Friday, May 6, 2011

Thermal probe diagnosis tool


I have made a direct application of my brand new thermal probe tool !

I had to repair a 60Go PS3 which was in YLOD.

I repaired it, and when I started it over, fan started to increase regurarly to reach full speed and to finally shut down the PS3 with the thermal warning message in the XMB.

So I probed the thermal probes of this PS3 and captured my tool's log in order to extract data. After some manipulations of the data I created 3 files for each probe of the PS3. Here are the graphics for each probe of the PS 3 :

Motherboard Probe graphic :

CELL Probe graphic :

RSX Probe graphic :

Measures are pretty relevant and thanks to them I clearly see that the CELL processor has its temperature rising and reaching the 85°C (which is the default thresholds of the thermal probes components) that generates the thermal alert interruption signal to the chipset, which leads the system to shut down with thermal alert message.

This confirms the fact that I used flux for the RSX during reflow and not for the CELL. I will have to run it again with flux on the cell this time, and I will check if there is some short circuit near the CELL.

So it is pretty fun to see that my tool works for diagnosis!


  1. I noticed the same thing, CBE gets hot much faster than RSX; even at XMB.

  2. Yes but in this case this is an abnormal behavior. Next time I get a PS3 working properly I will trace temperatures to see the normal behavior to have some kind of reference.

  3. Merci beaucoup pour ce blog très intéressant !

    J'ai du mal à comprendre en quoi l'utilisation de flux pour le reflow changerait quelque chose à la température du CELL ?

    J'ai lu que la vitesse de rotation du ventilateur était bridée pour que la console reste silencieuse. On pourrait donc imaginer de remplacer la thermorégulation de la ps3 avec un circuit externe qui utiliserait les sondes présentes sur la carte mère ?

  4. Ce que je voulais dire c'est qu'en ne mettant pas de flux, le reflow à peut être créé un court circuit entre 2 billes non critiques (donc pas de mise en sécurité), et que du coup cela fait chauffer le CELL anormalement.

    C'est une des évolutions possibles de mon outils.
    Je vais essayer de le porter sur quelque chose de plus simple !

  5. I think this whole mess were caused by the firmware itself (source codes bugs or intentionally; more YLOD, more sales), it should make the fan run faster. I noticed that the heatsinks were poorly designed, it should have much more vertical fins. I would prefer to remove the RSX heat spreader and replace it with the heatsink, same like ATI/nVidia's graphic cards design.