PS3 (Research/Experimental) - NEC/TOKIN Capacitors Replacement - YLOD

There are fluxes that dont need cleaning indeed and I don't need to ask anyone to know this, they are called "no clean flux"
I don't try to be always right just because i am.

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk
 
Ok, I opened it up again and the + to + wires literally burned black. I now installed 6mm² wires (Audio cables) and it seems to work fine now.
And yes, I cleaned it with IPA, but I have no pics of the "after" state... The RSX wires are red, as they are on the pics, the CPU wires were green and red, but were black XD
20200315_233308.jpg 20200315_233311.jpg 20200315_233316.jpg 20200315_233319.jpg
 
Ok, I opened it up again and the + to + wires literally burned black. I now installed 6mm² wires (Audio cables) and it seems to work fine now.
And yes, I cleaned it with IPA, but I have no pics of the "after" state... The RSX wires are red, as they are on the pics, the CPU wires were green and red, but were black XD
View attachment 24776 View attachment 24777 View attachment 24778 View attachment 24779
Oof. Might be worth upgrading to either 14 or 12 gauge for bridging since that's about what I recall being mentioned before. That and try and use the shortest run of wire you can.
 
Well, 14 gauge looks to be about 2,5 mm² and 12 about 4 mm², so I am now well above that. So it should be alright, I guess?
My wire should correspond to gauge 10.

And I had to solder in all 32 tantal caps. With 30, the ps3 instantly did the YLOD thing.
 
Well, 14 gauge looks to be about 2,5 mm² and 12 about 4 mm², so I am now well above that. So it should be alright, I guess?
My wire should correspond to gauge 10.

And I had to solder in all 32 tantal caps. With 30, the ps3 instantly did the YLOD thing.
Good to hear that everything seems to be rolling. My guess is the caps you got aren't functioning at rated uf so it took more to hit the right totals.

Don't suppose you ever measured the rated va actual uf values of any of them?

Edit: also what were the original gauge on the wires that burned up?
 
Well, I indeed measured every single tantal, I don't know how accurate my multimeter is, but everyone of them showed between 480 and 510 µF on my meter.

The burned wire was pretty thin, I would guess about gauge 24 or 22. 21 at a maximum.
 
The fateful day has arrived! I ran out of Sega CD fuses, so I tossed a bunch of caps in my Mouser cart. I had to narrow down the scope of the test since name brand, high voltage tantalum is CRAZY expensive. I'm not paying 20 bucks a cap just for some fun tests. Now, before we get started, I am obligated to once again repeat: THE HEAT TEST IS MEANINGLESS. I've also gone through 4 more consoles since my last update, and zero TOKIN failures. If you're counting along at home, good for you, 'cause I gave up, but it's somewhere around 3 failures in 75 consoles now. Buy the damn oscilloscope.

I was going to wait to do this until I got another TOKIN failure in since I don't like doing work that I don't have to, but I waited over a month to no avail. I then decided I'd just go ahead and do it on my CECHE01 unit that I use for drive testing, but that led to an important discovery: the adhesives I use to replace the IHS plus the suction forces from fresh thermal paste are way too goddamn strong. So I no longer have a functioning CECHE01 testing unit. Fast forward a week or two, and I accidentally bought a non-backwards compatible unit that I have no problem with sacrificing.

Mods: I don't know if this will get flagged as spam eventually, but the only way I'm going to be able to keep all these images straight is if I make a post for each set as I do it. There will be quite a few over the course of the night. If you want to merge them all in to one when I'm done to keep this thread tidy, have at it.

So, let's get started. This is a perfectly functioning CECHH01 with a DIA-001 motherboard. No work has been performed. No BGA issues. All scope images will be from the unit idling at the main menu. I don't want to do it under load since that introduces variables.

CPU, original NEC/TOKIN caps:
vCSGUVB.jpg


GPU, original NEC/TOKIN caps:
zzNXQVs.jpg



I have to poop, then I'll rip them off and get started.
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

CPU, 2640uF (8x330uF) 2.5V, 25 mOhm ESR caps:
Re1AoVm.jpg


GPU, 2640uF (8x330uF) 2.5V, 25 mOhm ESR caps:
SvxDmzn.jpg
 
Great,can we see some pictures of your work on both sides with the tantalums?

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

CPU, 5280uF (16x330uF) 2.5V, 25 mOhm ESR caps:
0ZZ5EAy.jpg


GPU, 5280uF (16x330uF) 2.5V, 25 mOhm ESR caps:
ioVngCI.jpg
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

CPU, 10560uF (32x330uF) 2.5V, 25 mOhm ESR caps:
oVjHBF0.jpg


GPU, 10560uF (32x330uF) 2.5V, 25 mOhm ESR caps:
0jbWvpC.jpg



Great,can we see some pictures of your work on both sides with the tantalums?
Yeah, I'm okay if nobody ever sees the soldering job I did to put 64 B case caps on one side of the board.
 
Well, 14 gauge looks to be about 2,5 mm² and 12 about 4 mm², so I am now well above that. So it should be alright, I guess?
My wire should correspond to gauge 10.

And I had to solder in all 32 tantal caps. With 30, the ps3 instantly did the YLOD thing.

YLOD is not just a reason that you don't have enough capacitors. I had the same case and I just soldered the capacitor badly in some place -> the console cut off the power supply immediately for safety. I soldered the capacitors again and the PS3 started working
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

I screwed up and only bought enough for one chip each on these ones, so you're only getting the GPU.

GPU, 5280uF (16x330uF) 2.5V, 1.1 mOhm ESR caps:
EtKC9SC.jpg
 
Cpu side was cleaner with the tokins vs tantalums. Were the ps3 doing something else than idle at the time?

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

I screwed up and only bought enough for one chip each on these ones, so you're only getting the GPU.

GPU, 5280uF (16x330uF) 6.3V, 25 mOhm ESR caps:
2MF69ss.jpg



YLOD and the waveform indicates bad caps. I don't know how to explain this one. Plenty of people have successfully used 6.3V caps here already, right? I have no idea. These are Vishay caps from Mouser, so I doubt I got a bad one in the bunch. I'll report back if I come up with anything, but....I'm stumped. One last try with some electrolytics and then I'm done.
 
Heat test is meaningless, the TOKINs are fine, and Epstein didn't kill himself.

I screwed up and only bought enough for one chip each on these ones, so you're only getting the GPU.

GPU, 6000uF (4x1500uF) 16V generic electrolytics:
phIqkFZ.jpg

l2R1SWc.jpg


This was weird. It gave a perfect 1.02 volts, then tried one time to ramp up and failed, then gave that tiny data looking burst, then YLOD. About 2 seconds elapsed.

Finally, for dirks and digglers, here's a dead short with no caps at all:
ig69V5Z.jpg


It gave that one little squiggly burst, and that's all she wrote.
 
Conclusion: you're probably not going to get anything cleaner than the original NEC/TOKIN caps. Another reason to please leave them alone unless you've checked them with a scope.

Otherwise, to the surprise of probably nobody, get as close as possible to all of the major original specs (capacitance, voltage, ESR). No good comes of any change in either direction.

Electrolytics do not work. If you had some success from replacing half and half with them, it's definitely because you removed the bad cap (or the microscopic shift from reassembly :butterfly:), not because you added electrolytics. Stop it.
 
Last edited:
Cpu side was cleaner with the tokins vs tantalums. Were the ps3 doing something else than idle at the time?

All images when the system is working are from just idling at the main menu. Images from when it isn't working are from whatever I can capture in the 2 seconds before it shuts back off.
 
Back
Top