Pages:
Author

Topic: Gigabyte GTX 1070 ITX - Problems? - page 2. (Read 1845 times)

legendary
Activity: 4326
Merit: 8899
'The right to privacy matters'
July 04, 2017, 06:42:50 PM
#16
Here we go...




lets see if that shows it.

it does and a nice fix on your part.
full member
Activity: 206
Merit: 100
July 04, 2017, 06:26:32 PM
#15
Here we go...

legendary
Activity: 4326
Merit: 8899
'The right to privacy matters'
July 04, 2017, 06:25:26 PM
#14
Oh yeah! I had one out of eight die on me on Sunday for no apparent reason. Didn't want to wait for newegg to RMA it and took a closer look at it.

Out of all cards I have, this one is the only 1070 that has a regular 10A FUSE (not polyfuse) marked F4 mounted on the back side right next to the PCIe fingers. It's a small white ceramic part (1206 size I think) with 10 written on it and it was blown open! No visible damage to it, just blown inside.

Today I soldered two AWG20 wires across the dead fuse and hooked up a cheap ass multimeter capable of measuring 10A DC current. The damn card works flawlessly and pulls 6.6A over that fuse (WTF!!) under full overclocked load. That's 80W of power via PCIe. I would never expect that much and now wonder how much does it pull from the 8-pin connector on the top... probably 2A?

So... to confirm, just get a multimeter and measure continuity of the F4 fuse. I think it is simply undersized for this design.

P.S. The risers I use are all V006C with 20A poly fuse onboard, but that obviously didn't help and the poor card blew itself up.

Let me know if any of you have this same fuse blown...

-scsi

nice find  let us know if it keeps working.
full member
Activity: 206
Merit: 100
July 04, 2017, 05:42:07 PM
#13
Oh yeah! I had one out of eight die on me on Sunday for no apparent reason. Didn't want to wait for newegg to RMA it and took a closer look at it.

Out of all cards I have, this one is the only 1070 that has a regular 10A FUSE (not polyfuse) marked F4 mounted on the back side right next to the PCIe fingers. It's a small white ceramic part (1206 size I think) with 10 written on it and it was blown open! No visible damage to it, just blown inside.

Today I soldered two AWG20 wires across the dead fuse and hooked up a cheap ass multimeter capable of measuring 10A DC current. The damn card works flawlessly and pulls 6.6A over that fuse (WTF!!) under full overclocked load. That's 80W of power via PCIe. I would never expect that much and now wonder how much does it pull from the 8-pin connector on the top... probably 2A?

So... to confirm, just get a multimeter and measure continuity of the F4 fuse. I think it is simply undersized for this design.

P.S. The risers I use are all V006C with 20A poly fuse onboard, but that obviously didn't help and the poor card blew itself up.

Let me know if any of you have this same fuse blown...

-scsi
legendary
Activity: 4326
Merit: 8899
'The right to privacy matters'
July 04, 2017, 05:22:54 PM
#12
mine are flawless.  most likely  you all over clock them. or over power them via tdp

Many people think  110%  and a high clock is okay.
the answer is it is not okay for these cards

what are your tdp settings?

 80%  = too high

70% = good

They all were at 65 - 70% tdp, never exceeded 75 C. I don't buy this explanation. On the contrary, ASUS 1070 Turbo @ 75% tdp are working fine since almost 2 months.


okay  the asus 1070 turbo has a big-ass cooler  not so much the fan  but the heat sink  is big.

i can not see  what happened to you but as I said  my gigabyte aorus  can slip  out of its msi afterburner settings  and shoot to 100% tdp.

So I think  your cards dropped the setting in msi afterburner  shot to 100% tdp and fried.

all my gigabytes have third party firmware clicked if in afterburner

if on nvoc0017  they have 100 watts as setting and they hold it.

One more question  did you manually set fan to 75% or 80%

if you did auto fans that can fail  and these cards will over heat much faster then  bigger 1070  cards..

I have more faith in nvoc0017

 then windows and  msi afterburner for my nvidia cards.


simply because  I have seen  cards  go into thermal over drive  on windows 7 and 10.


lastly did you do dual mining?  it is way harder then  single mining and if the msi settings fail these cards can fail.


Also  you have 5 or 6 in a row with   consecutive  serial numbers it could be a bad batch.


Did you use 1 card in a real slot to see if it works?

if you did and it did not work  did you try every card 1 at time in a real slot?

what is your mobo?  what is your psu?  what risers do you have?
legendary
Activity: 1498
Merit: 1030
July 04, 2017, 04:52:47 PM
#11
Seems like you got a bad batch - I've got about a half dozen of the Gigabyte ITX cards been working fine for months, bought one-at-a-time over the last 5-6 months ago.

 This can happen to any manufacturer, unfortunately.

sr. member
Activity: 428
Merit: 250
Inactivity: 8963
July 04, 2017, 04:24:10 PM
#10
mine are flawless.  most likely  you all over clock them. or over power them via tdp

Many people think  110%  and a high clock is okay.
the answer is it is not okay for these cards

what are your tdp settings?

 80%  = too high

70% = good

They all were at 65 - 70% tdp, never exceeded 75 C. I don't buy this explanation. On the contrary, ASUS 1070 Turbo @ 75% tdp are working fine since almost 2 months.
newbie
Activity: 4
Merit: 0
July 04, 2017, 03:17:45 PM
#9
The cards don't even get recognized by the OS, they're just dead. I think I just got a bad batch which is super unlucky but I can't attribute it to anything else...the serial numbers indicate that they were made one after the other so I think they're all from the same production run.
legendary
Activity: 4326
Merit: 8899
'The right to privacy matters'
July 04, 2017, 03:09:38 PM
#8
It can't be the motherboard or PSU because I've tried them on multiple PCs. I overclocked and undervolted them, but this hasn't caused any of my 20+ other cards (GTX 1070s, 1060s, RX 480s) to fail. I had them at +650 memory clock and 80% power using MSI afterburner.

80% is the beginning of high  but not ready bad.

did you dual mine eth  with another coin?

did you keep an eye on temps?

If you told me  100% tdp and dual mining I would be quick to blame you.

You need understand they are a specialty card they do not cool as well as most 1070's. But 80% tdp is not really bad.

I own
 the gigabyte       1070 itx
 the msi aero       1070 itx
 the galax katana 1070 slim

all three of them are fine but they do not cool as well as
the msi 1070 gaming red led
the asus 1070 strix

so I can get 400 sold for zec easy with the smaller cards zero issues
the msi gamer and the asus strix  will do 440 sols for zec easy

I have had msi afterburner drop its settings on a gigabyte 1080 ti aorus   from 70% to 100%  or 210 watts to 300 watts.   I wonder if the issue is gigabyte and msi afterburner

I fixed my issue in afterburner on a settings menu i check off third party cards as an option since then the aorus does not drop settings and go to 100%.

if any of the gigabytes you have left work  see if you can find the setting on the msi menus third party firmware settings overclock some thing like that
sr. member
Activity: 728
Merit: 256
NemosMiner-v3.8.1.3
July 04, 2017, 02:54:30 PM
#7
mine are flawless.  most likely  you all over clock them. or over power them via tdp

Many people think  110%  and a high clock is okay.
the answer is it is not okay for these cards

what are your tdp settings?

 80%  = too high

70% = good

i have 5 itx 1070's been running  flawless for 3+months i run default everything with 82% tdp (125w)
they = same hash rate as the G1 Gamiing with 72% tdp (125w)
newbie
Activity: 4
Merit: 0
July 04, 2017, 02:27:41 PM
#6
It can't be the motherboard or PSU because I've tried them on multiple PCs. I overclocked and undervolted them, but this hasn't caused any of my 20+ other cards (GTX 1070s, 1060s, RX 480s) to fail. I had them at +650 memory clock and 80% power using MSI afterburner.
legendary
Activity: 4326
Merit: 8899
'The right to privacy matters'
July 04, 2017, 09:49:34 AM
#5
mine are flawless.  most likely  you all over clock them. or over power them via tdp

Many people think  110%  and a high clock is okay.
the answer is it is not okay for these cards

what are your tdp settings?

 80%  = too high

70% = good
sr. member
Activity: 428
Merit: 250
Inactivity: 8963
July 04, 2017, 09:15:14 AM
#4
Yea, the are utter shit. I bought 12 a month ago and now i have only 6 working.
hero member
Activity: 1036
Merit: 606
July 04, 2017, 09:15:03 AM
#3
What PSU are you using and how many cards do you have connected per PCI-E connector?
full member
Activity: 187
Merit: 100
July 04, 2017, 09:14:37 AM
#2
more likely to be your motherboard to be perfectly honest
newbie
Activity: 4
Merit: 0
July 04, 2017, 09:02:40 AM
#1
Has anyone had problems with the short GTX 1070s? I bought 5 Gigabyte cards and they've all failed within a few days/weeks. They don't power on, not recognized by the computer, completely dead. I may have just gotten a bad batch but that seems unlikely so I wanted to put the question out there.
Pages:
Jump to: