Overclock.net › Forums › Cooling › Cooling Experiments › Ryzen 1800x with Nofan's CR-95C passive cpu cooler
New Posts  All Forums:Forum Nav:

Ryzen 1800x with Nofan's CR-95C passive cpu cooler - Page 4

post #31 of 56
Quote:
Originally Posted by Decoman View Post

So to try answer your question, the previous idle temp in bios was reaching 80+ deg C.

Looking at the first page in the overclocking thread, there's a tip for dealing with the temps, and so I re-enabled the "sense skew" and I set the value to 272 (previously set to "auto"):
http://www.overclock.net/t/1624603/rog-crosshair-vi-overclocking-thread

I also reseated the cooler, using less TIM, applying the time like I did the first time, with the "credit card" method. Given how weird it is to apply the noctua TIM this way, I think I am tempted to try that new TIM with liquid metal, if only to get to apply the TIM more easily (note, can't be used on aluminium surfaces I am reading on ThermalGrizzly's website).

So.. after enabling "sense skew" and setting the value to 272, I appear to have normal looking temps in bios, and in Windows.

These are the temps reported by HWinfo64, after a few minutes of staying in Windows 10 (only used for gaming).

CPU STATUS = 1.550 VID (?) (Heh, isn't 1.550 VID very high?, does this mean I have the most terribly binned 1800x???)

The following were the temps with an open case (panels taken off):
Cpu speed = about 3.7 GHz on all eight cores
CPU (TCtl) = max 51 deg C
CPU (Tdie) = max 27 deg C
CPU corevolt = 1.5 max (fluctuated between 1.381v, 1.494v, 1.285v)
Mobo temp = max 30 deg C
CPU socket = max 44 deg C
DIMM temp = 36 & 38 deg C
M.2 SSD temp = max 34 deg C
GPU core idle = 40-42 deg C
GPU VRM = 46 deg C

So depending on your ambient, if those are idle temps then you should be looking at tdie as an idle. Don't forget that you want medium/high mounting pressure for AMD and you want as little TIM as possible (AMD heatsinks are usually pretty damn flat and don't require much).

VID is set by AMD and everyones is set to that, mine even changes depending how fast i open hwinfo after boot up haha. The best way to get cool temps is to manually overclock the CPU, xfr will make the vcore boost up pretty high to get that one core to run at 4.0-4.1. You should check out the first page of the ch6 owners thread and the ryzen owners club, LOTS of info there.
    
CPUMotherboardGraphicsRAM
R7 1700X asus crosshair 6 Sapphire R9 390 Nitro CMK16GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Kingston HyperX 3K Seagate ST2000DM006 EKWB Supremacy Evo EK-XRES 100 Revo D5 - Acetal 
CoolingCoolingCoolingCooling
Alphacool GPX r9 390 m01 alphacool xt45 240 Alphacool xt45 360 Alphacool D5 PWM 
OSMonitorKeyboardPower
Windows 10 LG 29UM68 Corsair Strafe Corsair RM1000 
CaseMouseMouse Pad
Thermaltake X5 Corsair M65 Pro RGB Razer Goliathus Speed 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
R7 1700X asus crosshair 6 Sapphire R9 390 Nitro CMK16GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Kingston HyperX 3K Seagate ST2000DM006 EKWB Supremacy Evo EK-XRES 100 Revo D5 - Acetal 
CoolingCoolingCoolingCooling
Alphacool GPX r9 390 m01 alphacool xt45 240 Alphacool xt45 360 Alphacool D5 PWM 
OSMonitorKeyboardPower
Windows 10 LG 29UM68 Corsair Strafe Corsair RM1000 
CaseMouseMouse Pad
Thermaltake X5 Corsair M65 Pro RGB Razer Goliathus Speed 
  hide details  
Reply
post #32 of 56
Quote:
Originally Posted by Decoman View Post

Well ****. The new Seasonic fanless PSU I bought has bad coil whine. frown.gif How does this even get past "quality assurance"?
Temp at idle in bios, seem to be heh 80 deg C, though I suspect the temp is 60. Using bios 1002 atm iirc. I guess I have to tweak the sense skew if I figure out to do it.
They don't have any when it comes to noise.

60-80C idle means badly mounted cooler, at angle or not enough pressure.

You can use a lot of paste as long as you have enough pressure that will push the excess out.
post #33 of 56
Thread Starter 
I am super skeptical to any RMA for this (as I don't know what I will get in return), but it looks like the PSU is going that way. Having gotten a couple of emails from them, the people at Seasonic seem friendly enough, and sort of seem helpful. I have heard about other people complaining about the coil whine on this particular PSU, but for now I will hope that there are at least some good units and that I can get one that doesn't make any noise.

They gave me a list to check, and I managed to go though most of it this afternoon, I want to do the last thing as well, which is to test the psu with my other computer, though I don't expect anything to change.

UPDATE: I have arranged for returning the unit to the seller already the next day (today).
Edited by Decoman - 5/5/17 at 2:52am
post #34 of 56
Thread Starter 
Btw, I have successfully performed the IntelBurnTest on "standard", however that was with the front panel OFF. With max cpu temp of 73. Which seem great, and with only a single 700 rpm fan inside, and not even blowing at the cpu cooler.

With the front panel ON, however the IntelBurnTest failed, as it brought the cpu up to nearly 80 deg C, and with that the computer powered itself off.

I do have high hopes for this setup, and once I get the three temp controlled fans installed on the far right, I expect the high temps to be brought under control. smile.gif

Stay tuned for more! (One of these days!)
Edited by Decoman - 5/3/17 at 7:36am
post #35 of 56
Thread Starter 
I decided to install two more 140mm fans on the far right side of the cabinet, and the temps was suddenly improved, by about 10 deg. These were 1000 rpm fans. Apparently some airflow really helps a lot for keeping the passive cooler cool.

With front panel closed, I succesfull completed the standard IntelBurnTest a few times, however I ran into temp issue once I tried the setting above "standard".

Looking at the temps have me puzzled. It seems that the CPU SOCKET TEMPERATURE are generally six degrees higher than any of the other cpu temps. Which I think is probably why my computer shut down all these times, as I did not look at the cpu socket temp when stress testing.

I wonder precisely what parts of the mainboard are the "cpu socket".
Edited by Decoman - 5/4/17 at 7:17am
post #36 of 56
Thread Starter 
I think I made a slight mistake that I hope is correctable.

Not sure how this happened, but I am sure it was my fault. I have now applied liquid metal TIM for both the CPU and the GPU, however I also added a 1mm metal washer to the CPU cooler screws, big mistake I think. I can't tell for sure, but it is obvious that the metal AM3 back plate has a bend to it, obviously the pressure on the CPU is too much by the looks of it. The bend seem to match the 1mm washer height. SO DON'T DO WHAT i DID. Heh, hopefully I can avoid it being a disaster, by having loosened the four screws on the CPU cooler by one full turn.

The styrene washers on the graphics card, for the two screws there covering the VRM area, was replaced with a couple of silicon O rings. I think these would be delicate and won't easily melt. A reminder: rumor is that these two screws for the VRM area on the gfx card is supposed to have springs on them, but I think I lost them, so I had to use some kind of washer.

Another mistake: When applying the liquid metal TIM on the CPU heat spreader, despite having cleaned the surface really good, I forgot to use the cleaning pad that came with the Thermal Grizzly Conductonaut package. Hopefully that omission won't matter much. smile.gif I am looking forward to turning this machine back on and checking the temps. smile.gif

Btw, before applying the new TIM with liquid metal, when taking off the CPU cooler, I noticed that the old TIM had apparently been spread out in a uniform way from the center, which looked good.

I have received my new 140mm Arctic F14 TC (temperature controlled) fans, and have mounted them on the far right side in the case.

I have not yet received my 60x60x10 mm vrm fan that I will experimentally place behind the motherboard. I will try to tilt the fan a little so hopefully there will be a little bit of airflow going upwards. Not much space back there.


PROJECT STATUS: Very limited success (I did manage to fit all the parts, and the computer works, however it remains to be seen if I can keep the temps in check for when stress testing the CPU. Also, the computer should be fairly quiet, which is why I did things this way, not wanting 7 case fans blowing all the time.)

Btw, when handling the syringe for the liquid metal TIM, despite having "bumped/snapped" the syringe with my two fingers, and even when being very careful not to add much pressure on the end of the syringe, still, liquid metal TIM splotched out, and onto the floor. It seemed like a lot, which sucks. I have no idea how much is left in that syringe from Thermal Grizzly. One tip, would be to anticipate the splotching, so that you can "splotch" the TIM onto some surface that isn't contaminated, so that you can reuse that TIM for whatever you wanted to use the liquid metal TIM for, so that it won't go to waste.

Other news: The Seasonic Fanless 520W power supply has been RMA'ed, because of loud noise during idle. It might take some time to receive a replacement unit from the seller, though both the seller and Seasonic have been nice in handling this so far, very little stress on my part. The product was advertised as being completely silent, and how this got past quality testing I have no idea. None of my other power supplies has made noise when idle afaik.

Having removed the entire motherboard from the case, in order to have a better look at the previously installment of the CPU cooler, I could see more clearly that the fancy plastic on the mobo seemed to bump against the cooler, so I have sanded down maybe 2mm of that fancy plastic, though it looks like I should ideally have sanded off another millimeter. I don't expect this to matter much though, so I only sanded the mobo just a little bit more.
Edited by Decoman - 5/12/17 at 2:05am
post #37 of 56
I have used the cr-95 in my machine since christmas 2013. It cools a 3770k. I undervolted it and downclocked it to 3.8 to meet 70c max when doing prime95 for a couple of hours. That cpu is a 77watt part. My machine is completely passive cooled.

I wanted to chime in on the PSU issue. I use a Seasonic fanless 520w in my build and there is no sound from it. Before I made that computer passive it had a Seasonic platinum 850w supply with coil whine when the computer was off, super annoying. I got a x79 motherboard and used the 850w on that system and it is quiet. Sometimes coil whine is a freak of nature.

Good luck on your build, I'll be keeping an eye on it.
post #38 of 56
Thread Starter 
I think I can conclude now that the Nofan CR-95C isn't able to coole the 1800x at stock speeds/voltages when stress testing, leading to the cpu shutting itself down (presumably it is the cpu which does this). CPU socket temperatures goes out of control.

frown.gif

Here are some photos showing what the contact surfaces looks like after I removed the cooler. I have been using liquid metal TIM here, but I am not able to decipher the remains on the surfaces. Can't tell if I used too little liquid metal TIM or not.






The liquid metal TIM was applied to both surfaces, and with a minimum amount.

When I removed the cooler, I gave it a little twist, however the cooler came off right away. I wonder if maybe the cpu heatspreader isn't flat enough, or maybe I didn't use enough liquid metal TIM. Ofc, it seems as if maybe this TDP 95W cooler just can't handle the heat from the 1800x cpu.

I am guessing that it is the inadequate cooling on the cooling that makes the CPU socket temp go out of control. Though I can't explain why the CPU socket temp was 60 deg C and the Tdie was 30 when playing Civ5. Why would the CPU socket be that much more hot? Was it because I loosened the four screws a little bit? (I made the mistake of using washers, so the metal back plate on the back started to bend, so I loosened the screws a little when I saw that.)
Edited by Decoman - 5/12/17 at 2:08am
post #39 of 56
Would undervolting the 1800x down to 1700 levels be out of the question? It seems like a good balance of performance and lower TDP.

Also, subbed. Fully passive ryzen sounds glorious.
post #40 of 56
Thread Starter 
How do I remove this liquid metal TIM? Anyone know? It certainly doesn't rub off with alcohol.


I tried to undervolt the 1800x cpu with the Ryzen Master software, but I couldn't get it to work. I also didn't find any options for changing the cpu voltage in the bios, heh. I really feel like a newb here. This probably the only time things haven't worked out for me putting a computer together.
Edited by Decoman - 5/12/17 at 2:14am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Cooling Experiments
Overclock.net › Forums › Cooling › Cooling Experiments › Ryzen 1800x with Nofan's CR-95C passive cpu cooler