ANTMINER S2 Discussion and Support Thread

[dogie]

Here are the calcs.
1114W load @ 110V.
90.5% efficient @ 100% load @ 110V.
Internal load of 1008W.

And on 230V
92% efficient
Internal load of 1008W.
External power draw of 1095W.



PSU is specifically designed for 24/7 use @ 50C. Therefore there is a significant thermal safety factor on max load alone - even forgetting the usual one.
Tldr, its k.

[sahkan]

If you can measure the wattage at the wall, i would appreciate .

My meter is holding pretty steady at 1111w-1114w.  It's not the kill-a-watt, it's pretty similar but made by Keedox.   

Looks like mine is pulling 990W at the wall @110V on the 20AMP line, on the 15AMP line the PSU was shutting down every 2-3 minutes. After 30 minutes it's averaging 900GHs, so not a full 1TH. I will try the 1350W Platimax tomorrow to see if that makes a difference.

[seriouscoin]

If you can measure the wattage at the wall, i would appreciate .

My meter is holding pretty steady at 1111w-1114w.  It's not the kill-a-watt, it's pretty similar but made by Keedox.   

Looks like mine is pulling 990W at the wall @110V on the 20AMP line, on the 15AMP line the PSU was shutting down every 2-3 minutes. After 30 minutes it's averaging 900GHs, so not a full 1TH. I will try the 1350W Platimax tomorrow to see if that makes a difference.

Awsome please report back.

[dalenorman2005]

Well gentlemen (and ladies, perhaps??),

hope i'm not jinxing anything, so i'll say this with a hushed voice...

               they're working......

All 3 are up and running. 
Firmware updated to april 4 build. 
DHCP in the 192.168.2.x subnet working. 
max temp is about 60ºC in about 28ºC ambient temperature. 

all 3 needed their boards to be re-seated, ribbon cable to lcd connected (both ends sometimes), loose pcie power cable to power supply, etc.  All 3 power supplies are clearly used - lots of old dust.  but all the boards appear brand new. 

my "discarded" rates are sky-high, as is being discussed by others more knowledgable than I. 

Thanks to everyone who took the time to share their experience and insights. 

[klondike_bar]

If you can measure the wattage at the wall, i would appreciate .

My meter is holding pretty steady at 1111w-1114w.  It's not the kill-a-watt, it's pretty similar but made by Keedox.   

Looks like mine is pulling 990W at the wall @110V on the 20AMP line, on the 15AMP line the PSU was shutting down every 2-3 minutes. After 30 minutes it's averaging 900GHs, so not a full 1TH. I will try the 1350W Platimax tomorrow to see if that makes a difference.

Awsome please report back.

if you have a voltmeter check the levels - you could have a bit of voltage drop in the thinner 15A wiring.  105V vs 110V could be enough to drop the PSU efficiency a bit

[phobus]

the new firmware fixed the subnet issue for me   

[crazyates]

the new firmware fixed the subnet issue for me   

The 192.168.1.1 router issue?

[phobus]

the new firmware fixed the subnet issue for me   

The 192.168.1.1 router issue?

Yes. Before the update I had to edit cgminer.sh after every reboot, now I don't have to.

[cxboyminer]

So is everyone's S2 working perfectly?

[ZiG]

Well if you underclock the left too you can get your density as the right for NO additional cost .....  just saying.

I would rather have numbers work because convenience aint going to pay you back in the end.  To each his own.....but I like your style if you can afford it

No he wouldn't. He might get the same W/GH but that would make the already worse GH/U even worse.

Footprint is irrelevant, only power usage. I pay $50/mo per rack + Power. So if I underclock the S1's and managed to get the same W/GH then they'd only cost me < $100/mo more than the S2's

Where is that...that you pay $50 arack...NOT Florida, I guess...?

[innoman]

Is anyone seeing an issue where their S2 mines for about 10-15 or so minutes then stops and starts beeping?  This happens every time I power cycle and I have to stop/start cgminer or click "Save and Apply" under miner settings to get it going again.  (I only discovered the second method this afternoon).  After doing this once or a few times (varies), everything seems to work fine (for several hours, at least--still trying to get a good understanding of how often,etc...).

I've checked temps, re-seated boards, updated, checked all settings, power cycled, removed a board and let it run, switched the power cable (I'd initially put an older cable on it and it was getting a little warm in spots so I switched it out pretty quick) and checked all the other visible/accessible connections; all to no avail.  I've been working with Juan at 112bit who is also working with the Bitmain techs.  Just wondering if this is a widespread issue and if anyone has made any headway.

[1l1l11ll1l]

Is anyone seeing an issue where their S2 mines for about 10-15 or so minutes then stops and starts beeping?  This happens every time I power cycle and I have to stop/start cgminer or click "Save and Apply" under miner settings to get it going again.  (I only discovered the second method this afternoon).  After doing this once or a few times (varies), everything seems to work fine (for several hours, at least--still trying to get a good understanding of how often,etc...).

I've checked temps, re-seated boards, updated, checked all settings, power cycled, removed a board and let it run, switched the power cable (I'd initially put an older cable on it and it was getting a little warm in spots so I switched it out pretty quick) and checked all the other visible/accessible connections; all to no avail.  I've been working with Juan at 112bit who is also working with the Bitmain techs.  Just wondering if this is a widespread issue and if anyone has made any headway.

Yeah, I get that too, chasing ghosts right now. Try updating to the latest firmware (mentioned earlier in the thread) it's a firmware issue that'll be sorted out soon enough :-)

[innoman]

Yeah, I get that too, chasing ghosts right now. Try updating to the latest firmware (mentioned earlier in the thread) it's a firmware issue that'll be sorted out soon enough :-)

Hopefully!  It does seem that the save/apply or stop/start fixes the issue reasonably well as long as you don't power cycle and you wait around long enough for it to fail once or twice. 

[freebit13]

Yeah, I get that too, chasing ghosts right now. Try updating to the latest firmware (mentioned earlier in the thread) it's a firmware issue that'll be sorted out soon enough :-)

Hopefully!  It does seem that the save/apply or stop/start fixes the issue reasonably well as long as you don't power cycle and you wait around long enough for it to fail once or twice. 

Sounds like CGMiner is not initializing properly... clicking Save&Apply will restart cgminer.

[claudesdad]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

[nikolaz]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

There is not a single machine that would get your ROI unless you are playing the long term game. Price of BTC will go up to reflect that difficulty changes and cost of the miners. At the moment, BTC is going through some changes and government announcements having negative impact on the price.

Think this as an investment rather than regular monthly income.

Price of BTC could be $2,000 or $20 in the next 12 months.

Even you design your own asics and manufacturer your miners you will not make your money back unless you pre-mine and sell those miners few weeks / months later just like every other manufacturer.

[abracadabra]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

There is not a single machine that would get your ROI unless you are playing the long term game. Price of BTC will go up to reflect that difficulty changes and cost of the miners. At the moment, BTC is going through some changes and government announcements having negative impact on the price.

Think this as an investment rather than regular monthly income.

Price of BTC could be $2,000 or $20 in the next 12 months.

Even you design your own asics and manufacturer your miners you will not make your money back unless you pre-mine and sell those miners few weeks / months later just like every other manufacturer.

That reasoning is a fallacy.  A piece of hardware will generate a relatively finite amount of BTC in it's lifetime.  As difficulty increases, the amount of BTC generated by a piece of hardware asymptotically approaches zero.  Given the amount of BTC the hardware will generate and given the amount of fiat required to purchase said hardware, it usually makes more sense to buy and hold btc now if you assume that price of BTC will go up.

[nottm28]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

There is not a single machine that would get your ROI unless you are playing the long term game. Price of BTC will go up to reflect that difficulty changes and cost of the miners. At the moment, BTC is going through some changes and government announcements having negative impact on the price.

Think this as an investment rather than regular monthly income.

Price of BTC could be $2,000 or $20 in the next 12 months.

Even you design your own asics and manufacturer your miners you will not make your money back unless you pre-mine and sell those miners few weeks / months later just like every other manufacturer.

That reasoning is a fallacy.  A piece of hardware will generate a relatively finite amount of BTC in it's lifetime.  As difficulty increases, the amount of BTC generated by a piece of hardware asymptotically approaches zero.  Given the amount of BTC the hardware will generate and given the amount of fiat required to purchase said hardware, it usually makes more sense to buy and hold btc now if you assume that price of BTC will go up.

Yeah but where's the fun in that? You can't 'unbox' a bitcoin like you can a miner...

[klondike_bar]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

There is not a single machine that would get your ROI unless you are playing the long term game. Price of BTC will go up to reflect that difficulty changes and cost of the miners. At the moment, BTC is going through some changes and government announcements having negative impact on the price.

Think this as an investment rather than regular monthly income.

Price of BTC could be $2,000 or $20 in the next 12 months.

Even you design your own asics and manufacturer your miners you will not make your money back unless you pre-mine and sell those miners few weeks / months later just like every other manufacturer.

That reasoning is a fallacy.  A piece of hardware will generate a relatively finite amount of BTC in it's lifetime.  As difficulty increases, the amount of BTC generated by a piece of hardware asymptotically approaches zero.  Given the amount of BTC the hardware will generate and given the amount of fiat required to purchase said hardware, it usually makes more sense to buy and hold btc now if you assume that price of BTC will go up.

most of my equipment profitability calculations assume that the difficulty jumps will slowly decrease in size from the current 15-20%/jump to 10%/jump in the late fall, to 8%/jump around the end of the year.  At current prices that puts break-even sometime around august, and stops being profitable around december. however, if prices go up then it may remain profitable (if only marginally) to keep running through to february/march.

i know what you mean though, BTC returns should be calculated from a BTC price invested.  calculating this shows significant margins, but those margins are eaten up at the current BTC/electricity price ratio - espescially for those paying more than $0.20/kwh in warmer climates

[claudesdad]

what a complete trainwreck. geezus. considering s1's will be worthless in 3 months tops. it's time for an alternative. how about a 5Th for 6.5 btc. mmmmm hmmmmm

Yeah - I don't get where the ROI is coming from on some of these units. Given the difficulty level - and constant increases,  the market price of BTC, the cost of the mining rigs - and the upcoming boning from the tax man,  I just don't see it.

Must be that new math I've been hearing about.

There is not a single machine that would get your ROI unless you are playing the long term game. Price of BTC will go up to reflect that difficulty changes and cost of the miners. At the moment, BTC is going through some changes and government announcements having negative impact on the price.

Think this as an investment rather than regular monthly income.

Price of BTC could be $2,000 or $20 in the next 12 months.

Even you design your own asics and manufacturer your miners you will not make your money back unless you pre-mine and sell those miners few weeks / months later just like every other manufacturer.

That reasoning is a fallacy.  A piece of hardware will generate a relatively finite amount of BTC in it's lifetime.  As difficulty increases, the amount of BTC generated by a piece of hardware asymptotically approaches zero.  Given the amount of BTC the hardware will generate and given the amount of fiat required to purchase said hardware, it usually makes more sense to buy and hold btc now if you assume that price of BTC will go up.

Exactly.

Once you add in all of those things you mentioned - AND , if you're in the US you're now going to have to factor in capital gains taxes - I don't see the math in it.

I don't see BTC mining as a "long term" game - if the hardware doesn't pay itself back in a relatively short period of time - what you're looking at is potentially an ever extending payout period.

Quite honestly I think one of the only ways the mining starts to make any sort of "sense" ( if you want to call it that) - is if you're some sort of IT guy who has access to a budget and a lab infrastructure - and can buy the machines on the company's dime, put them in the company's lab (where they pay for electricity and cooling) - and then direct the BTC output to your own wallet.

That way you basically have no upfront costs and you get all the profit - so it's spectacularly profitable FOR YOU. And if you're in one of those companies that brings in equipment by the pallet load and ships out old equipment by the pallet load - they're likely not even going to notice the miners showing up on the budget.

If you're stuck with all of the costs - I don't see the mining working. 

As for this being off topic to this thread - I think it relates directly - because even though Bitmain is sending out one of the better products at a decent price (at least compared to others) - for the mining to "work" IMHO - their prices need to be less than half of where they are now.