I am not against overbuying on a quality PSU, well I am to a point, but linking some benchmark showing some extreme wattage usage does not help a beginner in any way and that was the main point to my rebuttal. Also, I think if you are getting into mining you do need to also learn how to tweak and optimize your rigs. Just throwing an overpriced excessive wattage PSU at non-optimized settings may have worked in April, May, June, July, and perhaps August for ROI purposes, but going forward people getting into the game are going to need to learn to optimize, optimize, optimize.

As far as the ROI question, I am with you 100% but I am tired of pointing out the obvious to people who only want to know how to get into mining. So I stopped pointing that fact out and now just offer advice on how to build rigs. But this again empathizes my goal of not overspending where you do not need to. Spending an extra $50-$75 on a oversized PSU will only lengthen that period.

So I recommend to the OP once he buys his gear, which I stand behind in my listing above, that he also comes back and asks about how to tweak it for optimal performance. There is also no harm in buying the base components that will support a a 5-6 GPU rig and only starting with 3-4 GPUs to learn these intricacies. Once you have this down you can always upgrade your rig to its maximum potential. This way you will have the best of both worlds, a reasonably priced and sized rig and you wont burn your house down in the interim period of learning how it all goes together and works.

Th only bad advice around here is yours. I have personal experience with many such builds and you can run 5 or even 6 RX580s off a 1000 watt PSU easily. If you are pulling over 200 watts per card you are definitely doing it wrong, and those benchmarks you linked are showing extreme cases, which again if you are running your mining rig like this you are doing it wrong. The guy is asking advice on how to build a rig right and mindlessly telling him to buy a larger PSU than he needs because you link some extreme un-optimized use case benchmark is foolish, especially when the price starts to rapidly climb past the 1000 watts range.

A RX 580 properly tuned meaning under-volt mod applied with clocks and voltages properly adjusted will draw 130 watts when mining ETH only and ~150 watts if dual mining. So for ETH only we could do 5 x 130 = 650 Watts, dual mining it would be 5 x 150 or 750 watts, well within reason. Even if he went 6 GPUs (which I only speced 5 for the AMD RX 580, the 6 was for Nvidia 1060's) it would be 780 and 900 watts respectively. While a 6 x RX580 dual mining configuration might be starting to push the PSU, it is still well under the maximum. However, a 6x RX580 just ETH only would be well within reason at 780 watts, or under the 80% load guideline I like to run around.

With Nvidia, the 1060's run under 100 watts each when mining ETH alone, so that would be 600 watts with 6 cards. And for his budget he could get 3 1070's which might be around 200 watts each, so again 600 watts. So tell me again where is the bad advice?

Yes, $2000 is probably about the average investment in mining equipment. With that amount you could build a nice rig of 4 to 6 cards and the associated support hardware. Much of the choice will come down to which coins you want to mine. Ethereum typically favors AMD and Equihash (Zcash) favors Nvidia, but they both can mine the other algorithm in a pinch.

Typical ETH build:

5 x RX580 8 GB - $310 ea or $1,550 total
1 x mobo with support for 5 or 6 GPUs - ~$120
1 x CPU (Intel G4400) - ~$50
1 x RAM 4 GB - ~ $30
1 x 120 GB SSD - ~$50
5 x PCIe USB risers - $50
1 x 1000 watts PSU - $150

This would put you right at your $2,000 budget. You would also need some type of frame to hold everything so you might need another $50 or so in materials.

A Zcash (Equihash) build would be much the same except you would pick some Nvidia cards such as the 1070's or perhaps the 1060's instead. The 1070's run around $500 so you could run 3 x of them and the 1060's go for $200 for the 3 GB version and $350 for the 6 GB version so you could run 5 or even 6x (if your motherboard supports it).

Actually I feel quite to opposite, it is probably a good thing it caught fire but did not burn his house down. This will be a very strong lesson and reminder for him going forward to do things correctly and not cheaply. Imagine if that had started on fire in the middle of the night and say he has young kids in the house. While it may suck to lose a few pieces of hardware, the result could have been much worse.

Anyway, I also agree with the consensus that you should not cheap out when building a mining rig. Buy a quality name brand PSU and only run it at most to 80% of its capacity. Also same is true with the circuits feeding your miners. If you are in the US that means a circuit can only continuously operate at 80% of its breaker rating, thus a 20 amps breaker/circuit can only supply 16 amps of current. On a 120V line this would only allow about 1900 watts of power, so this could feed only two 1000 watt PSUs running at 800 watts (to system) and will probably draw 850 watts (at wall) or about 1700 watts. Keep this in mind as well when adding rigs to your home to help minimize fire risks.

I have been through a couple of winters already with many miners in my garage and a shed in the back yard. Both areas are unheated, well except by the miners of course, and where I live the wintertime temperatures often drop below 0F.  As others have already mentioned, it is not the temperature as much as it is moisture in the form of humidity and perhaps condensation that you need to worry about.

As long as the miners stay running you should not have any issues as their natural heating will keep the area somewhat warm and since warm air holds moisture better than cold air, the humidity will typicality be lower inside your mining area than the outside anyway. The only time you would need to worry is if the power would go off for an extended period of time, thus chilling your equipment and room close to the cold outdoor temperature. In that case you would need to take some precautions when bringing the equipment back up to prevent condensation from forming.

Since condensation can occur when your bring a cold object into a warm environment, you need to watch out for that possibility. If the area is enclosed enough you could try to bring a regular space heater out there first to warm up the room a little before turning a miner on. Be sure to check for condensation on any of your miners first though as the danger will mainly be at the beginning of this process. Bring the miners up slowly with one or two at a time to gradually raise the temperature. This will help prevent further condensation from forming on other surfaces within the room that could drip on your exposed miners.

Again, I never had such concerns as my power is pretty reliable where I live, but if you did have an outage of over 12 hours in bitter cold temperatures it is better to take precautions. Waiting for a warmer day with low humidity would be the best option and as I mentioned earlier bring an external heat source into the room would help warm and dry it out a bit before trying to turn on all your expensive gear.

There are bots scanning popular websites Bitcoiner's frequent that will scoop that money up in literally seconds after it appears. Even if you caught the mistake as soon as you did it and tried to move the money off the wallet it would already be gone since bots will do this nearly instantaneously. And it is more than one, the bots are in competition with one another as well so its not like you would stand a chance. It also doesn't matter which currency as I assume most of them are programmed to work with most of the popular coins.

Using the park bench analogy above, it would be more akin to leaving your wallet on a park bench during a crowded thieves convention being held at the same park.

And so is Jamie Dimon and JPMorgan/Chase despite what he publicly says. There does seem to be a concerted effort to bundle a bunch of so call bad news in packages at times perhaps in an effort to drive price down so big players can load up. I am not saying this was his intent, but if there is money to be made then you can be sure bankers will be interested in it. The crypto space doesn't yet quite have the volume that these big players need, but that will come as time goes on and I am sure in 5 years or so if he is still at the helm his outlook will have changed a bit.

Yes, and this is why step one of any transaction of size should be to first send a small test deposit to ensure everything is working as expected. This is even more true when you are new, but even for experienced people, it is easy to make a mistake. I would rather risk 1 small $10-$20 test transaction then end up losing $1,000's because of a silly mistake.

I would also suggest upgrading your PSU, especially if your current one is a no name or generic brand. I have never heard of Powerlogic brand before so it may very well fall into that category.

Some of the cheaper PSU's won't supply their full wattage, at least for any length of time, before going poof. Stick with decent name brand ones (Corsair, Evga, Seasonic, etc.) as mining puts stress on your PSU like no other computing task. Even when gaming you might push it once in awhile, but mining is full throttle 24/7 non stop abuse and your PSU needs to be built to take it.

Expanding on the advice you already got from the other posters, you also want to size it appropriately. Normally the recommendation is to keep your maximum wattage load to about 80% of your PSU's maximum output. So for example, a 1,000 watt PSU could comfortably put out 800 watts continuously.

If you plan to add more GPUs in the future, you will need something that can dish out the power so I would say buy the biggest PSU you can afford within reason. From your setup a 850 watt unit might fit the bill, giving you room to grow and still providing the 20% cushion. I say 850 watts simply because there are quite a few decent brands in that capacity right on either side of the $100 mark ($90-$110). With an reputable 850 watt PSU you will have about 680-700 usable watts. This would power as many as 4 x 1070's along the rest of your system with ease.

The only option to prevent dumping is what some of the people above you already stated. There should be no rush to get IRD listed on an exchange. I said this a few weeks in this same thread and will say it again; this coin still has a lot of polishing to do before it is ready to be listed on an exchange.

I also agree with the sentiment that the people pushing the exchange issue just want to dump their early "pre-mine" coins and run. I say pre-mine in quotes because while it wasn't technically a pre-mine in the traditional sense, you did have very limited number of people mining it early on getting most of the blocks to themselves, it to me it is the equivalent of a pre-mine. Instead they should be helping donate to a fund that will pay various bounties to promote the coin, if the coin generates enough interest it will be a lot easier to get listed on a decent exchange something in the future.

Actually I think you have that turned around, as the hashrate the pool shows, or more accurately the share rate (which the pools derives a hashrate estimate from) is what you are being paid for. I do not care what your console displays, if the pool thinks you are only submitting x shares per hour then that's what they are paying you for.

Again, the pool's displayed hashrate is simply doing a reverse calculation on your actual submitted (valid) shares and just ball-parking what an estimated hashrate would need to be to generate that many shares per hour based upon the current network difficulty. So even if this is off a little the important point is the number of valid shares you submit to the pool over a period of time, such as over an hour. So regardless of what the miner console displays it still comes down to what the pool received and that is what your payout calculations are based on.

Maybe he was approaching it from a "dip your toes into the water before going big" angle, but even then I think it would be better to buy a single 1080Ti than say 3x 1060's, price and performance is about the same along with all the other advantages I listed earlier.

I do not know what he is saying but I would prefer going with the 1080TI and run a single rig with 5-6 or them than run 3 times that many rigs with lesser cards. Unless there was some other advantage like less power consumption per hash rate or something the higher density the 1080TIs would provide will always work out better. Every new rig also incurs a new motherboard, processor, RAM, SSD, risers, etc. so the more powerful a card you can put in a rig the better, again assuming all else is equal. As far as price it works out about the same too, as while the 1080Ti is more expensive it is not more expensive than 3x lesser cards.

4 GB of RAM is plenty for a rig to run on when it is in production. I sometimes buy the two packs though 2 - 4GB sticks or 8 GB and split them between two rigs one now and one in the future if I am sure I will need it as you can save a few $$ that way. One other tip, while you do not need 8 GB for production, if you do have two sticks I have used them both for the initial install, drivers, and configuration, as everything seems to run faster with more memory,e specially when you are right there waiting for the install to complete. But once everything is setup and working fine, remove 1 stick and it will run just fine on 4 GB.

Also, be sure to setup your virtual memory or swap space in Windows for 16 GB as recommend by Claymore, I never actually seen it use this much but with a 60 GB drive you will have spare capacity anyway and with a mining rig it is not like you need the extra space anyway.

I really don't know why this story keeps coming back up again and again and why people cannot grasp one simple concept. He did not pay millions of dollars for a pizza, he paid roughly $40 or the going rate at the time for two pizzas delivered with tip. He also did not even pay the pizza place, as I recall it was more along the lines of he traded 10,000 BTC to someone who could get a pizza delivered to his house, hot and ready to eat. Someone took him up on the offer, ordered two pizzas from a deliver joint and had them deliver it to the BTC guy's house. After he got the pizzas he kept his end of the bargain and sent the buyer his 10,000 BTC. So at the time the BTC was worth approximately $40/10,000 or roughly 4 tenths of 1 cent each.

At the time you could mine up a bunch more BTC just using your CPU and active trading wasn't even a thing. So yes I agree that actions like this help spread the word about BTC and its eventual rise, but to keep saying he lost or wasted million of dollars is just absurd. Thinking like that you may as well horde every item real or virtual you encounter as it could one day become worth a lot of money. Further, had all the early adopters horded it, it would be worth nothing today as their is not much of a market with just a few dozen people trading virtual tokens back and forth. It gets its value from be distributed to the masses, not by a handful people hording it all.

Actually 60 GB is plenty for a Windows install running just a miner program. I have several instances of Win 10 installs that only use around 45 GB of space, including a 16 GB swap file, drivers and a few copies of mining programs. One storage hog you may run into is the hibernate feature which will consume as much space as your physical memory. I usually turn this off from the command prompt since a mining rig will not be hibernating. The command is a simple:

Also 80 GB is not a real common size in SSDs, seems the popular sizes are 32 GB, 64 (60), and 128 (120) and so on. The price difference between a 60 GB SSD and a 120 GB SSD is often only $10. I suspect the manufacturers really don't make the small units but instead bin the parts and any with excessive read/write failures they can simply mark that part of the SSD bad and half the size.

For example if they only run production on 120 GB drives as a minimum, maybe if the failure would peg it too far under the threshold they would simply make it a 60 GB drive. Most SSDs have unallocated extra space anyway for normal failures over its life span where the drive logic can map the spare space as needed, so it would be a simple matter to also program them to recognize only 60 GB even if it had 100 GB usable, but less than the 120 GB.

So since the production costs are the same, they simply sell the better binned parts at a higher cost since they can use the full storage space. This is similar to CPU and GPU core binning practices where cores that don't pass the higher quality threshold may still run fine at a lower frequency or with less cores activated.

I prefer smaller rigs as well for the easier sourcing of parts and just general ease of administration. Like mentioned above, I consider a smaller rig to be the 6-7 GPU variety with one PSU powering it all. While the big rigs do have the WOW factor they also have their share of technical and administrative issues as well.

I have built a grand total of two > 7 GPU rigs and I think I will stop there as while they are nice to show off to friends they are a big pain in the rear to do just about anything with. The savings really isn't there either as usually the motherboards that support more than 7 GPUs are fairly expensive and since most rigs are built with a $50 CPU and a $25 stick of RAM, the extra cost really isn't much savings anyway.

Also as the poster above me said, if a rig does go down, this way you only have 5 GPUs out of commission instead of all 10 being down at the same time.

For one, what is reality? To assert that something has no use in the real world, one has to first ask what constituents the term real?

By your post I assume to you it means the current state of affairs. This is a pretty narrow and not very enlightened way of looking at it though. If people lived by your model of utility in a "real" world we would probably all still be using stick and stones.

No, the world and thus our reality is instead brought about by our thoughts, dreams, inspirations, aspirations which all lead to better ways to do things, better tools and inventions to help us mold and create our environment. Throughout history people have resisted change and new technologies often citing they had no practical use in this perceived real world of theirs. Well sorry to burst your illusions, but the world (thus reality) does not stand still and it will change and what you deem useful today may soon become obsolete as progress marches on.

I hate to be harsh but why even bother to ask for advice on this forum if you are going to ignore it? Obviously you have it set in your mind what you want to do already, just go for it.

Plenty of posters have already suggested not buying more cards as the $20-$30 per day per rig times are over.  Right now (this moment) with your setup (5x 1070 and 1x 1080ti) and assuming 8 cent kWh power you can earn maybe $10 a day with your entire rig while mining Zcash. A few months ago that was probably $40 a day. A few months from now it will probably be $5 a day, a few months later it might be $2 a day.

I do not know what you paid for your system, but looking online a cheap 1070 is around $400 and a cheap 1080Ti around $700. So just considering the investment in cards alone you have around $2,700 locked up in hardware. It may even be closer to $3,000 if we add in the other components, but since you haven't indicated if you already had the rest of the system I will let it go for sake of argument.

So if everything stays rosy and we do not drop below $10/day in profit you are looking at 270 days to get your money back, or ROI/breakeven. However, if we consider in 3 months time we seen a decline from $40/day to now $10/day in profitability, it is not unreasonable to think that it is possible daily profits could continue to go down even more in the future. So maybe in another 3 months you are looking at $3/day in profits and that 270 days turns into 365+ days. Basically you are forever chasing a ROI that may never happen.

Even if the price of a coin goes up, you may find a temporary relief from the loss of profits, but as profit goes up so does the number of miners who are adding more hash power to capture a little piece of that pie. So over the longer term, even with increasing coin prices, since hash power is also increasing your profit will continue to erode.

However, I don't expect you to read or even consider any of this. I would not be surprised to see in another week you proudly inform us that you now own 2 rigs of such a setup. Well I can only wish you good luck in your endeavors.

As has been said an SSD is not required for mining, but it does help with boot times and is probably more reliable over a longer term than a mechanical HDD. Plus if it is in a rig in your living space you do not have the HDD noise to contend with.

My advice would be if you already have some spare HDDs laying around by all means use them first. If you are buying new, a bare-bones OS and miner image does not take much space and even with Windows you could sneak by with a 60 GB drive, but the price difference is fairly small so 120 GB SSD is common. I have bought many 120 GB SSD (I believe they were Toshiba) for as little as $30 on sale. Another option, although more of a luxury item than a need, is a M.2 drive.

My last couple of rigs I have bought M.2 drives and eliminated dealing with an external (to the motherboard) drive completely. This way I can eliminate using SATA power cables completely and do not need to deal with a HDD or SDD flapping around on the rig. I probably would not recommend this route though, as even on sale they are fairly expensive as a 120 GB drive is still around $60-$70, or about double that of a normal SSD.

If you go Linux you can even get by with a USB flash drive. These are fairly economical and with a standard image easy to duplicate so you can deploy new miners with only minor adjustments to the image file.