Question Why do we only put a "pea" of thermal paste?

Nov 11, 2020
93
5
35
0
I had to repaste my CPU two weeks ago. This time, I only put a "pea" of paste, instead of covering the whole cpu with a layer (this, by the way, didn't spill to the sides)

I've had SIGNIFICANTLY worse temperatures because of this - compared to when I covered the whole CPU with a layer of paste

My temps before were around 23-24 c (even lower) with my windows closed. Now, with the "pea" method, I am averaging at at least 30 c, it hasn't gone lower

Is the "pea" method one of those myths about PC building, like static electricity killing components when handled improperly. I'm not denying that, maybe, at the some point, that was the case, but components have since improved and so those rules to PC building/handling components are now irrelevant

thoughts?

my theory is that:

With the pea method, the paste dries up quickly, and drives the temps higher; in comparison, with a layer of THICK paste, only a small amount of paste dries up and there is still plenty to be utilized by the cpu

edit:

at full load, my temps are 40c and rising. this is ridiculous, and on idle it is 30c (MINIMUM)

tldr; using a "pea" of paste is wrong; putting a thick amount of paste outweighs ANY risks it poses (provided that you dont overdo it and paste doesn't leak all over the place)

this is on a xeon system; on my i7 its a lot worse (like 50-55c worse)

View: https://imgur.com/WSrTXUs
In contrast, these are my temps when using the "layer" method (yes, i know loads are different, and this is when its in indle but just to give you perspective, at the moment, when idle, my temps are at least 30c, and only goes up the longer i have the system on- this is a lot worse on my i7) Edit: This is air-cooling, by the way, and I am using Kyronaught Extreme (2021) paste

View: https://imgur.com/6NqDFEf

Edit:

This is air-cooling, by the way, and I am using Kyronaught Extreme (2021)
 
Last edited:

USAFRet

Titan
Moderator
Mar 16, 2013
140,080
7,627
174,040
21,556
With paste, less is more.

An actual "pea" sized lump might even be too much.

Your "theory" is incorrect.
You want the least possible amount.
It only needs to fill in the microcavities between surfaces. More than that becomes an insulator.
 

jay32267

Illustrious
The pea method....or any other method is all pretty subjective.
I work in an industry where we routinely use thermal paste.
Ideally you want just enough thermal paste...where when you tighten the system up....it covers the entire area....with no air bubbles....but none squeezes out. How you do this is up to you. From a practical standpoint....I like a little to squeeze out all the way around which is how I know I have complete coverage.
 
Nov 11, 2020
93
5
35
0
Your displayed temps of 15-20C....what is the ambient temperature in the room?

If above 20C, those temps are physically impossible. You cannot have temperatures below ambient. Just cannot happen.

And temps at idle are a poor way to measure.
hahaha yeah, at the moment, I am in , and that was taken in January. it gets pretty cold here, at the moment, it is 11c in my city (and they consider that "warm" weather; it doesn't get hotter than 19c during the day, maybe it does, but it drops fast or doesnt go past 20c), so I am assuming temps in my room are 13-16c

at the time that photo was taken, windows were open, and it was freezing, even though it was already summer; i would say 16-19 degrees (probably lower, around 14-15c as this was taken at night)

edit:

just to add, the temps of 40-43c were taken today, and it is 11c (13-16c inside, although I have windows open and it is freezing)

while, the lower temps were taken, in january, at the peak of the summer, so it was significantly "warmer", around 16-19c (inside my room, that is)

edit:

i would even say my ambient temps then (in january) were colder, like 11c cold inside my room. it was definitely warmer weather than now, as we are in autumn now.....
 
Last edited:

rubix_1011

Contributing Writer
Moderator
A 'pea size' drop of compound is normally far too much, but does depend on how large you determine a 'pea' to be.

A rounded drop approx 3-4mm in diameter is about perfect for most CPUs, unless we're talking Threadripper. Then, it is multiple dots in a pattern rather than just one on most desktop Intel/AMD chips.
 

LolaGT

Proper
Oct 31, 2020
109
63
170
1
If it was an actual size of a pea that stuff would be all over the place. I already don't like cleaning up paste anyway. I don't want it squeezing out past the die if I can help it.
A kernel of rice is closer to the right size glob for a CPU.
 

hotaru.hino

Respectable
Sep 1, 2020
1,843
575
2,340
82
I don't recall anyone seriously recommending using a "pea" sized drop unless it's The Verge or Apple. Everything I've heard from said to use something more like a "rice grain" sized drop on the middle. However this isn't really sound advice anymore since the past 10 years.

For example Gamer's Nexus found the best application depends on the processor:


This is mostly because of where the CPU die is and its shape. Intel's processors are rectangular. AMD's Ryzens are also rectangular for Zen and Zen+, all over the place for Zen 2 and Zen 3. Threadripper just needs a ton everywhere.

Though I think you still can't go wrong with applying some and spreading it about with a credit card over until you cover the IHS. That covers basically every die configuration.
 
Reactions: barkersofgeraldine

Zerk2012

Titan
Ambassador
I had to repaste my CPU two weeks ago. This time, I only put a "pea" of paste, instead of covering the whole cpu with a layer (this, by the way, didn't spill to the sides)

I've had SIGNIFICANTLY worse temperatures because of this - compared to when I covered the whole CPU with a layer of paste

My temps before were around 23-24 c (even lower) with my windows closed. Now, with the "pea" method, I am averaging at at least 30 c, it hasn't gone lower

Is the "pea" method one of those myths about PC building, like static electricity killing components when handled improperly. I'm not denying that, maybe, at the some point, that was the case, but components have since improved and so those rules to PC building/handling components are now irrelevant

thoughts?

my theory is that:

With the pea method, the paste dries up quickly, and drives the temps higher; in comparison, with a layer of THICK paste, only a small amount of paste dries up and there is still plenty to be utilized by the cpu

edit:

at full load, my temps are 40c and rising. this is ridiculous, and on idle it is 30c (MINIMUM)

tldr; using a "pea" of paste is wrong; putting a thick amount of paste outweighs ANY risks it poses (provided that you dont overdo it and paste doesn't leak all over the place)

this is on a xeon system; on my i7 its a lot worse (like 50-55c worse)

View: https://imgur.com/WSrTXUs
In contrast, these are my temps when using the "layer" method (yes, i know loads are different, and this is when its in indle but just to give you perspective, at the moment, when idle, my temps are at least 30c, and only goes up the longer i have the system on- this is a lot worse on my i7) Edit: This is air-cooling, by the way, and I am using Kyronaught Extreme (2021) paste

View: https://imgur.com/6NqDFEf

Edit:

This is air-cooling, by the way, and I am using Kyronaught Extreme (2021)
23C you need to turn the heat on in your house LOL.
I see nothing listed showing the i7, anyhow you check CPU temp under load not idle, the type of paste used is almost irrelevant to anything their very little difference.

You can always buy you one of these and use it for paste.
View: https://i.imgur.com/tlAApJK.png


Joking but check the temp under load only.
 
Reactions: barkersofgeraldine

ASK THE COMMUNITY

TRENDING THREADS