thermal paste Can thermal paste in cpu pins or Mobo makes PC act wierd even if it's not conductive?

[Fracteller]

Hello all, Happy new year everyone!

I have a little question that I couldn't find an answer to; either it's about cleaning the cpu or mobo, my question is:

Can thermal paste that got into the cpu pins and/or mobo socket makes the pc act wierd even if it's not conductive (BSOD, freezes, this kind of instability)? What I want to know if it's something that happens if there's too much thermal paste.

My pc was acting really wierd this month, after alot of troubleshooting and testing with Memtest86 (+100 hours of testing all of the 128g rams in each 4 slots), Aida96, OCCT and alot of tests/benchmarks I couldn't find the solution. For example in a Blender Benchmark (Classroom and BMW) the GPU rendering couldn't even start, was stuck in the first phase where it says "this can take some time", usually it starts after 2 seconds (I have a 3950x). When I took it to a pc repair store, they told me that they found some thermal paste in the cpu pins and mobo, they totally cleaned them and when they tried the blender benchmarks they started immediatly. 

So I searched for this and I couldn't find any similar cases, of a thermal paste making the pc act wierd. I even saw pcs boot with alot of paste in them. So I would like some help to find an answer, to also know if I got bamboozled from the repair shop.

Thanks!

[OhioYJ]

You'll notice some pastes / compounds have warnings about being not conductive, but (and I've seen it worded differently) slightly capacitive, which could do odd things.

 

I wouldn't leave thermal pastes on a pin or in a socket.

[Alex Atkin UK]

In my experience once you have thermal paste in a socket, its game over.  If its a pin-based socket its nearly impossible to remove without breaking the socket.

If its conductive it will short something, if not conductive it will prevent good electrical connectivity of some contacts on the CPU.

[Bartholomew]

Yes it can.

 

Since it's non conductive, it can prevent proper contact if in-between the cpu and the socket for one or more pins for example.

To much paste "on top" of the ihs, or even splattered around the pcb is OK. But not on sockets (cpu, ram dimms) or connectors; that can prevent contact et-all, or form a resistor or even some slight capacitance etc. That can do bad / weird things to signals.

 

 

[Fracteller]

13 minutes ago, Bartholomew said:

Yes it can.

 

Since it's non conductive, it can prevent proper contact if in-between the cpu and the socket for one or more pins for example.

To much paste "on top" of the ihs, or even splattered around the pcb is OK. But not on sockets (cpu, ram dimms) or connectors; that can prevent contact et-all, or form a resistor or even some slight capacitance etc. That can do bad / weird things to signals.

 

 

 

14 minutes ago, Alex Atkin UK said:

In my experience once you have thermal paste in a socket, its game over.  If its a pin-based socket its nearly impossible to remove without breaking the socket.

If its conductive it will short something, if not conductive it will prevent good electrical connectivity of some contacts on the CPU.

 

17 minutes ago, OhioYJ said:

You'll notice some pastes / compounds have warnings about being not conductive, but (and I've seen it worded differently) slightly capacitive, which could do odd things.

 

I wouldn't leave thermal pastes on a pin or in a socket.

Thanks y'all for the responses!

[Alex Atkin UK]

I learned the hard way when I tried to remove a small piece from an LGA socket and bent the pins.  With hindsight, I might have been better off risking it blocking contact than trying to fix it.

[Fracteller]

11 minutes ago, Alex Atkin UK said:

I learned the hard way when I tried to remove a small piece from an LGA socket and bent the pins.  With hindsight, I might have been better off risking it blocking contact than trying to fix it.

oh damn, hope it wasn't an expensive one! 
 

[Alex Atkin UK]

1 minute ago, Fracteller said:

oh damn, hope it wasn't an expensive one! 
 

It wasn't, but it was unique back then as it was a transitionary board that supported DDR2 and DDR3 (not at the same time, as an upgrade path) so it was really useful when most boards at the time only supported one or the other.

[Fracteller]

6 minutes ago, Alex Atkin UK said:

It wasn't, but it was unique back then as it was a transitionary board that supported DDR2 and DDR3 (not at the same time, as an upgrade path) so it was really useful when most boards at the time only supported one or the other.

I always try to see videos and forums before manipulating my pc parts, and this situations is totally new to me. 

I have one more question please: The thermal paste is applied on the CPU along time ago, is it normal that it starts to make my pc act wierd after a long time? Is it like the thermal paste is dripping slowly because of the heat and getting into the cpu / mobo ? The thermal paste I have (Cooler Master Mastergel Maker) is kinda sticky, can it "melt" and become liquid'ish and drips?

[Alex Atkin UK]

17 minutes ago, Fracteller said:

I always try to see videos and forums before manipulating my pc parts, and this situations is totally new to me. 

I have one more question please: The thermal paste is applied on the CPU along time ago, is it normal that it starts to make my pc act wierd after a long time? Is it like the thermal paste is dripping slowly because of the heat and getting into the cpu / mobo ? The thermal paste I have (Cooler Master Mastergel Maker) is kinda sticky, can it "melt" and become liquid'ish and drips?

If its non-conductive then it getting anywhere other than in the CPU socket or RAM slots, is absolutely fine.  The trick is lifting a CPU with overhanging paste without it dropping in the slot, which is where I went wrong.  Generally now if its too close to the edge I will clean any overhanging off, even if I'm feeling lazy and not replacing the paste.

 

I'm not familiar with that paste, you're probably better off asking Cooler Master.  But if its quite fluid then I can see how it getting hot will force some out.  Its also not unusual after a few years to need to re-apply new paste, though some people are a bit too religious about this to me - I only do it if I'm seeing higher temperatures than usual.

[Fracteller]

3 minutes ago, Alex Atkin UK said:

If its non-conductive then it getting anywhere other than in the CPU socket or RAM slots, is absolutely fine.  The trick is lifting a CPU with overhanging paste without it dropping in the slot, which is where I went wrong.  Generally now if its too close to the edge I will clean any overhanging off, even if I'm feeling lazy and not replacing the paste.

 

I'm not familiar with that paste, you're probably better off asking Cooler Master.  But if its quite fluid then I can see how it getting hot will force some out.  Its also not unusual after a few years to need to re-apply new paste, though some people are a bit too religious about this to me - I only do it if I'm seeing higher temperatures than usual.

Alright thank you for your time. 

Last time I replaced the thermal paste is 6 month ago after 2 years of using/rendering with it. Dropped like 20°. 

At least I learned something new today

[Alex Atkin UK]

1 hour ago, Fracteller said:

Alright thank you for your time. 

Last time I replaced the thermal paste is 6 month ago after 2 years of using/rendering with it. Dropped like 20°. 

At least I learned something new today

I've honestly never noticed problems after many years, but it rather depends what quality paste was on there to begin with.

[Godlygamer23]

2 hours ago, Alex Atkin UK said:

I learned the hard way when I tried to remove a small piece from an LGA socket and bent the pins.  With hindsight, I might have been better off risking it blocking contact than trying to fix it.

Depending on how much is on the pins, installing the CPU can displace it enough that you have enough contact. It is also possible to remove thermal paste(or any material) from the socket, as I have done, but it takes a microscope, a steady hand, and light pressure. 

[Beerzerker]

3 hours ago, Alex Atkin UK said:

In my experience once you have thermal paste in a socket, its game over.  If its a pin-based socket its nearly impossible to remove without breaking the socket.

If its conductive it will short something, if not conductive it will prevent good electrical connectivity of some contacts on the CPU.

Not really.
I'm not saying you didn't have the experience you claim but I can say TIM removal isn't THAT difficult to do either on average.
Normally a good spray cleaner for electronics (Electronic cleaner) does a good job as long as it's "Plastic Safe" so it doesn't damage any plastic components/parts/pieces on the board.

For bad cases I either use as much cleaner as needed OR I just run it through the dishwasher.
Cans of cleaner aren't cheap and for bad/messy jobs that normally get it done without all the expense. Thing is with that you MUST let it dry completely/thoroughly before connecting back to power, usually about a day sitting in front of a fan drys it out completely.

[Godlygamer23]

23 minutes ago, Beerzerker said:

Thing is with that you MUST let it dry completely/thoroughly before connecting back to power, usually about a day sitting in front of a fan drys it out completely.

Depending on your water supply, this can be really dumb to do. I don't know if people that respond with "I just throw it in a dishwasher" are completely serious. If you have really hard water, and the water completely evaporates, then you can have mineral residue bridging certain leads or pads together. Or if the water becomes trapped somewhere in the socket, or goes where it can't evaporate quickly, then you are much more likely to have bridging occurring.

 

My workplace does blast populated PCBs with water, but it's deionized, and a specific process is followed. Water cannot be left on the board because it'll pick up ions over time, or in the case of this conversation, it'll already have them, and if the water doesn't evaporate quickly enough, then you just shorted leads together. If someone is really convinced to rinse it in the sink, or apparently throw it in a dishwasher, then they need to be ABSOLUTELY certain that there is no residue or moisture left before powering on the board at all. 

Edited January 8, 2023 by Godlygamer23

[Beerzerker]

23 hours ago, Godlygamer23 said:

Depending on your water supply, this can be really dumb to do. I don't know if people that respond with "I just throw it in a dishwasher" are completely serious. If you have really hard water, and the water completely evaporates, then you can have mineral residue bridging certain leads or pads together.
 

I don't have that problem here, that is with hard water.
No, just doing it once is not going to result in you having stalagtites in your board's socket to worry about.

 

23 hours ago, Godlygamer23 said:

Or if the water becomes trapped somewhere in the socket, or goes where it can't evaporate quickly, then you are much more likely to have bridging occurring.

I'm aware of that and know it can become trapped in a component but at the same time you have guys that just LOVE the shit out of blasting stuff with compressed air and that can inject moisture into a component.

I've never seen water get trapped in a socket and normally leave my boards out to dry fully as in FULLY.
Usually for a couple of days set in front of a fan or if there will be some humid days ahead I'll give it even more time.... And if there is a need to get it dry quickly inspite of such the oven can be warmed up and used - Note I said warmed up.
I'll set mine to it's minimum temp of 170f/84c and let it get to that temp, then cut it off and pop the board in. This way the board doesn't get too hot and drying takes place at the same time.

 

23 hours ago, Godlygamer23 said:

 

My workplace does blast populated PCBs with water, but it's deionized, and a specific process is followed. Water cannot be left on the board because it'll pick up ions over time, or in the case of this conversation, it'll already have them, and if the water doesn't evaporate quickly enough, then you just shorted leads together. If someone is really convinced to rinse it in the sink, or apparently throw it in a dishwasher, then they need to be ABSOLUTELY certain that there is no residue or moisture left before powering on the board at all. 

That's why I do my drying process as I do it.
Haven't had a short or problem because of it yet from over 10+ years of just doing it.

 

[Fracteller]

 

21 hours ago, Beerzerker said:

Not really.
I'm not saying you didn't have the experience you claim but I can say TIM removal isn't THAT difficult to do either on average.
Normally a good spray cleaner for electronics (Electronic cleaner) does a good job as long as it's "Plastic Safe" so it doesn't damage any plastic components/parts/pieces on the board.

For bad cases I either use as much cleaner as needed OR I just run it through the dishwasher.
Cans of cleaner aren't cheap and for bad/messy jobs that normally get it done without all the expense. Thing is with that you MUST let it dry completely/thoroughly before connecting back to power, usually about a day sitting in front of a fan drys it out completely.

22 hours ago, Godlygamer23 said:

Depending on how much is on the pins, installing the CPU can displace it enough that you have enough contact. It is also possible to remove thermal paste(or any material) from the socket, as I have done, but it takes a microscope, a steady hand, and light pressure. 

I personnaly won't try this, maybe only maybe with osmosis water (throught a machine to filter all the minerals in it) or Isoproylic alcohol at 95% + a soft paint brush or soft toothbrush, saw some videos about it that works well. For a sticky paste (like the one I have), I don't think that compressed air can remove it

[Beerzerker]

2 hours ago, Fracteller said:

I personnaly won't try this, maybe only maybe with osmosis water (throught a machine to filter all the minerals in it) or Isoproylic alcohol at 95% + a soft paint brush or soft toothbrush, saw some videos about it that works well. For a sticky paste (like the one I have), I don't think that compressed air can remove it

 Contact cleaner can get it out so if nothing else, I can suggest using that. A toothbrush if used correctly will get it out but if dealing with a board that has pins down in the socket BE CAREFUL using the brush, you can bend pins before you know it!

However you do it, take your time because it's going to take however long it takes to get it done.
There are no trophies for finishing in record time - You have more time than $$ anyway so use it to whatever extent you need to.