Topic: Question for experienced "semi-large" scale bitcoin miners (Read 10287 times)

 The Evap(s) should be placed at the input of the "duct", but otherwise that's somewhat similar to my current setup where I am at now.
 I'm doubtful it would work AS WELL as where I'm at, as the air around here tends to be very dry - but it should help some.
 DEFINITELY invest in a humidity meter to keep an eye on the RH at the intake of your miners when you try it out - I rarely see anything over 40% RH (and it has to be actively RAINING outside to get close to 50) but the humidity in my area tends to be very low.

 I don't bother with the "duct" as such - I put my evaps in the "hot output air" area and the output of them pointing at my air intake area.

 I currently have 2 Brisa "window" type units (BW3004 units IIRC, both bought during "end of season" sales from the local Ace Hardware) stacked on top of each other with some boards for spacing, and enough space that I can add a 3'd one on top if I need it in the future.

I forgot about this thread haha, I just recently got my mining operation up and running and it's coming along well! I'll post a photo of my mining operation here for anyone else curious, but I'll link the main thread below where you'll find the rest of them. Funny that you're in central Florida, what is probably only a few hours from me. Message me if you have any questions or just want to chat!

https://bitcointalksearch.org/topic/antminer-farm-now-operational-2359425

I am in a very similar boat here in Central Florida. I am no HVAC expert but was wording if the following would be possible……take a portion of my existing warehouse and build two small insulated rooms divided by Server Racks. Then use an evaporative cooler set up in a closed loop configuration, similar to the attached diagram. The humidity output by the miners should be low enough for the evaporative cooler to work properly. The server racks would pull cooler air from the room with the evaporative cooler and push the exhaust to the other room. There would then be a duct that connects the two rooms together to feed the evaporative cooler.

Am I totally crazy or would something like this actually work?

https://image.ibb.co/merfiG/diagram.png

I'm not actually planning to operate at 9c/kWh personally; that is the price I'd be paying for electricity that *isn't* generated by solar, if extra is needed. I will be competing with huge farms with power costs of under 3 cents/kwh, of course; however, I'm already in a state that is dedicated to increasing the renewable energy share, and through the SREC program my state is basically forcing the utility companies to subsidize private installations of solar energy. I'm planning to take advantage of that to install solar energy at a cost that will be fully amortized over time, though it will require a fairly large initial investment of up to $200k or more. That being said, renewable energy will continue to have applications even if bitcoin mining proves unprofitable with my farm operating entirely self-sufficiently (aside from repair/replacement costs for the S9 units, which are not terribly reliable from what I've read) - decentralized computing isn't going away, I think, and even if there are no profits available in that I could retreat to the safety of running server farms for others. There are several people on here who seem to offer hosting for 3-6c/kWh, and I'm sure there are others out there as well. The internet's a big place.

When I say the solar power will pay for itself within 5 years that is perhaps a bit overly optimistic, you're right about that; I've done some calculations for 20% lower efficiency than rated rather than the 10-25% higher which is apparently the norm for a well-functioning system. Even if it's coming in at 20% under for 80000kWh/year, that's still enough to earn $12-18,000 a year selling SREC's - and that does not actually factor the cost of electricity in, only the credits I'll be able to sell to local utility companies. I do still need to do some research on pricing for ground-mounted systems, though; they might be priced higher than I was thinking, the average I've come up with from browsing NREL files is about $2.30/watt installed for commercial solar projects of 100-200kW with a premium of $.40-50/watt for ground-mounting. I still have yet to ask my solar contractors what their price point would be, since I'm finalizing another contract with them now for a roof-mounted system to provide power to my house. After comparison shopping with 8 other installers in the area, the one I'm going with did eventually come in with a bid using premium equipment at about $.20/watt cheaper than the national average for roof-mounted residential systems, so I may be able to get the ground system for a better price as well. If I pay the price that the NREL files estimate, I'd be paying $280k for a 100kW system and then receiving a federal tax credit of nearly $100k, so I could be almost presidential and dodge my taxes for the next 30 years - the remaining $190k or so would be amortized over 13-15 years at worst, based solely on selling energy credits and leaving me with 1,200,000 kWh produced over those 15 years. Even if I couldn't find anything to do with all that energy, I would be able to sell it back to the utility companies at wholesale rate due to my state's favorable net metering laws. If all these worst-case scenarios come to pass, I'll still end up with a large solar installation and about a $40k profit from selling energy at wholesale costs. That also assumes that bitcoin mining, other cryptocurrency mining, and all decentralized computing ends up being a total bust even at those wholesale electricity rates and that I'm therefore better selling the energy for wholesale rates rather than using it - which seems like worse than a worst-case scenario, really. In that case, I'll *still* make a profit, just at an opportunity cost of not being able to pursue other profits. Find me an investment that doesn't have potential opportunity costs, though.

*n.b. Most solar panels are actually warranted to produce at a minimum of 82% efficiency for 25 years, so the assumption that they'll operate at 80% efficiency is again worse than a worst-case scenario, unless we're getting truly fanciful and assuming that my contractors, the solar panel manufacturers, the inverter manufacturers, and the warranty insurers all go out of business at the same time my system fails and that I can't find anyone to repair them at semi-reasonable costs. If that's all the case, I think we should also assume that hydroelectric production will be interrupted because all the water has dried up due to global warming; and that world governments have descended into anarchy due to rising civil unrest, overpopulation, limited food and water supplies, and the disruption caused by the Martian war against the Illuminati, the Bilderbergers and the lizard people.

 The problem is that 9 cents / kwh is very iffy for making ROI on - and that solar power investment will probably take longer than you're thinking to recoup it's investment.
 Keep in mind that you are competing with HUGE farms with power costs of under 3 cents/kwh (one specific well-known example, MegaBigPower aka GigaWatt in East Wenatchee, WA) - they're going to be profitable with the S9 for quite a while after the "next gen semi tech" makes it into miners, while you'd be struggling to break even WITH the new tech.

 If you're serious about mining, you move into Cheap Hydropower areas - central Washington state, most or all of Quebec, part or all of Labrador being the "hot spots" in North America due to large hydropower projects that subsidise low local rates through selling off excess production to HIGH cost power areas - though there are a few other places you might get to 5c/kwh on "industrial" scale mining, usually ALSO close to large hydropower projects.

 Funny note - the Grand Coulee Dam output almost all gets shipped quite a long ways from the dam itself, as the "local" county PUDs own their own Columbia River power dams with lots of excess capacity and doesn't NEED power from the Grand Coulee.

 *IF* Bitcoin resumes a price climb, and that price climb stays ahead of or stays very close to network hashrate increases, then you should make good money even with your "high cost by mining standard" power.
 Keep in mind though that sometimes "IF" is the biggest word in the English language.

Maybe you're right and I'm just being overly cycnical. I'm planning to make this a long-term venture rather than just banking on being able to capitalize fully on having better technology for the first few months it's out, though. If the L3+ really perform even close to how you're saying, the pre-order list will probably be very long already and it won't be possible to get one at a good price for several more months, much less 20 to spend $40k on. Bitmain does mining itself as well as selling parts, after all; and even now, trying to order Antminer S9's from Bitmain itself would be a 2-month wait according to their store page. That's with the unit currently in production, having been out for over a year, and with competitors launching soon. If we're just considering the L3+ now, we're probably late to the party, and considering it'll take months to be able to order them anyway I think it's worth waiting to see what sort of results people actually get before putting down a large unrefundable deposit. Personally I'll be moving forward planning on 5-10 S9's and trying to figure out how to get the whole mess powered by solar, if I have enough room for that many panels. The initial investment cost will be higher but once the solar has paid for itself I'll be working with mostly free electricity, giving me a decent competitive advantage on efficiency compared to all except the largest mining firms or other renewable miners. If BTC somehow ends up not reliably returning a profit as time goes on, even as the 16nm barrier remains and S9's stay top-of-the-line, I can cut my losses and sell the miners while still keeping the solar as a framework for some other more-solid decentralized computing work. The solar itself is virtually guaranteed to hit the break-even point after 5 years or so, at least in my state, and after that will continue to generate income for another 10 years as well as free electricity for probably 20+ years more. Even if the S9's only end up producing $100 a month, reflecting the fact that running them at a cost of $90/month would be the new standard, that's still a 1-year ROI on the machines themselves and the rest would be at profit rates nearly as good as what I'm currently earning with my standard investments.

Hey, the concept of cryptocurrency to me is too good to be true, so I think anything is possible. When you hear people say that people are becoming millionaires overnight because of bitcoin and crypto, they're not making that stuff up, it's actually happening. In terms of real investments (stocks, real estate) 50% ROI per month obviously is absolutely nuts. Even 5% per month like you said is insane.

If you run the numbers, I figure the worst case scenario is that the difficulty of LTC quadruples when people start firing up their L3+ miners in September and the price remains where it is now without rising at all. Even in that very unlikely scenario a $40k investment into L3+ miners right now would net about $11,000 a month which is still nuts. Now cut that in half in case you think my worst case scenario wasn't worst case enough. Now you're looking at $5500 a month which is 13% ROI per month. Even if the worst of the worst happens (relatively speaking) you're left with 13% per month which is still considered unobtainable in the real world of investing.

Not a single time out of ten would I pass up an opportunity like that.

Fair point, they might not even stay profitable for long with grid-provided electricity, depending on how long it takes the technology to become obsolete or commonplace. The price also might continue to rise as the difficulty goes up due to more money being invested in the infrastructure. Bitcoin might be made illegal in more countries as Ethereum continues to attract governments, which could cause a drop in prices due to the currency being less real-world-applicable. I'm not necessarily anticipating that things would stay the same in Bitcoinland for two years, though at this point there's no technology available to surpass the S9 that seems to be coming up on the horizon any time soon from what I've read so there should be some small degree of stability. It seems the price has consistently risen in the past based on difficulty increases, which is only logical; more people trying harder to generate the same limited resource. If the people trying to generate that resource (hypothetically) don't invest more in it, the difficulty shouldn't rise appreciably. Obviously it will continue to unless something happens to stop it, at this point, but the price should rise as well making the mining continue to be at least marginally profitable even with average electricity costs.

That being said, my plan is to install solar power and use renewable energy to power the miners for the vast majority of the time. If there are overages at the end of the annual "true-up" period it would cost $.09/kWh, if I've generated more than they use to run I'd be refunded somewhat less based on wholesale prices. I'll probably aim to have miners using slightly more energy than I generate for that reason, so that I wouldn't have to sell it back at wholesale, but would therefore end up paying $.09/kWh for some of the electric costs. As for the solar power, it'll probably take about 5 years to pay for itself from SREC's and then continue to generate more income on top of the electricity generated, so it's pretty much a no-brainer as long as I have enough investment capital. Clearly this isn't something that would happen with 20 miners for a $50k investment, as GMPoison was originally talking about, so I may be straying off-topic somewhat discussing my own plans; I wanted to discuss the possibilities of doing it entirely with grid-provided electricity as well though. It would be possible, in a state with good solar laws, to do the same on a smaller scale with a $50k initial investment so it's not entirely off-topic, just more limited in size. Either way, it should be able to continue to generate a profit as long as the S9's remain viable technology at all.

 Difficulty rise as more S9s (and other units) get sold is what you're missing.

 Most calculators don't even allow you to ESTIMATE that and factor it in - that's why I stick with BitCoinWisdom even though they haven't fully updated it for a while.

 The only serious limit to how fast the difficulty went up for most of the last year is how fast Bitmain could get the chips to make more miners with - but it looks like Bitfury is FINALLY getting into the game on the current node generation, and at least 2 other companies have publically entered the field with their own chip designs (and who knows how many miners BW.COM has been pumping out for their internal usage after they decided to NOT sell their current-node-gen miners to the public after all, along with at least one other large private farm I'm aware of that's working on their own chip for internal use, and what the heck is MegaBigPower AKA GigaWatt planning to fill THEIR new mining halls with?).

 The question isn't "is difficulty going to double", the question is "how long will it take for difficulty to double".


 While "many" and possibly even "most" commercial spaces have 200 amp 220V service in the US, I've been in quite a few that did NOT - retail shops and warehouses in particular are known for sliding by with lower power service since they're intended mostly for storing stuff.

If it sounds too good to be true, it probably is; if it isn't already, it probably will be soon. A 50% ROI per month, to me, certainly sounds too good to be true. Even that potential 5% ROI per month from S9's is exceptional, in terms of real money. Consider what that $22k/month is dependent upon and how much of a possibility there is for those factors to change quickly, because if they can they very likely will.

That would be really cool to build out a property for mining specifically with ventilation and power in mind, but that's way above and beyond what you would ever do for a $50k investment into mining, no? I also can't imagine they would be much of a fire hazard so long as the proper power supplies are used, proper cables are used, a proper electrician installs everything, and I get a very good system for airflow going. Also now looking at the L3+ rather than the S9 or T9, the profits off a $40k investment aren't looking like $2,200 per month but more like $22,000 per month, even at $0.10c/kWh.

This is very nicely described. Indeed the problems you described are very common and I think it is really not a good idea to put all the set up into residence or commercial building. There is safety concern in regard to our mining love as they get heated up quickly and many times fire hazards has been reported with them around the globe. Its better to build your own small shade away from city. This would be good for cooling as the global warming effect is lesser in the area away from city. You can let the noise disappear into empty space around. With open space and no tall buildings around you can set up three to four solar panels to harvest electricity from the solar energy. This will atlas cover the 30-40% of your energy cost. This will give you push factor to gain profits quicker. Nice concept and ideas have been implemented already and you can try searching over the internet for best cost effective project. If sat up properly it is really profitable thing. 

The numbers I ran didn't account for how much power it would eat up to keep the space quiet and cooled, and I came up with an estimated $90 spent per month per S9 to generate .1 BTC... if I were actually paying for the electricity. Assuming I went fully with air conditioning rather than exhaust cooling, that'd cost another 1 kW/S9/hr and go up to $180 per month per S9 to generate .1 BTC. As you said, fairly marginal but still profitable, unless I'm missing something. The price on bitcoin would have to fall by another 33% for it to be unprofitable. Certainly possible, considering we just saw a 20% drop, but at a profit of only $70 per month per S9 that would still take under 2 years to fully pay for itself. I wouldn't be investing for a little while anyway, which would give me some time to see how the upcoming halving of the reward rate will affect the profitability.

The addition of solar into the equation does help either way, though. My state has net metering laws that restrict interconnection on solar projects exceeding the site's annual needs, which means that the bitcoin mining should enable me to install/interconnect more solar and profit on that as well. If bitcoin mining ends up so difficult that I'll be operating at a loss, I can disconnect the S9's and move to Ethereum or something else profitable, or just generate electricity and sell it back to the grid at wholesale price while earning an extra $200-225 per 1000 kWh on SREC's. Even that is a profitable, though long-term, investment - so the electricity coming at $.09/kWh is more of a worst case scenario. Ideally I'll end up able to pay all the electric costs for this with the solar panels; though 20 miners with air conditioning might be a bit of an overambitious initial investment on solar, I don't have $1.2 million lying around that I'm willing to spend on a 400 kW solar grid. If I go with 5 of the S9's and use cheaper cooling methods, I could bring that down to a 100 kW or lower ground-mounted system and expand it depending on how profitable it is. A 100 kW system would generate $20k a year or so just in SREC's as well as providing electricity for several miners, and should cost $2-3/watt installed worst-case - a 10-15 year ROI if you completely discount the tax credits and electricity generated. If it's unprofitable with Bitcoin, I still have a decent renewable energy setup to turn into some other sort of server farm to rent out for decentralized computing in a setting other than Bitcoin, perhaps another cryptocurrency if they continue to be popular for speculation. Anyway, assuming that the solar will pay for itself and then some over time, the bitcoin mining will then be happening at wholesale electric rates for the most part and only going to $.09/kWh for overages. It seems like that should be enough to turn a tidy profit on the whole operation, better than the stock market and less risky anyway.

9 cents / kwh is going to be VERY marginal on if you will achieve ROI on the S9 - depends on how many of them Bitmain can pump out in the next year coupled with how many competing units the competition can manage to pump out.

 I've never owned a S9, but have been told they're somewhat louder than the S5's I HAVE operated.
 They're NOT quiet.
 With that said, it shouldn't take much baffling to keep the sound getting out enough to bother the neighbors - most of it is fairly high pitched, you're not dealing with a "boom car" deep bass type of sound that's HARD to block out.

 Heat generation is easy to figure - 1000 watts of power used 24/7 is appx 3400 BTU/hour.

 S9's use somewhat more than 1 kilowatt (depends on the specific batch how much more per miner) - 20 of them is going to add up to a LOT of heat.

I'm considering an investment of about the same amount for 20 S9's, though I may also have to look into L3's after reading the last few posts. I'm not particularly experienced, but I've run preliminary numbers on the S9's and they seem like a solid investment as long as they last for longer than a year. Does anyone have any info on their longevity, or is it too early at this point? They've been released for near a year now, so if they're not sturdy enough to consistently hit the break-even point I'd think that should be starting to become clear by now.

I'm not sure about renting a warehouse, it'd be a bit expensive in my area and more of a hassle to operate. Just how hot and noisy are these miners? I'd like to put them in a building separate from the main house on my own property if possible, as that would also let me install solar panels and generate a portion of the electricity for the system as well as generating income from SRECs to offset the remaining electricity cost. I'm in the Northeast, so the climate should be a bit more friendly to this kind of thing, but might need to look into insulating a building for sound if they're extremely noisy. My electric costs are roughly $.09/kWh, but with solar panels I'll be able to yield an extra $200 in SRECs for every 1000 kWh generated that way. At that rate, having air conditioning might be feasible if necessary; noise would definitely still be a consideration though, and in fact even more so. It looks like the S9 runs at about 77 dBa, which would put 20 of them at about 92 dBa. That might not even be legal in a residential area without some serious noise reduction. https://www.youtube.com/watch?v=LvzP4g1bfh0 seems like it could be a solid strategy, though 20 in that kind of setup might not be possible without expanding the space I was thinking of using. Any thoughts on all this?

Ah. Yeah if I go with a warehouse it must have not only the large garage doors but a small standard sized one as well, because I'm trying to come up with a solution to stack fans on top of each other to maximize the space of the door and not just have one row of fans on the floor. I would find a way to chain them all together along with the door to the ground, so you wouldn't be able to get in or out of the large doors for better security. I'm sure there is some industrial fan out there that is meant to be stacked for what I'm looking to do. If anyone has heard of anything like that, post it here!

 One space I was looking at to move my operation into was a smallish warehouse with a garage door on each end (part of a "warehouse/shop" building complex).

 The primary reason I decided against it was that there was no "man door" at all, and my knees don't handle stairs reliably any more - both garage doors were at "pickup/semi trailer" height.

 Airflow would have been trivial to set up though - louvered "inserts" under each door and straight-through airflow.

lol. just as I assumed Thanks for the tip.

On your project if I may give a different point of view, I am planning to open soon in Cadiz. Spain. Hot and rather humid too. My approach is not to avoid and get rid of the heat but to reuse it. I have contacted some research groups in the local university to explore the possiblibilty of using the heat to grow some useful algae, catalize some profitable organic reactions, produce some pressure batteries. I have to check too if this can be applied for some industrial process, in particular membrane water desalination.

I will be happy to share it here if I find something useful or the things I try that fail. by now is merely some exploratory approach.

Yeah, just run the numbers on the Antminer L3+. Haven't done much research into LTC mining with GPUs and how profitable that would be, but even if you (and should) account for a huge bump in difficulty when those L3+ miners ship out August/September, the profits are mind blowing. Before doing some research on LTC a few weeks ago I was convinced that I'd be going with S9's but no way now, if anything I'll mix some in for diversification. I thought LTC was dead so I didn't even consider it, haha.

That´s a curious perspective that I haven´t measured. I am going all the way now for some 20 s9 after testing profitability for 3 months but I haven´t tried litecoin mining.