Shoud i worry kepping my Graphic card running like this?

[krymancer]

I've had this GTX 1050 2GB from PcYes for about 3 years, I never had any problems with it, but a few days ago I noticed that the fans weren't working anymore, at first I thought the card just wasn't hot and didn't need to be cooled , but then I noticed running a program to stress the GPU that the fans weren't running at all. Reaching 90ºC.

 

 

So I went to investigate and very scared I put a multimeter in the GPU fans header and nothing, no voltage, I was intrigued, but I had the idea of soldering the fans directly to a normal fan cable and connected it to a SYS_FAN available on my motherboard:

 

I ripped the old connector and soldered it to the wires of an old fan I had. So far that the temperatures i got:

Idle:

 

 

Load:

 

 

So my question is, will I have a problem leaving it like this? (It makes more noise now, but it doesn't really matter to me)
OBS: in my location the average temperature is 30ºC

 

 

 

[Jorban]

4 minutes ago, krymancer said:

I've had this GTX 1050 2GB from PcYes for about 3 years, I never had any problems with it, but a few days ago I noticed that the fans weren't working anymore, at first I thought the card just wasn't hot and didn't need to be cooled , but then I noticed running a program to stress the GPU that the fans weren't running at all. Reaching 90ºC.

 

 

So I went to investigate and very scared I put a multimeter in the GPU fans header and nothing, no voltage, I was intrigued, but I had the idea of soldering the fans directly to a normal fan cable and connected it to a SYS_FAN available on my motherboard:

 

I ripped the old connector and soldered it to the wires of an old fan I had. So far that the temperatures i got:

Idle:

 

 

Load:

 

 

So my question is, will I have a problem leaving it like this? (It makes more noise now, but it doesn't really matter to me)
OBS: in my location the average temperature is 30ºC

 

 

 

Please do not use HWMonitor as it has been known to misreport statistics.

Use something along the lines of HWINFO64
and report back the results that you get from that application.

However, I would not be running the GPU like that.
The cheapest thing you can do is take a case fan and place it where it's pushing air to the GPU.
but those are unsafe temperatures.

EDIT:

I totally misread the post.
I applaud your enginuity in terms of rigging up something that works.

In my mind, I see no problem with this.
If your temps are low, there's no problem with running your system like that!

[GuiltySpark_]

The fans on the GPU stopped working and you thought to not only probe for voltage but wire the fans to a fan header on the motherboard? And its working? 

 

I dunno, I think this is the best thread I've read here today and applaud you for your ingenuity and execution. Theoretically, there is nothing wrong with that at all.

 

I'm not sure what the guy above is on about but assuming those temperatures are correct under load, 66c more than safe. About the only thing I agree with is the use of HWMonitor, just get HWInfo.

[8tg]

If it works don’t worry about it. 
The fans are moving, the card is cool, that’s all you need however it is you get to that point.

[krymancer]

12 minutes ago, Jorban said:

Please do not use HWMonitor as it has been known to misreport statistics.

Use something along the lines of HWINFO64
and report back the results that you get from that application.

However, I would not be running the GPU like that.
The cheapest thing you can do is take a case fan and place it where it's pushing air to the GPU.
but those are unsafe temperatures.

EDIT:

I totally misread the post.
I applaud your enginuity in terms of rigging up something that works.

In my mind, I see no problem with this.
If your temps are low, there's no problem with running your system like that!

I really dont know about HWMonitor, thanks for that. here are the prints i got from HWINFO64:
idle:


load:

 

[Whiro]

17 minutes ago, krymancer said:

I've had this GTX 1050 2GB from PcYes for about 3 years, I never had any problems with it, but a few days ago I noticed that the fans weren't working anymore, at first I thought the card just wasn't hot and didn't need to be cooled , but then I noticed running a program to stress the GPU that the fans weren't running at all. Reaching 90ºC.

 

 

So I went to investigate and very scared I put a multimeter in the GPU fans header and nothing, no voltage, I was intrigued, but I had the idea of soldering the fans directly to a normal fan cable and connected it to a SYS_FAN available on my motherboard:

 

I ripped the old connector and soldered it to the wires of an old fan I had. So far that the temperatures i got:

Idle:

 

 

Load:

 

 

So my question is, will I have a problem leaving it like this? (It makes more noise now, but it doesn't really matter to me)
OBS: in my location the average temperature is 30ºC

 

 

 

You’ll be fine. I’ve done something similar few years ago and card was running alright for a year in my pc (untill I’ve sold the card to get new one)

 

[Jorban]

31 minutes ago, krymancer said:

I really dont know about HWMonitor, thanks for that. here are the prints i got from HWINFO64:
idle:


load:

 

For a 1050 those hotspot temps ar kinda hot, however still within spec of the core. Good work on thinking out of the box, seriously, I would have never thought to do something like that to repair the GPU.

[Sarra]

I actually did this with a Noctua 92mm fan. I literally just zip tied it to a 550 TI, and ran it off a motherboard header.

 

That was the opposite of this, though, as the fans were dead, but the fan header on the board was inconvenient to access, and I didn't know if the GPU's header could power a 92mm Noctua fan without burning it out.

[Falkentyne]

6 hours ago, Jorban said:

For a 1050 those hotspot temps ar kinda hot, however still within spec of the core. Good work on thinking out of the box, seriously, I would have never thought to do something like that to repair the GPU.

Hotspot temps are fixed (Not Operational) on pascal and should be completely ignored.  The Delta is always 8C-13C depending on SKU.  A program called "Thermspy 3.3.0" (or 3.0.3 which is on techpowerup forums buried somewhere) can show the minimum hotspot delta by firmware (Even Ampere has a hotspot, which actually works properly, but it still has an absolute minimum delta which won't be exceeded, anywhere between 7C to 10C, but can go up to 20-30C if there are some huge thermal pad (VRM) or GPU paste/contact issues!

[Rysters Tech]

I would keep it like that if you are happy with the performance, through some Noctua fans and high-quality zip ties and will likely be both cooler and quieter than stock. but with the redneck engineered connector working fine, keep it like it is.