ANTMINER S3+ Discussion and Support Thread

[soy]

It's ONE per household per rebate offer/campaign and not for eternity and they occur pretty frequently; and I'm certain there are CREATIVE ways around such limitation.  I've purchased CX500s at the Egg at a final cost as low as $9.99 ($39.99 list price minus $20 rebate minus $10 instant promo code discount).  I have purchased several CX500s this way since the BFL Single SC days through the Cube era.  I, in turn, had used the rebate VISA cards to purchase Egg gift cards and apply them to purchase more CX500s with ZERO out-of-pocket cost.  That's the most bang for the buck for me.

If you use it for any length of time, the bronze will cost you more in the long run.  Won't take long to recoup the $40-$50 you'd save having a gold 1300w PSU.  Plus you need more outlets.

M

That's a fair assessment.  It would be nice if you could illustrate the details of the figures though.

Off the top of my head, the difference between PSUs with 90% and 85% efficiency would be around 20WAC at the wall with a stock S3+ which equates to about 0.48kWh per day.  At $0.12/kWh power rate, that's around $0.06/day difference.  Therefore, $50 / $0.06 = 833 days = 28 months = 2.28 years to recoup a $50 difference in price.  Is this what you had in mind?

Re needing more outlets, that's what power strips are for; and I've got enough of them already and I'm almost certain that most of us do.  It could also be argued that using a single/dedicated PSU for each S3 eliminates a single point of failure of three S3s powered by a single PSU.

The CX500 is "at least 82%" according to NewEgg.
The Lightning 1300 is "at least 87%", and "up to 92%".

That 5% span is what you estimated above. 

In typical manufacturer garble, the spec sheet lists the CX500 as "> 80%" and the Lightning as "> 87%".  How that translates into "at least 82%" for the CX500 is anyone's guess.

Also remember you do get what you pay for.  The golds are higher quality, will last longer, and will feed better power to your units.  2.28 years may seem like a long time, but I know I've had my PSUs around for mining longer than that.  And if the day comes when I'm no longer mining, I'll likely find a new home easier for my gold PSU than your bronze PSU.

M

Heck, maybe someone with two of those power supplies listed above can fire up a miner, get its kill-a-watt stats then change power supplies and see if the difference agrees with the chart.

[soy]

Hi all, i have question.  i had an s3 with bad asics on chain 2.  Bitmain RMA'd my faulty pcb chain 2 and sent me an RMA board.  when i received it i noticed right away that it had a burnt spot on the pcb (not sure what the component that has been melted is ?)  it appears to have had some sort of heat or flare up issue prior to them sending it to me.  i continued to put thermal paste and mount it to my s3.  right away i noticed that my fans were spinning at 2800-3000 (within the first 30 seconds of it hashing).  normally my s3's fan spun anywhere from 1800-2150 (tops).  my s3 is not overclocked, it's running at 218.75.  

anyone have any clue what's up with it ?  i cannot have this s3 spinning its fans at 3000.  I live in a tiny apartment.  I contacted Bitmain right away and sent them the pictures that i am posting and listed the issue.  

i forgot to mention that the temperature of chain 2 is 9 degrees higher than my chain 1 (again within the first minute of hashing)

additional details...this device had a faulty control card that kept it from hashing for 1 month, and had a bad chain 2 that had it missing 140gh/s for 2 weeks.  All in all i have missed damn near of 1.5 months of mining with device.  the device has been a lemon from the start and continues to get worse.  as much as i feel like Bitmain has been helpful, I am upset because they knowingly sent me a faulty PCB as my RMA.  

That gray discoloration next to the center cap?  If I were to guess it might be that they had to remove the cap and used a hot air desoldering tool.  Those don't just heat up the solder joint but the area near it as well.  I'd be more interested in keeping the ASIC pins free from debris.

[Biodom]

Hi all, i have question.  i had an s3 with bad asics on chain 2.  Bitmain RMA'd my faulty pcb chain 2 and sent me an RMA board.  when i received it i noticed right away that it had a burnt spot on the pcb (not sure what the component that has been melted is ?)  it appears to have had some sort of heat or flare up issue prior to them sending it to me.  i continued to put thermal paste and mount it to my s3.  right away i noticed that my fans were spinning at 2800-3000 (within the first 30 seconds of it hashing).  normally my s3's fan spun anywhere from 1800-2150 (tops).  my s3 is not overclocked, it's running at 218.75.  

anyone have any clue what's up with it ?  i cannot have this s3 spinning its fans at 3000.  I live in a tiny apartment.  I contacted Bitmain right away and sent them the pictures that i am posting and listed the issue.  

i forgot to mention that the temperature of chain 2 is 9 degrees higher than my chain 1 (again within the first minute of hashing)

additional details...this device had a faulty control card that kept it from hashing for 1 month, and had a bad chain 2 that had it missing 140gh/s for 2 weeks.  All in all i have missed damn near of 1.5 months of mining with device.  the device has been a lemon from the start and continues to get worse.  as much as i feel like Bitmain has been helpful, I am upset because they knowingly sent me a faulty PCB as my RMA.  

<snip>

That gray discoloration next to the center cap?  If I were to guess it might be that they had to remove the cap and used a hot air desoldering tool.  Those don't just heat up the solder joint but the area near it as well.  I'd be more interested in keeping the ASIC pins free from debris.

after 2-3 mo of use, I had white discoloration on S1 boards-it never affected anything.

[eoakland]

Hi soy, thanks.  what has me most concerned is the noise.  the fans spin at 3000 !  it's ridiculously loud, and i am running it stock at 218.75.  Bitmain tried to tell me this is normal.  Unreal.  my s3 used to have the fan run at 1800-2100 at stock.  i have a jet in my living room now.  unreal.   

[cloh76]

The Ax1200i really is a top notch PSU. I didn't actually buy them for the S3 units, I had 4 left over from my Litecoin GPU rigs that I've long since shut down.

But running 3 S3 units has been no problem. The PSU fan doesn't even seem to be at full speed. They've been running for weeks.

Room temperature may also affect the draw. Our Ants are in a cool server room, which helps keep them and the PSU cool. I'm also on filtered AC Power. So voltage off the wall might also factor into it. If off the wall you're Voltage is 116 vs 110, that's a 5% increase in power input.

Our large order of S3+ Units come in Tuesday (BitMain Shipping Confirmed Yesterday).  We've bought a dozen EVGA 1300 G2 Units. I am going to see how they compare.

Hoping I may be able to squeeze in 4 Units per 1300 even if I have to underclock. I'll post results and use my kill-a-watt to get you all numbers.

Strato

The Ax1200i may be top notch, but I don't understand the desire to throw money at it unless it's for fun.  

For less than half of that I can get a Rosewill Lightning Gold 1300W PSU.  I know it can run three S3s because I've used it to drive three OC'd S1s without a problem.

I think the only think the Ax1200i has going for it is it is pretty.  Seriously.  Well, alright, you can connect it up to a PC to see real time values.

Otherwise, they look identical, except the Lightning has 108A on the 12V rail, and Ax1200i has 100.4A.  Everything else important looks the same to me .. fan size, warranty, efficiency (what is Platinum supposed to get you?).  Except price of course.

Again, I ask why?

(NewEgg has Lightning at $149.99 AR, only $10 more $20 less than the equivalent eVGA.  The Ax1200i is at $329.99.)

M

I couldn't agree with you more.

Alternatively, I think the Corsair CX500 has the best bang for the buck especially when timed with the regularly-occurring $20 rebate offer in conjunction with additional promo code discounts of up to $10 at the Egg, bringing down the cost to as low as $10-$20 each depending on the list price.

Corsair will only pay ONE rebate per household per item. The cheapest I ever bought corsair cx500's was $45 after coupons etc. They are the best PSU for the price and they never burn wires. DO NOT buy the modular Corsair's though. The plugs they use are cheap and THEY WILL MELT  even with just one S1 running at 200gh.

I ran 2 sets of 3 S1s overclocked on one Corsair PSU for about a year and never had a problem with any parts melting or overheating.  Maybe Ill try the lightning out as well with my next batch of S3s

I find that hard to believe since you couldn't even buy them until the end of November and it is only September.

Hence the word "about."  My point is, I never experienced an issue with these PSUs even when I had them installed in my PCs.  They are versatile in my experience

[philipma1957]

Hi soy, thanks.  what has me most concerned is the noise.  the fans spin at 3000 !  it's ridiculously loud, and i am running it stock at 218.75.  Bitmain tried to tell me this is normal.  Unreal.  my s3 used to have the fan run at 1800-2100 at stock.  i have a jet in my living room now.  unreal.   

 I am going to test these in my s-3

http://www.amazon.com/gp/product/B00A460TK6/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1


they should fit and be quiet

[Count_Frackula]

I messaged Bitmain about my order that didn't confirm and got the below responses:

Hi, we'll confirm the order manually, no worries.

Hi, Confirmed your order, will arrange shipping as schedule. We are waiting for the bitcoins transferred from blockchain.
Thanks!

Mine is now showing as confirmed, as I imagine are just about everyone else's.

[IITravel01]

Hi soy, thanks.  what has me most concerned is the noise.  the fans spin at 3000 !  it's ridiculously loud, and i am running it stock at 218.75.  Bitmain tried to tell me this is normal.  Unreal.  my s3 used to have the fan run at 1800-2100 at stock.  i have a jet in my living room now.  unreal.   

 I am going to test these in my s-3

http://www.amazon.com/gp/product/B00A460TK6/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1


they should fit and be quiet

Are those holes on the 140mm fan at the 120mm fan placement for the screws?  If not, then I don't think that fan will fit.  I see on BitMain's site that the fans listed are supposed to be 14038's, but I just took the fan off one of my S3+'s that arrived today and they're actually 120mm x 38mm fans.

[visdude]

It's ONE per household per rebate offer/campaign and not for eternity and they occur pretty frequently; and I'm certain there are CREATIVE ways around such limitation.  I've purchased CX500s at the Egg at a final cost as low as $9.99 ($39.99 list price minus $20 rebate minus $10 instant promo code discount).  I have purchased several CX500s this way since the BFL Single SC days through the Cube era.  I, in turn, had used the rebate VISA cards to purchase Egg gift cards and apply them to purchase more CX500s with ZERO out-of-pocket cost.  That's the most bang for the buck for me.

If you use it for any length of time, the bronze will cost you more in the long run.  Won't take long to recoup the $40-$50 you'd save having a gold 1300w PSU.  Plus you need more outlets.

M

That's a fair assessment.  It would be nice if you could illustrate the details of the figures though.

Off the top of my head, the difference between PSUs with 90% and 85% efficiency would be around 20WAC at the wall with a stock S3+ which equates to about 0.48kWh per day.  At $0.12/kWh power rate, that's around $0.06/day difference.  Therefore, $50 / $0.06 = 833 days = 28 months = 2.28 years to recoup a $50 difference in price.  Is this what you had in mind?

Re needing more outlets, that's what power strips are for; and I've got enough of them already and I'm almost certain that most of us do.  It could also be argued that using a single/dedicated PSU for each S3 eliminates a single point of failure of three S3s powered by a single PSU.

The CX500 is "at least 82%" according to NewEgg.
The Lightning 1300 is "at least 87%", and "up to 92%".

That 5% span is what you estimated above. 

In typical manufacturer garble, the spec sheet lists the CX500 as "> 80%" and the Lightning as "> 87%".  How that translates into "at least 82%" for the CX500 is anyone's guess.

Also remember you do get what you pay for.  The golds are higher quality, will last longer, and will feed better power to your units.  2.28 years may seem like a long time, but I know I've had my PSUs around for mining longer than that.  And if the day comes when I'm no longer mining, I'll likely find a new home easier for my gold PSU than your bronze PSU.

M

I just didn't pull random numbers out of a top hat.  In fact, you didn't exactly provide supporting data with your "Won't take long to recoup the $40-$50 you'd save having a gold 1300w PSU" statement so I took the liberty of providing one in a jiffy.  You made a statement without offering details to support such a statement; I merely tried to make sense out of it.  Having said that,  I don't go by what the Egg says or what Wikipedia says or what the spec sheet says in this case because there are more accurate PSU data in the form of actual 80 Plus certification test results.  That's the whole point of the 80 Plus ratings, isn't it?  The following links are actual test data of a CX500 and a Lightning-1300.  Take note of the "actual" efficiency values and tell me if 5% efficiency difference between a "Gold" and "Bronze" is far off as you alluded, especially at around 80% load where we usually operate our PSUs at.  Use the efficiency values at 50% load and at 100% load to interpolate and determine efficiency values at a given load in between using ratios and proportions:

http://www.plugloadsolutions.com/psu_reports/CORSAIR_75-001667_ECOS%203258_500W_Report.pdf
http://www.plugloadsolutions.com/psu_reports/ROSEWILL%20INC._LIGHTING-1300_ECOS%202330.1_1300W_Report.pdf

If you were gonna nitpick to squeeze out another 1% efficiency difference between them, then by all means, lets make it a 6% difference instead of 5%.  That adds 0.1kWh/day -- from 0.5kWh/day to 0.6kWh/day.  Again, based on $0.12/kWh, that's an increase from $0.06/day to $0.07/day; recouping a $50 price difference from 28 months to 24 months .  Big deal!

When you admonished others in here about unnecessarily wasting money on expensive PSUs and suggested a cheaper alternative PSU (Rosewill Lightning-1300), I actually agreed with you:

https://bitcointalk.org/index.php?topic=671189.msg8825455#msg8825455

However, it is funny that when I suggested a much cheaper alternative PSU (Corsair CX500) over your suggestion, you don't seem to agree with it which actually contradicts with your original point of saving money as far as PSU purchase is concerned.

I agree.  You get what you pay for.  I have a couple of XP boxes.  I built them almost 15 years ago using a made-in-Taiwan-special generic PSU that I got from Fry's really dirt cheap.  I still fire 'em up on a regular basis.  I guess that counts as "getting what I paid for".  Of course, I also have an assortment of PSUs including workhorses such as AX1200, AX850, GS600, CX500 (some have been mining continuously since the BFL days), some Antecs and more generic stuff.  They are all still working...so, I guess I got what I paid for.  In the end, it's all relative.

[Gyrsur]

nice product... but paid within an hour and order is expired but BTC are on the address. they should increment the time of payment to two hours. not nice!  

EDIT: they fixed it manually.

[dance191]

Assembled tube is $400 plus shipping, so $430-440.
So, both cost approximately the same $/Gh ($0.55/Gh for not overclocked tube and $0.554/Gh for overclocked S3).
Tube deployment costs $34/mo more than of S3 (tube not overclocked, S3 overclocked) per deployed Th if you plug in $0.12/kw electricity cost.
The difference would be even bigger if you overclock the tube.
In addition, S3 has a solid resale market (important!) and known reliability and ease of use.
I say, advantage to Bitmain, unless you are very technically adept and will assemble tube yourself and use it only a couple of mo.

Just use BTC prices (instead of converting from BTC to $ to BTC), as it is an apples to apples comparison.  The assembled tube is .79 BTC each (0.7648 not assembled).  Shipping in the past was free, are you saying it changed?  If it hasn't changed, .79 BTC gets you 820 GH/s shipped, so 0.00096341 BTC per 1 GH/s.  The S3 is 0.58 BTC at 453 GH/s shipped, which is 0.00128 BTC per 1 GH/s.  That is not the same per $/GH/s (the S3 is about 33% more).  Maybe your conversion to $ and back has some rounding or something, I don't know.  If FriedCat charges for shipping, let me know, but last I heard it was included.

The S3 is a little more energy efficient that is for sure.  I have very cheap electricity rates, so this personally doesn't concern me, but it is probably a factor to others.  I would estimate at 1 TH/s the tube is about 200 watts higher energy usage.  At 0.12 per kWh, 1 TH/s would result in $17.28 more electricity per month using the tube.  However, the tube is about $150 cheaper per 1 TH/s so in about 8 months the $0.12 kWh would make then the same price.

I don't know about resale, hard to predict the future.  I think if you use it for the life of the miner, in about 5 months or so these will be priced very similar, but again hard to tell.  The S3 has a much larger following right now, that is for sure.  I don't think the tube has caught on (for numerous reasons).  Personally, I pretty much only care about the initial cost per 1 GH/s.  However, for others factoring in electrical costs and resale is important as well.

[baykan]

Bitmain corrected my payment as "paid" and "valid" now.
it looks OK. Waiting for the shipment.
Thank you for all replies...

[mdude77]

I just didn't pull random numbers out of a top hat.  In fact, you didn't exactly provide supporting data with your "Won't take long to recoup the $40-$50 you'd save having a gold 1300w PSU" statement so I took the liberty of providing one in a jiffy.  You made a statement without offering details to support such a statement; I merely tried to make sense out of it.  Having said that,  I don't go by what the Egg says or what Wikipedia says or what the spec sheet says in this case because there are more accurate PSU data in the form of actual 80 Plus certification test results.  That's the whole point of the 80 Plus ratings, isn't it?  The following links are actual test data of a CX500 and a Lightning-1300.  Take note of the "actual" efficiency values and tell me if 5% efficiency difference between a "Gold" and "Bronze" is far off as you alluded, especially at around 80% load where we usually operate our PSUs at.  Use the efficiency values at 50% load and at 100% load to interpolate and determine efficiency values at a given load in between using ratios and proportions:

I'm sorry, I was not implying or trying to imply that your estimate was off.  It seemed to be pretty spot on to me.

Thanks for taking the time to elaborate what I did not.

M

[Tupsu]

Spondoolies SP20 vs Antminer S3... which has better power efficiency and which has the better price? Discuss!

It might be a fair comparison if the SP20 shipped immediately. But it doesn't. There's no comparison. SP20 is far less profitable.

shipping for 3 SP20  is $480.00.
shipping for 1 SP20  is $160.00.

shipping for 20 Antminer S3+ is free .

And B9 is probably last Bach of Antminer S3+
I just purchased a 20 pc S3+ B9

Hi,
0.58btc per unit and no discount is offered. place order in our website
directly.

Best Regards

BITMAIN

No discounts are  offered, even though I've bought from them miners in the last 6 months  for more than 70BTC

Unfortunately, for me, is  the the SP 20  more expensive them 4x Antminer S3+ 
4x S3=905 euro+ 20%   
1x SP20=1350 euro + 20%

[DrG]

Spondoolies SP20 vs Antminer S3... which has better power efficiency and which has the better price? Discuss!

The Spondoolies miner comes in at $.70/gh right now. Not even close to BMT.
I'm not sure what else there is to discuss.
At 1100W it's not even as efficient as the S3+.

0.64<0.78-true statement
0.64>0.78-false statement

0.67-for Sp-30

355 (w)/453GH=0.78 S-3
3000(w)/4500GH=0.67 SP-30
1100(w)/1700GH=0.64 Sp-20

Price-wise, S-3 has a ~14% advantage over Sp-20 (~10% if giftcard was used)

That S3 rating is with 120V.  It gets more efficient when using a 240V feed.  Not enough to alter the ratio significantly though.

[Moria843]

I hope they don't void the warranty if you add a WiFi pigtail, drill the hole, and mount the antenna connector and antenna. Has anybody heard if Bitmain has a problem with us modding to allow WiFi?

[atledenin]

Hi all,

Hope someone can help in response to a problem I'm having with an RMA.  I'd been through all the diags with Tim from Bitmain and he'd agreed to do an RMA for one of the hashing boards as its only hashing at 170-176Gh, the other is fine.  I'd explained to Tim that I'd bought it from a reseller and was happy to purchase a board from them and then send the other back for a refund.

They advised that there was a 0.66BTC charge  (I know!) for a hashing board but would be refunded when the other is returned so I was happy to do this.  I filled the form in without the order number and reiterated that I bought from a reseller and I thought they'd be ok as I was paying for the board.  But my request got rejected at the final hurdle and was told to go back to my reseller.

Sooooooo I emailed my reseller who advised that I need to send a screenshot in, which I did.  It shows the unit on for about 14 hours with a hashrate of 387Ghs at stock speed.

The reseller has now been in touch and advised me that Bitmain are now refusing to do an RMA as it's within a +/- 10% threshold!

Anyone had anything similar happen? So frustrating!

[bspurloc]

upgrade kit ordered 9/11 shipping today. nice

[soy]

Hi all,

Hope someone can help in response to a problem I'm having with an RMA.  I'd been through all the diags with Tim from Bitmain and he'd agreed to do an RMA for one of the hashing boards as its only hashing at 170-176Gh, the other is fine.  I'd explained to Tim that I'd bought it from a reseller and was happy to purchase a board from them and then send the other back for a refund.

They advised that there was a 0.66BTC charge  (I know!) for a hashing board but would be refunded when the other is returned so I was happy to do this.  I filled the form in without the order number and reiterated that I bought from a reseller and I thought they'd be ok as I was paying for the board.  But my request got rejected at the final hurdle and was told to go back to my reseller.

Sooooooo I emailed my reseller who advised that I need to send a screenshot in, which I did.  It shows the unit on for about 14 hours with a hashrate of 387Ghs at stock speed.

The reseller has now been in touch and advised me that Bitmain are now refusing to do an RMA as it's within a +/- 10% threshold!

Anyone had anything similar happen? So frustrating!

Something similar tho I didn't try to return it as: a) I ripped it apart and changed the thermal conductive paste and b) although that improved it considerably it is still running slow but less than 10% off.  An inspection of the package, which I insisted be shipped USPS tho he wanted another carrier, shows the bottom of the package had been opened.  When it was closed  up he didn't even try to hide the fact it had been previously opened, just using the priority mail tape not the clear tape.  I had thought he had used a bigger box but I was wrong.  All the boxes are the same size.  He really didn't want to ship USPS even tho it was cheaper than the carrier he was suggesting.  I thought that was probably due to the long lines at the post office but perhaps he was worried about US Postal Mail Fraud because if the miner was a return from an other customer I doubt he could sell it as new.  I would suggest that when buying from a reseller, ask if the package had ever been opened and insist you want a miner in a box that hadn't yet been opened.

[BITMAIN]

Dear Everyone,

We are glad to share the news of ANTMINER S4 2TH/s, Batch 1 will be on sale this week, shipped before Sept. ends.

BITMAIN