ANTMINER S3+ Discussion and Support Thread

[Phosphorous]

Does anyone has any experience with using 2 different psu's (one for each blade) on an S3. I have a couple of corsair CX430's laying around and was hoping i could use 2 to power one S3.

One of those supplies should work fine for one S3 at stock frequencies as consumption is only around 14A per blade. Even overlocked to 250Mhz you'd probably be OK but your supply would be loaded to the max.

[1714017875]

1714017875

[1714017875]

1714017875

[9600]

I'm sure many of you know how much trouble I was having with my S3s from B3 last week.
Since I've been running the cgminer restart every 1/2 hour, these are the results.



S31 is clocked at 218, up from 212.
S32 is clocked at 250, up from 218.

24-hour figures are still on the rise.

I am going to re-torque the heatsink on S31 and clock back up to 250. It holds, but always loses one chip about 10 minutes in.

They KEY here, is the restart the cgminer process whenever your problem miners start losing hash. It seems to be VERY different for each miner. But this has been the answer for me.

It ghash.io stats page. Post your Antminer stats page.
In ghash stats page I get even 800 for single S3, these are not reliable.
Also when restarting average of 30minute even at miners stats is not so real.
I mean sometimes when cgminer us started the avg will just show a high value like 500, 600 sone times it shows actual value.
So, do you have a figure for every 1hr restart? Why you are restarting at 30minutes? Due to chip shutoff?

Those are 5 minute, 15 minute, 1 hour, and 24 hour averages. Its from my pool, they pay me according to accepted shares. This is what I'm most concerned about. cgminer stats on these S3s are veritably useless.

Yes I know I have asic errors, but I don't really care because in 15 minutes, I won't.
My hashrates are up, my payouts show it.

This is it for me. No one ever said it was an elegant solution.

47 minutes and one restart later, no ASIC errors. Same accepted share rate. This works.

[allcoinminer]

[MG]https://i.imgur.com/MyVl7y5.png[/img]

[MG]https://i.imgur.com/zgaxTrP.png[/img]

47 minutes and one restart later, no ASIC errors. Same accepted share rate. This works.

Have you noticed actual bitcoin payout increase with this?

[9600]

[MG]https://i.imgur.com/MyVl7y5.png[/img]

[MG]https://i.imgur.com/zgaxTrP.png[/img]

47 minutes and one restart later, no ASIC errors. Same accepted share rate. This works.

Have you noticed actual bitcoin payout increase with this?

That's really hard to say. I run a pretty big farm and the pool luck has been VERY high lately. I'm inclined to say yes but I have no real empirical data.

What I can say for SURE, is that I am seeing more accepted shares from both units.

I use a modified version of this https://bitcointalk.org/index.php?topic=222632.0 script to query and store share data in a MongoDB for analysis.

[Lowell904]

Does anyone has any experience with using 2 different psu's (one for each blade) on an S3. I have a couple of corsair CX430's laying around and was hoping i could use 2 to power one S3.

The CX430 should be rated for 32A on the 12+ rail. If you're feeling brave you could probably run a single S3@218 on one power supply. Would definitely work @212.

Also after bitcoin discount the corsair CX430 is only ~$15 from newegg. Very cheap solution for an underclocked S3

http://www.newegg.com/Product/Product.aspx?Item=N82E16817139026

You can't run a S3 on one CX430 because it only has one PCI-E.

I was thinking you could too.

[RiseOfTH]

Does anyone has any experience with using 2 different psu's (one for each blade) on an S3. I have a couple of corsair CX430's laying around and was hoping i could use 2 to power one S3.

The CX430 should be rated for 32A on the 12+ rail. If you're feeling brave you could probably run a single S3@218 on one power supply. Would definitely work @212.

Also after bitcoin discount the corsair CX430 is only ~$15 from newegg. Very cheap solution for an underclocked S3

http://www.newegg.com/Product/Product.aspx?Item=N82E16817139026

You can't run a S3 on one CX430 because it only has one PCI-E.

I was thinking you could too.

Can I run two Bitmain S3 on 950 watt PSU (http://www.amazon.com/gp/product/B003U29C40/ref=oh_aui_detailpage_o03_s00?ie=UTF8&psc=1)? It has four PCI-e, +12V 83A with max load of 1000W?

I have another 850W seasonic which has two PCI-e. But I have 2-6 pin mole connector to 1-pci-e connector coverter. So, I can make another two PCI-e connectors. Is is safe to use such configuration for two Bitmain S3?

Also, how long it takes to reach US once Bitmain ships the product from China (https://bitmaintech.com/product.htm)?

I have ordered two Bitmain s3 through a Ebay seller.

Thanks,

[MoreBloodWine]

Still no movement on he order page or emails for my two orders of 2 B5 S3's each ;-(

And here I thought they wee shipping before he 7th like they said ;-/

[9600]

Does anyone has any experience with using 2 different psu's (one for each blade) on an S3. I have a couple of corsair CX430's laying around and was hoping i could use 2 to power one S3.

The CX430 should be rated for 32A on the 12+ rail. If you're feeling brave you could probably run a single S3@218 on one power supply. Would definitely work @212.

Also after bitcoin discount the corsair CX430 is only ~$15 from newegg. Very cheap solution for an underclocked S3

http://www.newegg.com/Product/Product.aspx?Item=N82E16817139026

You can't run a S3 on one CX430 because it only has one PCI-E.

I was thinking you could too.

Depending on your confidence, you could always cut the 8-pin CPU connector and splice in a PCIe connector if you have an extra adapter lying around. All of mine are running on similarly adapted server PSUs.

[padrino]

Still no movement on he order page or emails for my two orders of 2 B5 S3's each ;-(

And here I thought they wee shipping before he 7th like they said ;-/

Am I missing something, it's only the 4th right now, still a strong possibility as far as I can tell..

[MoreBloodWine]

Still no movement on he order page or emails for my two orders of 2 B5 S3's each ;-(

And here I thought they wee shipping before he 7th like they said ;-/

Am I missing something, it's only the 4th right now, still a strong possibility as far as I can tell..

Wishful thinking they ship before the 6th.

[buy4crypto]

Still no movement on he order page or emails for my two orders of 2 B5 S3's each ;-(

And here I thought they wee shipping before he 7th like they said ;-/

Depends when you ordered. Between 4 orders of mine and some friends we have them at all various stages for batch 5. Our first time stamp is in china, while our second order is in ohio already. Our third order is moving along and the fourth order has yet to start the process.

[MoreBloodWine]

Still no movement on he order page or emails for my two orders of 2 B5 S3's each ;-(

And here I thought they wee shipping before he 7th like they said ;-/

Depends when you ordered. Between 4 orders of mine and some friends we have them at all various stages for batch 5. Our first time stamp is in china, while our second order is in ohio already. Our third order is moving along and the fourth order has yet to start the process.

[Bitkoinzz]

BITMAINTECH.COM OFFLINE. Anyone know what they are doing?
edit.. disregard. back up and no changes

[MoreBloodWine]

BITMAINTECH.COM OFFLINE. Anyone know what they are doing?

Running for the hills with this ? lol

[RoadStress]

BITMAINTECH.COM OFFLINE. Anyone know what they are doing?

http://www.downforeveryoneorjustme.com/bitmaintech.com

Seems to be working just fine!

[squashpile]

What are the error rates?

I'm sure many of you know how much trouble I was having with my S3s from B3 last week.
Since I've been running the cgminer restart every 1/2 hour, these are the results.



S31 is clocked at 218, up from 212.
S32 is clocked at 250, up from 218.

24-hour figures are still on the rise.

I am going to re-torque the heatsink on S31 and clock back up to 250. It holds, but always loses one chip about 10 minutes in.

They KEY here, is the restart the cgminer process whenever your problem miners start losing hash. It seems to be VERY different for each miner. But this has been the answer for me.

[-droid-]

S3 not hashing correctly?

I've got 4 units and only 1 hashes at 440-441 at stock freq, the others get 414-420ish.

In working with bitmain over email I tried re applying thermal paste and swapping psu's around and still nada.  Then they asked me to swap control board from the one good miner to one that is underperforming and lo and behold the underperforming one is now hashing at 440-441 with the other hashing at the lower speed.

I've swapped it on then others and it's the same result.  Since the static IP is saved on the control board it moves from swap to swap with the good control board, so it constantly thinks it's the good unit and hashes at adv. spec

[taipo]

Since mine has the same probs as yours, what solution did Bitmain offer to you to rectify the faulty control board.

[-droid-]

Since mine has the same probs as yours, what solution did Bitmain offer to you to rectify the faulty control board.

Nothing yet , waiting to hear back

[9600]

What are the error rates?

I'm sure many of you know how much trouble I was having with my S3s from B3 last week.
Since I've been running the cgminer restart every 1/2 hour, these are the results.



S31 is clocked at 218, up from 212.
S32 is clocked at 250, up from 218.

24-hour figures are still on the rise.

I am going to re-torque the heatsink on S31 and clock back up to 250. It holds, but always loses one chip about 10 minutes in.

They KEY here, is the restart the cgminer process whenever your problem miners start losing hash. It seems to be VERY different for each miner. But this has been the answer for me.

Check https://bitcointalk.org/index.php?topic=671189.msg8169759#msg8169759 here.