PS3 Do some PS3s run hotter than others?

R

raidriar

Member
I have a bunch of A01, B01, and E01 units. They are all delidded, all running the same PSU, all running same fan. I have spent like $50 on different thermal compounds trying to find the best one, including liquid metal. No matter what combination I used, I have no difference after testing MX4, AS5, Antex Formula X, liquid metal, Chinese paste. Within each individual unit, they run the same temperature/fan speed in the same test loop. Some units will run TLOU title screen at 28%, some will run at 31%, I have one particularly bad unit that will run it at 34% (under 68C limit).

Is it time to stop chasing paste and start accepting the heatsink differences between units may be to blame, or binning on the chip may require different voltage to operate? I have one particular unit that never cracks 30% fan speed when playing TLOU which is crazy to me, while I have another that's always bouncing between 31 and 36%. I always apply thermal paste the same way: thin initial layer then another couple dots so that it spreads nicely when the clamps go on.
 
Hey! Im new to ps3 hardware but from the little I have seen, yes, some models give about less or more heat than others. I have seen some models with thermal and electrical paddings and others without on certain components. So, that directly means that either Im looking at ps3's that have been opened before or that as the general system temp is lower on some models there is no actual need for thermal pads on certain components or that sony simply said to the pit with it and began producing without them possibly due to general hardware design. Of course a different heat sink will also comply.

Also, never trust fan speed measurements, they are never spot on.
 
I have/had two PS3 Slim 2501x (A and B), same manufacturing date. The B one is dead at the moment, got it cleaned up and applied Arctic MX-5 with good results. TLOU never got the fan to throttle full speed.

The A on the other hand, is a different story. Same procedure, same thermal paste (although I've run out so could have less amount than recommended), with acceptable results. TLOU got the fan to throttle full speed in some occasions during complex scenes of the game (explosions, fire, zombies).

I know that my experiences are inconclusive but I believe the amount of thermal paste is key. The heat sinks are to be blamed too, but I don't have other models to test.
 
Glad you guys have chimes in. It seems there is variance among units even if the same paste and method are used. I just don't know if I will blame the heatsink or the actual chips themselves. My IHS are very secure to the heatsink, so that should not be a variable when swapping the heatsinks, and repasting the cell and RSX is very straightforward. I will swap the clamps and screws as well just to rule those out as variables as well.

I hate variance like this, it was my goal to eliminate it completely. However it seems it may not be possible
 
raidriar said:
Glad you guys have chimes in. It seems there is variance among units even if the same paste and method are used. I just don't know if I will blame the heatsink or the actual chips themselves. My IHS are very secure to the heatsink, so that should not be a variable when swapping the heatsinks, and repasting the cell and RSX is very straightforward. I will swap the clamps and screws as well just to rule those out as variables as well.

I hate variance like this, it was my goal to eliminate it completely. However it seems it may not be possible
Click to expand...
I am right there with you in blaming the heatsink... the contact that is. I've been saying this for a second. The better you can improve the pressure to the heatsink, the better your temps will be.

To answer your other question, I do believe that some consoles definitely run hotter than others. There's a little bit of a temp difference in the early fat boys vs. the K, L, and P models which run slightly cooler. Then I feel like the 2000-2001 slims run about the same temps as the K, L, and P from my experience. Then the 2500-2501 units run a little cooler than those. Lot of that is due to design i'm sure and other factors such as the caps being used (mainly changed in the 2500-2501 units from my understanding). Not sure about the higher models as I pretty much only mess with CFW units.

But at the end of the day, I truly think heatsink contact is the number one thing one should be looking at for reducing temps. That's just my opinion that I've came to from working on several different models. Best reductions in temps i've gotten, happened when I improved contact with the heatsink doing things like the "eraser mod", although I prefer to use thermal pads in place of an eraser. I'd love to get more ideas floating around the scene with how we could improve the heatsink contact. I've seen incredibly drastic changes with this. Had a G model I was fighting a few months back, it was pushing 80% fan speed, and 75-80 deg temps on the cpu even after a de-lidding. I decided I'd try the eraser mod even though I was reluctant to do so, and I miraculously saw those temps go down to the 66-68 deg range, and the fan chilling between 30-40%. Improving heatsink contact is the way! I stand by that!
 
For reference, my fan speed variance to maintain under 68C is not very wide. Only one unit is bad, the majority are within 1-3% of each other.
 
Ok so small test: I swapped JUST the heatsink assembly. Same clamps, same screws, repasted above and below IHS. I even reused the same GPU heatsink. So really only the CPU cooler changed. I can confirm it is indeed related to the heatsink and not the actual chips. This console still ramps up to 36% at TLOU title screen, BUT CPU temp is now at 63 instead of 66, indicating there was something wrong with the heatsink contact area or heat pipe.GPU cooler is causing the fan to ramp up to 36% to hold under 68C, so I know it's not the paste or the chip. There must be some deformity or failure where the block and heat pipe connects or heat pipe has gone bad somehow because this is the only unit I have that ramps all the way up like this. Everything else I own will do 29-31% fan speed on the title, but his lemon heatsink needs 36% to accomplish the same temperature hold. Guess I'll buy a new heatsink assembly to see if it does any better, but it's really roulette at this point
 
raidriar said:
Ok so small test: I swapped JUST the heatsink assembly. Same clamps, same screws, repasted above and below IHS. I even reused the same GPU heatsink. So really only the CPU cooler changed. I can confirm it is indeed related to the heatsink and not the actual chips. This console still ramps up to 36% at TLOU title screen, BUT CPU temp is now at 63 instead of 66, indicating there was something wrong with the heatsink contact area or heat pipe.GPU cooler is causing the fan to ramp up to 36% to hold under 68C, so I know it's not the paste or the chip. There must be some deformity or failure where the block and heat pipe connects or heat pipe has gone bad somehow because this is the only unit I have that ramps all the way up like this. Everything else I own will do 29-31% fan speed on the title, but his lemon heatsink needs 36% to accomplish the same temperature hold. Guess I'll buy a new heatsink assembly to see if it does any better, but it's really roulette at this point
Click to expand...

I swapped the PSU on one of my BC PS3's from a later 40gb model. It helped as well to keep the temps down a tad. Also did the 19 blade fan swap but honestly I do not think that actually does anything it just came from a donor system.
 
atreyu187 said:
I swapped the PSU on one of my BC PS3's from a later 40gb model. It helped as well to keep the temps down a tad. Also did the 19 blade fan swap but honestly I do not think that actually does anything it just came from a donor system.
Click to expand...
for me PSU made no difference at all between APS 226 and APS 231, the temperatures are the same on cell and rsx, but with aps 231 of course it is cooler to the touch. The fan also had no impact on thermal performance, only on acoustics. 19 blade fan noise is more acceptable to my ears. Even thermal paste is not the problem. I have tried so many, they all perform within a degree of each other so I will call that margin of error. I think the biggest problem is heatsink fit and actual heatpipe problems. I tried with pressure paper before, none of my PS3s were consistent with each other.
 
Last edited:
raidriar said:
Ok so small test: I swapped JUST the heatsink assembly. Same clamps, same screws, repasted above and below IHS. I even reused the same GPU heatsink. So really only the CPU cooler changed. I can confirm it is indeed related to the heatsink and not the actual chips. This console still ramps up to 36% at TLOU title screen, BUT CPU temp is now at 63 instead of 66, indicating there was something wrong with the heatsink contact area or heat pipe.GPU cooler is causing the fan to ramp up to 36% to hold under 68C, so I know it's not the paste or the chip. There must be some deformity or failure where the block and heat pipe connects or heat pipe has gone bad somehow because this is the only unit I have that ramps all the way up like this. Everything else I own will do 29-31% fan speed on the title, but his lemon heatsink needs 36% to accomplish the same temperature hold. Guess I'll buy a new heatsink assembly to see if it does any better, but it's really roulette at this point
Click to expand...
Check the flatness of the heatsink surfaces, and also if the surfaces have some mark from the mechanizing process (scratches either in paralell or concentric)
And check also the "thermal adhesive" used to glue the copper heatpipes to the aluminum and all the metal pieces together, that glue is bad quality, sometimes it gets dry and could crack, sometimes there are also some "void" areas that was not filled with glue because the amount of glue used in factory was insufficient or because was applyed by a machine a bit randomnly

Edit:
The flatness of the surfaces usually affects the IHS (not the heatsink)
The scrathes depends of the heatsink revision (some of the heatsinks for PS3 slims looks horrible)
And the glue is a lottery... it seems they was not doing an strict control of his application in factory
 
sandungas said:
Check the flatness of the heatsink surfaces, and also if the surfaces have some mark from the mechanizing process (scratches either in paralell or concentric)
(...)
The scrathes depends of the heatsink revision (some of the heatsinks for PS3 slims looks horrible)
(...)
Click to expand...

True that. The heatsink surface is very rough, unpolished and irregular. The thermal paste can't compensate that kind of irregularities, not even by applying a lot of compound, so I would say a little sanding will be required in order to improve the contact surface.
 
I used a metal pad to lightly sand down the contact area on GPU and CPU, no improvement in temperature. I'm convinced it's something in the heat pipes themselves or how they were affixed to the rest of the heatsink.
 
You must log in or register to reply here.

Similar threads

Back
Top