well no matter what i try, even putting the s3 in front of an AC, the hash rate drops to about 400 after a few hours, not matter the cooling, the cleaning and re paste, heat sinks on the dc-dc and freq i use.
i need to restart it, usually restarting cgminer gives me back my top hash about about 30mins, than it drops slowly,
anyone know how i can kill and restart cgminer every X amount of minutes. ?
Topic: BITMAIN Antminer S3 support and OverClocking thread - page 39. (Read 158132 times)
You're right! Sorry. I was looking at the paper catalog.
On the PDF, the 220W/mk one, is on page 78!
I just think, the PAD would be easyer then the paste to apply.
OK, I checked, they are aluminium foil with extremely thin layer of heat compound on both sides, that could explain the high w/mk value.On the PDF, the 220W/mk one, is on page 78!
I just think, the PAD would be easyer then the paste to apply.
If you can have them for free or almost, why not.
I'm curious to see if it really improves the cooling of the chip.
You're right! Sorry. I was looking at the paper catalog.
On the PDF, the 220W/mk one, is on page 78!
I just think, the PAD would be easyer then the paste to apply.
On the PDF, the 220W/mk one, is on page 78!
I just think, the PAD would be easyer then the paste to apply.
What about using Thermal-PAD's instead of Thermal-paste ??
Check this PDF catalog: http://www.heatmanagement.com/local/media/pdfs/Kunze_ProductCatalogue-2014.pdf
I was looking the pad on page 80 !! 220w/mk !
There are even bether ones on page 94 mage with carbon !
What y'all think about ?
The one on page 80 is rated at 3w/mK
Some phobya pads are rated 7w/mK
And phobya nano grease extreme is 16w/mk.
When looking at heat pastes tests, you will also notice that at a point, it will not be cost efficient at all. Almost no gain with price doubled or tripled.
Good pastes in a 7-12w/mk range will be more than enough for our use in asics.
As a side note, not many materials will give you 200w/mK or over when mixed to form a paste/pad.
Maybe diamond that is 1000w/mK, but t the paste would cost way too much
Anyway, if it ever exists, anything over 200 will be useless since aluminium (heatsink material) is around 220w/wK.
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.
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.
What about using Thermal-PAD's instead of Thermal-paste ??
Check this PDF catalog: http://www.heatmanagement.com/local/media/pdfs/Kunze_ProductCatalogue-2014.pdf
I was looking the pad on page 80 !! 220w/mk !
There are even bether ones on page 94 mage with carbon !
What y'all think about ?
Whats the frequency that if you OC you need to connect 4 PCIE?? I mean, more than 218,75? 225?
And if you dont OC, its recommended to have only 2 PCIE connected?? Or if you have 4 it doesnt matter?
And if you dont OC, its recommended to have only 2 PCIE connected?? Or if you have 4 it doesnt matter?
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.
('too much of a good thing'.....
Cheers
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...
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.
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 !
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.
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.
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.
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.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
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
yeah both of mine prefer 212.5 over any other number, so why fight it. they get 430 at 328 watts
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.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
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
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.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
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
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. 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
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
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
Jump to: