Desperate Nerd Looking for HALP! Ryzen 1700x w h115i

Ok, so I just got hit by a Ransomware Virus (Phobos) the other day, and can I just say holy cow! In my 30+ years of working with PC's I've never been hit by a virus, and never thought I would be!!! But alas, I stupidly left one of my virtual machines fully vulnerable up overnight from testing with a network mounted drive, and between using my system and before my PC went to sleep (15 min default power settings) I was penetrated HARD. Granted only my tertiary drive was affected, I figured it was time to use it as an excuse to do my annual format.

Long story short, I've reinstalled my OS, and in the midst accidentally hit my reset CMOS button on the rear of my I/O panel while re-arranging some cabling (one shot swoop among all things necessary), so I've had to get back to "normal."

However, now that I'm back up and running on a near-stock BIOS, I'm hitting temps 10-15C over what I used to hit and I'm at a loss as to why...

A bit of info:

BIOS Changes:

- Virtually stock (virtually.. not explicitly)

- Modified XMP profile to hit 2399mhz on my 3khz RAM (see specs)

- Switched all mobo fan headers to DC, using stock FN-2 NZXT fans (3-pin) and set to FULL SPEED (max at 1100rpm? specs are hard to find, they try to reach a "peak dbm" according to marketing info)

- Disabled Cool n Quiet setting

- Literally everything else is stock

iCue:

- Managed to restore my existing profiles that I know for 100% fact I was running before format.

- Being a hardware/software dev, I opted to create a new profile matching similar curve just in case the restoring of the profile was somehow faulty or facading the fact it wasn't properly working.

- Using Quiet profile for pump, custom fan curve for .. well.. fans (see attached pics)

h115i:

- Noticing it was running hotter than normal, I opted to reseat (with full re-lapping of the block and fresh Arctic Silver 5) the cooler and double check all connections.

Other than that, I don't have much going on yet (as evidenced by the pics) but somehow my temps are running significantly higher.

I used to average 35-40c while idle, and now it seems to be running 48-53c (even after re-seating the cooler). Just starting up chrome with NO extensions causes my package temp to spike up to 60-63c for a few seconds.

My ambient room temp is pictured with an Omega Thermocoupler (1337 speak for a multi-hundred-dollar bad-mofo of a hardware probing thermocoupling thermometer [aka accurate beyond belief]) and have also verified the surface of the CPU is within 1 degree of what the motherboard is reporting.

It appears my coolant temp is REALLY low, which almost makes me feel like contact between the copper base and the CPU lid is detached, but the spread of a rice-dot thermal paste implies otherwise. When I re-seated, I actually did it twice - the first time I verified that the pressure of the mount screws were causing the paste to spread evenly, and it definitely looked good and well-covered.

From my analysis, everything is working 100% as expected, except my temps are what I consider TRASH.

HELP!?!?!?!?!? PLEASE!? I rarely ask for help, but I'm feeling a bit nervous as I continue to type this in fear some task will run and shoot my temp up to crippling effects.

---

Quick note, I did the risky thing of doing a Prime95 test with large FFT, and it peaked at 75c Tdie (95c Tctl), but apparently was under-volting the cpu at 0.85v instead of the typical 1.12-1.35 (iirc). This is auto managed by UEFI, as I have not locked it in (nor will I probably do so).

(Forgot to upload ambient temp pic [shown in Fahrenheit])

One thing of note is that when i put the pump to EXTREME (which I never have before, as proof as having an export of my iCue profile that I've restored) it does go down to a respectable 40c idle.

I don't see any kinks in the hosing, it looks like a fairly clean curve in the hose from radiator to the block... The unit is just barely over a year old, I would find it hard to believe the liquid in the unit would be coagulating already?????

It appears my coolant temp is REALLY low, which almost makes me feel like contact between the copper base and the CPU lid is detached, but the spread of a rice-dot thermal paste implies otherwise. When I re-seated, I actually did it twice - the first time I verified that the pressure of the mount screws were causing the paste to spread evenly, and it definitely looked good and well-covered.

If your two iCUE screen shots with that 41C coolant temperature (H115i Temp) are at the desktop, your coolant temp is not cold, but too warm. With some adjustment for core count and power settings, your coolant should sit about +4-7C over the room temp and have a load range of +6-10C with case temp changes also influencing. Coolant temp is effectively minimum CPU temp, so it is the warm coolant causing the high idle and that has a knock on effect with everything else. Starting at 40C, it's easy to jump into the 60-70s with activities that would normally be 20C lower.

The H115i is SATA power, so the BIOS settings are clear from blame and I can see your iCUE settings are fine. Little changes in fan or pump speed will not get you to 40C. The only thing that should get you to 40C coolant is a combination of high room/case temp + extended combination CPU/GPU load.

During the reconfiguration process, did you need to take the H115i radiator down or out of the case? The most common issue people seem to run into on that model is a partial blockage from anti-corrosives and adhesive/sealant forming little gummy balls (which of course they should not). Anyway, the typical story goes..."I took my cooler down for routine cleaning and when I put it back on, my temperature went to XXX!!!". The little gummy balls block the coolant flow and H115i Temp starts of kind of low, but slowly builds and builds never coming down. The take down shakes things up and the little gummy balls wind up in very disadvantageous places.

If you want to test this premise, take a mild stress test like Intel XTU (no Prime) and initiate the test with iCUE up and running. Note the starting coolant temp. I will estimate a typical coolant temp rise for a 1700X would be +4-6C over 10 minutes before leveling off. If you have a blockage, you won't have to wait that long. It will go 40-41-42-43...49-50-51 and up right in front of your face in a matter of 30-60 seconds. After ceasing the test (which you should do immediately), the coolant temp will stay at this elevated level and have difficulty coming back down. A 280mm cooler should be able to shed interior heat at about 1C every 30 seconds or so. Other dead giveaways are cold exhaust flow and/or one hot tube and one cold tube. Exhaust temp from the CPU radiator should be equal to the coolant temp. 40C is plenty warm and you will notice. If it feels like normal room temp air, then heat is not getting there. Typically the radiator can also reduce coolant temp by 1-1.5C per pass through, so that means tube temp should be indistinguishable. If one is clearly hot and the other not, than there is a blockage. It is also common to feel heat common up both tubes about half way if the blockage is at the pump.

@c-attack, thanks for the response.

To answer some of the questions/comments:

Little changes in fan or pump speed will not get you to 40C.

I don't know if it qualifies as "little", but putting the EXTREME profile on the pump has brought it down to a steady 34c while idle, and doing mundane tasks such as browsing the web do not affect the temp any more like it was. It was going upwards of 60c with the pump on low while browsing. I know for a fact I never had it set to EXTREME, and left it on QUIET before the reformat as I have the original Profile exported from iCue just before the format, and can see it reads Quiet inside the profile.

During the reconfiguration process, did you need to take the H115i radiator down or out of the case?

Only AFTER I noticed temps were a bit high. I took the block off, cleaned it, re-lapped it, and reapplied the AS5. It did come down ~5c overall, but not enough to be comfortable at.

...take a mild stress test like Intel XTU

I've never used Intel XTU, but the product page implies it's explicitly for Intel processors. Would it work with my AMD Ryzen processor? When I tried the Prime95 test, the processor package speed shot up immediately to 75c, while it did take a little bit of time for the coolant to go up.

However, I did notice the tube was significantly warmer than the radiator itself, which was very cool.

So... ok, it sounds like blockage... What's the RMA course of action for something like this? I bought it in 11/2017 (black friday sale), so it's been roughly 14 months since purchase, and all I can really provide is a screenshot from Newegg showing the order number.

Would this suffice? Is Corsair service typically pretty good with this stuff? Not sure if it's evident, but I'm freaking out a little bit.

Sorry, I forgot what processor we were talking about. Any mild stress test will work, but you don't want to be running small FFT type stuff. You want to provide a steady moderate load to see how the cooler handles heat, rather than how fast the cold plate can conduct heat away from the CPU. However, it's probably not necessary at this point. It seems pretty clear its a partial blockage of some type. That huge drop in temp when you cycled the pump to high is clear indicator. It should make absolutely no difference at idle when there is a paltry 20-40W to handle. It doesn't usually even show at full load. The cold exhaust with coolant temp at 35-40C is the other clear indicator. A lot of end symptoms can have multiple causes, but those two cut straight to heart of the matter and I would lead with that on the support ticket to minimize any back and forth.

If you ordered from Newegg, there should be an order history under your account. Go to the item and select "View and Print Order details". It's a grey button in the product header. That will create a more familiar looking invoice and that should do it. No fingerprint or DNA confirmation required.

Sorry, I forgot what processor we were talking about. Any mild stress test will work, but you don't want to be running small FFT type stuff. You want to provide a steady moderate load to see how the cooler handles heat, rather than how fast the cold plate can conduct heat away from the CPU. However, it's probably not necessary at this point. It seems pretty clear its a partial blockage of some type. That huge drop in temp when you cycled the pump to high is clear indicator. It should make absolutely no difference at idle when there is a paltry 20-40W to handle. It doesn't usually even show at full load. The cold exhaust with coolant temp at 35-40C is the other clear indicator. A lot of end symptoms can have multiple causes, but those two cut straight to heart of the matter and I would lead with that on the support ticket to minimize any back and forth.

 

If you ordered from Newegg, there should be an order history under your account. Go to the item and select "View and Print Order details". It's a grey button in the product header. That will create a more familiar looking invoice and that should do it. No fingerprint or DNA confirmation required.

HAHA to the fingerprint/dna bit :P

Ok so I initiated a support call, which the customer service rep was very dismissive and argued about which product I have - but did ultimately create a support ticket.

I recieved a canned response on trying a different fan header and SATA connection (which I've done) and of course to no avail.

Actually putting the system under a light load jumps it all the way back up to 60c super quick (just having a few youtube tabs open in chrome caused a ~20% load).

I haven't gotten a response back from the support ticket giving my results back, hopefully I hear something soon. Having downtime like this is not fun!

One thing I'm slightly worried about was they wanted my serial number, which of course it must be on the side that is on the backside of the case that I cannot see, so the rep suggested taking a picture of the sticker on the box.

Of course, what I didn't think of when I grabbed the box, is that I actually have TWO - one which is on my wife's identical build.... I wonder if they're going to raise a fit if I try to change the serial number on them once I remove it and find it's not the one I sent as an attachment to the support ticket.

Oh man that stinks! Fingers crossed for you the delivery system speeds up for you, and they don't have to go through the polar vortex thing we have going on.

So I called in this morning and got it moving along, express RMA it is!

Thank goodness, because I am not a fan of the waiting game! Uhhh... No pun intended, really!

Thanks for your input/help - now I get to exchange this mint-condition cooler (literally only been used for about 3 months) and hope my other one in my wife's computer isn't [going] bad. But, her monitor went out (way out of warranty) so that's holding us up on figuring that one out... I guess luckily we still have almost 4 years left on the warranty to find out!