watercooled with corsair H100i - they are taking air through wall holes from outside and blowing towards inside the house.. noise is acceptable, I mean they are not quiet, but after a while they are not bothering either.

Packaged into MiniITX case:


now I bought 2 more boards to build heater-miner to my parents with rPi. I like it - nice and compact and liquid cooled!

Although its not winter now, it would have been nice to get a kilowatt to try and quantify the .5 we are estimating.

the cost of the miner should make it worthwhile. if the electricity is 'free' because you'll be using the heat anyway, then the cost of the miner should easily be less than the value of the bitcoins it mines. its a no brainer!  (i was planning to do the same for my home in england.. no boiler.. just bitcoin miners)

-- Jez

Hi Dogie,

Thank you very much for the reply. I used 2 x 1.5 kw heaters last winter, running 14 hours per day (factory is open two shifts).  So, 3.0 x 14 x .5 = 21 kwh/day.

The S3 claims .34 at wall so I can now factor in cost of the miners and see if this is a worthwhile endeavor.

Cheers,

Les Cow

I think you'd be surprised how little you'll need to heat a space. You probably had a 1.5KW or so heater in there, which was 'on' 8 hours a day, but which was only actually heating for half of that time. So you use 1.5 * 8 * 0.5 = 8kwh per day of heating. You then know how much capacity you have for mining, adjusting numbers for your specific need.

In terms of noise,  I prefer the Antminer S3s.  The fans on it are pretty low speed,  and don't sound like a like a hairdrier or jet engine like the SP machines do.  I'd just get a high quality, single rail 1200w+ psu (you can buy the corsair AX1200s for reasonable prices on ebay) throw two or three S3+s on it,  and you are good to go. You could also buy a PSU designed for quiet operation as well (The RM1000 is the only unit I'm familiar with,  which can easily support two overclocked S3+s)

If that turns out to be too loud for you,  you could look into soundproofing the area that you put your miners in,  but I don't think you will have any problems in a noisy office environment... it should just blend in with the ambient noise.

Curious to find out what your setup will be and how it goes... keep us updated!

Hi All,

In real life I manage construction at a manufacturing facility.  For an office I occupy a 24 foot by 8 foot enclosed trailer.  It is stored outside in the North East US.  In previous winters I have heated the space with small electric heaters.  I am now thinking, why not use miners to generate heat, support the bitcoin network, and generate some dust.  I understand mining is not profitable, but I am already burning electricity to heat the space.

Before anyone starts with the angle that I am stealing the electricity - I own the business, factory, trailer...

In order, my priorities are:

1 )Adequately heat the space
2)  Not excessively loud as it is my office
3)  Ease of use.  I have not mined before nor do I write code or develop software.  I do have an engineering degree and I am pretty handy, but computer science is not my background.  Ideally I would like as close to plug and play as possible.  I do not have much spare time.
4) Generate dust.

Any suggestions on hardware to achieve my goals would be welcome.

Thank you.