I don't know about other countries, but in the US it is definitely NOT a scam. In fact within the past 6 months I sent in two Corsair PSUs that I had that were about 4 years old, and they would randomly reboot the rigs they were in. These were 850W TX units which Corsair doesn't make anymore and in about 8 days turnaround time I had two brand new RM850x fully modular replacements. Cost me $20 in shipping for both, so I do value a long warranty.

As far as the calculations, it of course varies on the exact efficiency, power costs, etc, but even that $0.15 cent a day saving so many of you are poof-poohing, it still adds up to $54.75/yr. So even at only 4 years it would be nearly a $200 savings. I think spending an extra $40-$60 upfront is worth the extended payback.

Yes. Put in in your batch file like below (for Windows):

I wire my miners up using NEMA 6-20R outlets and NEMA 6-20P plugs, an example receptacle is shown here: http://www.homedepot.com/p/Eaton-Commercial-Grade-20-Amp-Straight-Blade-Single-Receptacle-with-Side-Wiring-Light-Almond-1876LA/203492424

I use 20 Amp double-pole (240V) breakers and usually run about 4 outlets off of each breaker as my miners only pull about 3A-3.5A each, so my total circuit load is at most 14A which is well within the 80% NEC code specifications which allows a continuous 16A load.

I also use NEMA L6-30R receptacles (fed by 30 Amp double-pole breakers) in my garage to power my APC PDU power strips, and I can run about 24 Amps (7 rigs) off of each one of these: http://www.apc.com/shop/us/en/products/Rack-PDU-2G-Switched-ZeroU-30A-200-208V-21-C13-3-C19/P-AP8941

These PDUs work great for the rigs I mount on wire-racks. You can usually find used PDU's online for $50-$100 depending on capacity, features, etc., as they are too expensive to buy new at retail. The main benefit of using a PDU is you can run fewer higher amperage circuits to power many rigs (especially if your circuits will terminate far from the breaker panel.) Some of the better ones also have individual breakers for each receptacle and can remotely monitor power draw as well as switch on/off the outlets via a web-page. This can be a handy feature if you need to remotely reboot or power cycle a rig.

Yes, this is exactly what I need to do as well. On a 4-GPU rig I use 600 watts mining either ETH and ZEC, but I need to change ETH hash-rates to a more reasonable 108 value and ZEC to 840. So the defaults are definitely way off.

Using my values from above, ETH currently shows it is earning about $0.75 more per day on this rig than it would on ZEC. The interesting thing is the ZEC network hash rate has kept rising over the past 24 hours, so not sure if site like What-to-mine influence this trend or not. I would think most miners shoudl be able to figure this basic error with the default values.

They weren't drenched, but it looked like someone took one of those spray bottles to them on a mist setting. I think it was more how the water came in through the top and ran down in-between the heat-sink and GPU. When I said washed out, it wasn't like it was scrubbed clean, more like the consistency of the paste was loosened a bit and it ran down and out.

You are probably right on the purity of rain water that allowed it to continue to run, at the time I was sure it was going to short out any second as I raced to unplug it. For obvious reasons I wasn't just going to reach in and switch it off by hand, so I shut it down via GUI and unplugged it via the surge suppressor cord some feet away.

I have no doubt that over time it will probably fail before the other rigs do. Its only been about 6 months now since "The Event" , so I will see if it goes another year or two.

Yes, it is well worth it if you go with a name brand (Corsair, eVGA) as some of those PSUs now carry 10-year warranties. If you plan to mine long-term you will more than get your money back. I have also been watching Titanium supplies (when they go on sale) as they can get close to 96% efficiency when fed with 240V, which expanding on the above chart would be:

750W load - 96% PSU - 781W at wall

The price premium is currently pretty high so you need to catch these on sale, but it is getting to the point where the PSU is becoming almost transparent as far as additional power consumption. With a 70W difference between the most (Titanium) and least (Bronze) efficient models, at $0.10 power this can be up to a $60/yr savings, so looking at a 10 year life span this would equate to a $600 savings, I think it will more than pay for itself.

Usually there is a main breaker in the panel (fed from the meter) to cutoff all the other circuits, the rating of this breaker determines the size of the service. But yeah, 100A sounds about right for a dedicated garage or shed service. I might look into that option myself as I would rather not have to push my house circuits right to the limit all the time.

I think I heard of this before too. I hope I never have to "De-moisturize" my electronics again, but I'll keep it in mind for the future. My other thought at the time was to put them in the oven on low for some hours, but the couple of days of air drying and blow dryer seemed to work.

For your first rig I would try to buy a gaming motherboard (MSI Gaming 7) I believe supports 6 GPUs and a better quality PSU (eVGA or Corsair). This way if it does not work out for you, you can either convert it to a gaming rig for your own use or resell it. The GPUs you will probably have no issues reselling, but again stick with the better name brands that are in demand.

Right now for mining Ethereum or Zcash, I would also recommend either the RX470 or RX480 series cards. Since these are still new, the resale won't be much of an issue at all, you can probably get 75-80% back should all else fail.


I also agree 100% with this advice. Buy a mobo that can support 6 cards like you want, but there is no need to buy all 6 GPUs day 1. Get it setup and running and once you know your values, such as hash-rates, power draw, equipment costs, you can better calculate if adding the additional cards will make sense or not. Most of us started out this way and not just jumped in with a fist full of dollars and no real plan.

I am running a significant number of cards this winter in my garage, so I am mainly double checking my logic as I do have concerns if all will go well.

I started running them in the garage this past summer because of too much heat in the house, and the worse experience I have had so far was when I left the back garage door open (for venting) and a storm hit with strong winds that practically had the rain coming in horizontally. It rained a bit on one of the rigs about 12' in from the door. Surprising, when my wife asked me if I left the door open because it was raining so hard, I went out and seen the rig was still running although there was visible water on it. I shut it off via the GUI and then unplugged it. I took everything apart and let it dry for a couple of days, as well as used a blow dryer to try and get the deep down moisture.

Anyway, I re-assembled it and it ran, but the GPU temperatures were very hot. Upon shutting it down and disassembling the heat-sink, I found the thermal compound had pretty much washed out. So I took all the affected GPUs apart and cleaned/reapplied thermal grease and tried it again. It is running for 6+ months now with no ill effects, but I do suspect it might end-of-life on me quicker than rigs that weren't rained on.

So I am not saying moisture or humidity is not a hazard, but I think modern electronics are a bit sturdier than we give them credit for.

My bigger concern with running out of the garage in the winter is a blast of super cold air when the garage door is opened on sub-zero days, but so far everything has been working out. In another thread a guy posted that lives in Alaska and claimed he ran his rigs in the garage will no issues, so I am banking on that experience as well.

Here is an interesting calculation based off of that idea. Most residential homes in the US come wired for 240v and somewhere around a 150-250 amp breaker panel. This would yield a maximum capacity of 36,000 VA to 60,000 VA, which pretty much equates into watts for sake of this argument. So using 60% of these values, you need 20% for other things in your house, and the other 20% is for the safety overhead (no more than 80% total), a typical house could host 22,000 - 36,000 watts of mining equipment. This would put they survey's medium and lower categories firmly into sizes you can run out of your home, and the larger categories would start to require additional power typically found in more commercial environments. So using watts or GPUs, I think the groupings are still sized appropriately enough for this simple survey.

Why would this be an issue? Condensation mainly happens when a cold item, such as the cards are brought into a warm humid environment. So if your cards are in garage and running, then the power goes out, the cards will cool off slower than the surrounding air, which is probably not that humid to begin with in the Wintertime. So say they are off for 12-24 hours and the cards are equal with the now cooler garage air temps, then the power comes back on, the cards will heat up before the surrounding air. In both cases the cards should be at or above the surrounding air temperature, never below.

I could see to definitely not store them in such an environment, but if they are running it shouldn't be an issue.

Can you not make some profit with Ethereum and/or Zcash mining? This should even be more so if you can offset your heat bill somewhat by leaving them run, which it sounds like you can from your description.

I like your watt idea too for classification. I put everything discussed so far into a little summary for comparison:

Hobby 1-12 GPUs, < 1,800 watts, < $2,500 investment, $1-$12/day expected income

Micro Farm 13-25 GPUs, 1,800-4,000 watts, $2,500- $7,000 investment, $13-$25/day expected income

Small 26-50 GPUs, 4-8 kw, $6-12k investment, $26-$50/day expected income

Medium 51-150 GPUs, 8-24 kw, $12-40k investment, $51-$150/day expected income

Medium-Large 151-300 GPUs, 24-50 kw, $40-80k investment, $151-$300/day expected income

Large 301-600 GPUs, 50-100kw, $80-150k investment, $301-$600/day expected income

Professional 601+ GPUs, > 100kw, > $150k investment, $600+/day expected income

I agree, I usually figure 150w per GPU now with both Ethereum and Zcash mining. This figure also takes into account the rest of the system as I measure at the wall, so with 6 cards @ 150 watts is 900 watts total, so you will want at least a 1000 watt Gold or better PSU.

Well as I said it is all arbitrary as what constitutes small depends on the person and situation. If I tell someone off the street who knows nothing about mining that I have 50 GPUs running in my house, you're right they are going to think 1:) that's a lot of GPUs and 2:) that I am crazy.

Now if I tell a fellow miner the same thing, I think most of them will consider that still "small" as it would only consist of 8-10 dedicated rigs with an average of 5-7 GPUs each. Putting this into the context of a mining "farm", I think that is still a small amount as many here are likely to fall into the medium to large categories.

I also considered the investment requirements and likely income aspects, as 50 GPUs will net you around $50/day or about $1.00 each (after power) in income (depending on many factors of course) which while a nice side income I still consider this small for most people. I am basing this on a US viewpoint, and realize in some parts of the world this may be a lot, but in the US at least it is probably not enough to live solely on.

As far as the investment aspect, simplifying this as I know there are other hardware expenses, but saying a roughly $200/GPU investment on a 50 GPU farm would be around a $10,000 investment. While this is a good deal of money, as far as investing it is probably considered small as you would not expect to make enough off of a $10k investment to live off of or retire on.

While with a medium farm you are looking at a $10,000-$30,000 investment, med-large $30,000-$60,000, large $60,000-$120,000 and so on. Again these are simply arbitrary buckets to lump investments into, but I think they will work for our purposes.

He's already stated the days of expecting a new release every 2-3 days are over as the miner is getting close to maximum performance. This fact can be further evidenced by the increase in power draw on the last few releases meaning the GPU is starting to be fully utilized.

Except for bug fixes and new features, expect further performance increases to being in the low single digit percent range with the releases becoming more spaced out with weeks/months in between, much like it is with his Ethereum version.

From my limited observations with the new driver so far it doesn't look like there was any change in power draw (up or down) while mining versus the 16.11.5 version. For me I am staying with the older driver for now as it is too much of a hassle to update all my rigs for little to no benefit.

This is the point I am currently at. All through the slow start I sold every ZEC as I mined it as the profits were good. I am just now starting to transition from selling everything to keeping and holding some ZEC.

I am currently at a 75% sell to 25% hold ratio, but over the next month as I think the price will start to stabilize more and I will convert to selling enough to cover power costs and holding on to the rest, probably ending up with something like 75% keep and 25% sell.

I am curious of the makeup of the forum users and what size farms they have and how many GPUs they are current running.

The sizes are arbitrary but I think are still fairly representative of what the expectations would be if someone described themselves to someone else who was also a miner. I am counting the number of GPUs only as rigs can come in many sizes from 1 GPU for a gaming rig that mines on off-hours, to a dedicated 7x GPU 24/7 miner.

For example, I feel a hobbyist is someone new to mining cryptocurrencies and may be using their main rig with 1 or more GPUs to mine with. They may also have additional rigs they cobbled together from old equipment, or possibly took the plunge and setup one or two dedicated rigs.

A Micro Farm is someone who is taking a little bit larger stance and dedicating 3-6 rigs full time with 4-7 GPUs in each rig.

Small is someone who is starting to invest seriously into mining as it is doubtful at this point you would still be building out of old hardware and your purchases are geared toward ROI and making a small amount of side income.

Medium is again someone going further and looking to seriously start supplementing their income and treating mining as a side job. At this level you are probably also starting to dedicate areas of your house or property specifically for mining and either upgrading your power and maybe considering switching to 240V if you haven't already. Cooling also starts to become a concern at this level.

Medium-Large I broke this out as a separate category as I think you would still be a bit too small (at the low end) to rely on mining for a full-time income, but you are definitely looking to seriously supplement your income and invest heavily in mining. At this point even with 7x GPU rigs, you are maintaining 20-40 rigs and have probably addressed and mastered issues such as cooling, power, as well as automated much of your management duties such as monitoring, software updates, restarts, etc.

Large is someone who may very well be relying on mining for a large portion or all of his/her income. At this point you are probably beginning to add rigs in batches and starting to negotiate your purchases with wholesalers. You buy in bulk and take advantage of this to lower your build costs. You may even been looking to host your rigs in a dedicated location outside your home.

Professional is where you have mastered everything below and most likely are not only are earning all your income from mining, but perhaps hiring help as well to administer and maintain your farm. At this point you are also likely to be setup in a dedicated location to host your farm to take advantage of lower power costs and well as other infrastructure issues such as cooling, security, and a redundant internet and possibly power providers.