View previous topic :: View next topic |
Author |
Message |
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 6:27 am Post subject: Swapped thermal grease on laptop cpu |
|
|
Was wondering if I applied a rather thick coat, would the excess just be squished out thus not hurting overal thermal conductivity severely? Idling at ~35°C is this temp good?
Used arctic silver 5 wanted liquid metal but I read the stuff drips and eventually leaks away also causing short circuits on the mobo. |
|
Back to top |
|
|
bunder Bodhisattva
Joined: 10 Apr 2004 Posts: 5940
|
Posted: Sat Mar 31, 2018 7:15 am Post subject: |
|
|
It was tested that overdoing thermal paste usually accounts for 3-5c worth of temperature. An idle temp of 35c might be fine depending on what cpu you have. _________________
Neddyseagoon wrote: | The problem with leaving is that you can only do it once and it reduces your influence. |
banned from #gentoo since sept 2017 |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 7:42 am Post subject: |
|
|
It actually idles at ~32°C on low cpu clocks and I pressed down about 2.5mm on the syringe on a 3.5 gram tube of grease. I think it CAN be better if I redo it, but it seems fine and redoing it would really be a waste. Im sure the OEM grease was way worse anyways because there was a lot of it all around the CPU die, seems like they used too much and the thermal conductivity is probably way lower also. |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54667 Location: 56N 3W
|
Posted: Sat Mar 31, 2018 9:25 am Post subject: |
|
|
The_Document,
Thermal grease is a very poor thermal conductor but its a lot better that the air that would otherwise be in the gaps.
Here's the physics.
Two planes (the top of the CPU and bottom of the heat sink) can only ever touch at three points.
Consider how a three legged stool is unconditionally stable and a four legged chair will rock between several stable states.
Further, a 'point of contact' has a vanishing small surface area, so very little heat transfer can take place.
With two clean practical surfaces in contact, there is effectively a very small air gap.
Lets ignore the flatness for now except to say that minimising the air gap by improving flatness helps. You can't beat the three points of contact.
Thermal paste aims to replace the air in the gap with a material with better thermal properties.
When its applied as a 'sandwich filling', the original three point contact is lost and one poor thermal conductor is replaced by another.
It won't squeeze out enough, or it would run out over its intended useful life.
If you think you have put too much in, you probably did.
The right amount is no more than the size of a grain of rice. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 1:56 pm Post subject: |
|
|
NeddySeagoon wrote: | The_Document,
Thermal grease is a very poor thermal conductor but its a lot better that the air that would otherwise be in the gaps.
Here's the physics.
Two planes (the top of the CPU and bottom of the heat sink) can only ever touch at three points.
Consider how a three legged stool is unconditionally stable and a four legged chair will rock between several stable states.
Further, a 'point of contact' has a vanishing small surface area, so very little heat transfer can take place.
With two clean practical surfaces in contact, there is effectively a very small air gap.
Lets ignore the flatness for now except to say that minimising the air gap by improving flatness helps. You can't beat the three points of contact.
Thermal paste aims to replace the air in the gap with a material with better thermal properties.
When its applied as a 'sandwich filling', the original three point contact is lost and one poor thermal conductor is replaced by another.
It won't squeeze out enough, or it would run out over its intended useful life.
If you think you have put too much in, you probably did.
The right amount is no more than the size of a grain of rice. |
Great explanation but I read ppl say pea sized amount which is right? Im afraid of detaching heatsink might damage to processor processor can it happen? Is it common?
Just bought Thermal Grizzly Kryonaut will do it over anyways. Might as well much better than AS-5 too. |
|
Back to top |
|
|
Fitzcarraldo Advocate
Joined: 30 Aug 2008 Posts: 2056 Location: United Kingdom
|
|
Back to top |
|
|
P.Kosunen Guru
Joined: 21 Nov 2005 Posts: 309 Location: Finland
|
Posted: Sat Mar 31, 2018 3:52 pm Post subject: Re: Swapped thermal grease on laptop cpu |
|
|
The_Document wrote: | Used arctic silver 5 wanted liquid metal but I read the stuff drips and eventually leaks away also causing short circuits on the mobo. |
IIRC Arctic Silver lasts very long, it is still good after 5 years so not bad choice at all. |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6069 Location: Removed by Neddy
|
Posted: Sat Mar 31, 2018 4:15 pm Post subject: |
|
|
NeddySeagoon wrote: | The_Document,
Thermal grease is a very poor thermal conductor but its a lot better that the air that would otherwise be in the gaps.
Here's the physics.
Two planes (the top of the CPU and bottom of the heat sink) can only ever touch at three points.
Consider how a three legged stool is unconditionally stable and a four legged chair will rock between several stable states.
Further, a 'point of contact' has a vanishing small surface area, so very little heat transfer can take place.
With two clean practical surfaces in contact, there is effectively a very small air gap.
Lets ignore the flatness for now except to say that minimising the air gap by improving flatness helps. You can't beat the three points of contact.
Thermal paste aims to replace the air in the gap with a material with better thermal properties.
When its applied as a 'sandwich filling', the original three point contact is lost and one poor thermal conductor is replaced by another.
It won't squeeze out enough, or it would run out over its intended useful life.
If you think you have put too much in, you probably did.
The right amount is no more than the size of a grain of rice. | and on the flipside... it is a VERY good thermal insulator if you use too much...
When I give a damn I will take a razor blade to both faces to smooth them of any additional bumps, "pea size blob" in the middle and then smooth that out evenly with the razor. Now... I just do the pea blob.
* note I still do the razor method but for 100kVA inverters I am hand-building. _________________ #define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 5:39 pm Post subject: |
|
|
P.Kosunen wrote: | The_Document wrote: | Used arctic silver 5 wanted liquid metal but I read the stuff drips and eventually leaks away also causing short circuits on the mobo. |
IIRC Arctic Silver lasts very long, it is still good after 5 years so not bad choice at all. |
Does the paste I just bought have shorter life?
Naib wrote: | for 100kVA inverters I am hand-building. |
Any sot-227 in there? Send pics I like big IC packages. |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54667 Location: 56N 3W
|
Posted: Sat Mar 31, 2018 6:42 pm Post subject: |
|
|
The_Document,
Too little is safer than too much thermal compound and CPUs have got bigger heat spreaders too.
Thermal compounds come in two sorts.
Those that set over the first few hours of use and those that don't.
It can be very difficult to separate the CPU and heatsink if you use the first sort. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6069 Location: Removed by Neddy
|
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6069 Location: Removed by Neddy
|
Posted: Sat Mar 31, 2018 7:03 pm Post subject: |
|
|
yup that is an IGBT module. Coincidently I am designing an inverter right now using SiC MOSFETS (TO247 package) _________________ #define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 7:05 pm Post subject: |
|
|
Mentioned JFETs are probly better because higher voltage handling at lower resistance than MOSFETs. |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6069 Location: Removed by Neddy
|
Posted: Sat Mar 31, 2018 7:39 pm Post subject: |
|
|
The problem with JFET's is their "always on" default state. In the event of control powerloss OR failure of the gate PSU you are entering into shoot through territory. SiC MOSFETs are down in this ohmage as well so it is all good _________________ #define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Sat Mar 31, 2018 7:47 pm Post subject: |
|
|
Naib wrote: | The problem with JFET's is their "always on" default state. In the event of control powerloss OR failure of the gate PSU you are entering into shoot through territory. SiC MOSFETs are down in this ohmage as well so it is all good |
I PMed you. |
|
Back to top |
|
|
Goverp Advocate
Joined: 07 Mar 2007 Posts: 2191
|
Posted: Sun Apr 01, 2018 10:18 am Post subject: Off-topic physics |
|
|
NeddySeagoon wrote: | ...
Here's the physics.
Two planes (the top of the CPU and bottom of the heat sink) can only ever touch at three points.
Consider how a three legged stool is unconditionally stable and a four legged chair will rock between several stable states.
... |
All four legs on a square chair with the leg-ends in a plane can be arranged to simultaneously touch a continuous surface by rotating the chair by no more than 90 degrees. (I'm not sure of the exact proof, but effectively it's that if three legs touch, rotating by 90 degrees interchanges the legs, so at some point they must change from being in contact to not, and vice-versa, so by the definition of continuous there will be a point where all 4 will touch.)
This probably means that rotating the heat sink may find a closer fit to the chip below it. Note to chip makers: can we have circular chips and heat sinks. _________________ Greybeard |
|
Back to top |
|
|
Naib Watchman
Joined: 21 May 2004 Posts: 6069 Location: Removed by Neddy
|
Posted: Sun Apr 01, 2018 11:08 am Post subject: Re: Off-topic physics |
|
|
Goverp wrote: | NeddySeagoon wrote: | ...
Here's the physics.
Two planes (the top of the CPU and bottom of the heat sink) can only ever touch at three points.
Consider how a three legged stool is unconditionally stable and a four legged chair will rock between several stable states.
... |
All four legs on a square chair with the leg-ends in a plane can be arranged to simultaneously touch a continuous surface by rotating the chair by no more than 90 degrees. (I'm not sure of the exact proof, but effectively it's that if three legs touch, rotating by 90 degrees interchanges the legs, so at some point they must change from being in contact to not, and vice-versa, so by the definition of continuous there will be a point where all 4 will touch.)
This probably means that rotating the heat sink may find a closer fit to the chip below it. Note to chip makers: can we have circular chips and heat sinks. |
That only hold true for a perfectly flat plane and perfectly equal length vectors/legs. This isn't the case in reality. A 3-legged stool is always steady, but a 4-legged stool can be wobbly.
http://mathforum.org/library/drmath/view/53267.html
Now take two complete planes NOT points because we are dealing with a heatsink and a heatplane NOT a barstool.... However smooth you think these are they are not. If it was economically viable we would be seeing cold welding being used (nano-coating is making this viable)
The two surfaces are essentially random and thus three point will make contact which is in no way good enough to minimise the thermal resistance of this interface THUS some form of TIM is used. Paste is best as it will uniformly fill in the gaps (but relies on a control process as too little is bad, too much is bad), phase-change material is the next best as it will essentially fill in the gaps (consistent results). So now you will still have 3 points making contact BUT then additional thermal conductive path via the TIM
https://myheatsinks.com/wp-content/uploads/2013/12/gap_and_tim.png _________________ #define HelloWorld int
#define Int main()
#define Return printf
#define Print return
#include <stdio>
HelloWorld Int {
Return("Hello, world!\n");
Print 0; |
|
Back to top |
|
|
The_Document Apprentice
Joined: 03 Feb 2018 Posts: 275
|
Posted: Wed Apr 04, 2018 11:01 pm Post subject: |
|
|
Re did the thermal paste idling at 29C without X and 31C with KDE. I actually made it worse with Arctic Silver 5 because the temps were much higher a bit after applying idling at 41C, meaning I put too much, I put less than a size of a pea ammount of Thermal Grizzly Kryonaut paste, spreading it was very difficult. Works great. |
|
Back to top |
|
|
|