|
chalkboard17
|
 |
May 30, 2015, 01:57:07 PM |
|
new miner, my chains are 11 degrees apart in temp.  any solution to this? my other s5 is only 3-4 degrees apart. very same thing happened to me: https://bitcointalk.org/index.php?topic=1063402.0 (i removed the pictures but right now my miner is running at 53/45 - nearly winter in southern hemisphere + air conditioned room) found no solution to it i also had ~12c difference but i switched the side covers place with each other + put it over something with no hole. if u are placing it on something with holes under it to "let it breath" is actually worse. i put it over the foam that comes with the box to protect it so it doesnt shake a little bit the table. after doing these two things i now have avg temp diff ~9c. mine is also new model with heatsink. brand new 1300 g2 powering only one miner i will eventually put another fan in it to help with the noise + temperature i think. might eventually repaste thermal paste too |
|
|
|
|
gaalx
|
|
May 30, 2015, 03:38:08 PM |
|
Periodically drops Hashing Speed. It helps to restart. Details in the image. How to cure this problem?  Sorry, off topic. where are you getting these graphs from? i installed the nicehash firmware by smit and dont see the graphs eveyone was talking about.. |
|
|
|
|
gaalx
|
|
May 30, 2015, 03:39:53 PM |
|
Related to my post about the new miner being different from the older one, even though they are from the same batch. Here's the comparison of the heatsink on the old one (left) and the new one (right):
As a result the new one is always 3-5°C cooler than the old one.
where you bought the radiator? |
|
|
|
|
Crypto84
|
|
May 30, 2015, 04:21:09 PM
Last edit: May 30, 2015, 04:55:59 PM by Crypto84 |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
|
|
|
|
|
notlist3d
Legendary
 Offline
Activity: 1456
Merit: 1000
|
|
May 30, 2015, 05:15:52 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. |
|
|
|
|
|
Crypto84
|
|
May 30, 2015, 05:17:36 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. One blad would be 55 and the other would be 67.... |
|
|
|
|
notlist3d
Legendary
Offline
Activity: 1456
Merit: 1000
|
|
May 30, 2015, 05:21:03 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. One blad would be 55 and the other would be 67.... Ohh ouch that is not what I was expecting. Did you just get this miner or have you been mining a little bit? If just got it you might put in ticket with bitmain support about it and see what they say. With it being 67 I think they might give you opportunity to redo thermal paste on the higher blade. (I would ask them first as I'm assuming it's under warranty and don't want you to possibly lose that). |
|
|
|
|
|
Crypto84
|
|
May 30, 2015, 05:22:49 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. One blad would be 55 and the other would be 67.... Ohh ouch that is not what I was expecting. Did you just get this miner or have you been mining a little bit? If just got it you might put in ticket with bitmain support about it and see what they say. With it being 67 I think they might give you opportunity to redo thermal paste on the higher blade. (I would ask them first as I'm assuming it's under warranty and don't want you to possibly lose that). They already gave me the go ahead to redo the thermal paste. just seeing what i can do before i go that far. I think trying to power these 2 S5's with one power supply was a bad idea from the get go. they both just dropped to 700gh/s and 800gh/s. might have to split these up. im runnning out of ideas. |
|
|
|
|
notlist3d
Legendary
Offline
Activity: 1456
Merit: 1000
|
|
May 30, 2015, 05:29:17 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. One blad would be 55 and the other would be 67.... Ohh ouch that is not what I was expecting. Did you just get this miner or have you been mining a little bit? If just got it you might put in ticket with bitmain support about it and see what they say. With it being 67 I think they might give you opportunity to redo thermal paste on the higher blade. (I would ask them first as I'm assuming it's under warranty and don't want you to possibly lose that). They already gave me the go ahead to redo the thermal paste. just seeing what i can do before i go that far. I think trying to power these 2 S5's with one power supply was a bad idea from the get go. they both just dropped to 700gh/s and 800gh/s. might have to split these up. im runnning out of ideas. That sounds like possibly powering one blade on each. Which is weird as your 1600 should be able to power both easily. Thermal paste is messy and takes some time, but I have a feeling that is where it's heading. The good news is there is a decent shot that will be the fix. Bad thing is it's a pain and you need to buy a bit of thermal paste. |
|
|
|
|
|
gaalx
|
|
May 30, 2015, 05:58:51 PM |
|
Show a picture of how they are. s5 significant impact on the placement.
|
|
|
|
|
Crypto84
|
|
May 30, 2015, 06:02:49 PM |
|
This is really annoying. im on my second power supply. this power supply is 1600w titanium supernova from evga. all blades are directly plugged into the power supply. none are being powered by the second plug on a cord. they are all 1:1 (one plug/one connector. they both drop in hashrate by half intermittently (one of the blades drop in temp during this time to 35c). ive tried the latest firmware. one of the blades is running 12c hotter than the other. i thought since this was the 5th batch that I wouldnt have so many problems. guess i was dreaming.
edit: the blades be so far apart in temps is a cooling issue i think. put a box fan in front of it and now they are only 3c apart in temps
When you say 12 degrees hotter what temp is it getting? Is it 12 above 35 IE 47? If it is 47 that is not a bad temp on S5's. They do tend to run hot. As far as 12 difference It could be a bad temp sensor, airflow, hard to say. But main thing is you have one accurate one to stop it if it get's to hot which it sounds like you do. The hotter one if it's around 50 instead of 37 I would guess is the one closer to right on temp. One blad would be 55 and the other would be 67.... Ohh ouch that is not what I was expecting. Did you just get this miner or have you been mining a little bit? If just got it you might put in ticket with bitmain support about it and see what they say. With it being 67 I think they might give you opportunity to redo thermal paste on the higher blade. (I would ask them first as I'm assuming it's under warranty and don't want you to possibly lose that). They already gave me the go ahead to redo the thermal paste. just seeing what i can do before i go that far. I think trying to power these 2 S5's with one power supply was a bad idea from the get go. they both just dropped to 700gh/s and 800gh/s. might have to split these up. im runnning out of ideas. That sounds like possibly powering one blade on each. Which is weird as your 1600 should be able to power both easily. Thermal paste is messy and takes some time, but I have a feeling that is where it's heading. The good news is there is a decent shot that will be the fix. Bad thing is it's a pain and you need to buy a bit of thermal paste. So i have 1 S5 plugged into a 1300w supernova power supply and the other plugged into a 1600w supernova S5 plugged into 1300w supernova: pulls 603w S5 plugged into 1600 titanium rated superova: pulls 585w box fan at full speed: 150w using a titanium rated power supply is sooo not worth it. only saving me 18w? lol |
|
|
|
|
notlist3d
Legendary
Offline
Activity: 1456
Merit: 1000
|
|
May 30, 2015, 06:46:01 PM |
|
*made quote smaller so post uses less room. So i have 1 S5 plugged into a 1300w supernova power supply and the other plugged into a 1600w supernova
S5 plugged into 1300w supernova: pulls 603w
S5 plugged into 1600 titanium rated superova: pulls 585w
box fan at full speed: 150w
using a titanium rated power supply is sooo not worth it. only saving me 18w? lol
Part of it is your using a quality PSU all the EVGA supernova's are high quality. If you compared it to a cheap brand I would guess difference would be bigger. Not all brands are equal on quality. I am surprised it was only 18 watts though. I would have not bet on that being difference. |
|
|
|
|
alh
Legendary
Offline
Activity: 1844
Merit: 1050
|
|
May 30, 2015, 07:37:28 PM |
|
*made quote smaller so post uses less room. So i have 1 S5 plugged into a 1300w supernova power supply and the other plugged into a 1600w supernova
S5 plugged into 1300w supernova: pulls 603w
S5 plugged into 1600 titanium rated superova: pulls 585w
box fan at full speed: 150w
using a titanium rated power supply is sooo not worth it. only saving me 18w? lol
Part of it is your using a quality PSU all the EVGA supernova's are high quality. If you compared it to a cheap brand I would guess difference would be bigger. Not all brands are equal on quality. I am surprised it was only 18 watts though. I would have not bet on that being difference. I don't think anybody should be surprised by an 18W difference at a power draw of 600W. That's about a 3% difference. The differences between Gold/Platinum/Tinanium/etc aren't really all that large. The basic Wikipedia entry shows only a 4% difference: See: http://en.wikipedia.org/wiki/80_PlusSo if you have a high quality Gold that beats it's numbers by a bit, then you have a 3% difference. The big jump was from Bronze to gold. When you are at 92% efficiency, there isn't a lot left. One other small item. There isn't anything that says the two S5's are identical in terms of their actual power draw.If you switched the S5's between the power supplies, you might a larger or smaller difference. I think Crypto is in general correct that the differences between an EVGA Gold and an EVGA <better> are probably not worth the price, strictly in terms of efficiency. There may be other improvement such as cabling an wire thickness, though I don't really know. I wonder if PSU manufacturers test and rate and grade their power supplies like an ASIC vendor would. Might a Gold power supply be a Titanium that didn't quite make the grade for Titanium after it's manufactured? Wouldn't be hard to put a different label on it at the end. |
|
|
|
|
|
|
aarons6
Legendary
Offline
Activity: 1736
Merit: 1006
|
|
May 30, 2015, 08:17:56 PM |
|
yes i put one of these fans on as second fan in pull. its about the same as the original fan on loudness and air flow.. |
|
|
|
|
|
marvykkio
|
|
May 30, 2015, 08:35:45 PM |
|
yes i put one of these fans on as second fan in pull. its about the same as the original fan on loudness and air flow.. I substituted with the original, but they will not start |
|
|
|
|
aarons6
Legendary
Offline
Activity: 1736
Merit: 1006
|
|
May 30, 2015, 09:06:19 PM |
|
yes i put one of these fans on as second fan in pull. its about the same as the original fan on loudness and air flow.. I substituted with the original, but they will not start you didnt try to plug in 8 amps of fans into the s5 controller? |
|
|
|
|
|
dimke_yu
|
|
May 31, 2015, 10:54:43 AM |
|
I think he did... and as someone responded to him in another thread that fan need to be connected directly to the molex connector on the PSU not to the control PCB.
|
|
|
|
|
TrevorS
|
|
May 31, 2015, 03:26:39 PM |
|
@Bitmain
Is there any information as to what adhesive heat sinks would be appropriate to install on the S5 outer blades to help reduce temperature? It's getting to Summer and if you can share an effective modification, that would be great. I yesterday further under-clocked my six already under-clocked S5s from 312MHz to 300MHz so as to stay under 65C.
Thanks -- Trevor
|
|
|
|
|
Crypto84
|
|
May 31, 2015, 03:36:55 PM |
|
@Bitmain
Is there any information as to what adhesive heat sinks would be appropriate to install on the S5 outer blades to help reduce temperature? It's getting to summer and if you can share an effective modification, that would be great. I yesterday further under-clocked my six already under-clocked S5s to 300MHz from 312.5MHz in order to stay under 65C.
I feel your pain, i had mine inside in my second bedroom (right next to the one i sleep in) and my gf wasnt too pleased. so i had to put them in the garage. which gets REALLY hot. if bitmain enclosed the damn miner on all 4 sides air would flow better and I really wish they chose a fan that was more "home miner" friendly in terms of noise they could increase the size of the miner to accommodate the bigger fan and it would decrease noise. thats why they are in the garage gf couldnt sleep with the noise. i only have 2 S5's too. |
|
|
|
|
|
