CELL/RSX Brace

[TriForcer]

Is this something to have or 'needed'? I haven't seen the RSX variant before.
https://www.ebay.com/itm/306010086859

 

[Ganjo]

Ive never tried a RSX brace but the RSX isnt the source of heat in most units, its the CELL. I have tried CELL braces multiple times and it did not make ANY difference in temps. In BC models the CELL will hit 67C (if on webman) in 99% of consoles. Maybe this works better on later fats or slims but it does nothing in BC fats.

If you want to safely lower temps in your console undervolt the processors. They run them at a much higher voltage than it needs

 

[Joe Robertson]

Ive never tried a RSX brace but the RSX isnt the source of heat in most units, its the CELL. I have tried CELL braces multiple times and it did not make ANY difference in temps. In BC models the CELL will hit 67C (if on webman) in 99% of consoles. Maybe this works better on later fats or slims but it does nothing in BC fats.

If you want to safely lower temps in your console undervolt the processors. They run them at a much higher voltage than it needs

Currently there is no undervolt implementation that doesn't require soldering, delids are usually the most accessible way to lower temps when a repaste doesn't make a difference since most people don't know how to solder and have something like a raspberry pi to flash syscon, which is a shame since doing a delid on either chip poorly is very likely to permanently kill a console

Also in my experience the RSX is the hottest chip on 65nm CELL - 90nm RSX fats, well maintained 65-65nm fats since the RSX gets less airflow than the CELL, well maintained CECH2000 slims due to the 40nm CELL 65nm RSX config and CECH3000 slims thanks to the CELL having a soldered IHS while the RSX doesn't

 

[TriForcer]

I ordered a pair because I'm curious about them. Since I've already done 'delid' many times (have the right equipment) and also undervolting. Most interested in RSX brace against possible bump-gate problems.
Probably no difference at all - but you have to try it sometimes - for curiosity's sake.

 

[Joe Robertson]

I ordered a pair because I'm curious about them. Since I've already done 'delid' many times (have the right equipment) and also undervolting. Most interested in RSX brace against possible bump-gate problems.
Probably no difference at all - but you have to try it sometimes - for curiosity's sake.

The bracer is good to prevent warping and lower temps by a degree or two but it rlly wont help with bumpgate, 90nm RSX's are for the most part doomed to never have a runtime higher than 200-300 days no matter what you do

The 90nm RSX's underfill causes a tensile stress that's basically trying to push the die upward off of the substrate when it's cold and provides no support to the bga bumps when it's hot, causing a push and pull scenario that is perfect for fatiguing metals, a bracer cant do much against that

 

[TriForcer]

Well - they're on their way now, so I'll have to try them out on one of my PS3s. Thanks for all info.

 

[iKon]

The 3D printed bracers are not really a cooling mod. They are mainly there to help reduce motherboard flex and improve support around the CELL/RSX area.

The stock backplates only apply pressure through a limited number of standoff points. A properly designed bracer helps spread support more evenly and adds controlled center support under the die area, so the board stays flatter instead of flexing as much under clamp pressure and heat cycles.

So no, I would not advertise them as something that lowers temperatures. That is not the point. The cooling still comes from good thermal paste, clean heatsinks, proper IHS contact, fan/PSU choices, and good voltage setup.

The bracers are more of a mechanical support upgrade. I 3D print them in ASA, usually a plate full at a time, and I add them during Franky builds because they are cheap insurance against board flex. They are not magic, but I do not see a downside when they are fitted correctly.