[SOLVED] Computer starts only after warming it up with a heatgun ?

Status
Not open for further replies.
Jun 27, 2025
7
0
10
Last winter my computer (for specs: see below) would not start occasionally, especially on days with high humidity. Accidently, I found an online message from a person who solved that same problem by using a heatgun to blow warm air in to the case for a few minutes. I tried that and to my surprise, this worked for me also.

However the occasional start problem became more and more frequent over time and currently my computer needs this "warming up" every day, in spite of the fact that humidity is now very low in the summer.
As of last week, the computer does not fully start up anymore: I get the BIOS page and I have to run the SETUP.
Date and time are not remembered : they are back to 2019.
I replaced the CMOS battery twice already but that made no difference.

I tried to find the location of the problem by warming-up first the on/off switch, I tried the CPU, the memory , but that made no difference.
Just blowing warm air in the case for a few minutes makes the computer start.
After start-up, the computer works flawlessly the whole day (even with heavy loads).
During the day I can shutdown and restart without issues even after several hours , but the next morning the problem is there again.

Obviously, this very inconvenient, so that is why I decided to reach out to this forum.
How can I solve this problem?

PC Specs
Motherboard: ASUS TUF Gaming B550M – Plus
CPU: Ryzen 5 – 5600 X
RAM: 32 GB ( 4 x 8 GB Ballistyx))
GPU: Radeon RX 6700 XT
PSU: Corsair RM 650x
 
Last edited by a moderator:
My thought is that something is loose.

Warming things up causes expansion and whatever is loose tightens up and makes contact so all will work agaIn.

Until the next cool-down...

Power down, unplug, open the case.

Clean out dust and debris.

Verify by sight and feel that all connections, cards, RAM, jumpers, and case connections are fully and firmly in place.

Use a bright flashlight to look for signs of damage: bare conductor showing, melting, pinched or kinked wires, corrosion, cracks, moisture, swollen components, loose or missing screws, sagging supports.

Also: how old is the PSU? History of heavy gaming use? The problem could be in any number of places but an old or heavily used PSU would be a starting suspect.

Especially if you are blowing the warm air in via the PSU.
 
My thought is that something is loose.

Warming things up causes expansion and whatever is loose tightens up and makes contact so all will work agaIn.

Until the next cool-down...

Power down, unplug, open the case.

Clean out dust and debris.

Verify by sight and feel that all connections, cards, RAM, jumpers, and case connections are fully and firmly in place.

Use a bright flashlight to look for signs of damage: bare conductor showing, melting, pinched or kinked wires, corrosion, cracks, moisture, swollen components, loose or missing screws, sagging supports.

Also: how old is the PSU? History of heavy gaming use? The problem could be in any number of places but an old or heavily used PSU would be a starting suspect.

Especially if you are blowing the warm air in via the PSU.
Also probably a cold solder joint somewhere, possibly at VRM or PSU.
 
  • Like
Reactions: artk2219
Also probably a cold solder joint somewhere, possibly at VRM or PSU.
Has to be one heck of an air blower to fix a cold solder joint, it would melt anything plastic inside the case before fixing the joint.
It's either the bad connection that ralston mentioned or it could be a capacitor in the PS or on the mobo that is only a bit worn out and still works once it reaches working temps.
 
  • Like
Reactions: artk2219
It's probably a loose connection somewhere due to the system's age (by the components i assume its at least 4 years old). There are cases when the slots in the mobo start to widen over time or with repeated pugging/unplugging of things and they don't make good contact.
I had this happen with the ethernet port in an old mobo. It was losing connection at random and it would come back if i fiddled with the cable a bit because the port got loose.
 
  • Like
Reactions: artk2219
Has to be one heck of an air blower to fix a cold solder joint, it would melt anything plastic inside the case before fixing the joint.
It's either the bad connection that ralston mentioned or it could be a capacitor in the PS or on the mobo that is only a bit worn out and still works once it reaches working temps.
Doesn't have to melt or re-solder it,just heat enough to expand metal.
 
  • Like
Reactions: artk2219
Thanks for your replies.
To answer a few remarks:
- the system is indeed now 4 years old
- the heatgun is actually quite modest: it is only a 150 W hobby heatgun and I use it at a distance of around 30 cm for 2 or 3 minutes
- the PSU is located in a separate compartment at the bottom of the case and I apply the heatgun via the side of the case, so I think that the warm air from the heatgun will hardly reach the PSU
- when I put power on the system (via a separate switch between the computer and the electricity net) and before actually starting the computer, the RGB lights in the case already turn on: so the PSU is OK?


I will check all connections as advised by Ralston and I will give feedback on the result.

Is it logical to think that the problem is located near the BIOS chip given the fact that the computer does not keep date and time anymore (during the night)?
 
Last edited:
This:

"when I put power on the system (via a separate switch between the computer and the electricity net) and before actually starting the computer, the RGB lights in the case already turn on: so the PSU is OK?"

What is the reason for doing that?

Is that also how you turn the computer off?

Do you use the Windows shutdown menu to shut down the computer?

Windows requires some time to properly shutdown and prepare for the next boot. If power is prematurely cut then that can lead to problems via corrupted files.

There may be more going on than what seems to be....
 
After start-up, the computer works flawlessly the whole day (even with heavy loads)
To localize the problem area, buy a can of freezer spray and cool down individual components one at a time until the computer shuts down (or not). Unfortunately it's not cheap.
https://www.amazon.com/Techspray-1672-10S-Envi-Ro-Tech-Freeze-Spray/dp/B00Q882TBQ

Alternate use of a low power hot air source with a small nozzle to direct heat and freezer spray are standard techniques to isolate individual faulty components.

View: https://youtu.be/J9jROADpY2k


I used this method on a CD player which wouldn't work in winter until it had been warmed up. The culprit was a large controller chip which I replaced.

In your situation, if you find a problem on the motherboard, the easiest option is to replace it. Might be an idea to try a different CPU, new RAM and GPU first.

the computer does not keep date and time anymore
How sure are you that the "new" CR2032 batteries were healthy. It's a long shot because the shelf life of these batteries is measured in years, but I always check batteries with a multimeter when I get them out of the blister pack. If they're approx. 3.25V they're new. When they drop below 2.75V I replace, although they might continue working for another year.

I had one PC where the side contact inside the battery holder was not touching the (healthy) battery. It was bent out of shape. The result was the BIOS kept on losing the time and date, because the positive side of the battery was isolated from the BIOS chip.
 
To Ralston:
I started with this extra power switch between the computer and the electricity net after a PSU spontaneously caught fire some years ago, even when the computer was turned off.
I turn off the computer via the Windows menu and only after everything stops (and the monitors are in standby) I cut the power to the system. I never cut the power prematurely.

To Misgar: Yes, the CMOS battery was my first suspect also. That is why I bought a new 2032 battery a few weeks ago, but no improvement.
Last week I bought another battery from another store and another brand, but still no improvement.
I do not have a multimeter, but it does not seem very likely (although not impossible) that 2 new batteries (A-brands), bought from 2 different shops would both be broken.
But I will inspect the battery-holder for problems.
 
My bad, i guess I didn't read the whole first post, because trying to fix something electronic with a heat gun is not something i consider proper. Doing it couple of times, even maybe for a week or two - ok, but for months? - come on. I'm not trying to hurt anybody's feelings, but what has to happen in order to seek proper repair? Yeah, it's nice if you can fix it yourself, but that doesn't always work.
Anyway, if you have some time to waste, you could take the components out of the case, and heat them individually with the heat gun. And hopefully that way narrow it down.
 
To Logan: no, my feelings are not hurt.
In the beginning the problem occured only once every 2 or 3 weeks (so it was not really a problem).
The situation deteriorated gradually and only the past 2 weeks the computer did not start anymore without heating (but it would still boot into Windows), and since 1 week I get only the BIOS screen after the warming-up.
In order to improve my understanding of computers,I always try to find a solution myself , but I did not see solution for this problem. Therefore I decided to reach out to the great experts on this forum to hear and learn new things, instead of taking the easy road and bring the computer to a repair shop.
 
But I will inspect the battery-holder for problems.
This is where I found a multimeter useful. The faulty battery contact looked OK, but when I connected the positive meter probe to the side contact, it read 0V instead of roughly 3V. When I moved the probe to the top of the CR2032, the meter read 3.25V. A gentle tweak with a small flat bladed screwdriver and normal operation was restored.

Of course it's equally likely there's nothing wrong with your CMOS battery holder and something else has failed.

If you can fit a metal nozzle for your heat gun, try to localize the heat to one specific area, rather than heating the whole case. If necessary, divide the interior into "compartments" with peices of cardboard to stop the heat spreading.


when I put power on the system (via a separate switch between the computer and the electricity net) and before actually starting the computer, the RGB lights in the case already turn on: so the PSU is OK?
All this proves is the +5VSB Standby rail is working. It doesn't prove that the main +5V and +12V rails are fully operational. However, your computer runs fine for hours on end once it's warmed up, so the PSU is probably OK.

PSUs have a finite lifetime and secondary (low voltage) capacitors tend to age. The result is higher ripple voltage on the DC supply rails and degraded performance. You can't measure ripple voltage on a multimeter set to DC Volts. You really need an oscilloscope to check ripple. N.B. This is unlikely to be your problem.

It's possible your PSU is faultly, but I suspect the motherboard. The best way to check is by slowly replacing each component until the fault disappears (can get very expensive).

Do you have a spare PSU?

Have you tried booting the system with only two DIMMs (in slots A2 and B2)?

Are you running with XMP/EXPO/DOCP enabled. If so switch off memory overclocking.

Have you run MemTest86 (boot from USB) to confirm the RAM is OK. 4 DIMMs can cause problems at high XMP speeds, whereas 2 DIMMs are often more stable.

Are the 4 DIMMs from a single kit of 4 matched modules, or two pairs of DIMMs with potentially different memory chips and different timings?

Do you have a spare graphics card? It doesn't have to be anything fancy.

Do you have a spare (compatible) CPU?
 
After inspecting the computer interior closely, I found white corrosion on several aluminium items / connectors and on the frame of the case. Probably due to the high humidity (usually 80% or more) on my location (close to the sea).
The CMOS battery holder did not look bright and shining, so I cleaned it and placed (again) a new battery.
Furthermore I disconnected and reconnected all cables to the mother board and PSU (the Corsair PSU is modular).
Perhaps the slowly increasing corrosion caused accordingly increasing electrical connection problems.

Unfortunately, I will only know tomorrow morning if the situation is better.

To answer Misgar's questions / remarks:
- the 4 DIMM modules came in a single kit of 4 matched modules
- I did a few exchanges with the modules, but that did not make a difference.
- the memory is not overclocked
- no, I don't have a spare graphics card or PSU

I tried to find the location of the problem by warming up specific components, but I got no clear solution. Your suggestion to use cardboard to create a "compartment" is a good idea !
If the computer does not start tomorrow morning (in spite of the corrosion removal action), I will use that approach.
 
Good morning,

It certainly is a good morning, because my computer started immediately and without problems (and without using the heatgun) this morning!
So disconnecting, cleaning and reconnecting all plugs and cables solved the problem.
Many thanks for your advices, comments and recommendations !
 
Status
Not open for further replies.
OSZAR »