S3+ (BM1382) Overclocking with voltage setting

[intobit459]

No by regular value i am referring to the third value xxx:xxx:YYYY Where the ys are
15:218.75:yyyy

there are two different reg values in that link thats why i am asking which reg value should i try?  Im never home to edit settings

funny we are both editing our posts too

I just read another part you wrote I thought default voltage shipped was 0785 is it actually 0725 so i am overvolting going to 0750? I thought 0750 was undervolted

I guess a better question is am i using more power with a setting of 0750 and 262.5 than a factory s3 batch 1 with factory settings?  Im thinking not but im getting confused here.

[pekatete]

No by regular value i am referring to the third value xxx:xxx:YYYY Where the ys are

There's your reg value for 262.5 - I mentioned this to you earlier!

<snip> ... </snip>
There is a link in the OP in this regard (about adding frequencies), and extracting the 262.5 freq line I get: 15:262.5:0a02
<snip> ... </snip>

I just read another part you wrote I thought default voltage shipped was 0785 is it actually 0725 so i am overvolting going to 0750?

I would not say you are over-volting, simply setting the ideal voltage for the frequency you are running - look at the datasheet! Prior to the new firmware, all freqs (overclock and underclock) were being done on the same voltage that bitmain had hard-coded, thus they had to choose a middle voltage that would work for a range of freqs. With the voltage setting (and the datasheet), we can set the voltage to the freq we run the rigs at, thus less errors and better efficiency.

I guess a better question is am i using more power with a setting of 0750 and 262.5 than a factory s3 batch 1 with factory settings?  Im thinking not but im getting confused here.

You need to acquaint yourself better with the basics here; A setting of 0750 represents a voltage of 0.75v and that of 0725 is for 0.725v. The default frequency is 218.75 (I think) whereas you'd be overclocking to 262.5 (over is the operative word here). So yes, your rig will be utilising more power since the clock frequency is faster i.e shrinking a second in time as you'd know it, and also using a touch more volts per cycle, i.e 0.025v. Having said that, if say you are running a 650watt PSU, you shall not see a difference in power draw either way.

[intobit459]

I was under the impression the default voltage is 0785 which is apparently wrong it is 0725 according to you.  The datasheet just shows me using less power than a factory s3. Its all really confusing. What do i add to get actual watts like the control board uses added to the asics.  Im using a 750w gold rating thermaltake toughpower 2 pcie.  Guy i got it from said use 4 pcie but he gave me two pcie cables with 2 pcie ends on one side so i figured i shouldnt plug in the additional 2 since they probably are on the same wire

[pekatete]

I was under the impression the default voltage is 0785 which is apparently wrong it is 0725 according to you.  The datasheet just shows me using less power than a factory s3. Its all really confusing. What do i add to get actual watts like the control board uses added to the asics.  Im using a 750w gold rating thermaltake toughpower

You could very possibly be under the right impression, or I could indeed be correct on that, but in this thread's context, that's neither here nor there, more so the power consumption of the board, fans etc.

[intobit459]

We will figure it out but you've been very helpful. Thank you.  I will try the regvalue and see if it has a positive effect on hw errors.

For future reference I believe we still can use vi /etc/config/cgminer but theres no reason to as its in web interface.  Because that worked for me using vi when I changed my clock value and didnt touch timeout or regvalue which i will when i get home.  But I do have an extranonce copy of cgminer installed.  Though I dont think that matters.

[IITravel01]

I believe that the default voltage may be different for the current firmware.  I don't know if it's different for the S3 and the S3+, here's what I believe.  For the S3+ the default voltage for the Nov. firmware was 0770.  I suspect that the new Dec. firmware may have a default voltage for the S3+ is 0725 (I was told that was the temporary fix or work around for the hashing dropping to less than 50Gh/s on the Nov. firmware).  If anyone wants to use putty, you can see for yourself to verify.  Does the default voltage show up on the new December firmware when you go to the advanced screen or is it blank like before?  I can't tell on mine because I had 0725 set on my two S3+ and when I updated this morning to the new Dec. firmware I selected to keep settings, so after it rebooted I went to the advance tab and it still showed 0725.

[intobit459]

It might be in cgminer file
Try ssh in
Then
cat /etc/config/cgminer
Thats where clock settings are

I will find file when I get home to see where voltage is stored.

What is default in october firmware and batch 1?

is the december firmware stable for us to run 0750 and 262.5 overclock or should I stick to october?  November was disregarded

firmware is same for s3 and s3+ to my knowledge

I also emailed bitmain and will post answer soon

[IITravel01]

Firmware might be the same for the S3 and the S3+, but the settings aren't.  The S3 had a lower default frequency setting than the 225M that the S3+ does (which might mean the have different voltage values for each frequency setting).

[BTC-TMXSTAR]

Hi! My Antminer S3+ will shutdown itselft after 2minutes of running.
Any ideas?
Please help me


Thanks advance
and
God Bless!!!!!

[intobit459]

Data gathered from bitmain emails

default value for S4 is 0725.

they also said s3 is 0760 or 0770 and s3+ is 0750

idk... hard to know without investigating firmware without custom settings

S3 is 441gh/s and S3+ is 453gh/s


Edit. (OCFREQ/REGFREQ)xAMPSxVOLTS×#chips=WATTS

[pekatete]

I found another sweet spot for my / the S3+ at a frequency of 275 and a voltage setting of 0815. I had to change the timeout from the suggested 15 to 14 to reduce the HW errors and having run it overnight, it looked good. I've had to power cycle the rig (as I did not do that yesterday) and currently re-running the setting, but looks good thus far! (also updated the OP on the OC thread here: https://bitcointalk.org/index.php?topic=750220)

[IITravel01]

Does page 7 (labeled 6) of the new Version 2 help or is this for the upcoming S5 only (notice 275 is listed at 14 delay)?  https://www.bitmaintech.com/files/download/BM1384_Datasheet_v2.0.pdf

[pekatete]

Does page 7 (labeled 6) of the new Version 2 help or is this for the upcoming S5 only?  https://www.bitmaintech.com/files/download/BM1384_Datasheet_v2.0.pdf

Clearly labelled BM1384, so not for the S3 which has a BM1382 (as far as I know)!

[IITravel01]

Does page 7 (labeled 6) of the new Version 2 help or is this for the upcoming S5 only?  https://www.bitmaintech.com/files/download/BM1384_Datasheet_v2.0.pdf

Clearly labelled BM1384, so not for the S3 which has a BM1382 (as far as I know)!

Oops, didn't catch that, but did see that 275 was at optimal 14 delay so it's coincidence then (oddly on page 3 under Overview it states BM1382 probably in error).  Do the other delay settings match up to the suggested speed also?

Clock Freq. (M)         Ideal delay (ms)             Rec. delay (ms)
200                          21.3                              19
225                          18.9                              17
250                          17.0                              15
275                          15.5                              14

[pekatete]

Does page 7 (labeled 6) of the new Version 2 help or is this for the upcoming S5 only?  https://www.bitmaintech.com/files/download/BM1384_Datasheet_v2.0.pdf

Clearly labelled BM1384, so not for the S3 which has a BM1382 (as far as I know)!

Oops, didn't catch that, but did see that 275 was at optimal 14 delay so it's coincidence then.  Do the other delay settings match up to the suggested speed also?

You are right about freq 275 having an optimal 14ms delay and for the life of me, I do not know why I had mine set at 15! And by the way, that BM1384 datasheet still has most data relating to the BM1382 (including that work input timing table). I think the only thing they've changed thus far is the Power performance table!

If I recall correctly, the S3 does not ship with these freqs and I got them off a thread on this forum .... so that must have been wrong, or even more likely, I was tinkering around (as I do) and forgot to set it back!

[canford]

I continue to find this extremely frustrating.  I am running 6 S3+ units, each supplied by a 750-800W power supply with a single 12V rail and all four PCI-E connectors connected.  I tried the 275/0815 settings, and exactly the same thing happened as the last time.  Hour 1: fantastic, all units over 500 GH/s, holy crap why did I not do this earlier.  Hour 2: OK, but hashrates down both as reported by units and at the pool.  Hour 3: terrible, all hashrates reported by units and by pool way below stock results.  So I have yet again reverted everything to 225/231/243/237/225/237 with voltage left blank.

I do not understand the mechanism here.  If it is heat, why does it take 2 hours to show up?  With the OC settings, all chips show up as "o" and working.  But the hashrate results speak for themselves.  But why the very prominent jump in hashrate for the first hour, but then utter crap out in hour 3?  Can anybody explain this?

Which firmware are you using?

I am using the 1024 firmware, the only change on top of that is adding the new frequencies into cgminer.lua.  All six units stable again at the lower frequencies with no voltage setting.

Trying again on one unit, this time with ckolivas's S3 cgminer 4.6.1 build, also deleting --queue and adding --lowmem.  I can't think of anything else to try.  The S3's are in a ventilated space, so the ambient temperature over the three hour period is constant.  Why would I get great results for 1+ hours, but poor results over 3 hours?

[DarkRavenGamer]

I found another sweet spot for my / the S3+ at a frequency of 275 and a voltage setting of 0815. I had to change the timeout from the suggested 15 to 14 to reduce the HW errors and having run it overnight, it looked good. I've had to power cycle the rig (as I did not do that yesterday) and currently re-running the setting, but looks good thus far! (also updated the OP on the OC thread here: https://bitcointalk.org/index.php?topic=750220)

http://s7.postimg.org/4ko9o5wvv/S3_275.png

Amazing thread with amazing results!! I can't wait to get home to try out voltage OCi g my S3+ miners.

[jjiimm_64]

I found another sweet spot for my / the S3+ at a frequency of 275 and a voltage setting of 0815. I had to change the timeout from the suggested 15 to 14 to reduce the HW errors and having run it overnight, it looked good. I've had to power cycle the rig (as I did not do that yesterday) and currently re-running the setting, but looks good thus far! (also updated the OP on the OC thread here: https://bitcointalk.org/index.php?topic=750220)

Could you please post how many watts at the wall this is pulling...  when I did my tests of voltages I could not get the wattage at the wall to change by changing the volts..

[pekatete]

Could you please post how many watts at the wall this is pulling...  when I did my tests of voltages I could not get the wattage at the wall to change by changing the volts..

I do not have a tool to measure the pull at the wall, but if your tests showed no change, I have no reason to doubt it.

OBITER DICTA: It seems the new firmware that bitmain are testing (I got a copy from bitmain earlier today and you can ask to be sent a copy), handles voltage settings well using the default value of 0000 (I suppose that signifies auto select the voltage). I flashed one of my S3's with the firmware and in the first hour that I was able to observe, it seemed to be handling well. I will not be able to check on the rig as I am away for a couple of days, but will be able to on the 26th or 27th.

[jjiimm_64]

Could you please post how many watts at the wall this is pulling...  when I did my tests of voltages I could not get the wattage at the wall to change by changing the volts..

I do not have a tool to measure the pull at the wall, but if your tests showed no change, I have no reason to doubt it.

OBITER DICTA: It seems the new firmware that bitmain are testing (I got a copy from bitmain earlier today and you can ask to be sent a copy), handles voltage settings well using the default value of 0000 (I suppose that signifies auto select the voltage). I flashed one of my S3's with the firmware and in the first hour that I was able to observe, it seemed to be handling well. I will not be able to check on the rig as I am away for a couple of days, but will be able to on the 26th or 27th.

Thanks for the reply.. will said firmware be publicly avail soon? or now? would love to retest the wattage at the wall issue.