Post: #1

I’ve been interested in crypto for a few months, and while I have been using dollar cost averaging to build my investment, I wanted to accumulate more crypto than my stingy nature would allow. Thus, I started mining it myself with hardware I already possessed. In addition to the monetary rewards, it appeals to the side of me that likes tinkering with things for the sake of it. Spending hours of trial and error to improve hash rate by 5%, decrease power consumption by 2% or reduce operating temperature by 3 degrees might not make much sense on a time/returns basis, but fuck it, I’d be tinkering with something anyway, might as well do it with a single purpose in mind. I’m by no means an expert, but through a couple of months of trial and error I’ve come from no knowledge at all, to a reasonable level of understanding that I can share to save others some trouble.

There are 3 types of crypto mining devices:

1. CPU mining – the power of your CPU can be utilized, but this is the least powerful, and returns specifically from CPU mining are supplemental at best. In fact, many miners choose to leave their CPU idle, as the small returns are not worth the power costs and wear and tear.
2. GPU mining - the miners this guide revolves around. Graphics cards are far more powerful for mining than CPU’s while still being accessible to regular people with regular hardware. The more powerful the card, the higher the hashing power, but it tends to come at the cost of higher power consumption.
3. ASIC mining – the invention that “broke” bitcoin mining. Short for Application specific integrated circuit, these machines are designed to mine and nothing else, thus are incredibly powerful and efficient at it. They are however, beset by supply, noise and power consumption issues. Some coins have been developed to be ASIC resistant, but as far as mining specifically for bitcoin goes, if you’re not using ASIC, don’t bother.

Components and design

Frame – I started with just a normal ATX case but for any more than 2 GPU’s the limitations on cooling make it less than ideal. For greater than 2 GPU rigs you’ll want an open air frame – popular designs are like this:



They’d be easy to make for cheap, or you could waste $300 buying one. But being the cheap cunt I am, I made one based around an old milk crate. Not pretty, but every bit as functional and cost me nothing.

PSU – anywhere from 700 – 1400 watt PSU depending on load. This is the backbone of your system, so don’t cheap out. Check the power draw of the cards you plan to use, multiply by the number of cards and then add some for the motherboard, memory, HDD etc. ie If a card has a TDP of 135W and you want to use 6 of them – 6x135 = 810W just for the GPUs. You wouldn’t use anything less than a 1000W PSU for this. Alternatively you can power the GPUs from one PSU, and the motherboard from another. The important thing is the efficiency rating – you’ll want gold or better. Cut corners on the efficiency rating, and you’ll end up paying for it in wasted power costs.

Motherboard – Look for a full ATX motherboard with 6 PCI slots. To utilize them all you’ll need to use powered risers. Most gaming motherboards fit the bill, though dedicated mining motherboards are available too. Not sure what benefit they purport to offer, probably nothing.

CPU – most mining rigs just have the cheapest, lowest spec CPU to get the thing to run, as it’ll likely remain idle.

RAM – hardly used, so you can afford to cheap out. You could get by with 4GB with a linux system, but I’d probably go with 8GB since I haven’t got a fucking clue how to use linux.

GPU – the most popular cards are the AMD RX470/480/570/580 cards or Nvidia GTX1070, each with their own pros and cons. They all have similar hash rates, but for example the GTX1070 has lower power consumption but higher initial outlay while the AMDs are cheaper to buy but are more power hungry. For me, the ideal would be the RX570 on balance of its hash power, power consumption and purchase price.

Getting started

With the rising popularity of GPU mining, supply has outstripped demand for GPU’s, thus prices for new hardware are inflated and availability is scarce. For the most popular mining cards retailers typically have a limit on how many can be purchased at any one time, meaning you might need to build your rig piecemeal, or pay a considerable premium. Keep an eye out for specials, and snap them up when you can. I recently ordered an RX580 for $300 AUD, when the regular price was about $450. Problem is, there was a limit of 1 per customer, but I can wait until the next special comes around to grow my rig. You could look around for second hand cards, but this comes with added risk. You never know what the previous owner has done to the card – they could have been running it for their own mining operation, constantly at high temperatures and burnt out half the cores by the time you get your hands on it. Still, some older cards like the AMD R9 series or HD7970 still offer good hash rates at a much lower cost than new cards so you may want to take a gamble. Bear in mind they are less power efficient than newer models and will gradually deteriorate, so second hand cards will incur replacement costs sooner than brand new ones.

Running costs – the components will be drawing power 24/7 so it is a balancing act between hashing power and power consumption. For most miners, the increased purchase price and power consumption of the absolute most powerful cards like GTX1080(Ti) or RX Vega make them a little less desirable than slightly lower-spec components – the cards that are generally considered to sit in the “sweet spot” are the AMD RX470/480/570/580 or Nvidia GTX1070. I started mining with a couple of nvidia 1050ti’s which are one of, if not the most power efficient cards on the market, as I live in a city with relatively high power costs so I need to place a fair bit of weight on the power consumption side of things. I do have solar panels that provide enough to power it through the day, but the costs incurred at night will still be significant. If you’re lucky enough to have cheap or free power, load up with as much hashing power as you can and to hell with the power consumption. Make sure to get a kill-a-watt or similar energy consumption meter, so you can make accurate calculations.

Post: #2

Choice of coin/algorithm

This is where knowledge of the market, in addition to pure technical knowledge becomes important. Certain components are better suited to some algorithms rather than others, for example Nvidia cards are generally better for ethereum, while AMD cards are better suited to the cryptonight algorithm (Monero) so your choice will have to factor in the hardware on hand. The website whattomine.com is very useful for this – plug in what hardware you have and the cost of power and it will spit out a profitability calculation for a range of coins. Bear in mind this calculation is based on prices at the current time, and doesn’t take into account projected future profitability. Unless you’re planning to dump your mined coins as soon as they hit your wallet, immediate profitability should take a back seat to future projected value, especially bearing in mind that mining gives gradual accumulation of currencies rather than the instant fix of trading. It’s no point turning your miners to a coin on the pump, as by the time you’ve accumulated a reasonable amount, it’ll probably already have dumped. I happened to get lucky in that I chose to mine zencash for a decent amount of time before its price boomed, however I could just as likely have dedicated hash power and time towards a coin like sia, the value of which just keeps dropping.
The profitability will also be affected by the fluctuating difficulty level of the coins you choose – the website coinwarz.com tracks the current difficulty of many coins. I keep an eye on this, and if the difficulty of a target coin goes too high or too low, you can switch your miners accordingly.

Placement
Three things that can spell death for your rig are dust, heat and moisture, so you need to have a position that minimizes all three. In addition, keeping the components at a good operating temperature will likely require fans running at all hours, which can be quite noisy. Not a problem for the rig itself, but a problem for you and anyone else that shares the space. I’ve got a reasonable sized house so I can sit my rig inside without disturbing anyone at night, but this could be very difficult in an apartment. Fans can be set to run silently, but at the cost of diminished cooling power, so you mightn’t be able to run your machine as hard if noise is an issue.


Software/pool
Once you’ve got the rig built and a target coin selected, you’ll need to download the appropriate mining software and join a mining pool.
Software will depend on which algorithm and hardware you use, and some optimized miners like claymore and xmr-stak might incorporate a developers fee of 1 or 2% - well worth it in my opinion, as they more than make up for this in hashpower gains and/or power efficiency. These can be downloaded for free on github. The download page also contains the syntax in which they need to be executed to point them towards your chosen pool or wallet. The syntax might take a bit of trial and error to get it right, but once you’ve done it a few times you’ll be fine.

Mining rewards are gained by “finding” a block, with the finder being allocated a reward. When you join a pool, you merge your hash power with the pool, thus increasing the chances of the pool finding a block. While your rig is mining, it submits shares to the pool, and any blocks found by the pool are allocated within the pool based on the amount of shares submitted. Within a decent sized pool, you can be pretty much guaranteed a steady return, whereas if you were to mine solo, unless you have an incredible amount of hash power, there’s a good chance you’d never actually find a single block. However, if you did get lucky and find a block, the reward would be all yours.

Some popular pools are nanopool and suprnova, which encompass several coins. Considerations to take into account are the pool fee and the payout amount and frequency. There’s not a lot of difference in pool fees, which are typically 1 – 2% across the board, but the minimum payout can be a bit of a trap. For example dwarfpool, which is one of the biggest ethereum mining pools, has a minimum payout of 1 Eth. If your hash power is on the lower end, it might take many months to accumulate this much, and if you don’t see it through, whatever time and hash power you’ve already dedicated to it will be wasted. Nothing worse than dedicating a month of mining to one coin/pool and wanting to change targets with nothing in your wallet to show for it, so make sure you check it out before investing too much time to a pool. Unless you’ve got 100+ MH hashing power, it’s probably best to stick with some of the smaller coins as they tend to have lower payout thresholds – zencash, monero and aeon worked well for me with my low-end rig.

Fine tuning

Most manufacturers have programs to tinker with the settings on GPU’s that can make a big difference overall. The most popular program is MSI afterburner, but I use gigabyte xtreme gaming engine. You start your miner, check the hash rates, power draw and temperature, slightly alter the GPU settings such as GPU clock, memory clock, voltage etc, then check the miner again. Bear in mind that different algorithms respond differently to different settings. For example ethereum does best with a lowered GPU clock and elevated memory clock, while equihash is the reverse. It can be a tedious process, but once you find the optimal settings you can save several profiles to easily swap between them.

Profit!
Remember to keep a log of your earnings and costs. There undoubtedly will come a time when it’s no longer profitable to mine at home, and you want to make sure you don’t overshoot that mark. By accurately tracking profitability, when things become borderline you could still get out on top by selling off your hardware to less diligent miners.

Post: #3

Are you making money off this?

Post: #4

GPU yields on ETH are less than $7/day nett.

If you paid $2000 for a setup, thats around 10% yield a month, which takes 10 months of uniterrupted mining to get your capital back.

Or.... u could just pile on the ETH train and make 10% swings every month.

Post: #5

That's the $64 question.

Short answer: yes
Long answer: It's complicated.

From the first time I started up my first miner until today is less than 2 months.

At the end of the first month I took an audit of all my mined assets and at the time of auditing, they totalled about $47. This first month included all my fucking around, wasted time, mistakes and a period of 4 or 5 days when no mining was done at all.

The second month will be over in a week, and has seen much more consistentcy in mining than the first, with this months earnings from 3.5 weeks sitting at about $90 as of 2 days ago. But how much of this increase is due to increased mining productivity and how much is due to the massive price pump of bitcoin in that time?

My calculated power cost for the time is about $36. At this point - $47 + 90 - 36 = $101 /2 = ~$50 per month profit.

Bear in mind this has been done with hardware I already owned, so there was no upfront cost. The only thing Ive spent on this is $300 for a 'graphics card that will double my hashpower, and that hasn't even arrived yet, so its productivity hasn't been counted. I'll need data from another month or two before I can work out how long that will take to pay itself off. Íf one had a decent gaming PC sitting idle it is profitable to turn it to mining, but if one were to build a decent sized rig from scratch, the equation would be very different, and you'd have to count on it taking a year to pay off. Better to spend $2000 on a mining rig or $2000 a year on crypto? Who knows? But I do know that buying the crypto straight up won't keep me amused like mining does.

The returns on this scale are not going to make me a millionaire, but they do supplement my DCA. And if you lived in a country with cheap power and a very low monthly wage, it could be very attractive.

Post: #6

Its Better off to spend $2000 a year in crypto.

Definitely.

Eth mining is too low a yield to consider even if you have free electricity. Until it moons again, you probably wont make much.

Post: #7

Nice intro to mining, but I will note that you will need a lot more cards with difficulty levels the way they are now to make any decent money mining. I thought about making a mining thread myself, but ultimately decided against it because I wouldn't recommend it to anyone at this point if they were trying to make money.

I've been mining ETH now for about 4 months now. Started off with a 6x1060 rig just to test things out first, and then added another rig with 6x1070 not long after. I like NVIDIA over AMD because they are much easier to set up and get running. No messing with flashing your BIOS and what not. Just download MSI Afterburner and overclock and you're good to go.

With that set up I was getting about 313 mhz mining only ETH. I figured with my cheap power costs out here in China I had a competitive advantage and according to some mining calculators that factored in increasing difficulty, I was looking at 6 months or so to start seeing profits.

In total I spent about $6k on the 2 rigs. Later on I added another 1070 to my main desktop to mine with whenever I'm not using it. I also tried dual mining decred in addition to ETH for a month, but it basically turned my house into a sauna and the increased power from dual mining and running the AC wasn't worth it. My power bill went from about $15 a month to $150 in that one month for only 3 decred or so. Not worth it.

So where am I at now after 4 months of mining? So far I've only mined roughly 8 ETH and the 3 DCR. So I'm behind my initial 6 months timeline and it's looking more like 10 months before I see profit. Now, I don't mind this so much because I was always looking at it as a way to just generate some guaranteed side income. I also enjoy messing with computers and just building these and getting them working was fun for me. If that doesn't sound like you, don't mine.

And if you have enough cash to blow to build some decent mining rigs, you are almost certainly better off just outright buying some crypto instead. For proof let's look at what $6k could have gotten me instead back in July:

-BTC at 2.3k--->now worth $19.2k
-ETH at 200---->now $8900
-LTC at 45----->now $8100
-NEO at 6------>now $26k
-OMG at .50---->now worth 76k and double that if you sold at the peak for the ultimate moonshot

I think the average person who would be interested in mining is already probably buying and holding different cryptos. Unless you want to build a modest supplemental income stream or just like tinkering with computers, I wouldn't recommend getting into mining, especially for making money. That said, I think there's a good shot ETH moons again sometime in the future, and eventually when it does go to PoS you could always mine other coins.

Or also sell your cards as they will retain some resale value. Just remember to put 'lightly used' in the description.

Post: #8

I thought this was going to be a post about going back to the Coal Mines after going broke via risky investments (Bancor, ETN, REQKT, etc)

Post: #9

Been mining Electroneum (ETN) for the last week or so, and managed to mine a few hundred using just the one card, and its even worth mining with the cpu.

Price 3x against bitcoin and 5x against fiat in that time. Its a very profitable coin to mine while its new and the difficulty level is low.

Post: #10

You are a little fish in a big unforgiving ocean when you go about mining. There are some infrastructures that allow you simply to invest and forgo the noise, power consumption, and headache while getting a return. However quite a few are scams or are terrible about paying out.

I feel the ship has long sailed for profitable solo mainstream mining, especially with GPU prices where they are, unless you’ve already got a large farm with potentially solar installed.

Agreed with above poster that mining new currencies is likely the way to go for profitability, as there’s less adoption. But this is a moving target over time of course.

Interesting nonetheless.

Post: #11

It doesnt have to be a headfuck.

Ive now got my rig set up so it just runs 24 hours a day and doesnt require any attention. I still use it as a normal desktop.

Ive got the cards limited to 72 degrees - they could be run a lot harder but at this temp, the fans hardly make any sound at all.

The difference to my power bill is trivial. It turns out my earlier estimate of $36 /2 months was nearly double the real cost.

Im getting over $100 profit a month, without doing a thing. Why wouldnt you mine?

Post: #12

So you’ve already recouped the initial investment? It’s not profit until the investment is recouped, and generally you’re looking at quite a while for this. Couple with price fluctuations (things don’t always go up) and I don’t see why you would mine.

Post: #13

Well, at the minimum this is a great datasheet breaking down everything you would have to do if you did want to mine, and explains how low margin it is.

Post: #14

Agreed, thanks for the data sheet.

Post: #15

Do you mind sharing what number and type of GPU you use? And maybe the hash rate per card for the crypto you mine?

Post: #16

Ive got 2x 1050ti which each get 14mh/s on ethereum, 1x r7 370 which gets 450h/s on cryptonight and 1x rx 580 300 sol on equihash. All but the 580 were ones i already had rather than ones selected for their mining suitability. You can check the profitability of most modern cards on whattomine.com.

Room for 2 more cards on the motherboard that I may or may not fill.

Post: #17

I've been thinking of building my first ever rig for a long time. I thought the mining boom would be a perfect time to learn. I'm still on the fence though

Post: #18

It has got me researching this further, and looking at Antminers. Does anyone have experience with ASICs or is everyone just using resources from idle pcs/gpu rigs?

Post: #19

Have you tried water cooling with thermal paste for gpus? I would think that would lower heat consumption production a lot, although I'm not sure if it would draw much more energy with the pumps involved. A really interesting computer setup I have seen was using a fishtank filled with mineral oil. I don't think it use anymore energy and probably could be overclocked easier for this type of work.

Post: #20

great info thanks

Post: #21

Antminer lifespans ultimately works out much shorter than gpu setups.

They do come with sorta warranty but it takes a month to send back to china and another to come back. After that your warranty is mostly void.

Do check your yields before purchasing though, some wont return your investment before the warranty runs out.

Post: #22

I've been meaning to update this thread for a while.

I'm still mining, and it is still profitable to do so, but it's pretty hard to determine exactly how profitable, for a few reasons.

First, when I started mining , ETH was at something like $3 or $400 (AUD) whereas now it's about $1200, so while I was making about $50 a month early on, it's closer to $200 a month now. It depends how you calculate your profit - on immediate or future value? If you believe in a particular crypto, or the market in general, mining at a loss now might eventually become profitable if you HODL.

So the price fluctuations make it difficult, but another issue is that I don't really care to calculate in depth anymore. A few hundred worth a month is nice, and gives you a steady stream to consolidate or take new positions, but it's barely significant in comparison to trading gains. When your portfolio can increase by $1000 in a day, it's hard to dedicate much time towards acquiring $200 a month.

Thankfully, I don't need to dedicate any time to it. it just runs, and periodically gives me a bit more funds to sink into crypto, which has now taken the place of DCA. I haven't put any money into crypto for months, and have even withdrawn my entire principal (plus a bit more) while my balance continues ticking upwards.

It won't last forever though - the amount of ETH I could mine was steadily decreasing, which is why I switched to ETC a few weeks ago. It was just lucky that ETC happened to pump soon after, and if that trend continues, mining it will be profitable for a long time to come. And when it's not, I can switch again to lower difficulty coins, or turn the miner off and use the powerful PC I've built for what it's meant for - gaming and VR.

Post: #23

@kingblast

Surprised you are mining in Australia. Electricity is mad expensive here compared to everywhere else I've ever lived. Something like 34kw/h. I guess it's still profitable regardless?

Post: #24

26c kw/h where I am, so yes, power is a significant consideration, but I've based the design around keeping it pretty efficient.

I added a used 280x and found that it drew almost as much power as the other 3 cards combined. Decent hashpower, but it wasn't worth the power consumption, so I sold it again (at a profit) on gumtree.

AMD cards aren't really suitable for mining where power costs are high. I've got a current-gen one which isn't too bad, but anything earlier just drains the juice. The other 3 cards I've got are Nvidia ones which don't generally have as much raw power on most algorithms, but are miserly on power compared to their AMD equivalent.
I've also got solar panels to feed it, at least during daylight hours.

If I didn't have the panels and/or if my rig was all AMD, I could very well be losing money.

Post: #25

I've actually thought about buying cards and bringing them to Eastern Europe where electricity is 4-10kw/hr but then I remembered the import taxes on the cards