|
oskuro
|
 |
July 26, 2014, 11:28:40 AM |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way  thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine |
|
|
|
|
|
allcoinminer
|
|
July 26, 2014, 11:46:23 AM |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine Temperature getting from a particular point doesn't mean that one of your component is getting fried. But bitmaintech's technical team can contribute something here. Else we need to check it ourselves without the cover. |
|
|
|
|
|
oskuro
|
|
July 26, 2014, 11:48:28 AM |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine Temperature getting from a particular point doesn't mean that one of your component is getting fried. But bitmaintech's technical team can contribute something here. Else we need to check it ourselves without the cover. But i dont know why, the S3 that has less hashrate, is not estable with 237,5 now indicates its hashing at 367,66 .... i will have to lower frequency for that S3. arggggg. will test it with 225 freq other S3 also decreased from 475 GH/s to 460 GH/s (250 frequency) |
|
|
|
|
philipma1957
Legendary
 Offline
Activity: 4130
Merit: 7907
'The right to privacy matters'
|
|
July 26, 2014, 12:01:43 PM |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine Temperature getting from a particular point doesn't mean that one of your component is getting fried. But bitmaintech's technical team can contribute something here. Else we need to check it ourselves without the cover. But i dont know why, the S3 that has less hashrate, is not estable with 237,5 now indicates its hashing at 367,66 .... i will have to lower frequency for that S3. arggggg. will test it with 225 freq other S3 also decreased from 475 GH/s to 460 GH/s (250 frequency) that is the dc to dc convertor it warms up after time and works less efficiently . thus the hash rate drop. I have my 2 units set to 212.5 they are rock solid at about 430 gh they pull 328 watts each. based on my knowledge of electrical gear I believe these s-3's should be run at freq 212.5. I believe the oc camp will be reporting on burnt out s-3's in about 6 weeks. Just a guess. But these are very power efficient . So running it for 5 to 7 months and turning profits is possbile.  |
|
|
|
|
oskuro
|
|
July 26, 2014, 12:09:02 PM
Last edit: July 26, 2014, 12:21:06 PM by oskuro |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine Temperature getting from a particular point doesn't mean that one of your component is getting fried. But bitmaintech's technical team can contribute something here. Else we need to check it ourselves without the cover. But i dont know why, the S3 that has less hashrate, is not estable with 237,5 now indicates its hashing at 367,66 .... i will have to lower frequency for that S3. arggggg. will test it with 225 freq other S3 also decreased from 475 GH/s to 460 GH/s (250 frequency) that is the dc to dc convertor it warms up after time and works less efficiently . thus the hash rate drop. I have my 2 units set to 212.5 they are rock solid at about 430 gh they pull 328 watts each. 212.5 ok man thx, i will test with that frequency too (default 218.75 game me 412,72 and 434,55, lets see with 212.5, should be less GH/s i suppose...) bot at 212.5 are 690W how can you have 654W  LOL with 212.5 on my best S3 i get x!! haha i cant lower to 212.5 that unit xD |
|
|
|
|
|
allcoinminer
|
|
July 26, 2014, 12:20:16 PM |
|
It is. I pay 0.12 kwh. Even with OC I earn more / month and the prize increase for power is not that high. OK then I will leave it this way thx Edit: well but I don't think they can explode for OC. They almost run same temperature than without OC and watts increase is not that high. Mmmmm hard decision . For me it was mining stable at 440 using 218.75 but if I can get more with OC. .. I don't get any X or - works fine Temperature getting from a particular point doesn't mean that one of your component is getting fried. But bitmaintech's technical team can contribute something here. Else we need to check it ourselves without the cover. But i dont know why, the S3 that has less hashrate, is not estable with 237,5 now indicates its hashing at 367,66 .... i will have to lower frequency for that S3. arggggg. will test it with 225 freq other S3 also decreased from 475 GH/s to 460 GH/s (250 frequency) that is the dc to dc convertor it warms up after time and works less efficiently . thus the hash rate drop. I have my 2 units set to 212.5 they are rock solid at about 430 gh they pull 328 watts each. [im]http://imagizer.imageshack.us/v2/1280x1024q90/540/W4xRY8.png[/img] How about water cooling the DC-DC ? Funny idea. Drill holes in some small heat sinks and connect them in series with a small flex tube and glue. Fill them with cooling liquid and get the two tube endings to the intake fan side. Arrange some small copper tubings as radiator to cool the cooling liquid inside. The intake airflow will cool it a bit. The liquid can be rotated through the tubing using a small pump. That pump may be run using the hot air flow from the miner. |
|
|
|
|
|
ZiG
|
|
July 26, 2014, 03:02:45 PM |
|
generated a table to work with, based on the S1/S2 tables: ...package 'cgminer'
config 'asic-freq' 'default'
#option 'freq_value' '1486' #262.5M
#option 'chip_freq' '262.5'
#option 'timeout' '15'
#option 'freq_value' '1386' #250M
#option 'chip_freq' '250'
#option 'timeout' '16'
#option 'freq_value' '1306' #243.75M
#option 'chip_freq' '243.75'
#option 'timeout' '17'
#option 'freq_value' '1286' #237.5M
#option 'chip_freq' '237.5'
#option 'timeout' '17'
#option 'freq_value' '1206' #231.25M
#option 'chip_freq' '231.5'
#option 'timeout' '18'
#option 'freq_value' '1185' #225M
#option 'chip_freq' '225'
#option 'timeout' '18'
option 'freq_value' '1106' #218.75M
option 'chip_freq' '218.75'
option 'timeout' '18'
#option 'freq_value' '1086' #212.5M
#option 'chip_freq' '212.5'
#option 'timeout' '18'
#option 'freq_value' '0782' #200M
#option 'chip_freq' '200'
#option 'timeout' '20'
first off thanks to you. second can you find a number between 200 and 212.50 ( 206.25 ) comes to mind third can you find something for 215.625 my guess for 206.25 would be (use at your own risk): option 'freq_value' '1006' #206.25M option 'chip_freq' '206.25' option 'timeout' '19' With all my due respect for your time and effort, but this table is NOT very applicable to S3... You can NOT apply the same S1 math/calculations... to S3, because these are using different chips with different PFD (PLL Frequency Divider...I guess), subject to different conditions...due to chips redesign...I guess... BM frequency table - from BM1382 datasheet ...: Work Input timing Clock Freq.(M) Hash rate(G) Reg.value Ideal delay(ms) Recommenddelay(ms) 100 6.30 0783 42.6 38 125 7.88 0983 34.0 31 150 9.45 0b83 28.4 26 175 11.03 0d83 24.3 22 200 12.60 0782 21.3 19 225 14.18 0882 18.9 17 250 15.75 0982 17.0 15 275 17.33 0a82 15.5 14 300 18.90 0b82 14.2 13 325 20.48 0c82 13.1 12 350 22.05 0d82 12.2 11 375 23.63 0e82 11.3 10 400 25.20 08f2 10.6 10 These are the same in S3's "asic-freq" file... Conditions for S3 valid frequency divider's values now are different...check them here...: https://bitmaintech.com/files/download/BM1382_Datasheet_v3.0.pdf ... Pay attention to ..." For normal operation the following constrains MUST be satisfied..." I did NOT check Your value(s) on my S3's, but seems like they are NOT the same "animals"... Cheers ZiG |
|
|
|
|
|
oskuro
|
|
July 26, 2014, 03:18:27 PM |
|
Right now for me 713 W total for 2 S3 Frequencies: 1-->430 GH/s stable, 212,5 Freq, 41/42 temp 2-->439 GH/s stable, 218,5 Freq, 46/43 temp(cant run at 212,5 i get xx and -) I will keep them that way just to test, i dont like the OC its nos stable  |
|
|
|
|
jonnybravo0311
Legendary
Offline
Activity: 1344
Merit: 1023
Mine at Jonny's Pool
|
|
July 26, 2014, 07:55:04 PM |
|
Right now for me 713 W total for 2 S3 Frequencies: 1-->430 GH/s stable, 212,5 Freq, 41/42 temp 2-->439 GH/s stable, 218,5 Freq, 46/43 temp(cant run at 212,5 i get xx and -) I will keep them that way just to test, i dont like the OC its nos stable Huh... weird that under clocking causes you to see "x" and "-". Even weirder that running that same unit at 218.75 is stable. I have 1 unit that just simply won't run stable at stock clocks, but gets a pretty respectable 425-430 at 212.5. Another runs perfectly happily at 218.75 getting 440. Neither will over clock at all. |
Jonny's Pool - Mine with us and help us grow! Support a pool that supports Bitcoin, not a hardware manufacturer's pockets! No SPV cheats. No empty blocks.
|
|
|
|
ZiG
|
|
July 26, 2014, 11:17:51 PM |
|
Right now for me 713 W total for 2 S3 Frequencies: 1-->430 GH/s stable, 212,5 Freq, 41/42 temp 2-->439 GH/s stable, 218,5 Freq, 46/43 temp(cant run at 212,5 i get xx and -) I will keep them that way just to test, i dont like the OC its nos stable Huh... weird that under clocking causes you to see "x" and "-". Even weirder that running that same unit at 218.75 is stable. I have 1 unit that just simply won't run stable at stock clocks, but gets a pretty respectable 425-430 at 212.5. Another runs perfectly happily at 218.75 getting 440. Neither will over clock at all. Hey guys...Please, hook up all your 4 PCI-E connectors... You never know what the PCB designer had in mind...but there is a REASON for these 4 connectors...like traces current(s), parallel (redundant) supply...etc... Mine S3 are running @ 504 / 250MHz ...& @ 475 / 237.5Mhz now... ZiG |
|
|
|
|
jonnybravo0311
Legendary
Offline
Activity: 1344
Merit: 1023
Mine at Jonny's Pool
|
|
July 26, 2014, 11:28:24 PM |
|
Right now for me 713 W total for 2 S3 Frequencies: 1-->430 GH/s stable, 212,5 Freq, 41/42 temp 2-->439 GH/s stable, 218,5 Freq, 46/43 temp(cant run at 212,5 i get xx and -) I will keep them that way just to test, i dont like the OC its nos stable Huh... weird that under clocking causes you to see "x" and "-". Even weirder that running that same unit at 218.75 is stable. I have 1 unit that just simply won't run stable at stock clocks, but gets a pretty respectable 425-430 at 212.5. Another runs perfectly happily at 218.75 getting 440. Neither will over clock at all. Hey guys...Please, hook up all your 4 PCI-E connectors... You never know what the PCB designer had in mind...but there is a REASON for these 4 connectors...like traces current(s), parallel (redundant) supply...etc... Mine S3 are running @ 504 / 250MHz ...& @ 475 / 237.5Mhz now... ZiG I ran my S3s with all 4 connectors and it made no difference. They simply will not handle any kind of higher clock frequencies. One of them doesn't even manage to hash at the stock clocks and has to be run at 212.5. Congrats to you on getting "good" ones that can handle the higher frequencies, though. Wish I were as lucky. |
Jonny's Pool - Mine with us and help us grow! Support a pool that supports Bitcoin, not a hardware manufacturer's pockets! No SPV cheats. No empty blocks.
|
|
|
philipma1957
Legendary
Offline
Activity: 4130
Merit: 7907
'The right to privacy matters'
|
|
July 26, 2014, 11:30:05 PM |
|
Right now for me 713 W total for 2 S3 Frequencies: 1-->430 GH/s stable, 212,5 Freq, 41/42 temp 2-->439 GH/s stable, 218,5 Freq, 46/43 temp(cant run at 212,5 i get xx and -) I will keep them that way just to test, i dont like the OC its nos stable Huh... weird that under clocking causes you to see "x" and "-". Even weirder that running that same unit at 218.75 is stable. I have 1 unit that just simply won't run stable at stock clocks, but gets a pretty respectable 425-430 at 212.5. Another runs perfectly happily at 218.75 getting 440. Neither will over clock at all. Hey guys...Please, hook up all your 4 PCI-E connectors... You never know what the PCB designer had in mind...but there is a REASON for these 4 connectors...like traces current(s), parallel (redundant) supply...etc... Mine S3 are running @ 504 / 250MHz ...& @ 475 / 237.5Mhz now... ZiG I ran my S3s with all 4 connectors and it made no difference. They simply will not handle any kind of higher clock frequencies. One of them doesn't even manage to hash at the stock clocks and has to be run at 212.5. yeah both of mine prefer 212.5 over any other number, so why fight it. they get 430 at 328 watts |
|
|
|
|
WheresWaldo
|
|
July 27, 2014, 07:23:44 AM |
|
generated a table to work with, based on the S1/S2 tables: ...package 'cgminer'
config 'asic-freq' 'default'
#option 'freq_value' '1486' #262.5M
#option 'chip_freq' '262.5'
#option 'timeout' '15'
#option 'freq_value' '1386' #250M
#option 'chip_freq' '250'
#option 'timeout' '16'
#option 'freq_value' '1306' #243.75M
#option 'chip_freq' '243.75'
#option 'timeout' '17'
#option 'freq_value' '1286' #237.5M
#option 'chip_freq' '237.5'
#option 'timeout' '17'
#option 'freq_value' '1206' #231.25M
#option 'chip_freq' '231.5'
#option 'timeout' '18'
#option 'freq_value' '1185' #225M
#option 'chip_freq' '225'
#option 'timeout' '18'
option 'freq_value' '1106' #218.75M
option 'chip_freq' '218.75'
option 'timeout' '18'
#option 'freq_value' '1086' #212.5M
#option 'chip_freq' '212.5'
#option 'timeout' '18'
#option 'freq_value' '0782' #200M
#option 'chip_freq' '200'
#option 'timeout' '20'
first off thanks to you. second can you find a number between 200 and 212.50 ( 206.25 ) comes to mind third can you find something for 215.625 my guess for 206.25 would be (use at your own risk): option 'freq_value' '1006' #206.25M option 'chip_freq' '206.25' option 'timeout' '19' With all my due respect for your time and effort, but this table is NOT very applicable to S3... You can NOT apply the same S1 math/calculations... to S3, because these are using different chips with different PFD (PLL Frequency Divider...I guess), subject to different conditions...due to chips redesign...I guess... BM frequency table - from BM1382 datasheet ...: Work Input timing Clock Freq.(M) Hash rate(G) Reg.value Ideal delay(ms) Recommenddelay(ms) 100 6.30 0783 42.6 38 125 7.88 0983 34.0 31 150 9.45 0b83 28.4 26 175 11.03 0d83 24.3 22 200 12.60 0782 21.3 19 225 14.18 0882 18.9 17 250 15.75 0982 17.0 15 275 17.33 0a82 15.5 14 300 18.90 0b82 14.2 13 325 20.48 0c82 13.1 12 350 22.05 0d82 12.2 11 375 23.63 0e82 11.3 10 400 25.20 08f2 10.6 10 These are the same in S3's "asic-freq" file... Conditions for S3 valid frequency divider's values now are different...check them here...: https://bitmaintech.com/files/download/BM1382_Datasheet_v3.0.pdf ... Pay attention to ..." For normal operation the following constrains MUST be satisfied..." I did NOT check Your value(s) on my S3's, but seems like they are NOT the same "animals"... Cheers ZiG Does this mean I can push a 275 frequency on the antminers that allow? I would like to test and see the power consumption, just want to make sure I don't melt my units |
|
|
|
klondike_bar
Legendary
Offline
Activity: 2128
Merit: 1005
ASIC Wannabe
|
|
July 27, 2014, 12:06:31 PM |
|
Thought i would post this, since the bitmain chart with 25MHz steps keeps being posted instead, and for fine-tuning 25MHz is far to large of a jump in frequency. 218-243MHz is a good range to focus on, upping speed 6MHz at a time, running for 3-6hrs to check for errors, then increasing again generated a table to work with, based on the S1/S2 tables: ...package 'cgminer'
config 'asic-freq' 'default'
#option 'freq_value' '1486' #262.5M
#option 'chip_freq' '262.5'
#option 'timeout' '15'
#option 'freq_value' '1386' #250M
#option 'chip_freq' '250'
#option 'timeout' '16'
#option 'freq_value' '1306' #243.75M
#option 'chip_freq' '243.75'
#option 'timeout' '17'
#option 'freq_value' '1286' #237.5M
#option 'chip_freq' '237.5'
#option 'timeout' '17'
#option 'freq_value' '1206' #231.25M
#option 'chip_freq' '231.5'
#option 'timeout' '18'
#option 'freq_value' '1185' #225M
#option 'chip_freq' '225'
#option 'timeout' '18'
option 'freq_value' '1106' #218.75M
option 'chip_freq' '218.75'
option 'timeout' '18'
#option 'freq_value' '1086' #212.5M
#option 'chip_freq' '212.5'
#option 'timeout' '18'
#option 'freq_value' '0782' #200M
#option 'chip_freq' '200'
#option 'timeout' '20'
|
|
|
|
|
ZiG
|
|
July 27, 2014, 04:38:50 PM |
|
Thought i would post this, since the bitmain chart with 25MHz steps keeps being posted instead, and for fine-tuning 25MHz is far to large of a jump in frequency. 218-243MHz is a good range to focus on, upping speed 6MHz at a time, running for 3-6hrs to check for errors, then increasing again generated a table to work with, based on the S1/S2 tables: ...package 'cgminer'
config 'asic-freq' 'default'
#option 'freq_value' '1486' #262.5M
#option 'chip_freq' '262.5'
#option 'timeout' '15'
#option 'freq_value' '1386' #250M
#option 'chip_freq' '250'
#option 'timeout' '16'
#option 'freq_value' '1306' #243.75M
#option 'chip_freq' '243.75'
#option 'timeout' '17'
#option 'freq_value' '1286' #237.5M
#option 'chip_freq' '237.5'
#option 'timeout' '17'
#option 'freq_value' '1206' #231.25M
#option 'chip_freq' '231.5'
#option 'timeout' '18'
#option 'freq_value' '1185' #225M
#option 'chip_freq' '225'
#option 'timeout' '18'
option 'freq_value' '1106' #218.75M
option 'chip_freq' '218.75'
option 'timeout' '18'
#option 'freq_value' '1086' #212.5M
#option 'chip_freq' '212.5'
#option 'timeout' '18'
#option 'freq_value' '0782' #200M
#option 'chip_freq' '200'
#option 'timeout' '20'
... " 200 12.60 0 782 21.3 19 225 14.18 0 882 18.9 17 250 15.75 0 982 17.0 15 275 17.33 0 a82 15.5 14 " the pattern for 25MHz steps is obvious...in RED ... Still I did not get how the S1/S2 table values are applicable / usable instead of S3 values... Anybody checking them ... 225MHz - "1185" instead of "0882"... 250MHz - "1386" instead of "0982"... ZiG |
|
|
|
|
|
PaintShopPro6
|
|
July 27, 2014, 08:42:17 PM |
|
One of my batch 1 units will not overclock. In fact, sometimes it won't start up at all. I have it taken apart right now. I am running to the store to get some alcohol for cleaning (well, maybe some other kind to drink as well...). Several of the chips have this problem. I know that some people say that thermal paste does not conduct electricity. However, the thermal paste that is used by BitMain has some oil in it which I do believe conducts some electricity (as learned from my S1 experiences). I am hoping a cleanup job will fix this S3.  |
|
|
|
|
|
PaintShopPro6
|
|
July 27, 2014, 09:43:47 PM |
|
For the record, cleaning seems to have fixed my unit. It overclocks to 237.5 Mhz comfortably now whereas it would not OC previously and many times would fail to run at all.
|
|
|
|
|
HerbPean
Legendary
Offline
Activity: 1638
Merit: 1005
|
|
July 27, 2014, 10:18:41 PM |
|
For the record, cleaning seems to have fixed my unit. It overclocks to 237.5 Mhz comfortably now whereas it would not OC previously and many times would fail to run at all.
Thanks for the info !  |
|
|
|
|
|
daddyfatsax
|
|
July 27, 2014, 11:23:30 PM |
|
For the record, cleaning seems to have fixed my unit. It overclocks to 237.5 Mhz comfortably now whereas it would not OC previously and many times would fail to run at all.
I am still curious how they managed to get paste in the holes where the screws go. |
|
|
|
|
|
ZiG
|
|
July 28, 2014, 12:45:36 AM |
|
For the record, cleaning seems to have fixed my unit. It overclocks to 237.5 Mhz comfortably now whereas it would not OC previously and many times would fail to run at all.
I am still curious how they managed to get paste in the holes where the screws go. You don't get it ..."Chinese engineering" ... Just kidding... I don't understand either... |
|
|
|
|
|
