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

Looking closely, it looks like it's actually one pointing the right way, and the others pointing the wrong way... Which is fascinating in its own way, since it all works still. I guess it goes back to the old question of "how much capacitance do we really need?"

Oh my.... holy moly, what have i done?!!?!!

I told you guys i have dyslexia, but wow this time... i might go cry about it lol.

Well back to the soldering room for me... see you guys in "many moments later"

I might not be back for a while, i need to go drown my sorrows in Kit-Kats & Coffee (without any creamer, as a way of self mortification lol)

EDIT: :cheerful: Thank-You, guys so much, truely i mean it, i would never have spotted those stupid mistakes without your help, and i would have continued testing the PS3 for days like that. Now i can go sort that out without causing any more damage. Back soon, all being well.
 
Last edited:
Oh fudge!!, well spotted bro, arrrrrrrr! ....oh man, i can't believe i didn't see that! :crybaby:

...darn it, i have to take everything apart again. :oops:

But thanks so much for spotting that, it's really appreciated. On a side note, i saw Sony had reversed the polarity on some of the Tantalums they had put in, and they put some of them in "series" see here in the photo below.

See the zone where the power rail feeds the Tantalums, all that is positive (+) and the ground (-) is only on the outside of that zone, and yet we see Sony put the two middle Tantalums (seen in the photo) in "series" and reversed polarity, those 2 middle tantalums don't touch any (-) ground at all...

View attachment 26335

They are in parallel, I don't have a PS3 slim board here, but the same was done on PS4 boards and all negatives stay on ground. These boards has many layers and in one of them is connected with ground. See it with your multimeter.
 
They are in parallel, I don't have a PS3 slim board here, but the same was done on PS4 boards and all negatives stay on ground.
Looking closely, it looks like it's actually one pointing the right way, and the others pointing the wrong way... Which is fascinating in its own way, since it all works still. I guess it goes back to the old question of "how much capacitance do we really need?"
you're right! I saw only tantalum position, not wich line this position was soldered!
 
They are in parallel, I don't have a PS3 slim board here, but the same was done on PS4 boards and all negatives stay on ground.

Well there is seemingly no ground in that zone (see the photo) only positive. Where are the 2 Tantalums in the middle getting their (-) ground from?

The 2 Tantalums on the outside are in parallel, yep, but the 2 Tantalums in the middle here seem to be in series?

Red zone = (+)
Black zone = (-)

rev-2.png
 
Last edited:
Well there is no ground in that zone (see the photo) only positive. Where are the 2 Tantalums in the middle getting their (-) ground from?

The 2 Tantalums on the outside are in parallel, yep, but the 2 Tantalums in the middle here seem to be in series.

Red zone = (+)
Black zone = (-)

View attachment 26336

See the other side of board, probably you will see points but connect your multimeter on orange circle and test the resistance, you will see they are on ground.

Sorry my bad english!
 
See the other side of board, probably you will see points but connect your multimeter on orange circle and test the resistance, you will see they are on ground.

Sorry my bad english!
Yeah... i've been pouring through the service manual, and i've seen several cases where the connection point is a via under the chip, and it goes some other place or across the board. It's amazing how much stuff they managed to cram in here.
 
See the other side of board, probably you will see points but connect your multimeter on orange circle and test the resistance, you will see they are on ground.

Sorry my bad english!

The other side of the board is the same, it shares the same zones for (+) and for (-). Meaning it seems there is no ground (-) where the 2 tantalums touch in the middle. It's interesting for sure.
 
Last edited:
Yeah... i've been pouring through the service manual, and i've seen several cases where the connection point is a via under the chip, and it goes some other place or across the board. It's amazing how much stuff they managed to cram in here.

Ahh i see, that explains it then, thanks for the info guys, most appreicated.
 
Last edited:
The other side of the board is the same, it shares the exact same zones for (+) and for (-). Meaning there is no ground (-) where the 2 tantalums touch in the middle. It's interesting for sure.

But i need to go fix my own PS3 problems lol, and not worry about this just yet, it matters not because most of us are replacing NECs on older boards anyways.

I tested in a PS4 board and all negatives are connected on ground.
https://www.psdevwiki.com/ps4/images/8/8f/SAA-001_-_Backside_-_IMG_2996.JPG
 
I tested in a PS4 board and all negatives are connected on ground.
https://www.psdevwiki.com/ps4/images/8/8f/SAA-001_-_Backside_-_IMG_2996.JPG

Many thanks, now i can see those (-) zones, like islands, within the (+) positive zones, and i guess the PS3 has the same thing.

Well thanks for clearing that up for me, now i can stop thinking about it, i was wondering what Sony had done there, and i don't have a newer PS3 boards to check for myself. Anyways i still need to sort out my 3 reversed tantalums on the RSX, i was tired at the time, and i was excited to finish quickly so i could show you guys the results. So ciao for now amigos.
 
Last edited:
Hiya All, success!! ...with Tantalums + MLCCs combination, only on bridge wire.

This is my original PS3 from before, i took off the "new" NEC/Tokins and tested this, and the PS3 works great.

Each NEC was replaced with the following:
- 4 x 470uF - 2.5V (Size B) Tantalums
- 4 x MLCC array:- 1 x 10uF, 1 x 1uF, 1 x 0.1uF, 1 x 0.01uF

The larger MLCCs are X7Rs, and the smaller ones are C0G/NP0 type.

The photo below shows the stage just before i soldered them to the PS3, some Kapton tape was used.

(There are 8 of these, 4 on each side of the board)
View attachment 26331

...and here below is the PS3 working, and i have tested it many times since, and it works great so far.

View attachment 26332

EDIT:- please forgive the 3 x reversed tantalums on the RSX i was absent minded when i did those ones lol. I will aim to fix that soon and add an extra 1 x 22uF MLCC to each array (like the NEC datasheet shows - see link below) and report back the results, all being well.

https://web.archive.org/web/2007121...-tokin.com/product/cap/proadlizer/test03.html

what is the name of those tantalum and where did u put mlcc
 
Hiya All, success!! ...with Tantalums + MLCCs combination, only on bridge wire.

This is my original PS3 from before, i took off the "new" NEC/Tokins and tested this, and the PS3 works great.

Each NEC was replaced with the following:
- 4 x 470uF - 2.5V (Size B) Tantalums
- 4 x MLCC array:- 1 x 10uF, 1 x 1uF, 1 x 0.1uF, 1 x 0.01uF

The larger MLCCs are X7Rs, and the smaller ones are C0G/NP0 type.

The photo below shows the stage just before i soldered them to the PS3, some Kapton tape was used.

(There are 8 of these, 4 on each side of the board)
View attachment 26331

...and here below is the PS3 working, and i have tested it many times since, and it works great so far.

View attachment 26332

EDIT:- please forgive the 3 x reversed tantalums on the RSX i was absent minded when i did those ones lol. I will aim to fix that soon and add an extra 1 x 22uF MLCC to each array (like the NEC datasheet shows - see link below) and report back the results, all being well.

https://web.archive.org/web/2007121...-tokin.com/product/cap/proadlizer/test03.html

This is some really nice work here man. What did you make those little boards out of, and where did you get the thin copper sheathing? This is the best idea here I have seen so far. This thread has grown into an amazing community effort. You guys should really be proud of yourselves.
 
This is some really nice work here man. What did you make those little boards out of, and where did you get the thin copper sheathing? This is the best idea here I have seen so far. This thread has grown into an amazing community effort. You guys should really be proud of yourselves.

do u know the name of that tantalum?
 
ofst[ 76]:err_code:0xa0403034, clock:0xffffffff = RSX fault

So, have you reflowed the RSX? - if not try and reflow first.

The timing IC's could be causing you and issue as well, check their voltages. But im afraid your 44 ohms is below a working tolerance, its possible one of the ram chips on the RSX has died.

Here's the log, and a few other things I got out of it

> errlog
errlog
ofst[ 80]:err_code:0xffffffff, clock:0xffffffff
ofst[ 84]:err_code:0xa0403034, clock:0xffffffff
ofst[ 88]:err_code:0xa0403034, clock:0xffffffff
ofst[ 92]:err_code:0xa0403034, clock:0xffffffff
ofst[ 96]:err_code:0xa0403034, clock:0xffffffff
ofst[100]:err_code:0xa0403034, clock:0xffffffff
ofst[104]:err_code:0xa0403034, clock:0xffffffff
ofst[108]:err_code:0xa0403034, clock:0xffffffff
ofst[112]:err_code:0xa0403034, clock:0xffffffff
ofst[116]:err_code:0xa0403034, clock:0xffffffff
ofst[120]:err_code:0xa0403034, clock:0xffffffff
ofst[124]:err_code:0xa0403034, clock:0xffffffff
ofst[ 0]:err_code:0xa0403034, clock:0xffffffff
ofst[ 4]:err_code:0xa0403034, clock:0xffffffff
ofst[ 8]:err_code:0xa0403034, clock:0xffffffff
ofst[ 12]:err_code:0xa0403034, clock:0xffffffff
ofst[ 16]:err_code:0xa0403034, clock:0xffffffff
ofst[ 20]:err_code:0xa0403034, clock:0xffffffff
ofst[ 24]:err_code:0xa0403034, clock:0xffffffff
ofst[ 28]:err_code:0xa0403034, clock:0xffffffff
ofst[ 32]:err_code:0xa0403034, clock:0xffffffff
ofst[ 36]:err_code:0xa0403034, clock:0xffffffff
ofst[ 40]:err_code:0xa0403034, clock:0xffffffff
ofst[ 44]:err_code:0xa0403034, clock:0xffffffff
ofst[ 48]:err_code:0xa0403034, clock:0xffffffff
ofst[ 52]:err_code:0xa0403034, clock:0xffffffff
ofst[ 56]:err_code:0xa0403034, clock:0xffffffff
ofst[ 60]:err_code:0xa0403034, clock:0xffffffff
ofst[ 64]:err_code:0xa0403034, clock:0xffffffff
ofst[ 68]:err_code:0xa0403034, clock:0xffffffff
ofst[ 72]:err_code:0xa0403034, clock:0xffffffff
ofst[ 76]:err_code:0xa0403034, clock:0xffffffff

> bestat
bestat
(Error State) (Unknown Error)

> powerstate
powerstate
ATA Power : OFF
PCI Power : OFF
RSX Power : OFF
XDR Power : OFF
Eurus Power : OFF
SB Power : OFF
RSX Thermal Sensor : UNAVAILABLE
BE Thermal Sensor : UNAVAILABLE

> disp_err
disp_err
CheckStop: None
PLLUnlock: 0
RSX Int: None
PowerSeq: 00
 
ofst[ 76]:err_code:0xa0403034, clock:0xffffffff = RSX fault

So, have you reflowed the RSX? - if not try and reflow first.

The timing IC's could be causing you and issue as well, check their voltages. But im afraid your 44 ohms is below a working tolerance, its possible one of the ram chips on the RSX has died.
Oh no! That's not great.

Nope, never done anything besides the yearly thermal paste switcheroo. As far as the IC's, I'm just about done testing all of ones for the DC-DC diagram. So far, the only ones that came up bad were the 6021 and 6022. Their outputs are marked as 1.5_Bridge and 3.3_Bridge, respectively. Once I'm done with those, I'll be testing the VRM vor RSX and Cell.
 
Here in this photo below, the Tantalums are only on the bridge wire...

I want to test this a little further to document the results for you guys, first results were good and the PS3 booted and worked fine, with 8 of these proto-types. I did one for every NEC replacement.

It's for those using bridge wires:- if we look at the way the NECs work, and what their role is, then adding a bridge wire can be counter-productive for our Tantalums unless done correctly. Because if we give the current the option (a route) to leap-frog (by-pass) the filtering capacitors and go straight to the RSX / CELL, then the processors won't work correctly, because the current is fluctuating all over the place and the line is too noisey.

Adding a bridge wire (incorrectly) is like having little, to no capacitors at all...

The current will always take the path of least resistance, so we need to force the current through the filtering process of the capacitors, with no other options / routes, and only then arrive at the RSX or CELL once filtered. That's how those processors want it so they can function correctly.

I tried this concept on my 1st PS3, i did all 8 x NEC replacements with this idea, Tantalums only on the bridge wire(s), like this photo is showing for 1 x NEC replacement, it was all like that (x 8) and it worked, the PS3 booted up, and worked fine! Originally for that test i had used size B Tantalums, and in this photo it's those 'Panasonic' size D ones - this photo below is just to give an idea of how the concept / idea was installed / looks)

However, the same day i had just received the "new" NECs from China in the mail, and i was too excited to wait any longer, so i took out my test bridge-wire Tantalums and soldered the "new" 8 x NECs, and that worked too, but now i can't de-solder the "new" NECs and they are stuck there, so i am working on another (2nd) PS3 to hopefully show the test results again, with photos etc, of this Tantalum only on the bridge wire idea working.

I also want to add an array of MLCCs (0.01uF to 1uF) to this proto-type, will aim to report back the results soon, all being well, but it might be some time yet because i have some other commitments (work) that i need to sort out first.

View attachment 26227

Here are all 8 of them, that i installed together for the original test, the PS3 booted and worked.
View attachment 26228

i just want to ask what is the name of those yellow things?
 
So i had received a fat ps3 cechhk with the same fault code

[POWERSEQ] Error : BitTraining BE:RRAC:RX0:GLOBAL1:RX_STATUS <---- southbridge having issues talking to the CELL
[SSM] state: 0104 -> 0304
[SSM] ssmCb_AfterBeOn2() called.
[SSM] PowSeq Fail : Detected !
[SSM] state: 0304 -> 0700
POWSEQ] AV Backend Letup
[SSM] Shutdown mode : syspm_stat=00000000/00000000
[ERROR]: 0xa0404401
[ERROR]: 0xa0403034 <---- RSX powerint pin not triggering
[POWSEQ] PowerSeq_Letup called.
[SSM] state: 0700 -> 0600
(PowerOff State) (Fatal)

So all of the voltages checked out except one, the timing chip, which lead back to the RSX powerint pin.

This board had two faults, the southbridge and rsx - both had poor bga connection, fixed by a reflow.

Check your syscon log, if you hit enter key after typing 'bringup' command in internal mode, this will show the boot process.

Hopefully for you it just needs a reflow or reball.

Oh no! That's not great.

Nope, never done anything besides the yearly thermal paste switcheroo. As far as the IC's, I'm just about done testing all of ones for the DC-DC diagram. So far, the only ones that came up bad were the 6021 and 6022. Their outputs are marked as 1.5_Bridge and 3.3_Bridge, respectively. Once I'm done with those, I'll be testing the VRM vor RSX and Cell.
 
Back
Top