ANTMINER S4 Discussion and Support Thread

[dogie]

Ive been waiting patiently since JUNE for compensation which bitmain had finally offered to pay for four dead PSUs (actually i RMAd the first one, but the replacemnet which took 2 weeks was DOA & dented too, so its 5..) after MUCH MUCH struggle and effort

Can you post your order number so it can be looked at please? [or PM to Tim]

[1714194372]

1714194372

[1714194372]

1714194372

[bolehvpn]

Did anyone notice the new firmware up?

https://www.bitmaintech.com/support.htm?pid=00720140930114518599JXGHWWD80660

I'm pretty sure I didn't see it on the 16th October which is the date of that file.

fix a bug of share committing.
modify miner status web page.
MD5: BF5AB5F21AED999146800AB206A0CDA0

Dogie, can you verify that this fixes the issues and therefore we don't need to use CK's custom miner anymore?

[MrGreenHat]

I have not received a response from BITMAIN support in over 48 hours.

Will be answered soon, and / or Tim will expedite it.

In my email, I  apologized for getting so upset

Thats not a problem, won't affect anything.

you're going to need to do an RMA to replace the PSU for the second time

From what you suggested before, it didn't sound like a PSU issue as your non Bitmain PSU was also not able to power the miner, correct? If you're in the US then BITMAIN_warranty can do a unit swap for you.

I can't believe I'm saying this but I think I like you now! ;-) I am indeed in the U.S., and if this is possible (a full unit swap) that would be fucking AWESOME! Please let me know how I can go about this process so I can edit my posts and not harm your (BITMAIN's) reputation further.

[dogie]

...

I was attempting to explain why the PSUs didn't fail because they were a no name Chinese product, nevermind.

[dogie]

Did anyone notice the new firmware up?

https://www.bitmaintech.com/support.htm?pid=00720140930114518599JXGHWWD80660

I'm pretty sure I didn't see it on the 16th October which is the date of that file.

fix a bug of share committing.
modify miner status web page.
MD5: BF5AB5F21AED999146800AB206A0CDA0

Dogie, can you verify that this fixes the issues and therefore we don't need to use CK's custom miner anymore?

Firmwares are dated when they were completed and submitted for testing, not when they're released.

See a page or two back. Doesn't include CK's custom fixes, includes other fixes which provides a small hash rate boost. I don't know if compatibility with his patch remains as it was designed for a previous firmware.

[dogie]

I have not received a response from BITMAIN support in over 48 hours.

Will be answered soon, and / or Tim will expedite it.

In my email, I  apologized for getting so upset

Thats not a problem, won't affect anything.

you're going to need to do an RMA to replace the PSU for the second time

From what you suggested before, it didn't sound like a PSU issue as your non Bitmain PSU was also not able to power the miner, correct? If you're in the US then BitmainWarranty can do a unit swap for you.

I can't believe I'm saying this but I think I like you now! ;-) I am indeed in the U.S., and if this is possible (a full unit swap) that would be fucking AWESOME! Please let me know how I can go about this process so I can edit my posts and not harm your (BITMAIN's) reputation further.

Let me be clear, I'm not able to authorise a return on your behalf, its just a potential. PM / call BitmainWarranty with a link to your post and information on your order number, shipping address etc.

[bgibso01]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Oh yeah, and if they use CGMiner, isn't one of the requirements to disclose what modifications a company makes to it?

[dogie]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Being dated 10/16 means it was locked and submitted for testing around that date, so even if the engineers had been aware of CK's version (which they weren't until today), it wouldn't have gone in this firmware. I don't believe CK's fix fixes all pools / the difficulty problem entirely (), but it likely includes other small bug fixes that haven't been included yet. They're comparing code now.

[-ck]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Being dated 10/16 means it was locked and submitted for testing around that date, so even if the engineers had been aware of CK's version (which they weren't until today), it wouldn't have gone in this firmware. I don't believe CK's fix fixes all pools / the difficulty problem entirely (), but it likely includes other small bug fixes that haven't been included yet. They're comparing code now.

I submitted the fixes I had to a bitmain engineer and they were merged in this s4-fixup branch here:

https://github.com/jameshilliard/cgminer/commits/s4-fixup

So they know about all the tweaks and changes I did to make the binary work. Note the commits are dated 18th October. Whether they decide to use them in their master branch here:
https://github.com/bitmaintech/cgminer/commits/master

Is a different story, and it certainly does not look like they did.

Yes my changes do fix diff handling on ALL pools.

[bgibso01]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Being dated 10/16 means it was locked and submitted for testing around that date, so even if the engineers had been aware of CK's version (which they weren't until today), it wouldn't have gone in this firmware. I don't believe CK's fix fixes all pools / the difficulty problem entirely (), but it likely includes other small bug fixes that haven't been included yet. They're comparing code now.

How were they not made aware of it until today?

CK "Since they didn't fix it themselves, I've submitted fixes for them. You can grab my binary and install it manually in the meantime, or wait for them to include the fixes and create new firmware." 10/19/2014

At this link, https://bitcointalk.org/index.php?topic=796839.940 you posted right below CK who was referencing the fix.  Did you not inform Bitmain's support of the work around and issue?  If your duties don't allow a two way communication with Bitmain, what's the point?

[bolehvpn]

Well I upgraded to Bitmain's firmware on one of my S4s.

Several issues:

• Still doesn't retain settings upon firmware upgrade
• The reboot function from the web interface is still wonky. Miner won't start up unless power cycled.

I'm monitoring the hash rate. For the moment seems ok.

[bgibso01]

I submitted the fixes I had to a bitmain engineer and they were merged in this s4-fixup branch here:

https://github.com/jameshilliard/cgminer/commits/s4-fixup

So they know about all the tweaks and changes I did to make the binary work. Note the commits are dated 18th October. Whether they decide to use them in their master branch here:
https://github.com/bitmaintech/cgminer/commits/master

Is a different story, and it certainly does not look like they did.

Yes my changes do fix diff handling on ALL pools.

Thanks CK.  I appreciate Dogie's guides, but the postings in this thread are more in line with bcp19....too soon?

[dogie]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Being dated 10/16 means it was locked and submitted for testing around that date, so even if the engineers had been aware of CK's version (which they weren't until today), it wouldn't have gone in this firmware. I don't believe CK's fix fixes all pools / the difficulty problem entirely (), but it likely includes other small bug fixes that haven't been included yet. They're comparing code now.

I submitted the fixes I had to a bitmain engineer and they were merged in this s4-fixup branch here:

https://github.com/jameshilliard/cgminer/commits/s4-fixup

So they know about all the tweaks and changes I did to make the binary work. Note the commits are dated 18th October. Whether they decide to use them in their master branch here:
https://github.com/bitmaintech/cgminer/commits/master

Is a different story, and it certainly does not look like they did.

Yes my changes do fix diff handling on ALL pools.

Thanks, passed it on.

PS while you're, found the most minor description blip:

--rock-freq <arg>   Set RockMiner frequency in MHz, range 200-400 (default: 270)

--rock-freq <arg>   Set RockMiner frequency in MHz, range 125-500 (default: 270)

[FUBAR-BDHR]

Good luck it took me over 2 months to convince them I had a bad s1 then when the did finally send me replacement blades both of those had even worse issues but they wouldn't do anything about it.  Took weeks to convince them both my s3 units were bad as well.  Now I have 2 more that are starting to exhibit the same issues the other 2 did new.  It's just not worth trying to deal with their "support" and the extended down time for an RMA.  Glad I stayed away from the S4.

[MrGreenHat]

Rather disappointing on the support part of this thread regarding the firmware.  If I understand correctly from reading the past couple of pages, the firmware is dated 10/16, but not posted until 10/27.  CK has implemented a work around / fix in the mean time and the folks being paid to answer support on this thread can't say whether or not BM fixed the problem in their firmware that CK had to fix to get the machines to function properly on pools that use var diff.  Does that about sum it up?

Being dated 10/16 means it was locked and submitted for testing around that date, so even if the engineers had been aware of CK's version (which they weren't until today), it wouldn't have gone in this firmware. I don't believe CK's fix fixes all pools / the difficulty problem entirely (), but it likely includes other small bug fixes that haven't been included yet. They're comparing code now.

I submitted the fixes I had to a bitmain engineer and they were merged in this s4-fixup branch here:

https://github.com/jameshilliard/cgminer/commits/s4-fixup

So they know about all the tweaks and changes I did to make the binary work. Note the commits are dated 18th October. Whether they decide to use them in their master branch here:
https://github.com/bitmaintech/cgminer/commits/master

Is a different story, and it certainly does not look like they did.

Yes my changes do fix diff handling on ALL pools.

You are a kind, amazing person for doing what you have done.  I wish that I could power on my S4 and update the firmware to yours, but apparently that is never going to happen.

[Prelude]

Maybe, maybe not. A lot of places in the world suffer from brownouts which can bring voltage plummeting down to ~90v. For anywhere in Canada/USA to be a constant 100v instead of 120v to me is ridiculous. Detroit is a special situation though, I suppose.

I'm fortunate enough to have one of the best electrical utility companies in the world (Hydro Quebec) servicing me, so I don't have to worry about low voltage situations. I draw a constant ~100-120A from my 200A home electrical installation, and still show 118v/236v at the wall. That being said, I would gladly pay an extra $20 per miner if it meant having an actual quality PSU in the unit. When it comes down to it, these high wattage power supplies almost have enough current to arc weld. I have absolutely no faith in a no-name Chinese power supply company, and my worries were proven right with the S4. I don't care if it's made in China, as long as it caries a known brand name with specs and reputation that I can look up. I'll pay extra every day if it means getting a hassle free PSU from the likes of Seasonic, Flextronics, Delta, Super Flower, etc. Apluspower Co. doesn't inspire any faith at all, and is a fire hazard in my books.

I'd like to believe that the revised power supplies from B2 will solve the issue, but there have already been reports of burnt out/DOA PSUs. I'll be replacing mine with 2000w Deltas to be safe, and when I sell them they will still have working OEM PSUs.

Its worth noting that there is nothing wrong with those B1 PSUs for computer use - the problem was caused by the unusual loading pattern miners have. This turned would would be considered sensible cost saving [/earlier development] into a failure mode.

Unusual pattern or not, if it was a quality unit it would have handled the loads. There was a lack of engineering somewhere along the line at Apluspower Co. that resulted in them failing where a proper PSU would not have failed. The day my S4's PSU died, I ordered a replacement Delta DPS-2000 for it and also for my newer B2 S4. I know the Delta unit will handle the loading pattern without issue, as any well built PSU should. I would expect consumer grade PSUs to handle the load also, so long as they are quality units from Seasonic, Felx, etc.

To be clear, I'm not bashing Bitmain. They are not power supply manufacturers, and I'm sure they took A+ at their word that their PSU's would be sufficient to properly power the S4. A+ is to blame here, IMO.

I bought my DPS-2000 for $30 on ebay. If I'm not mistaken, they were introduced on the server market about 5 years ago and cost somewhere in the 4 digits range. While it's used, and requires cooling be provided externally, it's a much safer bet than anything A+ could ever muster. Maybe in the future Bitmain would be better off integrating an unconventional used server PSU in their units. It doesn't get any better than a server PSU from the likes of Delta, Zippy, etc. Even when used.

[dogie]

Maybe, maybe not. A lot of places in the world suffer from brownouts which can bring voltage plummeting down to ~90v. For anywhere in Canada/USA to be a constant 100v instead of 120v to me is ridiculous. Detroit is a special situation though, I suppose.

I'm fortunate enough to have one of the best electrical utility companies in the world (Hydro Quebec) servicing me, so I don't have to worry about low voltage situations. I draw a constant ~100-120A from my 200A home electrical installation, and still show 118v/236v at the wall. That being said, I would gladly pay an extra $20 per miner if it meant having an actual quality PSU in the unit. When it comes down to it, these high wattage power supplies almost have enough current to arc weld. I have absolutely no faith in a no-name Chinese power supply company, and my worries were proven right with the S4. I don't care if it's made in China, as long as it caries a known brand name with specs and reputation that I can look up. I'll pay extra every day if it means getting a hassle free PSU from the likes of Seasonic, Flextronics, Delta, Super Flower, etc. Apluspower Co. doesn't inspire any faith at all, and is a fire hazard in my books.

I'd like to believe that the revised power supplies from B2 will solve the issue, but there have already been reports of burnt out/DOA PSUs. I'll be replacing mine with 2000w Deltas to be safe, and when I sell them they will still have working OEM PSUs.

Its worth noting that there is nothing wrong with those B1 PSUs for computer use - the problem was caused by the unusual loading pattern miners have. This turned would would be considered sensible cost saving [/earlier development] into a failure mode.

Unusual pattern or not, if it was a quality unit it would have handled the loads. There was a lack of engineering somewhere along the line at Apluspower Co. that resulted in them failing where a proper PSU would not have failed.

This, I guess.

[ZiG]

Maybe, maybe not. A lot of places in the world suffer from brownouts which can bring voltage plummeting down to ~90v. For anywhere in Canada/USA to be a constant 100v instead of 120v to me is ridiculous. Detroit is a special situation though, I suppose.

I'm fortunate enough to have one of the best electrical utility companies in the world (Hydro Quebec) servicing me, so I don't have to worry about low voltage situations. I draw a constant ~100-120A from my 200A home electrical installation, and still show 118v/236v at the wall. That being said, I would gladly pay an extra $20 per miner if it meant having an actual quality PSU in the unit. When it comes down to it, these high wattage power supplies almost have enough current to arc weld. I have absolutely no faith in a no-name Chinese power supply company, and my worries were proven right with the S4. I don't care if it's made in China, as long as it caries a known brand name with specs and reputation that I can look up. I'll pay extra every day if it means getting a hassle free PSU from the likes of Seasonic, Flextronics, Delta, Super Flower, etc. Apluspower Co. doesn't inspire any faith at all, and is a fire hazard in my books.

I'd like to believe that the revised power supplies from B2 will solve the issue, but there have already been reports of burnt out/DOA PSUs. I'll be replacing mine with 2000w Deltas to be safe, and when I sell them they will still have working OEM PSUs.

Its worth noting that there is nothing wrong with those B1 PSUs for computer use - the problem was caused by the unusual loading pattern miners have. This turned would would be considered sensible cost saving [/earlier development] into a failure mode.

Unusual pattern or not, if it was a quality unit it would have handled the loads. There was a lack of engineering somewhere along the line at Apluspower Co. that resulted in them failing where a proper PSU would not have failed.

This, I guess.

Dogie...enough excuses ...please...

ZiG

EDIT...: Hope that you COULD do BETTER ...to earn your salary, buddy...

[Tim.Shao19]

yep, i will try to confirm if you can get your miner swapped with BitmainWarranty for diagnostics.  
There should be some normal procedures, but we will try to find a better solution for u considering your situation.
Please pm me about this.  

I have not received a response from BITMAIN support in over 48 hours.

Will be answered soon, and / or Tim will expedite it.

In my email, I  apologized for getting so upset

Thats not a problem, won't affect anything.

you're going to need to do an RMA to replace the PSU for the second time

From what you suggested before, it didn't sound like a PSU issue as your non Bitmain PSU was also not able to power the miner, correct? If you're in the US then BITMAIN_warranty can do a unit swap for you.

I can't believe I'm saying this but I think I like you now! ;-) I am indeed in the U.S., and if this is possible (a full unit swap) that would be fucking AWESOME! Please let me know how I can go about this process so I can edit my posts and not harm your (BITMAIN's) reputation further.

[edgar]

Ive been waiting patiently since JUNE for compensation which bitmain had finally offered to pay for four dead PSUs (actually i RMAd the first one, but the replacemnet which took 2 weeks was DOA & dented too, so its 5..) after MUCH MUCH struggle and effort

Can you post your order number so it can be looked at please? [or PM to Tim]

im not in the US

i have sent all relevant details to bitmain - as i stated we have exchnanged over 400 emails. most of them utter stupidity, usually repeatedly.

i am now at the point that only actual compensation as offered previously will suffice.

i have enough shit antminer junk in my home, i am able to read and as such will never accept another dangerous (3 month BS warranty) rig from bitmain.

four S2s = FOUR DEAD PSUs + 1 more PSU that was RMA'd and arrived dirty, dented and only able to power 9 blades.ffs!

multiple dead/failing boards have been RMA'd - some were ok, others not.

a Dead power board has been RMA'd - an equally shit & filthy one was sent

a dead control board has been RMA'd - it works but arrived absolutely FILTHY.

he's refused to replace the buckled fans & the dead LED screen and to be quite honest the last blade needing RMA'd has just sat gathering dust despite having been sent a shipping label. Its just more messing around to be sent yet another already RMA'd piece of shit.

so now ive repeated myself to you too, what do you suggest?

more order numbers?

more screen grabs?

Tell 'Tim'?

ZZzzzzzzz....