7820x temp issues after delid

[yazerak]

So I just sold my 6700k for a strix x299E and 7820x with monoblock, I didn't have a cooler on hand to test the before temps but I delidded my cpu following the guide on the gamers nexus youtube and all seems to work fine no spill over or anything(thank you for the nail polish tip btw) but I am worried about my temps i'm overclocking to 4.6ghz at a 1.2 Vcore -4 AVX and -6 AVX512 and on aida 64 stress test with fpu stress I'm seeing temps at around 85 on the package and spikes up to 90/91

these numbers seem high to me but I don't know how they would've been before hand to compare however online i see a ton of 7820x's sitting at 65-75 at 4.7/4.8 with a Vcore of 1.235/1.25 (if i try those settings i hit 100)

is there something wrong, should i delid again and hope maybe i didn't put enough LM or too much silicon on my reseal? or are these temps the expected temps? i cant find much online about the 7820x so id figure id ask here

more details:
noctua nh1 thermal paste
i7 7820x 4.6ghz 1.20 vcore -4 avx -6 avx512 3.0 mesh
asus strix X299E gaming board with EK nickle monoblock
gigabyte aorus 1080ti WB extreme
lian li pcO11 case
custom loop with 2 EK Pe 360mm radiators, 6 corsair ml fans 3 on each rad, top rad is pull bottom rad is push,1 d5 pump, loop order goes res/pump/gpu/cpu/top rad/bottom rad/ res
my ambient temp might be high since I live in south Texas but I have an ac next to the pc set to 70 at all times

temps were collected using HWinfo/aida64/realtemp (cam would give me notifications that my cpu was hitting 87)


TLDR: 7820x package temps hitting 85 with spikes to 90+ on aida64 after delid should i reapply LM or check my seal or is this normal, using custom loop 720mm rad space

[Airdragonz]

how's the liquid temps and contact with the waterblock?

[airdeano]

snap a couple of pix to see your loop arrangement.

this can shed some answers pretty quick.

 

as to the amount of LM used, you'll have to remove the CPU to re verify the contact of the DIE to IHS and the waterblock to IHS to make sure these are plentiful.

[Hugs12343]

91°C seems very high to me. I wouldn't like my cpu to be anything above 80°C. Try adjusting your fan curve on your radiator fans and case fans.

[KarathKasun]

If you re-glue the assembly together you will find that temps are worse.  IMHO I would use tiny amount of gel CA glue (superglue) just to make sure the lid does not move and immediately install the CPU so it applies pressure to the IHS and forms a better contact with the die.

[Hugs12343]

I feel like someone who owns a 7820x should know what they're doing. Remember X stands for extreme and is meant to be used by enthusiasts.

 

[yazerak]

1 hour ago, airdeano said:

snap a couple of pix to see your loop arrangement.

this can shed some answers pretty quick.

 

as to the amount of LM used, you'll have to remove the CPU to re verify the contact of the DIE to IHS and the waterblock to IHS to make sure these are plentiful.

I doubt it's the loop order though my 6700k was in the same loop but with an an supremacy evo 

 

[yazerak]

1 hour ago, KarathKasun said:

If you re-glue the assembly together you will find that temps are worse.  IMHO I would use tiny amount of gel CA glue (superglue) just to make sure the lid does not move and immediately install the CPU so it applies pressure to the IHS and forms a better contact with the die.

I'l try super glue Thursday that could be it I used some silicone to reseal  things seem fine on idle temps but I should be safe and try again

The cpu isn' hitting tjmax but I do think temps should be better especially with the LM

[KarathKasun]

You have to be careful when resealing, most skip this step altogether.  The silicone can cause the gap to be too large (AFAIK it changes volume when drying) making the contact very poor.

[For Science!]

7 hours ago, yazerak said:

-

Don't know what tools you have, but normally a delidding/relidding tool should let you apply a fair amount of pressure on the IHS over the die to during silicone drying.

 

[AngryBeaver]

I personally would stay away from super glue. I have always used something like RTV silicon gasket maker.

 

If you don't have a nice tool to remount there are several other ways to do it, but generally what I would do is apply the gasket maker, make sure to leave that little section open without any adhesive.

 

Then when it comes time to put the lid on.. put the chip in the socket, then position the IHS slightly higher up than center, because when you apply the clamp it will move a little bit. If you are worried about this messing with your LM application too much (haven't had this issue personally) then you can put it on and let it sit while having something sitting on top to apply force to the IHS.

[KarathKasun]

Not talking about doing the whole thing with superglue, Ideally you only put it at the corners to keep it stationary yet easy to break away if you need to do so.

[yazerak]

After re sealing and re applying the LM my package temps have dropped by about 5° and my cores are closer together however at 4.6ghz @1.2v with a -4 avx offset on both 512 and normal I'm still getting temp warnings of 87° and still spiking into the 90s on an aids 64 test with fpu enables

[KarathKasun]

Not enough LM or there is a flow issue with your loop.  The i7-7820X is a monster to cool from what I have seen.  If you are using a variable speed pump, make sure to increase pump speed considerably under load.

[yazerak]

On 4/15/2018 at 7:19 PM, KarathKasun said:

Not enough LM or there is a flow issue with your loop.  The i7-7820X is a monster to cool from what I have seen.  If you are using a variable speed pump, make sure to increase pump speed considerably under load.

im certain its either rad space or airflow when i take off my side panels temps drop a good 5c, but i tubes feel hot making me think im not getting enough cooling with my rads im trying to move cases into the 011 dynamic and add a radiator to help....... hopefully

 

[AngryBeaver]

12 hours ago, yazerak said:

im certain its either rad space or airflow when i take off my side panels temps drop a good 5c, but i tubes feel hot making me think im not getting enough cooling with my rads im trying to move cases into the 011 dynamic and add a radiator to help....... hopefully

 

From what I can see you have 2 360mm rads or 720mm of rad surface area. That should me more than enough to cover your current setup with some to spare. What pump are you using? Also if you are seeing a 5c drop from removing the panels you probably are having an airflow issue and are just recirculating how air to your rads.