PS3 PS3 thermal paste replacement did not help, fan still speeds up

[DoublesAdvocate]

Repairing them, basically.

Had a few come in with liquid metal, had dried up, so not good.

Various articles have shown not a good idea to use liquid metal as an alternative.

Liquid metal doesn't dry up that's the whole advantage of it, if you put it on bare copper or aluminum it will alloy with it and destroy the metal but there's no issue using it between the IHS or anything else with a coating that the LM doesn't like to alloy with.
What articles?

 

[db260179]

Yes correct, long term i can't see the advantages to using it, and eating away the metal is not a good thing.

What temp reductions do you get? with using this?

Liquid metal doesn't dry up that's the whole advantage of it, if you put it on bare copper or aluminum it will alloy with it and destroy the metal but there's no issue using it between the IHS or anything else with a coating that the LM doesn't like to alloy with.
What articles?

 

[DoublesAdvocate]

Yes correct, long term i can't see the advantages to using it, and eating away the metal is not a good thing.

What temp reductions do you get? with using this?

Correct, which is why you do what I said anyone competent does who actually reads the instructions and use it only with metals that the LM does not like to alloy with such as Nickle.
5-10C
What articles?

 

[Dariush Aghaee]

There is also issues that occur with the IC buck regulator chips - the two black chips above the nec tokins - Over voltage or low voltage can cause the CELL or RSX to overheat. A simple resistance check on the nec tokins postive points - a good resistance is 2.5 ohms and above, below means there is issues.

I've repaired CECHHK models with the 65nm and barely reach 60C when running a heavy game.

.

Could you please share more details about those regulators and where exactly to measure the resistance?

Correct, which is why you do what I said anyone competent does who actually reads the instructions and use it only with metals that the LM does not like to alloy with such as Nickle.
5-10C
What articles?

You mean replacing IHSs with another thing that made from Nickel? Can we find something like LM friendly IHS?

 

[DoublesAdvocate]

Could you please share more details about those regulators and where exactly to measure the resistance?



You mean replacing IHSs with another thing that made from Nickel? Can we find something like LM friendly IHS?

The IHS of the PS3 is already fine since it has a Nickel plating, a bare copper IHS could be unsafe to use however.

 

[Dariush Aghaee]

The IHS of the PS3 is already fine since it has a Nickel plating, a bare copper IHS could be unsafe to use however.

And you try LM ON PS3s of yours and they didn't dry out. Right?

 

[db260179]

https://www.psdevwiki.com/ps3/Regul..._Buck_Multiphase_Optimized_LGA_Power_Block.29

Modern day term call them VRM, they regulate current to the RSX an CELL. As like overclocking modern cpu's with more voltage can create silicone breakdown and overheating affects the running.

So a PS3 CECHC i've just fixed now, the voltage reading was 1.222V+ on the 1.0V+ line, on the CELL cpu - when you look at the schematics it has two voltage parameters - 1.0V+ and 1.2V+, the 1.0V+ was 1.222V+ and the 1.2V+ was ok at 1.3V+ reading.

The charactisics of this issue is turn on ps3 and within 8 minutes or more overheat starts to occur. If you follow the recent thread with talking to the syscon, you can use the internal command 'tmp 0' to get current running temperature of te CELL

tmp 0
TZone No:00
1st BE Primary Temperature:67.70(0x43b7)

I've tweaked the default syscon fantbl to run as follows:

fantbl get 0
fancon No:00
P0: TempD:0.0(0x0000) - TempU:64.0(0x4000) duty:30%(0x4f)
P1: TempD:66.50(0x4280) - TempU:72.0(0x4800) duty:32%(0x54)
P2: TempD:68.50(0x4480) - TempU:73.50(0x4980) duty:33%(0x56)
P3: TempD:69.50(0x4580) - TempU:74.50(0x4a80) duty:35%(0x5b)
P4: TempD:70.50(0x4680) - TempU:75.50(0x4b80) duty:37%(0x60)
P5: TempD:71.50(0x4780) - TempU:76.50(0x4c80) duty:38%(0x63)
P6: TempD:72.50(0x4880) - TempU:77.50(0x4d80) duty:40%(0x68)
P7: TempD:73.50(0x4980) - TempU:79.0(0x4f00) duty:50%(0x82)
P8: TempD:75.50(0x4b80) - TempU:85.0(0x5500) duty:63%(0xa3)
P9: TempD:80.50(0x5080) - TempU:91.0(0x5b00) duty:100%(0xff)

So with the replaced thermal paste and replaced IC buck regulators it runs a constant 67-69C.
Before its temperature after thermal paste done, 76C. and fan speed at 50%.

To measure the voltage, just probe the NEC tokins nearest to the CELL (same applies to the RSX) - it should read 1.1V+ as a start
Measuring current is another possible issue and is leaking too much. Adding a resistor for high current to lower to 1.1V+ if its reading greater than 1.2V+.

As a ball park figure on resistance, measuring the 5V+ to GND on the IC Buck chips should show 217ohms - but voltage check is a better indicator.
I'll add a photo as an example

Could you please share more details about those regulators and where exactly to measure the resistance?

 

[DoublesAdvocate]

And you try LM ON PS3s of yours and they didn't dry out. Right?

Correct, what some people think is "drying out" is actually some of the LM alloying with the top layer of the die and the IHS.

 

[Nator]

Liquid metal has too many risks to use, so avoid!

A pro tip who wants to do this, delid cpu and gpu, use a good thermal paste - i use GD900 from aliexpress as cheaper alternative to MX4

And search for 'copper shims 42x42mm 1.2mm ps3', creates a great heat bridge and i've found a reduction of heat by 2%

Unfortunatley the early CELL and RSX 90nm are power hungry, heat monsters - if you are getting 70c or below you are doing well and its the best you will get with them. the newer 65nm and 40nm will run at 60c or lower

There is also issues that occur with the IC buck regulator chips - the two black chips above the nec tokins - Over voltage or low voltage can cause the CELL or RSX to overheat. A simple resistance check on the nec tokins postive points - a good resistance is 2.5 ohms and above, below means there is issues.

Hi. After deliding do i have to put the IHS back in place or is it possible to make the die contact directly with the heatsink? I'm also trying to figure out if liquid metal can be used with IHS and heatsink materials and the PS3.