Cleaned Antminer S3+ thermal paste, now -------?

[_Crash_]

Hi,

I've just cleaned the right side of the Antminer S3+ (as if you was looking at it with the Ethernet port at the back) by blowing to get the dust out. I took off the metal heatsink and cleaned up the paste a little and reapplied some of my own (not too much at all). I made sure that none was overflowing onto the other components and then reapplied the heatsink and then turned on the miner. Throughout this process I was grounding myself by touching my PSU's metal casing, before, during and after this.

When it was on, I noticed my ASIC status showing ------- in the bottom left. I noticed a small RED LED illuminated (RF1) on the miner on the opposite side from where I was working on (the left side) as you can see In the photos below. I've tried restarting the miner, but still the same issue. I never had this issue before.








Any ideas?

Thanks,
Owen.

[philipma1957]

turn the psu off. unplug the psu for 2 minutes.

plug the psu in turn it on.

wait 10 minutes take a screen shot of it let us see the screen shot.


If you have the ---- chips at this point let me see the screen shot with at least 10 minutes run time. not 22 seconds.

[_Crash_]

Hi,

I let it run for nearly 20 minutes, and this is what I got;

[Sweminer777]

more voltage, less voltage.

[_Crash_]

more voltage, less voltage.

What do you mean?

[dimke_yu]

He means try with different voltage settings for chips in dashboard. Also try with lower clock settings. Did you tried to connect 4 PCI-e power cables or 2?

[_Crash_]

I connected2x PCI-E connectors and 4x. I also tried setting the voltage from what I had it at (0750) to 0000 (default). I did a reset also, but still the same issue.

[philipma1957]

Some how the cleaning you did broke something.  What thermal compound did you use?

Some are non conductive   but some are conductive.  Maybe  you used a conductive compound and shorted something.

If you do not know if the thermal compound is conductive  (Arctic Silver is conductive)  you can run a line of it on a piece of paper and put an ohm meter to

it.

If it is conductive replace it.

Here are some non conductive compounds

http://www.overclock.net/t/1371721/best-thermal-compound-thats-non-conductive

[_Crash_]

Well I was thinking about using alcohol wipes and cleaning everything around the chip. Leaving it for a good half hour and seeing if it fixes the issue. What do you think?

[_Crash_]

OK well that didn't work. It just took out the other row somehow. I'm thinking this has been really poorly built. The PCB's I work with are so much more sturdier than this. The IC's have had shorts all across them and yet they still work when you take the shorts out! Yet with an Antminer, you touch one and the whole fu*** thing breaks.

[notlist3d]

Well I was thinking about using alcohol wipes and cleaning everything around the chip. Leaving it for a good half hour and seeing if it fixes the issue. What do you think?

I may have went overboard on not being needed.  But I am going to be redoing thermal paste on a C1.  I got some remover and purifier.

It is a little extra time and a tad more money but figure I might as well do it right.  Also was pretty cheap on amazon.

[_Crash_]

Ok, I've just been on bitmain support yesterday all night.

The RED L.E.D that is lit up called "RF1" indicates that the hashboard is working. So the hash board in my case that isn't working, the L.E.D isn't lit up. There is a chance that if you clean the thermal paste off and make sure there are no short circuits, then the board will begin to operate normally, however this is only a chance. In my case, this didn't work, however I still think there is some shorts on a very small scale that I can't see.

[notlist3d]

Ok, I've just been on bitmain support yesterday all night.

The RED L.E.D that is lit up called "RF1" indicates that the hashboard is working. So the hash board in my case that isn't working, the L.E.D isn't lit up. There is a chance that if you clean the thermal paste off and make sure there are no short circuits, then the board will begin to operate normally, however this is only a chance. In my case, this didn't work, however I still think there is some shorts on a very small scale that I can't see.

I have been working on a C1.  I used cleaner and purifier and it still did not do the trick on some.  1 blade works perfect after it.   

I think I'm sitting it in parts unless I decide to give other 3 one more try.    I think from doing it 3 are good blades and one has one or two bad chips possibly if I could get them all to work right.

I have figured out it is a time consuming process to do it right to say the least.

[_Crash_]

Ok, I've just been on bitmain support yesterday all night.

The RED L.E.D that is lit up called "RF1" indicates that the hashboard is working. So the hash board in my case that isn't working, the L.E.D isn't lit up. There is a chance that if you clean the thermal paste off and make sure there are no short circuits, then the board will begin to operate normally, however this is only a chance. In my case, this didn't work, however I still think there is some shorts on a very small scale that I can't see.

I have been working on a C1.  I used cleaner and purifier and it still did not do the trick on some.  1 blade works perfect after it.   

I think I'm sitting it in parts unless I decide to give other 3 one more try.    I think from doing it 3 are good blades and one has one or two bad chips possibly if I could get them all to work right.

I have figured out it is a time consuming process to do it right to say the least.

Is that in terms of the ASIC status showing "-" for those chips you was unsuccessful with?

[notlist3d]

Ok, I've just been on bitmain support yesterday all night.

The RED L.E.D that is lit up called "RF1" indicates that the hashboard is working. So the hash board in my case that isn't working, the L.E.D isn't lit up. There is a chance that if you clean the thermal paste off and make sure there are no short circuits, then the board will begin to operate normally, however this is only a chance. In my case, this didn't work, however I still think there is some shorts on a very small scale that I can't see.

I have been working on a C1.  I used cleaner and purifier and it still did not do the trick on some.  1 blade works perfect after it.   

I think I'm sitting it in parts unless I decide to give other 3 one more try.    I think from doing it 3 are good blades and one has one or two bad chips possibly if I could get them all to work right.

I have figured out it is a time consuming process to do it right to say the least.

Is that in terms of the ASIC status showing "-" for those chips you was unsuccessful with?

Well the one that had that now shows only 8 chips and no "-".   It's been a really hit and miss project.   

And with C1 is a huge pain draining liquid to try again.

[_Crash_]

I spoke to Bitmain support and ordered a new Hash PCB for my S3+. Bought with Bitcoin and it came within a week (to the UK). They are really helpful and answer all your questions. Glad that it came today and installed it successfully - It's running fine now. They sent one of the newest hash boards out (V1.4 - not sure if it's the latest, but I think it is).

[dimke_yu]

I spoke to Bitmain support and ordered a new Hash PCB for my S3+. Bought with Bitcoin and it came within a week (to the UK). They are really helpful and answer all your questions. Glad that it came today and installed it successfully - It's running fine now. They sent one of the newest hash boards out (V1.4 - not sure if it's the latest, but I think it is).

I am glad that your device is hashing full speed now 

[goxed]

(Arctic Silver is conductive)  you can run a line of it on a piece of paper and put an ohm meter to


http://www.overclock.net/t/1371721/best-thermal-compound-thats-non-conductive

Are the new AS compounds conductive? because it used to be more or less non conductive.