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

[evaldas256]

It would be pure speculation at this point to say for sure.
The question too would be has anyone here been able to revive a SEM-001 motherboard with a tantalum cap replacement?
Reading through all of the posts, I don't remember if anyone has.

I fixed it, for a minute (the CECHG). This is the first time i saw it run.
The system now has 14 470uf 10V caps (yellow chinese cheapies) + 1 old Tokin per chip. I used a hair drier, hottest setting, on the whole board, both sides, focused on the last two Tokins especially and it actually fired up. Then it ran for a minute, shut down, ran only for five seconds, shut down, then two. And the same thing happens every time i let it sit for 10+ minutes.
So i guess i kind of revived the two tokins with heat? Or could this be not tokin related?
I would love to hear some opinions. And i guess, what guage wire and how many wires should i use for the bridging?
P.S. Video is green because i'm using RGB Scart.

 

[demetris]

Yeap one of those tokins you left on the board is toasted. Replace all with tantalums and bridge the + per instructions and you will be golden

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk

 

[Griggs]

I fixed it, for a minute (the CECHG). This is the first time i saw it run.
The system now has 14 470uf 10V caps (yellow chinese cheapies) + 1 old Tokin per chip. I used a hair drier, hottest setting, on the whole board, both sides, focused on the last two Tokins especially and it actually fired up. Then it ran for a minute, shut down, ran only for five seconds, shut down, then two. And the same thing happens every time i let it sit for 10+ minutes.
So i guess i kind of revived the two tokins with heat? Or could this be not tokin related?
I would love to hear some opinions. And i guess, what guage wire and how many wires should i use for the bridging?
P.S. Video is green because i'm using RGB Scart.

Perfect! It's coming back to life! Keep us updated on your progress.
I remember someone recommending #14 minimum. I remember reading that Squeept had a console come back on warranty with burned up wire. Can't remember the size he used though might have been #18 or 20awg? Too much to remember. Stranded or solid copper, your choice. Stranded would be easier to work with. You're going to want to bridge each set twice. I've seen most making the jumps on the top side of the board due to room constraints on the underside.

If I get mine to boot up, I'm really curious to see what firmware it has installed. This thing is a time capsule and hasn't seen the XMB in a decade. I'm guessing it's probably around 3.5? Just waiting on those caps! Should be here tomorrow!

 

[aminiani]

and about slim models which NEC is first suspect?
do you have any photo of slim motherboard please?

 

[XYZ]

and about slim models which NEC is first suspect?
do you have any photo of slim motherboard please?

I think they all have an equal chance of failing (though I'm happy to be corrected).
I'd recommend starting with the NECs on the underside (side B) as you've got a bit more room and are less likely to damage other components. Just be careful you don't knock off any of the micro capacitors.

Pictures of a DYN-001 MB attached:

 

[squeept]

Perfect! It's coming back to life! Keep us updated on your progress.
I remember someone recommending #14 minimum. I remember reading that Squeept had a console come back on warranty with burned up wire. Can't remember the size he used though might have been #18 or 20awg? Too much to remember. Stranded or solid copper, your choice. Stranded would be easier to work with. You're going to want to bridge each set twice. I've seen most making the jumps on the top side of the board due to room constraints on the underside.

The one that came back was 2x 26 awg per chip on a CECHE01. Took about 5 months to melt, though. A test after I realized I screwed up the calculations blew the 10 amp fuse in my multimeter immediately. Since the board/system gets so hot, you'll probably want to use the "transmission" column of this chart - https://www.powerstream.com/Wire_Size.htm. Until someone with a more robust ammeter can weigh in, the absolute bare minimum you're looking at is 1x 12awg or 2x 14awg per chip.

 

[joe musashi]

I fixed it, for a minute (the CECHG). This is the first time i saw it run.
The system now has 14 470uf 10V caps (yellow chinese cheapies) + 1 old Tokin per chip. I used a hair drier, hottest setting, on the whole board, both sides, focused on the last two Tokins especially and it actually fired up. Then it ran for a minute, shut down, ran only for five seconds, shut down, then two. And the same thing happens every time i let it sit for 10+ minutes.
So i guess i kind of revived the two tokins with heat? Or could this be not tokin related?
I would love to hear some opinions. And i guess, what guage wire and how many wires should i use for the bridging?
P.S. Video is green because i'm using RGB Scart.

can you show us the motherboard to see what u did

 

[evaldas256]

Yeap one of those tokins you left on the board is toasted. Replace all with tantalums and bridge the + per instructions and you will be golden

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk

Perfect! It's coming back to life! Keep us updated on your progress.
I remember someone recommending #14 minimum. I remember reading that Squeept had a console come back on warranty with burned up wire. Can't remember the size he used though might have been #18 or 20awg? Too much to remember. Stranded or solid copper, your choice. Stranded would be easier to work with. You're going to want to bridge each set twice. I've seen most making the jumps on the top side of the board due to room constraints on the underside.

If I get mine to boot up, I'm really curious to see what firmware it has installed. This thing is a time capsule and hasn't seen the XMB in a decade. I'm guessing it's probably around 3.5? Just waiting on those caps! Should be here tomorrow!

Oh, i'm a dumbass. I left the fan unplugged so it shutting off was just it overheating, hahaha. Now it runs just fine without shutting off. I guess i'll wait until it YLODs again. And i still have space for 4 more tantalums per chip, so maybe i won't even remove those last two tokins.

can you show us the motherboard to see what u did

I'd rather not, it's so ugly with multiple mistakes visible, like resoldered tiny caps that i knocked off, me scratching off the solder mask where it wasn't even necessary to do so, tantalums soldered not flush with the board despite my efforts Basically, definitely no example to follow, there are much better photos uploaded here.

 

[evaldas256]

Hi, folks. New to the Forum. Like many others, I've read the entire thread over the course of 4-5 days. It captured my attention and I am intrigued. I'm trying my first capacitor swap on a CECHG-01 (SEM-001) that's been sitting dormant for probably 10 years.
If I burn the house down, no big loss.

I have three other CECHA-01's, two of which have the original NEC's and still functional to this day. I have my original (A) launch console in all it's glory still fully working. The third A-Console is a YLOD purchased on eBay slated for a capacitor swap as well to see if I can get it to boot. I figured that I would practice on the G console first

So far on the CECHG I've replaced the top NEC's with 4 tantalum caps rated at 470uf/6.5v. Still YLODing with no change of behavior of the 2 second YLOD. I'm going to keep going and replace the bottoms as well, just waiting on another shipment of caps.
I'll make sure to jumper these also when I remove the bottom NEC's.

I've been out of the soldering game for a while, but was a huge hobbyist back in the day, so it's nothing new. But I found this pretty challenging.
I'll keep updating my progress as I go.

Welcome to the fight! We'd love to hear your results.

Looking at your picture, it looks like you're missing most of the micro capacitors below Cell. On your picture, they're in the bottom right side. For reference, look at this (my picture is flipped from yours, my Cell is on the top left.

By the way, i have to thank you guys. If it wasn't for you, i probably wasn't going to touch this PS3 anymore and just use it for parts because i certified it as dead
I replaced each tokin with 3x 470uf except one per chip, added 3 more tantalums per chip to compensate for the ones i'm not replacing, added 2 more per chip just for good measure and it was still dead. Maybe these chinese cheapies i'm using are actually just lower uF than they are rated. Won't know until the system starts YLODing again, and i add more tantalums.

 

[joe musashi]

Oh, i'm a dumbass. I left the fan unplugged so it shutting off was just it overheating, hahaha. Now it runs just fine without shutting off. I guess i'll wait until it YLODs again. And i still have space for 4 more tantalums per chip, so maybe i won't even remove those last two tokins.

I'd rather not, it's so ugly with multiple mistakes visible, like resoldered tiny caps that i knocked off, me scratching off the solder mask where it wasn't even necessary to do so, tantalums soldered not flush with the board despite my efforts Basically, definitely no example to follow, there are much better photos uploaded here.

what is the best replace for nec tokin that make the ps3 work so good without ylod my ps3 motherboard is same yours

 

[evaldas256]

what is the best replace for nec tokin that make the ps3 work so good without ylod my ps3 motherboard is same yours

Well i can't say for sure now, because i revived it, presumably, only temporarily with a hair dryer. Once the tokins die again, maybe i'll find that tantalums aren't enough and i need to add MLCCs to get it running as other people were talking about earlier in the thread.
What you could do is try what i did if you still have Tokins on your board. Try using a hair drier, on the highest heat setting, on both sides of the board two minutes each. I was focussing on the tokins and places where there were a lot of tiny caps, but it was probably just the tokins that needed the heat. If it starts up after that, then it's probably safe to assume you need the tokins replaced.

 

[joe musashi]

Well i can't say for sure now, because i revived it, presumably, only temporarily with a hair dryer. Once the tokins die again, maybe i'll find that tantalums aren't enough and i need to add MLCCs to get it running as other people were talking about earlier in the thread.
What you could do is try what i did if you still have Tokins on your board. Try using a hair drier, on the highest heat setting, on both sides of the board two minutes each. I was focussing on the tokins and places where there were a lot of tiny caps, but it was probably just the tokins that needed the heat. If it starts up after that, then it's probably safe to assume you need the tokins replaced.

i removed them all but i never add any tantalums i'm waiting any that make his ps3 who is same mine to work without any issues

 

[aminiani]

Little MLCC caps have (-) and (+) polarity or not?
i damaged 4 each of RSX MLCC..is it important or no?

 

[Griggs]

Little MLCC caps have (-) and (+) polarity or not?
i damaged 4 each of RSX MLCC..is it important or no?

MLCC's do not have polarity. If you damaged them, they will most likely need to be replaced/repaired.

 

[ElGris]

I recommend people to buy G/H consoles if the BC isn't an option. Those have the 90nm RSX that could be use on BC models, and they're more reliable imo than later fat models.

@Rinku

"After adding 6 caps the result was same but when i soldered 8 caps, the ylod delayed with black screen and system shut down when something is about to display on tv screen. i was very frustrated. i decided to let it go but don't know how it occurs to me that lets try to add 2 more caps. And i soldered 2 more caps i.e total 10 caps of 1500 µf each and the system boots up with xmb on the screen."

This is what I was talking about, this is the moment when you know the console is about to revive, the inflection point. The minimum gap between life and death. LOL.

Good you didn't give up, and those question might never got an answer since with caps like those, you are reducing the console performance, and we already know they're bad for this kind of applications. If you bought tantalum caps from chinese places, beware, those only work for weeks, then die. In this case, you can add the amount of caps needed to make the console boot, the capacitance required. This only to see the console booting, that's all.

https://www.psx-place.com/attachments/img_20200608_195529-jpg.26173/

P/S: This reminds the time I was playing MGS2 and left Fatman with his bald head full of darts

 

[Griggs]

Well.. No luck to report..
I tried changing out the NEC's one by one on my CECHA-01 (COK-001). Total of approx 1,300µF to start per NEC. All bridge wires installed. Still had YLOD.
Went on to add another 330uF cap to bring each to approx 1,600µF per NEC. Still dead. Using the 'B' case (small case) 330µF 2.5v.

Still continuing to test.

I'll post pics when I get a chance.

 

[marciolsf]

Well.. No luck to report..
I tried changing out the NEC's one by one on my CECHA-01 (COK-001). Total of approx 1,300µF to start per NEC. All bridge wires installed. Still had YLOD.
Went on to add another 330uF cap to bring each to approx 1,600µF per NEC. Still dead. Using the 'B' case (small case) 330µF 2.5v.

Still continuing to test.

I'll post pics when I get a chance.

do you have a scope, or a usb-serial adapter? if so, theres a few other things you could try to do further troubleshooting.... see this thread Im working on — https://www.psx-place.com/threads/fun-with-syscon-and-serial-commands.29913/

 

[Workz_777]

what volt mlcc capacitor should I use? thanks

Hmmm, i think Sony might have used 10V ones, i was looking at the schematics not so long ago, thanks to @marciolsf, and saw Sony had used 10V (0.1uF) MLCCs under the RSX, so i would guess 10V would be good, but they are a little more expensive than 25V or 50V ones.

The NECs are 2.5V, so 10V for the MLCCs would be good.

If anyone has a PS3 board with the official Sony version of Tantalum + MLCC combination, found on later board revisions (from CECH-21xx and above) they could measure the MLCCs like as seen in this photo below, thanks to "Littlebalup". Or if anyone has a link for the schematics of these later boards then we could know for sure what Sony put in.

The MLCCs could have a value between:- 0.01uF and 0.1uF, and at a guess:- would be 10V

 

[Workz_777]

I only buy backwards compatible models, so I'm also wondering if the higher nm process is more robust in terms of noise. I've run them missing half the caps with like 400mv peak to peak of noise and they ran GT6 and TLOU just fine. So a tantalum replacement is going to work better with them with no additional modifications to reduce noise. I think the console I was checking all the different tantalum specs on was a H or L, and it didn't like some of them.

The failed TOKINs aren't bad because they're allowing too much noise, though, it's something weird with their reaction to current flow. I never got to invite my EE buddy over that works in power electronics. Maybe next year when covid is over I can set him loose in my lab with the schematics and some beer.

Anyway, yeah, I'm on board with the idea that for whatever reason, later non-BC models might need some additional modification.

Yeah i agree, later non-BC models don't seem to re-act the same to a Tantalum replacement. It could be those newer (non-BC) models didn't suffer the same problem because they only consume like 100W (average) compared to the BC models 200W (average).

The BC models have 3 x power regulators on their RSX power rail, compared to the other later models which only have 2 x power regulators on their RSX power rail.

I think maybe the NEC problem / fix is more common with BC models than non-BC models. I feel non-BC models might see a result from a reflow / re-ball on the RSX / CELL.

But as we know capacitors only last so long anyway, sooner or later (depending on hours of use + heat) those NECs will fail, so it's worth replacing them one day. Maybe doubling up, a reflow (or fresh re-balling with lead solder) with 8 x "new" NECs could be a good way to fix / future proof / extend the life-span of a broken (ylod) PS3.

...having said that, i really want to test Sony's MLCC + Tantalum combination, just like they did on their later PS3 board revisions, to see what results that gives. It might help for newer non-BC boards.

 

[Workz_777]

you mean i should i use both of them do have photo for someone use both of them

ps: i have a question outside that thread i have a ps4 1st gen when i plugin it wont start up and even there is not a beep from what is that does it from the power supply or what?

Hiya, yep in an ideal world. Here is a photo when Sony first replaced NEC/Tokins with Tantalums + MLCCs. We can see in the photo, for this board revision Sony left the NECs on the RSX, but from the next board revision and onwards, Sony put Tantalums + MLCCs on both processors.

Photo from a DYN-001 board.

Would be good to know exactly what these little MLCCs measure, if i had to guess... i would say they are 0.1uF 10V each.

P.S. Sony only did this on non-BC boards, once the average Watts consumed by the PS3 was approximately halved, and they only used 7 x 470uF Tantalums in total per processor. But this would not be enough for BC models, those would need like double the Tantalums (14) compared to this photo.