Antminer S5 - HW Errors and ASIC (x)'s

[icemanncsu]

Hey guys I received my new Antminer S5 last night and completed the setup and wanted to report back to you all to see if these numbers are ok.

My PSU is a new EVGA 120-G2-1000-XR 80 PLUS GOLD 1000 W.  http://www.newegg.com/Product/Product.aspx?Item=N82E16817438010

PSU is attached to the Antminer S5 via two PCI-e cables, one per board, each cable with three connectors (PSU > Board1 Port1 > Board1 Port2 and PSU > Board2 Port1 > Board2 Port2).

I ran the box for about 6 hours over night and took two screen grabs, one after 3 hours of uptime and another at 6 hours.  I am concerned with the the number of HW errors that are occuring and the fact that there is a single ASIC (x) being reported on one of the boards.  The single ASIC (x) does not appear immediately, it takes about 5-10 mins of uptime before it changes from (o) to (x).

I am attaching two status screenshots to this message, can you please review them and give me your opinion as to what I should do.  Only thing I can think of is that maybe the PSU is not able to deliver the amount of current needed on those two PCI-e (VGA 1 and 2) ports.  There are 6 PCI-e (VGA 1-6) ports on the PSU.  I am thinking of using four of them and plugging in the Antminer S5 boards power inputs to one each.  Since I don't know the circuit internal on the PSU, the spec says that it's a +12V single rail and this should be fine and cause no harm.  Any thoughts?

As for the environmentals, this was running in my living room with an ambient temperature of 73 degrees fahrenheit.  The few times I viewed the status over night, it appeared that the boards peaked into the low 50's degrees celsius.  I am using the original stock fan for pushing and a NF-F12 in a pull configuration (running at full speed, wired direct to the PSU).

Thanks a bunch!

http://s24.postimg.org/9zojk5nqt/imageedit_1_7260019291.jpg

http://s24.postimg.org/adpvjr7ud/screenshot_10_0_20_60_2015_07_24_03_20_58.png

[-droid-]

so you have two cables coming from PSU that have 2 PCIE connectors on the end of each cable?

if you have more, try to plug all 4 PCIE connectors with individual cables, reboot the s5 by power cycling your PSU and let us know how it goes

[icemanncsu]

so you have two cables coming from PSU that have 2 PCIE connectors on the end of each cable?

if you have more, try to plug all 4 PCIE connectors with individual cables, reboot the s5 by power cycling your PSU and let us know how it goes

I made the cable change as you described and the the results are the same.  HW Error count is increasing to 50 already at 2mins of uptime.  The (x) on the ASIC status is the same as before.

Conclusion: No change


update:

As you can see in the image below the (x) changed back to (o) at sometime after the S5 was up for a while, but the HardWare Errors (HW) are quite high!?  What do I need to do?

http://s22.postimg.org/5akk8nohd/Screen_Shot_2015_07_24_at_6.jpg

[Philipma1957onvacation]

Drop freq try

350.00 I know you are here. So go lower.

343.75
337.50
331.25
325.00

[gablay12]

Why do you recommend frequency drop?Do you think a rail of PSU has been damaged?

[icemanncsu]

I have tried two brand new PSU's and the same thing happened.

Now, the entire board on that chain (#2) is only showing 23 ASIC cores and after 10 minutes of runtime its complete (x)'s.

I guess that board is toast.

[gablay12]

I also recommend to change/try cables.

[Nexious]

Have you tried dropping the frequency and different cables/psu?

[philipma1957]

Drop freq try

350.00 I know you are here. So go lower.

343.75
337.50
331.25
325.00

you are not doing this yet..   and as I read this again.  you are using the cables wrong…   use 4 separate cables  not 2 cables with 2 plugs each.