Separate names with a comma.
Discussion in 'Overclocking & Cooling' started by Kyle_Bennett, Oct 4, 2017.
Yeah this unit is fucked. Get refund
Just wanted to post another follow-up. I've installed a Noctua NH-U14S TR4-SP3 air-cooler and it works great. For 60% the cost I am getting 59C @ 3.9GHz full load. Also in the long run it will be nice not to worry about maintenance or possible failure. I would suggest for anyone else wanting to streamline their build with a sense of long term dependability to go with this low cost option.
The caveats are make sure your case is large enough, this particular model is very big and I managed to fit it in with less then 1 mm clearance on the side panel. Their is also the strong possibility it will block some of your ram slots. Not and issue in my immediate case as I am only using 4 slots, but if you need all 8 make sure you do the research.
Noctua AMD Threadripper Air Cooler Roundup
We take all three Noctua air coolers built for AMD Threadripper CPUs and put these to the test on both the 2950X and 2990WX. We run the NH-U14S, NH-U12S, and NH-U9 through the paces with both Precision Boost 2 and Precision Boost Overdrive using multiple fan configurations and compare these to the Silver Arrow and Wraith Ripper too.
Between the three air coolers the Noctua is the most readily available and most often gets discounted. The Silver Arrow performs the best in an absolute sense but that fan sounds like a tornado. If you strapped that fan to the U14S I think it will perform almost if not just as good.
Thanks for the review, wish I caught it earlier. I might have gone for the Thermalright Silver Arrow since it's going for only $10 more than the Noctua on Amazon right now.
Interesting thought. In the review the Thermalright achieves a 0.2 GHz gain over the Noctua in the 2950x overdrive test and a 0.3GHz gain over the Noctua (even with 2 fans) on the 2990WX overdrive test.
I think you can't ignore the number of heat pipes having a significant effect:
Noctua U14S: 12 pipes ($80)
Cooler Master Wraith Ripper: 14 pipes ($120)
Thermalright Silver Arrow: 16 pipes ($90)
Noctua a12x15 x3 on silver arrow might be the ultimate solution
You know what, I'm still within a couple weeks of the return period on the Noctua. Just ordered the silver arrow and going to give it a try.
Thinking about fans, I imagine the only real consideration in terms of performance is how much air gets moved (CFM). In terms of stacking fans I've actually seen tests where there is only a marginal performance increase when adding identical fans. When I think about it, if a fan is running say 1500rpm and moving 83 CFM, adding another fan running the same speed won't add any speed to the air, it will only add static pressure. When I think of 2 cars driving the same speed, they don't move any faster, but if they hit a wall there will be more force.
Essentially the performance increase in stacking fans is that the air will maintain speed when meeting resistance (cooler fins) but will not make the air go any faster than the rated CFM.
How about this for comparison:
3000rpm, 159 CFM, 41.3db
the specs on the Thermalright fan that ships with the Silver Arrow:
2500rpm, 130 CFM, 45db
I'm not a quiet pc nut, so I'm actually not too concerned about getting the quietest fan. Though it seems a lot of Noctua's claim to quiet fans is just to cripple the speed/CFM by adding their LNA wire which I think just reduces the voltage or something going to the fan. In that case I feel like you can just profile you fans manually by adjusting the fan curve. I'd rather get the most performance possible and if it's too loud with minimal performance gains, I'll dial the fan curve down.
The specs for the noctua fan that ships with their coolers is a bit different than the retail fans, iirc. I'd have to go check again.
Anyone know which side is the intake to the pump on the Enermax? I'm going to cut it and put in a reservoir.
If I remember correctly (95% sure), bottom one. Will check it later.
Ended up buying a Silver Arrow TR4, and idle temp wise it is about the same as the Enermax 360 lol. At load the Silver Arrow is actually better than the Enermax 360 I was using that is 7 months old. It was getting hotter faster at load than usual the past couple of weeks so I was thinking it was on it's way to failure.
The fan on the Silver Arrow isn't that great silence wise. It isn't loud, but it's a higher pitched whine sound when running full throttle and is very noticeable. It fits even with my GPU in the closest slot, but I moved my GPU to the next slot because it is very difficult to remove the video card (the little tab you have to press to release it from the slot needs a long thin screwdriver) and access to 2 fan headers is right next to the heat sink on my MB and can't be reached with the GPU in the closest slot.
Using small fft prime95 I still hit 85c @ 3.7ghz on my 1950x. I stopped it there, but for most of my other tests like 3d Rendering and Video compression the load was less and stayed in the 75c-78c range. That's with an ambient of around 21c. Overall a sideways move, but at least it should remain consistent.
I was right, top is outlet, bottom inlet. BTW hose I.D. is 1/4, O.D. 1/2.
I have but one upvote to give. Thanks!
For anyone looking for air cooling options I posted a detailed description over here:
The long & short, I got the best results with a Noctua NS-U14S cooler and replacing the fan with Thermalright's TY-143 140MM fan. So the extra heat pipes did nothing to improve the Silver Arrow, and even stacking 3 fans on it didn't do better than just having a faster fan on the Noctua.
Whelp, so I did it. I did a teardown of the Liqtech TR4 that I replaced after failed cooling performance on my threadripper:
I figure I would do my best to clean it to see if it's salvageable at all. I used an old watercolor brush & alcohol on the copper fins they were so narrow:
The main housing was horribly corroded, I used a bamboo skewer to scrape off some kind of encrusted residue. Whatever the issue, it had actually eaten into the metal block. I don't have any kind of material knowledge with this stuff, if anybody has an idea what the issue might be I'd be interested to hear it. Also wondering if it's worth salvaging this unit at all. I have another Liqtech II that was failing, and a third unopened Liqtech II that I received recently. I plan on opening all of them and taking a look. Is this a coolant issue? Could replacing the fluid in an unused unit prevent build-up and corrosion?
Any jelly like substance in the cooler? If there was it would suggest a lack of a biocide to kill bacteria build up.
Not sure if anyone has swapped out the fluid - may something to try out, see if temps drop back with the new fluid and report back to everyone
I intend to do exactly this. Now that you mention it, I'll do before and after benchmarks, take pictures, and maybe do a write up. Seems to be a lot of interest in this particular cooler.
Plan is: benchmark, take apart cooler, document, clean, replace coolant, cut intake line, add micro reservoir, replace coolant with diluted automotive coolant solution (75-25), benchmark.
I did the Liqtech II today. Not as bad, but still same signs of bacteria and corrosion. The blobby stuff really makes you wonder if this is some scifi thing where you die of a horrible disease from touching it:
I did the same routine and cleaned everything off. The cold plate doesn't look too bad:
The main block still shows signs of corrosion. I wonder if the exposed aluminum needs to be hit with rustoleum or some kind of marine enamel, or if the paint would just flake. At the very least I imagine the fluid needs some kind of corrosion retardant because of the aluminum copper contact? It looks like the jelly nibbled off one of the nubs I circled in the photo:
The blobby stuff is jelly that builds up when water and its bacteria are cooled down, we get it in the fridges where I work all the time, a good biocide should kill it off. I'm doing my 360 next week, going to use xspc EC6 clear pre mix as the new fluid, it says its got corrosion inhibitors and biocides mixed in.
Thanks for the tip, I think I might try it as well.
I'm also interested in splicing the tubing. At this point I have a couple extra radiators, wondering if it's possible to combine them in a loop to get better cooling.
Its taken a while to get this posted but I pulled my 360 apart back in march when i was off work. The disassembly process was easy, hats off to gamers nexus for their tear down video on this.
Colour of the fluid when i drained it all:
Now according to the video there should be over 200ml of fluid in the 240mm version - I drained 150ml out of my 360 and several chunks of paint and solid matter.
State of the cold plate and block before cleaning.
That yellow mass was encrusted onto the block hard, after cleaning it off there was a good 2mm chunk out of the aluminium. I've attributed the cause of it being so hard to the fact that my mobo has a fault with the on board chip that does all the hardware monitoring, if you leave a program open that constantly polls the chip, aura sync or HWinfo 64 for example, it will eventually lock up the chip which then shuts down all the cooling fans and aio pump, If you don't notice this the next thing that happens is a thermal shutdown - this happened a few days before i pulled the aio apart - it took me 30 minutes of a handheld battery fan blowing onto the block to cool everything down enough to allow a restart. Its a known Bios bug for my mobo but Asus don't seem to be in a hurry to fix it.
This also explains the heat discoloration on the cold plate. The jelly was easy to remove from the edges but it took a while to get it out the microfins - a hour with an old toothbrush helped here.
After cleaning everything off I filled the aio with deionised (no distilled available where i live) water, again the aio would only take 160-170mm so not sure how GN got 200ml out of a 240, plugged it into to a spare power cable and ran it for 3 days, changing out the water twice. The first change out produced water just as dark as the original fluid. The second swap out was much cleaner. On the 3rd swap out i put in the XSPC EC6 clear fluid that I am using in it to flush out all the water.
That ran for a day and then I drained and cleaned everything for the final time before refilling with more fluid.
i reinstalled the aio with no leaks or problems. The pump seems to be running about 100rpm slower according to HWinfo 64 but my temps are back where they should be, idle was showing 29.3c with a room temp of around 23c. Fully loading all the cores put the temps into the mid 50's but i'm going to do more testing on this in a few weeks.
Having just checked my phone for the pics I realised that i have no pics of after the cleaning which is a shame but both my block and cold plate did not come up as shiny as jhsu's. i'm going to give the fluid another week or so and i'm going to pull everything apart again over easter and see what teh fluid looks like. I'll also get some pic's to add to this post of after the clean and especially the damage to the block.
Looks like corrosion to my untrained eye, probably accelerated by the heat and stagnant fluid.