Overclock.net banner

Bad EKWB CPU Contact, please help.

28K views 194 replies 32 participants last post by  oreonutz 
#1 · (Edited)
Hey guys!
I recently picked up a EK Classic for my Ryzen 7 3700x, but I'm experiencing very high temps. I made a thread explaining the situation and asking for help. Here's a YouTube video I made for a quick recap:


So, I emailed EK's support, but they're taking too long to give me a definite answer on what I should do, so maybe you guys can help me out? It seems as if the IHS on the CPU and Cold plate are not making adequate contact. What should I do now?

UPDATE 8/19/2019
I received the new EK Cold plate, while it *seems* to have made better contact due to the better thermal paste spread, it still gets very hot. Rads and water are cold to the touch after a prime95 load, and changing fan speed doesn't seem to effect temperatures at all. I might have to bite the bullet lap the original Cold plate, and POSSIBLY Lap my CPU. I really don't wanna void the warranty on the most expensive and fastest CPU I've ever purchased.
 

Attachments

See less See more
1 2
#2 ·
Lap the block?

Sent from my GM1915 using Tapatalk
 
#31 · (Edited)
get a metal ruler type deal & see if the cpu IHS is derformed, if so lap. more likely the IHS than the block.

or maybe that one radiator simply isn't enough for that chip? idk enough about amd heat production these days.

if block is flat, & chip is flat... does your radiator get nice & warm after it runs for a bit? but based on that chip imprint, chip ihs isn't flat/bad crank job... you seem pretty competent though, so chip ihs.
Lapping the block is not recommended. Most IHS's are not flat these days and the blocks are designed to seat against their shape. Or rather, I should say that if you lap the block, I would lap the IHS as well.
 
#3 · (Edited)
get a metal ruler type deal & see if the cpu IHS is derformed, if so lap. more likely the IHS than the block.

or maybe that one radiator simply isn't enough for that chip? idk enough about amd heat production these days.

if block is flat, & chip is flat... does your radiator get nice & warm after it runs for a bit? but based on that chip imprint, chip ihs isn't flat/bad crank job... you seem pretty competent though, so chip ihs.
 
  • Rep+
Reactions: battlenut
#4 ·
Watch the youtube video please.. I have 2x radiators (1x 120mm and 1x 240mm) and they do not warm up at all during a sustained prime 95 load, my water barely gets warm. However if I do a GPU Stress test, the rads will heat up, which makes me deff think there's a heat transfer issues between IHS and Waterblock.

Does lapping void the AMD Warranty?
 
#5 ·
yeah, i ended it @ 3/4s thru.

yes, lapping voids warranty. Even scrubbing the info off the front voids warranty, at least @ Intel...

You may be able to talk them into sending another sample though, if you can prove its the IHS causing the temp issue. not sure. haven't dealt with team red in many moons, & now give intel the extra money to insure my chip past spec.
 
#6 ·
Alright. Thanks! I talked a lot in the video so I see why you ended it.
I'll see if I can get a straight edge and look up methods for testing flatness. I don't want any downtime with my PC, so hopefully AMD can cross-ship with me.
 
#7 ·
I'd lap the CPU. Use a high conductive paste like Prolimatech PK-3 too, but not liquid metal; copper will soak up the LM and you get worse thermals over time. have personally seen it 3 times. It's fine on nickel, but it will stain the laser-etched surface of a CPU IHS if you don't lap.
 
#8 ·
yeaaap, i found out the hard way what LM can do to a CPU warranty, though I never saw it increase temps... I can confirm it'll pit copper over time though. I believe its the galium or whatever?
 
#9 ·
Hi,
Take apart the water block see what thickness razor blade jet plate is in it and the block should of came with an extra
Swap for the thinnest
Block is suppose to bend to conform with any imperfections on the chip or so they say.

Otherwise do the spin test with the chip on a piece of glass see if it take off or stops pretty quickly
If it want to take off rma the chip.
 
#11 ·
I thought EK was shipping them per SKU these days... Meaning, assuming he purchased the "correct" kit, the correct plate should already be in there for his chip. Either way, jet plates affect temp by a few degrees these days & the newest intel series has a giant plastic insert instead of the razor plate!

I still love the sexiness of my all copper supremacy over this modern all metal black block i got.
Since it's an AMD block is came with 1 jet plate preinstalled. I tried testing it removed. The mouting hardware also comes preinstalled.
The CPU doesnt spin well on a glass surface. Maybe for a second or two.
 
#10 ·
I thought EK was shipping them per SKU these days... Meaning, assuming he purchased the "correct" kit, the correct plate should already be in there for his chip. Either way, jet plates affect temp by a few degrees these days & the newest intel series has a giant plastic insert instead of the razor plate!

I still love the sexiness of my all copper supremacy over this modern all metal black block i got.
 
#12 ·
Hi,
I don't thing ek has very good amd blocks but thought it was just thread ripper wise

What type of fluid you using not ek cryocrap I hope.
Water block looked clogged when you took it a part ?
 
#16 ·
Return the EK and buy a Heatkiller IV or Raystorm Pro. All we see around here these days is issue after issue caused by EK's total lack of QC.
What were your temps with the stock cooler for comparison? How do you know something is wrong?

If this is stated in the video I won't be able to watch it at my location.
I'm using XSPC fluid, but I've been running pure distilled water since the video. I took apart the system and cleaned it, no clogs at all.

I would return the EK block, but it was on clearance for $70 (maybe for a reason) and was a bargain for me.

I never ran the stock cooler, I've been on water since my 2600, which I even used a corsair AIO on.
 
#14 ·
I traded out my ek Intel block for a barrow rgb block. Equivalent temps, less $

Sent from my GM1915 using Tapatalk
 
#15 ·
What were your temps with the stock cooler for comparison? How do you know something is wrong?

If this is stated in the video I won't be able to watch it at my location.
 
#17 ·
This is after just a few seconds of Prime95 Smallest FFT running. Which one of these software is accurate? SUPPOSEDLY I read that Asrock has a weird bug with the temp monitoring that they have not fixed.
 

Attachments

#34 ·
Damn, maybe I should be checking my Block too. Keep in mind I am running a 3900x, so more cores and power. But even with my Core Voltage at 1.3v, specifically under Prime 95 Small FFTs, I hit 110c INSTANTLY. If I don't end the process immediately, the second my Tdie hits 116c the PC shuts off to protect itself. I couldn't believe it the first time I saw it. However, under a realistic workload, like Blender, even when using an AVX instruction set, I top out at 82c under the same manual 1.3v. I can run a 10 hour render (Something that used to take me 16 hours on the 2700x) and My Max Temp will still be 82.5c.

So far the 2 things I have found that I can't run or the temp gets crazy like this is Prime 95 small FFTs and Aida64's FP 64-Ray Trace test.

Right now I am running just the EKWB Phoenix 360 Kit with the AM4 Block, and the pump Maxed out. Definitely about to add a second 360 Rad and D5 Pump, but with my current setup at that voltage should I be seeing crazy temps like that during Prime 95 @EKJake?
 
#18 ·
Hi!

First start by disabling PBO.

Then confirm which version of prime95 you are running.

If you are running 29.8 build 5, then continue as you have been, if not download the latest version.

For temperature monitoring, choose one app, either Ryzen Master or hwinfo64.

If using hwinfo64, your CPU temperature is the one that is reading from "SVI2" (look for the CPU Core Voltage).

Fire up prime95 and select the second option (I think its the one you have already used).

For comparison sake, look at the cooling in my signature and the CPU I am running.

My peak temperature after running prime95 for 10 minutes is 84C, CPU voltage is 1.366v.

I am using conductonaut, I have tried in total 3 mounts, 2 with conductonaut and 1 with kryonaut, I could not improve my temps.

Also, looking at your spread of thermal paste, keep in mind the CPU core is found at the bottom right from the little golden arrow thats in one of the corners of your CPU.

The spread looks OK to me, not optimal (can you tighten the screws further ??), but I dont think its going to make difference in temps.

I would have the pump going to the CPU first (top inlet) it looks like you are going from the pump to the vga then to the cpu (bottom inlet).

Please post back your temps after 10 minutes, if they are around 80-85C then contact is not the issue and you are within the correct temperature range
 
  • Rep+
Reactions: skupples
#19 ·
You should maybe put an air cooler on for time being. If you are looking for new block, heatkiller will be making new block for Ryzen3000. No Eta but they said on its thread here (in watercooling part) end of summer.

If you want to get it, you will probably have to order directly from watercool store (prices include vat and will be taken out at checkout). Shipping is flat so if you add more items, it doesn't increase shipping costs.

EKWB has a dedicated rep here, EKJake (used to be caselabs rep) and you can maybe contact him for help.
 
#21 · (Edited)
don't run pure distilled, you'll kick yourself in 3-6 months when your blocks look 100 years old. Skip the kill coil as well, and ALL OTHER MANUFACTURERS COOLANTS! The proof is in the pudding, Mayhem's just signed deals to start filling AIOs, and many companies now just buy & rebrand Mick's stuff BECAUSE IT IS THE BEST. Hands down, no longer and discussion about it. It's mayhems, or its garbage.

invest in mayhem's biocide & inhibitor drops. Cheap, and last forever. It'l also keep all your parts looking brand new, forever, if properly used.


as to advise to change your flow pattern, don't listen. You've gotta tiny loop for that pump, flow isn't the issue. flow doesn't really change or "slow down" like some people envision (discussions of serial gpus/parallel gpus aside).

like we already discussed, if 1GPM, or north, you're fine on flow.
 
#22 ·
The distilled is temporary at moment, as I'm currently flushing it out getting to the bottom of my high temps.
 
#23 · (Edited)
Good ideas everyone, I'll attempt everyone's diagnosis, and even throw my AIO on temporarily to see if temps change.
 
#27 ·
You said you tried spinning the CPU. Have you tried the same on the EK CPU block?
 
#29 ·
Hi,
Save you the trouble by design the water block should have a hump in it the razor blade jet stops it from going too far
Copper/ nickle plated base as I said before will flex when tightened to conform to the cpu cap.
 
#32 ·
What is your voltage set at? I would double check the thumb screws used and start there. If you disassembled the block for any reason, also double check the jet plate to make sure it's in the correct orientation still and hasn't shift for any reason.
 
#36 ·
I have everything stock in the BIOS, when Prime95 is running it's getting like 1.25 volts. No LLC obviously, this is totally stock.
How should the jet-plate be oriented? It only goes in 1 way doesn't it?
 
#35 ·
Ive posted on the youtube video before I found your post here. Im having the exact same issues only with a ryzen 3600 so prob not as bad as 8 or 12 core but temps are too high. Just upgraded my son's 2600k to a ryzen 3600/b450 tomahalk.

Narrowed issue down to ek water block/cpu ihs. The water block def has a high spot in the center but ek jake said its suppose to be this way. Ive remounted the block so many times now as well as dissassmbled and im frustrated at this point. tried diffeernt paste and etc but never had issues like this before.

Custom loop
EK supremacy evo amd rgb block(comes preassembled with jet plate and hardware installed. no way to mess up install)
360 radiatpr
ek ddc pump res combo

tools used: Fluke 50D K/J Thermometer

water temp at start of stress test: 25.1C
20min end water temp: 27.0c
case air temp: 23.2C
Cinebench r20 hits 80C+ with cpu at stock. thermal paste tried: gelid extyreme, conductonaut and kryonaut.

You can see the water isnt even getting warm because of the poor transfer of heat to the waterblock.
 
#37 ·
Ive posted on the youtube video before I found your post here. Im having the exact same issues only with a ryzen 3600 so prob not as bad as 8 or 12 core but temps are too high. Just upgraded my son's 2600k to a ryzen 3600/b450 tomahalk.

Narrowed issue down to ek water block/cpu ihs. The water block def has a high spot in the center but ek jake said its suppose to be this way. Ive remounted the block so many times now as well as dissassmbled and im frustrated at this point. tried diffeernt paste and etc but never had issues like this before.
I have everything stock in the BIOS, when Prime95 is running it's getting like 1.25 volts. No LLC obviously, this is totally stock.
How should the jet-plate be oriented? It only goes in 1 way doesn't it?
I know the 2700X was a hot chip, but that seems high to me then. As for the jet plate, I haven't had a Classic block in hand, yet. However, it was possible to install the jet plate in an incorrect orientation on previous blocks. The jet plate should be perpendicular to the fins, not parallel with them. Essentially, the slot in the jet plate should stretch across multiple fin channels.

Do me a favor, and both of you please submit a ticket to our support team and let me know what they are. I'd like to follow up on this and see if it's a one off issue or if there's something causing poor fitment.
 
#41 ·
Another user having issues:

He tries to improve the contact by lapping. He ends up going back to air cooling and gets lower temps with a be quiet dark rock pro air cooler he talks about here:

At the end he talks about what I suspect and that is the cpu block only cools the center of the cpu and since ryzen 3000 the chips are not in the center we are seeing the high water cooled temp issues. Kind of like how they had the redesign the threadripper water block years ago so add more fin coverage on the block.
 
#44 ·
Totally. That and I really try to avoid EKWB at all costs now. The build I do for my friend, I make sure they do not use EKWB parts.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top