Swedish ASIC miner company kncminer.com

[idee2013]

EUREKA!
I just discovered something.
They don't like to be OVERCOOLED either!

funny i was having the same thought after seeing them cooled and no real change, as well as others mentioning that they ran their unit hot on .90, that it might allow chip to seat better in solder or something like that....

what stats u got phoenix1969?

its too early to tell because it will take time for my averages to come up, but it looks to be the final tweak, like it may end up 275, 280+, 280+
im optimistic, but definitely see an increase, and within 2 mins...
all I did was shut my boxfan off
I was all proud of my 32C temps, lol...they like 55

can confirm that with my saturn but not with my jup. the jup is slower without a big fan.

[1713889344]

1713889344

[1713889344]

1713889344

[1713889344]

1713889344

[eurgbp2011]

2 cores is practically nothing.  That shouldnt cause significant performance issues. That's 1% of cores not working on a die, or a couple of GH/s

but this one is performing 25% worse than normal one
Good One shares 7630336 (99.78%)
This Bad One shares 6264832 (99.64%)

How to find out where the problem is ?

[xyzzy099]

2 cores is practically nothing.  That shouldnt cause significant performance issues. That's 1% of cores not working on a die, or a couple of GH/s

but this one is performing 25% worse than normal one
Good One shares 7630336 (99.78%)
This Bad One shares 6264832 (99.64%)

How to find out where the problem is ?

In your post in the other thread you posted more data which showed a >17% HW error rate, which accounts for a lot of the under-performance.  By contrast, my Saturn has well under 2% HW errors.  I also noticed that ASIC board #3 is getting a lot hotter than the others - ~74C vs ~54C on the others.  That might be where your problem lies.
 

[ElGabo]

EUREKA!
I just discovered something.
They don't like to be OVERCOOLED either!

funny i was having the same thought after seeing them cooled and no real change, as well as others mentioning that they ran their unit hot on .90, that it might allow chip to seat better in solder or something like that....

what stats u got phoenix1969?

its too early to tell because it will take time for my averages to come up, but it looks to be the final tweak, like it may end up 275, 280+, 280+
im optimistic, but definitely see an increase, and within 2 mins...
all I did was shut my boxfan off
I was all proud of my 32C temps, lol...they like 55

can confirm that with my saturn but not with my jup. the jup is slower without a big fan.

I've tried the same, but when the chips went up to 50C I've lost 20Ghs.

So now I try to cool as much as I can...

It gives between 550-560. 560 if I can cool down enough. So what the heck?

Oh yes, the top of the box is on....

[eurgbp2011]

2 cores is practically nothing.  That shouldnt cause significant performance issues. That's 1% of cores not working on a die, or a couple of GH/s

but this one is performing 25% worse than normal one
Good One shares 7630336 (99.78%)
This Bad One shares 6264832 (99.64%)

How to find out where the problem is ?

In your post in the other thread you posted more data which showed a >17% HW error rate, which accounts for a lot of the under-performance.  By contrast, my Saturn has well under 2% HW errors.  I also noticed that ASIC board #3 is getting a lot hotter than the others - ~74C vs ~54C on the others.  That might be where your problem lies.
 

This one is underperforming.
Is this RMA case or I can fix something on my own?

Shares are being counted from exact single moment.

bad one 470.35 GH/s   6320640 (99.64%)   22528 / 24064 / 0

cgminer version 3.4.0 - Started: [2013-10-24 20:50:57]
--------------------------------------------------------------------------------
 (5s):526.5G (avg):514.5Gh/s | A:7174144  R:56320  HW:1933296  WU:7887.7/m
 ST: 2  SS: 0  NB: 190  LW: 8989291  GF: 0  RF: 0
 Connected to ubtcguild.com diff 512 with stratum as user
 Block: 0000c6e08eb048ec...  Diff:268M  Started: [16:06:23]  Best share: 83.3M
--------------------------------------------------------------------------------
 [P]ool management ettings [D]isplay options [Q]uit
 KnC 0:                | 529.6G/514.5Gh/s | A:7174656 R:56320 HW:1933383 WU: 7888.0/m
--------------------------------------------------------------------------------

 [2013-10-25 16:08:22] Accepted 0055cf53 Diff 763/512 KnC 0
 [2013-10-25 16:08:23] Accepted 00007e76 Diff 133K/512 KnC 0
 [2013-10-25 16:08:24] Accepted 006596f9 Diff 645/512 KnC 0
 [2013-10-25 16:08:24] Accepted 001fde5d Diff 2.06K/512 KnC 0
 [2013-10-25 16:08:26] Accepted 0074d20b Diff 560/512 KnC 0
 [2013-10-25 16:08:27] Accepted 004af4d2 Diff 874/512 KnC 0
 [2013-10-25 16:08:28] Accepted 007634ca Diff 554/512 KnC 0



this is the good one

542.43 GH/s   7697920 (99.78%)   16896 / 38400 / 0

 cgminer version 3.4.0 - Started: [2013-10-24 00:42:22]
--------------------------------------------------------------------------------
 (5s):581.3G (avg):548.2Gh/s | A:17935872  R:132096  HW:958395  WU:8023.7/m
 ST: 2  SS: 0  NB: 362  LW: 19464588  GF: 0  RF: 0
 Connected to btcguild.com diff 512 with stratum as user
 Block: 0000c6e08eb048ec...  Diff:268M  Started: [16:06:23]  Best share: 21.8M
--------------------------------------------------------------------------------
 [P]ool management ettings [D]isplay options [Q]uit
 KnC 0:                | 533.4G/548.2Gh/s | A:17936384 R:132096 HW:958411 WU: 8023.9/m
--------------------------------------------------------------------------------

 [2013-10-25 16:09:22] Accepted 0073fba7 Diff 565/512 KnC 0
 [2013-10-25 16:09:25] Accepted 00378c46 Diff 1.18K/512 KnC 0
 [2013-10-25 16:09:25] Accepted 0077b6c6 Diff 547/512 KnC 0
 [2013-10-25 16:09:30] Accepted 006cac73 Diff 603/512 KnC 0
 [2013-10-25 16:09:49] Accepted 000b1fc9 Diff 5.89K/512 KnC 0
 [2013-10-25 16:09:50] Accepted 003f691f Diff 1.03K/512 KnC 0
 [2013-10-25 16:09:53] Accepted 001d375e Diff 2.24K/512 KnC 0
 [2013-10-25 16:09:54] Accepted 002e6e7d Diff 1.41K/512 KnC 0
 [2013-10-25 16:09:57] Accepted 00590299 Diff 736/512 KnC 0
 [2013-10-25 16:10:03] Accepted 00720e3f Diff 574/512 KnC 0
 [2013-10-25 16:10:05] Accepted 000abdf9 Diff 6.1K/512 KnC 0
 [2013-10-25 16:10:05] Accepted 002c0fcb Diff 1.49K/512 KnC 0
 [2013-10-25 16:10:10] Accepted 0070fc78 Diff 580/512 KnC 0
 [2013-10-25 16:10:14] Accepted 006d02a7 Diff 601/512 KnC 0
 [2013-10-25 16:10:14] Accepted 00307a06 Diff 1.35K/512 KnC 0

[soy]

2 cores is practically nothing.  That shouldnt cause significant performance issues. That's 1% of cores not working on a die, or a couple of GH/s

but this one is performing 25% worse than normal one
Good One shares 7630336 (99.78%)
This Bad One shares 6264832 (99.64%)

How to find out where the problem is ?

Find out at what page users started reporting reception of miners and what they did that was or was not successful at improving performance.  (chuckles)

I'd suggest shutting down, then putting light pressure on the tops of the VRMs one at a time.  Don't expect to put enough pressure to get a snap sound.  Then boot and install firmware 0.90, run enablecores.bin, shutdown, wait 30 seconds, startup.  Open a new browser, go to your webpage address, sign in, hit check status, wait a few seconds, give it a control+F5, the browser refreshes, at this point sometimes it will only show 2 or 3 cores.  Close IE, open a new instance of IE, repeat the process, you should now see all cores and at what percentage they're running.  If there's an improvement you can move up in firmware versions or stay with 0.90.  Anything over 0.94 will use less energy.  Then decide if the increased power usage warrants the additional GH/s and go with the firmware you feel is best.

 going back to 0.90, running enablecores.bin,

[sbfree]

seems like you box has same 20% HW error rate as my saturn....it is firmware related i believe, then again, i don't know diddly.

[eurgbp2011]

How fast will KFCMiner do RMA?

[eurgbp2011]

2 cores is practically nothing.  That shouldnt cause significant performance issues. That's 1% of cores not working on a die, or a couple of GH/s

but this one is performing 25% worse than normal one
Good One shares 7630336 (99.78%)
This Bad One shares 6264832 (99.64%)

How to find out where the problem is ?

Find out at what page users started reporting reception of miners and what they did that was or was not successful at improving performance.  (chuckles)

I'd suggest shutting down, then putting light pressure on the tops of the VRMs one at a time.  Don't expect to put enough pressure to get a snap sound.  Then boot and install firmware 0.90, run enablecores.bin, shutdown, wait 30 seconds, startup.  Open a new browser, go to your webpage address, sign in, hit check status, wait a few seconds, give it a control+F5, the browser refreshes, at this point sometimes it will only show 2 or 3 cores.  Close IE, open a new instance of IE, repeat the process, you should now see all cores and at what percentage they're running.  If there's an improvement you can move up in firmware versions or stay with 0.90.  Anything over 0.94 will use less energy.  Then decide if the increased power usage warrants the additional GH/s and go with the firmware you feel is best.

 going back to 0.90, running enablecores.bin,

well then I'll try your idea ....
Back to the Bunker!

[soy]

seems like you box has same 20% HW error rate as my saturn....it is firmware related i believe, then again, i don't know diddly.

I was getting >50% HW errors with firmwares over 0.94.  Moving down and enabling cores, using the higher voltage I get 4.5% (my maht hw*100/accepted).

[canth]

How fast will KFCMiner do RMA?

However long it takes to ship to them + 1 day of testing  + shipping back. Realistically, it's around 4-5 days.

[Biffa]

Crap,  everyone is reporting their chips running in the 45 degree range.  I have one at 48, two at 55 and one at 64.  I'm hashing at a pretty stable rate of around 540-550.  Should I be concerned about the temps?

No, mines the same. People are chasing ghosts around here.

[eurgbp2011]

Crap,  everyone is reporting their chips running in the 45 degree range.  I have one at 48, two at 55 and one at 64.  I'm hashing at a pretty stable rate of around 540-550.  Should I be concerned about the temps?

No, mines the same. People are chasing ghosts around here.

73C -75C here

[eurgbp2011]

How fast will KFCMiner do RMA?

However long it takes to ship to them + 1 day of testing  + shipping back. Realistically, it's around 4-5 days.

you unscrew & peel off fan from CPU and send board?

[soy]

somebody doesn't like me helping competition - i post and pool stops responding, distributed denial of service attack?

[Biffa]

somebody doesn't like me helping competition - i post and pool stops responding, distributed denial of service attack?

Slush? Yes, my Jupiter just switched to my backup pool

[eurgbp2011]

seems like you box has same 20% HW error rate as my saturn....it is firmware related i believe, then again, i don't know diddly.

Mining Status

CGMiner Status Running (pid=11787)
Last Checked Fri Oct 25 16:37:22 UTC 2013
Avg. Hash Rate 507 Gh/s
Difficulty Accepted 16,896
Difficulty Rejected 0 (0.0 %)
Hardware Errors 6,792 (20.8 %)
WU 8,131
Difficulty Stale 0
Network Blocks 1
Pool Rejected 0.0 %
Pool Stale 0
Best Share 95,956
Found Blocks 0


HW Status

ASIC slot #1 48.0 ℃
ASIC slot #2 -
ASIC slot #3 60.5 ℃
ASIC slot #4 51.0 ℃
ASIC slot #5 49.5 ℃
ASIC slot #6 -

[mwarrior]

How fast will KFCMiner do RMA?

However long it takes to ship to them + 1 day of testing  + shipping back. Realistically, it's around 4-5 days.

Good luck with that.  I am a day 1 order 454 with a unit I can't use.  I can't get in touch with KNC.. have left messages on their voicemail.  Nothing.  I have been able to hash for 4 days with a day 1 order that showed up on day 4.  

[RoadStress]

Firmware 0.97 is better than 0.96 which is better than 0.95 and so on. Good job KnC!

Hmm i praised the 0.97 firmware, but now i have to change my statement. For me the 0.97 firmware had better speed in cgminer, but in the pool i had lower speed. Changed it back cu 0.96 and now in the pool i get close results as in cgminer. So be sure to check both.

[canth]

How fast will KFCMiner do RMA?

However long it takes to ship to them + 1 day of testing  + shipping back. Realistically, it's around 4-5 days.

you unscrew & peel off fan from CPU and send board?

Yep. Torx #15 (IIRC) + philips #1 and then the heatsink comes right off the chip. Send in the entire chip/board w/o the heatsink in a static bag if you have one.