[SOLVED] Longest Lasting Video Cards

[nazareneisrael]

Hi. I am editing my post, to shorten it up a lot.
My video card (MSI GTX1660) is failing after 18 months.
Video card prices are outrageous right now, but I don't have a choice, as I need a card.
Is there a "most reliable" brand of video card?
I read that ASUS was considered one of the top manufacturers for longevity on video cards. Does that seem right? (Or is that all propaganda?)
Thanks.

 

[TomKraven]

I have never had a GPU fail on me. Then again, I have never owned any GPU that uses more than ~100W.

I agree with Grobe's general impression that failures do appear to be far more common on higher-end/higher-power cards. If you have a part that has 2-3X the complexity and power draw, you can expect it to be 2-3X as likely to fail since there are that many more potential failure points and that much more power available to dump into any fault to make it fatal.

The most you can do about it is keep the card clean so conductive dust doesn't cause a malfunction or overheat that could nuke it. I clean my case's air filters every 1-3 months and do a thorough clean (I even open the PSU to check up on capacitors that may need replacement, remove dust clumps that got sucked in because the stupid case scraped them from the filter during removal and add silicon grease to fan bearings) once every 1-2 years.

Dang man, opening a PSU is just for the pros in my book. All i do is remove the grill, clean the fan and dust everything i can reach from the outside without having to open the case. Any mishap and you could fry your PC or get a discharge from a charged and faulty capacitor.

 

[ohio_buckeye]

DO NOT open the psu. If you aren’t sure what you are doing, those are like a giant capacitor and have enough juice to kill you or make you have a really bad day if you don’t know what you are doing.

 

[InvalidError]

DO NOT open the psu. If you aren’t sure what you are doing, those are like a giant capacitor and have enough juice to kill you or make you have a really bad day if you don’t know what you are doing.

Any PSU built by any real company doing serious businesses in civilized countries will have circuitry of some sort to discharge capacitors orders of magnitude faster than you can open it after pulling the plug since all civilized countries have electric, fire, occupational, etc. codes that require it. That said, you do need to be aware that the bleed circuit can fail or be defective so there is a non-zero chance of shock.

For PC PSUs though, the same circuit that provides 5VSB will bleed the primary capacitors to 25-50V within a few seconds of input power getting removed, so you'd need a double failure (bleed + 5VSB) to have a realistic chance of shock.

 

[nazareneisrael]

Hi Tom. Hi Ohio. Thanks! But I don't want to open it. I just want to test the 12VDC, 5VDC, and 3.3VDC, to make sure the PSU is outputting what it should. Because I don't know why else the GTX1660 would fail in system A but not in system B, whereas the GTX1070 works fine in both systems. So I came back to Lutfij's comments about the PSU, and thought I should learn how to test them.

I watched some YouTube videos last night, and I think I know how to test it. It doesn't seem too hard. The bummer is that the multimeters in the hardware stores here are all really cheap here (southern Chile), so I ordered a good one (which will take three or four weeks to get here). So in the meantime I am glad that I have a backup system!

InvalidError, thanks for the advice. I plan to leave the PSU cases closed. My plan is simply to test both PSUs, and if the 12VDC, 5VDC and 3.3VDC are not within +/-4% of their respective values, I plan to swap the PSU out. (I am expecting the PSU on my primary to be bad, and the PSU on the other system to be good, but we will wait for the meter, and see.) So now I either have to wait until the good multimeter gets here, or see if I can find someone who will loan me one, or break down and buy a junky one at the local hardware store (which seems like a waste, but I absolutely hate having down equipment, so if I can find a cheapie that works I might just break down...).

Thanks for the advisories!

 

[nazareneisrael]

Ahh, but actually, I do have some Noob questions, now that I think about it.
Do I need to take both PSUs completely out of their cases? Or just disconnect the 24-pin connectors and short out the green and a black?
Or what has to be disconnected to do the testing, and what can remain plugged in?
Thanks!

 

[TomKraven]

Any PSU built by any real company doing serious businesses in civilized countries will have circuitry of some sort to discharge capacitors orders of magnitude faster than you can open it after pulling the plug since all civilized countries have electric, fire, occupational, etc. codes that require it. That said, you do need to be aware that the bleed circuit can fail or be defective so there is a non-zero chance of shock.

For PC PSUs though, the same circuit that provides 5VSB will bleed the primary capacitors to 25-50V within a few seconds of input power getting removed, so you'd need a double failure (bleed + 5VSB) to have a realistic chance of shock.

You are giving too much credit to the manufacturer, so much that you are basically discounting everything that could happen after that unit has left the store. You don't know the context of every specific PSU there is even if its a TIER S+++ psu from the greatest manufacturer ever, the user might have been using it or storing it in subpar conditions. If you want to open or recommend opening a PSU, that person should be aware of the risks AND all the methods of mitigating said risks like measuring the charges with a multimeter, isolating your body and the working space and knowing how to manually discharge the caps in case the bleeder has failed. Otherwise, you or others might die.

 

[ohio_buckeye]

Also, you have no idea what users do or don’t know. So until you know that you should pretend users don’t know anything.

 

[InvalidError]

You are giving too much credit to the manufacturer, so much that you are basically discounting everything that could happen after that unit has left the store.

I'm not discounting anything. If a PSU has a known-working 5VSB rail, such as one that was in a working system and is only being taken out for cleaning, that alone will drain the input capacitors to a safe voltage in a matter of seconds.