Topic: ANTMINER S3+ Discussion and Support Thread - page 190. (Read 710164 times)

realy? chain 2 is very bad?

but temp is ok

i had to turn my miners off for a bit and when i turned them back on cgminer says
This section contains no values yet

on both miners..

i updated both to the new cgminer..

would anything special need to be done to start them back up?

i do have the freq set to s3 default, so its not that..

edit, it seems to be a DNS issue. keyed googles dns in manually and its fixed now..

Don't be surprised if it lifts off the ground and flies out the window.

hello

fan speed is normal? s3.

If you aren't too concerned about noise or a little extra power, you could do the "blue wire mod" - i.e. clip it - and it'll run full speed all the time. 

Missed my B8 S3+ delivery today but I'm picking it up from UPS in an hour,  so stoked!  Can't believe they sent these out 2nd Day Air from China at their expense, How Sweet!   Bitmain you rule.... well except for your preorder on the L1.

I don't get why they don't have both fans run the same speed as dictated by the exhaust temperature.  To have the intake fan run slower because it's cooler means, I'm guessing here, less efficient air flow.  If the slower fan is getting help from the draw of the faster fan it's just a bigger load on the faster fan.

when I added secondary fans to my s1 I think the temps actually went up... the fans in the back couldn't spin fast enough

I agree... I used to have extra exhaust fans on my s1's til I realized how pointless that was and just wasting more power. I don't know why people waste so many hours re-applying thermal paste,etc.. for a few ghs and also the wasted mining from downtime.

I do think there is nothing wrong with oc'ing, upgrading firmware, and updating cgminer. Those things are really easy and don't require too much tinkering. Also with the updated cgminer, the webui is much more responsive due to reduced load.

This constant angst over a few GHs is so pointless. Just let these things run and quit wasting your time and energy trying to squeeze out (literally) a couple of extra pennies. Downtime from messing with settings and firmware and reapplying thermal paste is doing more harm than good to your earnings.

So, oc'd 2 b8's for 5hrs and 1 is running pretty stable at 503gh, the other keeps falling to around 480-490 and showing 1 x. It goes away after restarting, but comes back after a few hours. Might try lowering it's clock a lil to see how it runs. Bought to upgrade cgminer on both.

Discarded shares don't mean anything on the S3. Ckolivas has confirmed this himself.

im getting a lot of discarded shares with my new s3s as well. people say they don't matter, but it seems to make a difference...

the s1 seem way more consistent.

one of my s3 bounces around 460hg to 400gh with lots of discarded... 2x more discarded than accepted...

other one stays steady at 450-460gh but with 2x more discarded than accepted shares...

Before you do all that "-" is not a dead a ASIC "x" is and even those can come back (I've seen it) power off the supply try moving your PCIE cables around. I assume your only using 2 one on each board, if your using 4 unplug 2 (if no plans to OC) and then power the supply back on. Reboots alone don't always fix S3's. I agree discards don't matter.

Discards do not matter. CK said they mean nothing and should never have been reported and tracked in the first place. The "-" you see is a malfunctioning ASIC. It should still be under warranty and BitMain can help you fix it if the chips is bad. I got someone good when I had a under performing board and they replaced it.

5 hours in... hashing is slowly dropping. at 491gh/s now. 495 is the the set average for old cgminer.
yall sure I dont need to install the newest firmware?

My upgrade kit order (48 kits) is "in shipping status" as of late last night (Eastern US timezone for me), waiting for the tracking number.  I expect them to arrive by early next week if they follow the same timing as my B5's.

Edit: for reference, my order was placed 8/14.
Edit 2: also for reference, we went through 112bit.com, not through bitmaintech directly, but I don't think that's really much of a difference.

Where are we at with order deliveries guys? Anyone shipped who ordered at end of August?

Also, I haven't noticed any posts from people about the upgrade kits so assuming these are being posted last

hi
my antminer s3 at stock settings started to low performing

freq
218.75
   
temps
37
41

asic status   
oooooooo oooooooo
oooooooo ooo-oooo

so one chip is dead?
reboot does not solve this...


Elapsed   GH/S(5s)   GH/S(avg)   FoundBlocks   Getworks   Accepted   Rejected   HW   Utility   Discarded   Stale   LocalWork   WU   DiffA   DiffR   DiffS   BestShare
                                                
5h 40m 10s
406.93
428.80
0
429
15,852
5
42
46
45,253
0
3,601,572
6,084
2,029,025
639
0
2,530,253

that a lot of discarded...
is it normal ?


there is polmine raport
https://polmine.pl/getgraphdatasrv.php?gid=BTC_worker5567092517e0ed42d94cf318195e0737

The problem S3 which I changed out the thermal paste was down about 430 this morning.  I had moved both into the back room last night as after a warm day it got uncomfortable in my living room.  So, I restarted the problem S3 this morning and it has been getting warm in there, 91° 31%rh.  It has been running well since rebooting, the warmer room and low humidity.  6hr50min 440.91GH/s, 0 HW, but utility is 15, fewer shares tho the score seem equal to the other.  It's temps read 47 & 46 while the #1 unit is reading 44 & 42, running at 441.60 HW 160 in 15 hours but its utility is 16.

Having learned the art of building heatsinks for high voltage FETS in the late 1970's, what you want for perfection is two perfect surfaces in contact but since perfection isn't possible you want as little thermal paste as possible filling any air gaps between the surfaces.

When I took the inner heatsink off a hashing board, a good board by the way, the thermal paste seemed applied by machine in 4 rows along the top ASICs and 4 more rows along the bottom ASICs.  It was not well spread, just flattened.  There were air gaps between the rows.  Thermal paste conducts heat less well than the copper on the board or the aluminum.  So, essentially the heat went from the copper plane to the rows of thermal paste to the aluminum on the other side of the thermal paste.  I removed it all and applied an even thin layer of thermal paste to the surface and tightened from the center out.  When doing the outer heatsink I suggest you carefully remove it after removing the 4 spring'd screws but not to slide the heatsink while removing.  If you lift it up and off you won't build up thermal compound on the very fine ASIC pins as might happen if you slide the heatsink to the side.  I'm sure the stuff doesn't conduct but dust might adhere to it at the pins over time.  Big mistake is thinking more paste more conduction.

Also, as you remove the silver thermal paste from the inner heatsink, you may notice a couple or more small lumps of solid thermal compound stuck to the heatsink.  At least that was the case of the bad hashing board that had new paste applied in addition to the original.  No way that heatsink could have been flush, it stood away from the copper plane at least the depth of the hard lumps.