ANTMINER S3+ Discussion and Support Thread

[edonkey]

if you calibrated your pricing so that miners made a little money, you would sell a lot more units.

You guys need a new hobby.  S3's are flying off the shelf.  There is exactly ZERO incentive for Bitmaintech to do anything to 'bump' sales.  To even make such a statement is laughable.

They did about $22,000,000 on day one.

Wow. I'm in the wrong end of this business ;-)

Assuming a 4 or 5 times markup, Bitmaintech is the only one who is a real return out of this deal.

[allcoinminer]

I've got a Corsair RM850 which is running two over clocked S1's.  I was thinking of getting two S3's, and was wondering what PSU to get, I would like to get a PSU that has 8 PCI-e connectors.  What would you guys suggest for such a PSU?

If you like your Corsair and want to stick with the brand, the RM1000 and the AX1500i, AX1200i all satisfy your requirements.

Yes, I do like Corsair.  Obviously with 4 PCI-e per S3 we are overclocking, so which of those three would be suitable for overclocking the S3's?

If we consider the maximum Overclocking, One S3 may consume upto 554Watts.
Then two S3-s it will add up to 1108Watts, Since it will be running 24hrs yu need to add some margin above the maximum.
So, AX1200(100.4A*+12=1204.8W) will be enough no need to go to AX1200i, its a little more expensive.
From the Corsair AX1200 Modular PSU you will get 8 PCI-E ports and a 90% efficiency.

[pumaro]

if you calibrated your pricing so that miners made a little money, you would sell a lot more units.

You guys need a new hobby.  S3's are flying off the shelf.  There is exactly ZERO incentive for Bitmaintech to do anything to 'bump' sales.  To even make such a statement is laughable.

They did about $22,000,000 on day one.

Wow. I'm in the wrong end of this business ;-)

Assuming a 4 or 5 times markup, Bitmaintech is the only one who is a real return out of this deal.

Granted they actually might not have incentive to lower their price when people don't realize they will not get 0.74 BTC out of these things. Difficulty has been increasing ~10-20% every 12 days.
If difficulty stopped increasing now it would take over 50 days to get your BTC back but when it rises 20% every 12 days you are screwed to the nth degree...

[-droid-]

you should get 4 850 watt psu's your psu's should only be running about 60% to stay in the efficiency range.

thats if you get poor quality psu's

[edonkey]

If these units were only $50 cheaper, then the cost of 2 units plus a decent power supply (est. $150) and free power would mean a possible fiat ROI in a couple hundred days and maybe a small profit in 12 months. This is with the coinish calculator, 1.2% per day difficulty rise, 1% pool fee, and 30 day delivery (because these things never ship on time).

Too pessimistic. With free power these will easily break even and make a nice profit.

I don't see how. If anything a 16.8% difficulty adjustment every 2 weeks (the default on the coinish calculator currently) is reasonable, or somewhat on the low side. And given that any purchase made now would be in batch 2 (which is supposed to ship on 7/20) it could easily slip to the end of the month before receiving the units (which is on the other side of the next difficulty increase).

What I come up with (after buying a decent power supply for $150; yeah I can go cheaper, but PSU's are not a good place to scrimp) is no break even point. The loss after 12 months is $74.

With over clocking to 1 THs for two units the loss after a year is $25. Tantalizingly close, but not really there...

Of course this assumes that Bitcoin prices stay the same, which is unlikely. Hopefully a year from now we all hope Bitcoin is up from where it is now.

At this stage, I'm probably better off sitting on my Bitcoin and watching my existing Scrypt ASIC rigs ride into the sunset. I have free power, so these will never get unprofitable. They will just get really boring ;-)

[Phosphorous]

With an S1 overclocked to 400Mhz, each PCI-e connector consumed just under 16A. With 16Awg wires (x3), this is about the most you want to run through a single PCI-e connector. This is not something to take chances with. If you're going to overclock the S3, use 4 PCI-e connectors, don't skimp and try 2. Regardless of specifications, you never know what shortcuts your PS or cable manufacturers may have taken.

[pumaro]

If these units were only $50 cheaper, then the cost of 2 units plus a decent power supply (est. $150) and free power would mean a possible fiat ROI in a couple hundred days and maybe a small profit in 12 months. This is with the coinish calculator, 1.2% per day difficulty rise, 1% pool fee, and 30 day delivery (because these things never ship on time).

Too pessimistic. With free power these will easily break even and make a nice profit.

I don't see how. If anything a 16.8% difficulty adjustment every 2 weeks (the default on the coinish calculator currently) is reasonable, or somewhat on the low side. And given that any purchase made now would be in batch 2 (which is supposed to ship on 7/20) it could easily slip to the end of the month before receiving the units (which is on the other side of the next difficulty increase).

What I come up with (after buying a decent power supply for $150; yeah I can go cheaper, but PSU's are not a good place to scrimp) is no break even point. The loss after 12 months is $74.

With over clocking to 1 THs for two units the loss after a year is $25. Tantalizingly close, but not really there...

Of course this assumes that Bitcoin prices stay the same, which is unlikely. Hopefully a year from now we all hope Bitcoin is up from where it is now.

At this stage, I'm probably better off sitting on my Bitcoin and watching my existing Scrypt ASIC rigs ride into the sunset. I have free power, so these will never get unprofitable. They will just get really boring ;-)

The difficulty jumps every 12 days not two weeks, two weeks or 14 days would be a stand still for difficulty. Basically every time the 2016th block is solved difficulty is re-assessed. When we complete 2016 blocks before the 14 days we get smacked hard with a new higher difficulty.

Don't forget the fun of RMA'ing these back to china or wherever for repair. All of my miners had or have some die failure out of the box, this stuff isn't easy folks and I'm afraid to say if you haven't made your profit in BTC by now you won't do it by mining. Your new window of opportunity is buying BTC directly, but you don't have to take my word for it!!

[hurricandave]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

[iglasses]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Well now we have really jumped the shark.  Unless we calculate LUCK the math is unreliable?
ugh.

[dropt]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Yes, I too have been raped hard by luck over the last three months, but this is independent of hashrate distribution.  Slush has had a 1 month luck of 107% while BTCGuild has been at 91.9% even though BTCGuild is 2x-3x larger than Slush by HR.

[hurricandave]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Well now we have really jumped the shark.  Unless we calculate LUCK the math is unreliable?
ugh.

The guys at bitminter just went over 24hrs without a block, where is that in your calculations. The pool returns are diminishing and always have been, but we are at a tipping point where the drops are nearing an exponential loss for the public if the manufacturers don't back off the network and get the gear distributed to public miners.

[pumaro]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Yes, I too have been raped hard by luck over the last three months, but this is independent of hashrate distribution.  Slush has had a 1 month luck of 107% while BTCGuild has been at 91.9% even though BTCGuild is 2x-3x larger than Slush by HR.

Its all Luck some good some bad but it all averages out, although SLUSH and ELIGIUS seem to payout better than BTCGuild, maybe they are just inherently unlucky.

[theonegilly]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Yes, I too have been raped hard by luck over the last three months, but this is independent of hashrate distribution.  Slush has had a 1 month luck of 107% while BTCGuild has been at 91.9% even though BTCGuild is 2x-3x larger than Slush by HR.

Its all Luck some good some bad but it all averages out, although SLUSH and ELIGIUS seem to payout better than BTCGuild, maybe they are just inherently unlucky.

Maybe BTC Guild is actually running at 101.1% and 1 of every 10 blocks goes straight to the Owner.

Who know.

[Threader]

I see post after post where mathematicians calculate profit or loss based on speed and difficulty, but the biggest factor that will get in the way is LUCK, the POOL LUCK has been destroyed by manufacturers and the private farms. Bitmain is a contributor to this as well. The difficulty is rising in leaps and bounds based on network speed but most of the public pools combined hash rate is getting left in the dust! Oh except when Bitmain or one of the other Manufacturers decides to raid a pool by jumping in with a Ph/s or two for a couple of hours while "testing". Crunch the numbers all you want, till you start adding in the Variance and LUCK of your favorite pool, those calculators are meaningless.

Yes, I too have been raped hard by luck over the last three months, but this is independent of hashrate distribution.  Slush has had a 1 month luck of 107% while BTCGuild has been at 91.9% even though BTCGuild is 2x-3x larger than Slush by HR.

Its all Luck some good some bad but it all averages out, although SLUSH and ELIGIUS seem to payout better than BTCGuild, maybe they are just inherently unlucky.

Maybe BTC Guild is actually running at 101.1% and 1 of every 10 blocks goes straight to the Owner.

Who know.

Sure I have been guilty of thinking this a few times during shitty luck days

[iglasses]

Maybe you guys are joking and my sarcasm radar is out of tune but if you are serious I think that's a pretty wild accusation.  If Guild, or any pool for that matter, was posting results that didn't jive with what was in the blockchain don't you think someone would figure it out?  BTC has a lot of really smart, dedicated people and if that isn't enough it also has people with tens of thousands of dollars invested who pay attention to stuff like that.
JMHO.

ig

[visdude]

you should get 4 850 watt psu's your psu's should only be running about 60% to stay in the efficiency range.

The Corsair CX750M (with 4 PCIe connectors) is I think the best deal around (especially with promo code discounts plus rebates) as a dedicated single PSU solution for an overclocked S3.  Alternatively, two CX500/CX500M (this would be my power configuration if I decide to overclock since I already got a few of them) would also be a good buy which can be had for as low as $19.99 if timed with regular promo code discount offerings up to $20 plus $20 rebate at the Egg.  In fact, I was able to get some of my CX500s just by applying the prepaid rebate cards from past purchases...and so on and so forth .  Such a solution would offer plenty of headroom (a total of 1000W) at a dirt cheap price and modularity (i.e. presumably, if one PSU goes out, only one blade would be out of commission with the other continuing to hash).

Current deals at the Egg:

CX750M - $20 rebate
CX500M - $15 promo code discount plus $20 rebate
CX500 - none

[shaxs]

Did I read correctly that it only needs 2 PCI-e cables connected to run as normal (not overclocked)?

Correct. 2 for regular, 4 for OC'ing.

What I do not understand is this sentence: "Don’t parallel connected different DC input from different PSU into the same Hashing board." So does this mean you need two PSU if you're overclocking?

I understand they mean dont use 2 PSU for same hashing board. Anyways, with 1 PSU u can connect 4 PCI-e so no problem

yeah this is most likely exactly what they mean.    


but with 2 psu's  each one doing one board each even thou it seems okay I will let some one else do it first.

Depending on how the shipping goes I might be the guinea pig here. My understanding is that the reason, the only reason, to use two PSU's is to overclock. If you are not planning to OC one PSU is all that you should need. In my case my PSU army is a mix of   Corsair 750's and Thermaltake 850's so if I want to OC I am going to have to use the dual config.

I've still got a couple of these here:

https://bitcointalk.org/index.php?topic=518013.0

So basically, I could use one for two S3s if not oc'd, correct?

I am in the same boat. I have a bunch of those too. However, since there are no longer screw terminals, I do not know how I will connect them to the S3. How do you plan on doing it?

[CryptoCrane]

How many S3s will there be in the first batch?

Edit: Batch 1 units for sale on Amazon. Great if you only want to buy 1 or don't have enough bitcoin.

Hey, thanks for linking to our batch-1 S3s on Amazon! Our primary market are people who want bitcoin but don't feel comfortable with the exchanges and therefore don't already have enough bitcoin to purchase one directly from Bitmain. 

[phillipsjk]

I am curious: is the cover made out of plastic or metal? It should save on the cost of a case (which is one reason I have been a little reluctant to get an S1 (still haven't payed off my old blade either). Cases block EMI, which helps prevent interference with Over-the-air television.

I was thinking these things just need a 500W power supply: I suspect (but have not tested) that you can plug the 4Pin CPU connectors into the spare ports for an extra safety margin. Edit: bad idea (as in flames shooting out of one of your expensive components if you manage to force it).

[ZiG]

I am curious: is the cover made out of plastic or metal? It should save on the cost of a case (which is one reason I have been a little reluctant to get an S1 (still haven't payed off my old blade either). Cases block EMI, which helps prevent interference with Over-the-air television.

I was thinking these things just need a 500W power supply: I suspect (but have not tested) that you can plug the 4Pin CPU connectors into the spare ports for an extra safety margin.

CPU connector(s) are different pinouts...Don't even try...

ZiG