VRM heatsinks

[cc_moreau_cc]

Hey guys,

Just yesterday night I installed my Arctic Accelero Xtreme 3 cooler on my 6970. I installed the heatsinks on all the VRAM but didn't think it was necessary to have them on the VRMs. I was also confused since I couldn't seem to locate them. The question for you guys is, are heatsinks on the VRMs really important to the point where my card will blow up without them or will the cooling from the fans be enough?

Thank you.

 

[cc_moreau_cc]

Just to add a little info, I played Battlefield 3 for 2 and a half hours last night at stock clocks (880/1375) and didn't seem to get any problems. The GPU temps never went above 55C and the fan speed was at auto. That's what I want to know. Will VRMs blow after prolonged periods without actual heatsinks on them?

 

[A Bad Day]

VRMs manage the current, and the more you OC, the more current the GPU pulls. If the VRMs start to heat up while having to put out more current, they will become less efficient and put out more heat, thus resulting in unstable currents/voltage when they become overheated. And then your GPU would artifact or crash from the unstable power supply.

VRMs are typically located on the right side of the GPU.

http://i1140.photobucket.com/albums/n575/Yokes29/SS201.jpg

I'm assuming that the large heatsink is covering the VRMs. I'm not sure about the two groups of two capacitors to the very right, but it wouldn't hurt putting VRM heatsinks on them as well as a precaution.

 

[cc_moreau_cc]

Thanks for replying. Yeah, sort of understand where the VRMs are but the picture you linked me is different to my card. The long heatsink you're talking about isn't on my card. Here is what my PCB looks like, it's a reference model.

http://i1-news.softpedia-static.com/images/news2/Radeon-HD-6970-Gets-Pre-Launch-Teardown-PCB-Pictured-2.jpg

Now, I didn't crash at all last night while playing for 2 and a half hours of BF3 yesterday so does that mean I don't have to put the heatsinks on? It would be really annoying to have to take it off again and put the heatsinks on. Also, if anyone can highlight the VRMs that need to be cooled, that would be much appreciated.

Is there also a program that will let me monitor my VRM temps?

 

[A Bad Day]

I don't think VRMs have their own temperature monitoring. However, there was a GPU (don't remember which brand or model) that only had half of its VRMs covered with a heatsink, the rest were bare. Many of them were poor overclockers and typically died fast, especially with OC.

Again, I would play it safe since VRMs supply current to your power-hungry GPU, thus making them potentially dangerous if they burn out. And hot VRMs always have shorter lifespan even if they don't fry.

I found this interesting thread that talks about VRMs and MOSFETs. Although it primarily focuses on CPUs and motherboards, it should also apply to GPUs since they are essentially tiny motherboards with very hot chips: http://www.overclock.net/a/about-vrms-mosfets-motherboard-safety-with-high-tdp-processors

 

[cc_moreau_cc]

Thank you for your feedback and thanks for the article, it was really informative. I guess I'll put them (it's just a pain in the ass to get the cooler off again) on but using MSI afterburner I checked the temps and 1 hour of Battlefield 3 resulted in VRM 1 running at 84C and VRM 2 running at 76C so, clearly the fans themselves are doing the job.

 

[cc_moreau_cc]

Does anyone else have useful feedback regarding VRM heatsinks? I want to be absolutely sure it is necessary since taking the cooler of again would be really annoying.

 

[nukemaster]

I would use then personally, but those temps are NOT bad at all for VRMs

Based on the image you link

 

[cc_moreau_cc]

Hey thanks for the answer nukemaster. I'm looking at the image and it's what I thought but, I'm really afraid that, since those chips are really close to the board, any heatsink interference with the board will blow my 6970 to heaven and then some. lol. My plan is to monitor the temps for a whole week and then, Friday next week, I'll decide whether I put them on or not. I'm not sure what these VRMs are rated at but if you say 84 and 76 isn't bad, then that makes me a little more confident.

Also, anyone mind telling if these guys were overreacting about this?

http://battlelog.battlefield.com/bf3/forum/threadview/2832654490059610048/

 

[A Bad Day]

Maybe they heard a lot of horror stories about cooked VRMs. Or they cooked their own VRMs before.

There are some low-profile heatsinks available on the market, so there shouldn't be any spacing conflicts.

 

[cc_moreau_cc]

Best answer selected by cc_moreau_cc.

 

[cc_moreau_cc]

Such a sensitive subject. I'm asking all sorts of people everywhere and getting mixed results. Guess it's just a personal preference. I've been hearing about life span but honestly, when the 8000 series come out, I'll upgrade and that will be within a year so no worries there. With the stock cooler, those chips only had stupid dingy thermal pads. Anyway, thanks to the both of you for your feedback, much appreciated. 😀

 

[nukemaster]

Most VRMs are rated over 100c for sure, but cooler is always better.

You are right to worry about contact with anything. This is why MANY heatsinks had heat transfer pads(Just as an example would be the Accelero made for the 5870 cards. http://www.overclock3d.net/reviews/cases_cooling/arctic_cooling_accelero_xtreme_5870_review/3). They do not work as well, but have the advantage of reducing the chance of the heatsink it self shorting something out.

You do have to be careful with installation because if I remember right, you GLUE those suckers on so they have to be in the right place from the start. You must also ensure you clean the VRM's very well(90% alcohol for most people) before installing the heat sinks.

Cards that already had a non stock cooler may already have VRM heatsinks(or a plate) installed.

 

[cc_moreau_cc]

I will install them at this point, I'm just going to take an extreme amount of time to make sure I do it right. I know a friend who fried a card, a hole, right through the card because of an improper installation. I am looking to overclock the card so having them on will be better, definitely.

 

[nukemaster]

Ok.

Yeah....Take your time clean it up REAL(any oil will cause them to not stick and fall off) well and then let that stuff harden(if it is a thermal glue, almost sure it is.) over night if you can.

 

[cc_moreau_cc]

Hey, nukemaster, will I need to put new thermal paste on my GPU when I take it off since it'll be all smudged or do I just leave it?

 

[nukemaster]

Yes, clean and new paste.