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

[sandungas]

I removed the 4 tokins of the cell right now. There is no more continuity between v_in anfd v_out. So my yellow stipe assumption is a shit...

Good to know, can you check for confirmation if the 2 tokins of a side are connected with the other 2 tokins of the other side ? (by the VIAS next to v_in & v_out)

You know, is tricky, to check this is needed to place the motherboard in vertical (holding it with the legs, or the chin, lol), and touch the tokins edges with the mutlimeter tips with one hand at each side of the motherboard

 

[littlebalup]

Good found btw, the suggested schematics shown in that link seems to match fine with the modern PS3 models (without tokins), i wonder what effect does to have all that amount of small 10uf ceramic capacitors together... but seems to be something generally used in other devices

My understanding is the ceramics and tantalum capacitors do not have the same reaction time. So they use both in parallel to smooth the voltage waveform as most as possible to stabilize the voltage vs the cell/rsx consumption peeks.
Not using ceramic capacitors in add of tantalum may result of voltage variation. Too much variation may result in an in game YLOD.

Good to know, can you check for confirmation if the 2 tokins of a side are connected with the other 2 tokins of the other side ? (by the VIAS next to v_in & v_out)

Yes they are. The vias cross the entire board, bottom to top.

 

[sandungas]

Yes they are. The vias cross the entire board, bottom to top.

Thanks, i did some drawings today assuming there are 2 groups of tokins, and i would hate to have mistakes in them (because in that case i would need to remove or update them, but im not keeping a copy of the original .psd files ops, i was rushing a bit)

 

[Naked_Snake1995]

Thanks, i did some drawings today assuming there are 2 groups of tokins, and i would hate to have mistakes in them (because in that case i would need to remove or update them, but im not keeping a copy of the original .psd files ops, i was rushing a bit)

So in conclusion, this is much larger than i though [emoji23]

Sent from my G8141 using Tapatalk

 

[Revak3115]

I did this morning on the CELL side, and i must say Holy $h@t [emoji23]

No more annoying beeps on the shutdown, guess even i have to learn something new each day [emoji23]

Thanks for sharing the info, ill make sure to add it to the tutorial as a troubleshooting for this specific Partial/YLOD problem.

Sent from my G8141 using Tapatalk

Lol I guess I'm glad I helped you out in a way. I hope yours doesn't beep anymore and lives on for years to come

 

[Naked_Snake1995]

Lol I guess I'm glad I helped you out in a way. I hope yours doesn't beep anymore and lives on for years to come

It will, i am constantly looking at it, with that menacing stare, so it won't dare to die on me [emoji23]

Sent from my G8141 using Tapatalk

 

[SEGA-CD]

Has this been done on a system that has a full ylod? As in there is no getting the PS3 to boot at all no matter the amount of finicking you do, it just immediately yellow lights when you hit the power button and flashes red until you shut it off? Most of what I've seen so far have been partial YLODs where people can still use the system just not reliably/without the system being a certain temperature/for a long period of time.

EDIT so I got a CECHH01 with the YLOD and heat gunned the NEC caps on the top and bottom of the board, and it does indeed work now, though sometimes it will YLOD the first time you turn it on, but the second time it kicks on. This is more proof to me that the capacitors are at fault here.

 

[Delita]

Did I bridge a via or two during my cap removal? I removed all caps top and bottom and I am checking continuity on the pads after ive cleaned them with fresh solder and wick and I am getting continuity between the three closest to the cell and none on the farthest pad. If I perform the wire jump as suggested this would result in all the pads being shared? I will keep inspecting and report back, im not trying to start a fire here...

 

[Naked_Snake1995]

Has this been done on a system that has a full ylod? As in there is no getting the PS3 to boot at all no matter the amount of finicking you do, it just immediately yellow lights when you hit the power button and flashes red until you shut it off? Most of what I've seen so far have been partial YLODs where people can still use the system just not reliably/without the system being a certain temperature/for a long period of time.

EDIT so I got a CECHH01 with the YLOD and heat gunned the NEC caps on the top and bottom of the board, and it does indeed work now, though sometimes it will YLOD the first time you turn it on, but the second time it kicks on. This is more proof to me that the capacitors are at fault here.

Yes, this was preformed on full YLOD and Partial YLOD machines, the success rate and feedback has been positive so far, everyone had they're machine working again.

Sent from my G8141 using Tapatalk

 

[Naked_Snake1995]

Did I bridge a via or two during my cap removal? I removed all caps top and bottom and I am checking continuity on the pads after ive cleaned them with fresh solder and wick and I am getting continuity between the three closest to the cell and none on the farthest pad. If I perform the wire jump as suggested this would result in all the pads being shared? I will keep inspecting and report back, im not trying to start a fire here...

Excellent, a report and a picture would be useful for users in the future who want to replace the NECs for Tantalums, as this procedure its quite a pain the the arse, ive seen full NEC replacement that only required one bridge connection, others with multiple bridge connections between the Tantalums, at first i though the CELL didn't share a line with the RSX, then @sandungas pointed out that the course its shared between the CELL and the RSX, if i am not mistaken, but ill leave him to explain, as hes more experienced than i am.

Sent from my G8141 using Tapatalk

 

[Delita]

so this is what I meant earlier, just to kind of better explain what I meant. I get continuity between these following pads color coded, and it doesn't matter which pad, top or bottom, they all share.

Is this okay? Can someone else confirm? I feel like the pictures shared in earlier threads make it sound like the outer pads are positive and the inner are negative (ground?) which seemingly means the three light blue pads are really off? Or do the tantalum caps not care about this lol.

Id appreciate others confirming, I am good at following directions, not so much knowledgeable on electrical engineering.

Cheers

https://i.imgur.com/gkCx4GC.png?2

 

[Revak3115]

So today I replaced one cap for each Chip. So one (4 small caps) for RSX and one (4 small caps) for the CELL and it fixed these issues for me. It no longer has any issues. thanks for the awesome tutorial and guide.

*UPDATE*
When I rebuilt my database through the XMB via Rebug it crashed with the 3 beeps. This concerned me so I tried it 3 more times and it didn't Crash at all. Welp I guess more NECs are bad good thing I have the remainder of the new caps coming in the mail on Monday.

EDIT: rebuilt database again after playing a ps1 game on 100% fan speed for about 20 min and when I clicked rebuild database through XMB it gave me the YLOD crash.

 

[Fin9ersMcGee]

Got a 'China Special' bag of 100 tantulums on the way.
3 fat ps3s to fix. One being a CECHC. Ps2 BC, here I come

 

[Vishera]

Got a 'China Special' bag of 100 tantulums on the way.
3 fat ps3s to fix. One being a CECHC. Ps2 BC, here I come

I only buy Japanese capacitors,you want them to be good quality so they last as much as possible without problems.

 

[Naked_Snake1995]

*UPDATE*
When I rebuilt my database through the XMB via Rebug it crashed with the 3 beeps. This concerned me so I tried it 3 more times and it didn't Crash at all. Welp I guess more NECs are bad good thing I have the remainder of the new caps coming in the mail on Monday.

EDIT: rebuilt database again after playing a ps1 game on 100% fan speed for about 20 min and when I clicked rebuild database through XMB it gave me the YLOD crash.

Mine still holds up, ive been torturing it all week long, so far no crashes or anything suspicious.

But yes, as ive mentioned before this kind of repair process its a shot in the dark, because you don't know which NECs are good and which are bad,so you'll either replace all of them, or go by your guess or instincts.

Sent from my G8141 using Tapatalk

 

[Revak3115]

Mine still holds up, ive been torturing it all week long, so far no crashes or anything suspicious.

But yes, as ive mentioned before this kind of repair process its a shot in the dark, because you don't know which NECs are good and which are bad,so you'll either replace all of them, or go by your guess or instincts.

Sent from my G8141 using Tapatalk

Mine has Held up pretty good also.

This seems to be the only time it has crashed since i replaced 1 CELL NEC.

if you're running CFW could you rebuild database in the XMB as a test for me?(For me the system has to be cool i.e around 50C)

i think i'll remove all the NECs and bridge all the pairs individually(the positive points), that shouldn't hurt it right?

 

[sandungas]

so this is what I meant earlier, just to kind of better explain what I meant. I get continuity between these following pads color coded, and it doesn't matter which pad, top or bottom, they all share.

Is this okay? Can someone else confirm? I feel like the pictures shared in earlier threads make it sound like the outer pads are positive and the inner are negative (ground?) which seemingly means the three light blue pads are really off? Or do the tantalum caps not care about this lol.

Id appreciate others confirming, I am good at following directions, not so much knowledgeable on electrical engineering.

Cheers

https://i.imgur.com/gkCx4GC.png?2

Thats wrong, look at the picture most at top-right corner in the datasheet
https://file-databank.netdna-ssl.com/wp-content/uploads/2018/09/NEC-TOKIN-0E907-DATASHEET.pdf

Every tokin have 4 "solder lines" under them... the 2 lines closest to the center are both ground
Then you have 1 line V_IN.... and another line V_OUT (the V_OUT lines are the ones closest to CELL/RSX)

My guess is all the V_IN lines (of all the tokins) are connected together (like the red lines you painted in your drawing, this is fine)
All the ground lines are connected together too
And the V_IN lines are not connected in bewteeen CELL and RSX

In other words... there are 4 tokins dedicated to CELL... and another 4 tokins dedicated to RSX

 

[sandungas]

Mine has Held up pretty good also.

This seems to be the only time it has crashed since i replaced 1 CELL NEC.

if you're running CFW could you rebuild database in the XMB as a test for me?(For me the system has to be cool i.e around 50C)

The "rebuild database" should not be related... most probably your unestable behaviours are caused by small temperature changes, as mentioned before when a capacitor is partially damaged becomes extremelly sensitive to temperature

i think i'll remove all the NECs and bridge all the pairs individually(the positive points), that shouldn't hurt it right?

Using several wires doesnt hurts, but is not really needed, look at the picture posted by @Delita (and check yourself with a multimeter to verify it)... all the tokins are connected together in 2 groups (and every group is composed by 4 tokins)

Basically is this same circuit repeated 2 times, one for CELL and another for RSX
The V_IN lines (most at left in the drawing) are connected together (and ground too)
But at right the V_OUT are separated... one V_OUT is dedicated to CELL and the other V_OUT is dedicated to RSX

 

[Sixx]

Good job. I'm not familiar with electronics much.
Can I use this type of capacitor?
10V 470uF, 477A
16V 330UF, 337C

 

[Revak3115]

Using several wires doesnt hurts, but is not really needed, look at the picture posted by @Delita (and check yourself with a multimeter to verify it)... all the tokins are connected together in 2 groups (and every group is composed by 4 tokins)

Ok, so just to be clear here.

using 8 wires (where the old NECs are) for Bridging that will be fine? i'm quite new to this component repair, i've only done 3 soldering jobs of my own so far. i have a Multimeter somewhere but do i need it? even if the Tantalum Caps don't need a bridge wire does it matter if i just go ahead and put one anyway? i don't want to fry my system by bridging when its not needed.

Thanks

Sorry for the confusion on my end, like i said i'm very new to this stuff and i don't want to break anything