CECHC04 (almost) instant overheating

Bozo

Forum Noob
Hello guys, i have a CECHC04 that i'm working on.
When I got it it was just instantly overheating after turning it on - since it's such an old model I knew that delidding was going to be necessary. So I did that.

After delidding the console worked and the temps were under control, but soon after that it started getting NEC related YLODs. So I replaced some NECs and it started working again.

Since then I decided to give it to my cousin, but before that I decided to replace all the remaining NECs to make sure that it doesn't spontaneously die on him. So I did that.

Since the replacement the console acts like it's not delidded again. The first startup it will work (booting up normally, display included) for like 20 seconds before it starts getting increasingly loud and then dies. Any subsequent attempts to turn it on will result in the same happening but this time after around 5 seconds. For it to work long enough to display anything I need to wait some time before turning it on again.

Does anyone have any ideas? I made sure to clean out the old paste and apply new (even better brand) before putting it back together. Thanks in advance!
 
I would dissassemble and assemble it again keeping attention at how all the parts fits with each others, maybe there is something out of place "pushing" the heatsink so it doesnt makes a good contact
The tokins replacement went fine and doesnt seems related with the actual overheating

Edit:
I mean... pushing the heatsink in a wrong direction (lets say... laterally), in that case the heatsink doesnt applyes the correct pressure
 
Well I tried just that and it didn't help. So i thought that maybe i should tighten the screws more (i also tried bending back plates of the cooler to make them apply more pressure) and i think i overdid it - now the console has instant ylod. I'm afraid I might've applied too much pressure :(

Edit: Seems I was wrong - The console was acting dead no matter how many times I tried to turn it on, but then when my dad came by to see what I was working on i turned it on to show him it was dead and then it turned on, loaded webman (i saw the yellow light flash a few times before turning back to green) and overheated once again.

So now I'm completely at a loss as to what is going on with this thing.

Edit #2: Now it started acting differently again - after it's been off for some time it starts to boot up (display lights up) and then it dies after 6-7 seconds after powering on. The fan doesn't spin up gradually however, but it shoots up near max from being very quiet for only half a second before dying. I don't understand this at all. If I didn't replace them all i'd blame NECs, but now - I have no idea.
 
Last edited:
Replace all the remaining NECs. Also, didn't you check the temps on webman?

I've replaced all of the NECs, although I've replaced them with 4x330uF and not 4x470uF, since 470uF ones cost more and the NECs are 1200uF anyway.

And about webman - the console remained on for long enough to check temps only once, and they were like 37°C on CELL and 83°C on RSX, so there's something wrong with RSX's cooling - I don't know what though. I've applied fresh thermal paste both under and on the IHS. I've also tried realigning the IHS several times, but it didn't help.
 
This looks like a bad contact between the IHS and the CELL die. That 90nm die warms almost instantanly, so after a delid you need to add some stuff to get the pressure you had when the IHS was glued.


Watch that video and see what you need to add. Also, use good quality thermal c. on both IHS and attach them to the heatsinks with pressure, so the next time you open up the console, you won't need to touch the IHSs. Just like in the video, and that's the correct way from now on.

I use little piece of plastic, like a credit card for that little amount of pressure you need on the CELL. You should upload pics of the mobo, and a video of the actual shutdown.
 
This looks like a bad contact between the IHS and the CELL die. That 90nm die warms almost instantanly, so after a delid you need to add some stuff to get the pressure you had when the IHS was glued.


Watch that video and see what you need to add. Also, use good quality thermal c. on both IHS and attach them to the heatsinks with pressure, so the next time you open up the console, you won't need to touch the IHSs. Just like in the video, and that's the correct way from now on.

I use little piece of plastic, like a credit card for that little amount of pressure you need on the CELL. You should upload pics of the mobo, and a video of the actual shutdown.


You may be right, but after I replaced just the top NECs the console worked just fine (it was after the delid). Perhaps I was lucky with the alignment?

Like I said, that one time console booted up for long enough to check temps the temps on CELL were perfectly normal (37 degrees), but RSX was melting (83 degrees), so I wouldn't point fingers at CELL for these shutdowns.

Thanks a lot for the input, maybe it will be possible to solve this mystery - I'd be really sad to se a BC console go.

As for the video of the shutdown and pics of the mobo I'll post them as soon as I can.
 
Woah thats wayyy too much paste especially on GPU but CPU even tho its too much should still work ok. clean it all off with IPA and start again.

See how the paste is spilling over the sides of the dies and the rams? u only need to put enough on there so it only Just spills out and i means JUST. I basically only use a grain of rice size on the dies and a pea size amount on the IHS's and thats heaps. The pressure from the clamps will spread the paste perfectly as is so no need to spread it by hand. The rams only need a tiny about, almost like match head size if that. I dont agree with how the video above does his consoles but everyone will have different methods that work for them so really there is no kinda wrong way as long as u dont use too much and u get perfect contact when the IHS's, Heat sink and dies all make contact with each other properly.

See the photos i circled is where it looks like there is hardly any good contact being made see if u can improve this when u redo the paste. Make sure u have no thermal glue left over on the rams, the top right rams looks like theres some left in the corners??? i cant see until u remove the paste properly but its possible u have some still left thats keeping the IHS to not sit flush.

20200503_195437 - Copy.jpg
Inked20200519_233157_LI.jpg
 
Okay, I've added pics with the paste cleaned up to the google drive folder that I've linked. I've also added my thermal paste appllication pics there.

Weirdly, now the fans don't gradually ramp up, but they just go loud for a second before it dies.
 
What thermal coumpound is that? Even if it were bad quality, you shouldn't get a YLOD due overheating that fast if the RSX is not so well attached with the IHS, the CELL's the one that overheats instantanly if it's not well attached.

Try putting some plastic like I said, but in the RSX part, on the right side.

https://drive.google.com/file/d/1BCdymlEz4lXxs4ip1D_03GsgV3bjzZnF/view

Then, try again and check the time with the system totally cool. Also, don't put too much, and don't put it on the rams, there's no need.
 
Thats heaps better the amount u used, a little much on the rams but thats good u have them on there they do need some sorter protection as well but they aren't as crucial as the dies.

Try removing all the silicon glue from the CPU that we can see and the under side of the IHS, if ur not sitting the IHS properly this can cause it to sit off the die so removing it ensures even if u misaligned it slightly, it wont be hinded by these. u should be able to at least scratch the majority of it off just with ur finger nail, then clean off with IPA.
 
Thanks for all your input, I'm really confused by they way this console acts.

Thermal compund that I'm using is thermal grizzly cryonaut, so I wouldn't blame it for the overheating.

I've tried the plastic pieces like in the video on both rsx and cell, but sadly it had no effect whatsoever.

After trying to realign the IHS multiple times I started to suspect that maybe the cooler itself is at fault - maybe the heatpipes broke inside or something like that. So I swapped it for a cooler from a CECHA06 i have around, but the outcome was exactly the same.

So now I really have no idea what might be going on. Last thing that comes to mind is that maybe some sensor is damaged, but those are probably built into the chips so no way to check/replace those. Anyone have any other ideas? I'm ready to try anything at this point :p

EDIT: I've managed to capture webMAN temps on video - It's most definitely RSX that is overheating. I'm adding the link below.

https://drive.google.com/file/d/193...LBMl9yBdXM3GIhpBkVJEcpORMS9hlDWWjsBffRV5uOzYo
 
Last edited:
Nop, i can only insist in what i said before, it looks the heatsink + IHS + RSX DIE... are not seating well with each other
It could be caused by a deformation in the metal shields + heatsink assembly (what i suggested before), or there is some residues from the previous thermal paste (what wrx884 suggested), or maybe you are not screwing the bolts to the bottom, or dunno

Best thing you can do is to dissassemble and assemble the PS3 several times until you find where is the problem... but dont waste thermal paste while doing it
 
Now that you're mentioning this, one of my two A's die because of this problem. I have two A's that I revived with the tantalum replacement, and both needed a delid on CELL, after that, one was still instant overheating, just like your problem, so I delid the RSX on that one too, but with the messed up I damaged the die when removing that STUPID GLUE on the rams. So that one needs a RSX change.

Since you already delid the RSX, can you get another heatsink from other BC? If not then check with extreme detail what could be the contact problem. The console boots, so there's no bga issues. You could clean the PCB with Iso. alcohol on every thermistor too. Something is not right :/
 
@ElGris I've already tried to swap the heatsink - exact same effect. I've found another RSX IHS lying around, so I'll try swapping the IHS next.
 
@ElGris I've already tried to swap the heatsink - exact same effect. I've found another RSX IHS lying around, so I'll try swapping the IHS next.

If the A unit u used the Heat sink from worked fine use the same IHS from that system.

Do u have a straight edge u can use to put on the original heat sink and the IHS?? if so use that and see if u can find any concave spots on both. IHS's are not always 100% flat this is another reason why using thermal adhesive on the rams again will help keep it flat once back under pressure when it goes back together, but u shouldn't need it while just doing tests.
 
Thats showing and issue of heatsink binding issues.

Do a finger heat test on both cpu and gpu - need to take the blueray off to do this.

If they are cold then three possibilties are happening -
1. heatsinks are not binding properly causing the core to heat up and then shutdown
2. the thermal sensors are broken , damaged
3. the internal diodes are broken or short circuited causing a false postive

Send some photos of close ups of your heatsink and general board area

Thanks for all your input, I'm really confused by they way this console acts.

Thermal compund that I'm using is thermal grizzly cryonaut, so I wouldn't blame it for the overheating.

I've tried the plastic pieces like in the video on both rsx and cell, but sadly it had no effect whatsoever.

After trying to realign the IHS multiple times I started to suspect that maybe the cooler itself is at fault - maybe the heatpipes broke inside or something like that. So I swapped it for a cooler from a CECHA06 i have around, but the outcome was exactly the same.

So now I really have no idea what might be going on. Last thing that comes to mind is that maybe some sensor is damaged, but those are probably built into the chips so no way to check/replace those. Anyone have any other ideas? I'm ready to try anything at this point :p

EDIT: I've managed to capture webMAN temps on video - It's most definitely RSX that is overheating. I'm adding the link below.
 
@ElGris I've already tried to swap the heatsink - exact same effect. I've found another RSX IHS lying around, so I'll try swapping the IHS next.

This will not be the issue, you need to verify if the cell and rsx are getting hot at all? if they are not then the heatsink is not binding or worst the thermal sensors or internal core are naffed!

If u have a multimeter check for 3.3v voltage on pin1 of the thermal sensors (under the cpu and rsx - 6 pin)
 
Back
Top