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

[sandungas]

1, - This is the top side of the PCB, is this correct?

Is something conceptual, but i will explain it for curiosity sake
If you take a very simple circuit board with only 1 layer and all the components using pins that are mounted in the board "through hole" is obvious what is top and what is bottom
At top side you have the components (and no solders), also is printed in withe the board name/serial, other info for "jumpers" or things like that
And at bottom side you have all the solder points (and no components)

In the PS3 motherboard there are components soldered at both sides, also the name/serial of the motherboard sometimes is printed at both sides too
In this case the rule we should follow is to focus in which side is located CELL/RSX, thats top side
Additionally, at bottom side are located all the "testpoints" intended to connect the motherboard with a jig-pin-machine at factory to program it (some of this testpoints are used to solder the teensy flasher btw)

2, - I see some discoloration on the tokens'. Maybe that is a sign of the problem, or that is how they look like all the time?

Yep, some people mentioned it before in the thread, that internal layers of metalic materials gets burned at he corners, is a sign that was weared out

 

[Maroon Storm]

So based on your answer (thank you for that), I think that is the correct side, where I need to start replacing both tokens (and because my observation looks like was correct about the discoloration). That is the part where the components using the shield only to cool down aka. passively cooled by the shield. I've tried to read thru the thread thoroughly, but I didn't see mention it, but I'm sure you right.

 

[SiegHart93]

So based on your answer (thank you for that), I think that is the correct side, where I need to start replacing both tokens (and because my observation looks like was correct about the discoloration). That is the part where the components using the shield only to cool down aka. passively cooled by the shield. I've tried to read thru the thread thoroughly, but I didn't see mention it, but I'm sure you right.

I revived my ps3 fat cechl01 that was yloding me in some games like bioshock infinite thanks to this post just replacing the nec tokins at top side of the board with proper tantalums 330uf 2.5 4 per each tokin. But i think the best way to test if the problem is tokins related you just need to open ur ps3 and with a heat gun set at 350 degrees focus only on the tokins not the gpu or cpu just the tokins in a interval of 3mins for each side of the board. i did it myself with a ps3 ceche01 that was yloding me at start of boot and sometimes when i get to the xmb it ylods me with every game i try to play. i did the trick of the heat gun to that ps3 just focusing in the tokins for 3 mins and 20 seconds in each side of the board and the result was that this ps3 came back to life again but now it ylods with the most demanding games like the last of us or bioshock infinite (a good start because before it ylods me at boot and when i tried to play every game even the less demanding, now i can turn it on flawesly and play not demanding games). so the problem was tokin related and not gpu or cpu that often people think... good luck with the tokin replacement bro

sorry for my english btw :P

 

[squeept]

I now have a spare lead reballed CECHE01 unit finished, so I can chop up my drive testing unit and not care if I botch it to smithereens. I'll be a few days since I need to Dremel some jigs so that I can show missing parts, keep it running cool, and play games all at the same time. I'm going to get some baseline images from my oscilloscope to show yet another way you can prove that the TOKIN's are fine. This time all you'll need is an o-scope, and you can leave the board completely intact.

Sincere question here no matter how much it sounds like I'm being a dick: I'm beginning to suspect that literally nothing will change some of your minds about this. If I manage to have it running Gran Turismo 6 and the Last of Us while missing any amount of TOKIN's..... you guys are just going to move the goalposts again, right? Like.... is there even any point to me trying this? I'd rather not waste a few hours setting it up and recording it if everyone is just going to close their eyes then plug their ears and hum.

 

[Edwired]

I revived my ps3 fat cechl01 that was yloding me in some games like bioshock infinite thanks to this post just replacing the nec tokins at top side of the board with proper tantalums 330uf 2.5 4 per each tokin. But i think the best way to test if the problem is tokins related you just need to open ur ps3 and with a heat gun set at 350 degrees focus only on the tokins not the gpu or cpu just the tokins in a interval of 3mins for each side of the board. i did it myself with a ps3 ceche01 that was yloding me at start of boot and sometimes when i get to the xmb it ylods me with every game i try to play. i did the trick of the heat gun to that ps3 just focusing in the tokins for 3 mins and 20 seconds in each side of the board and the result was that this ps3 came back to life again but now it ylods with the most demanding games like the last of us or bioshock infinite (a good start because before it ylods me at boot and when i tried to play every game even the less demanding, now i can turn it on flawesly and play not demanding games). so the problem was tokin related and not gpu or cpu that often people think... good luck with the tokin replacement bro

sorry for my english btw :P

i had the same problem before where gran turismo 5 works fine but with some sort of stuttering then gran turismo 6 just shut down with 3 beeps flashing red light only played it for less than two minutes. i may say if something else may be causing shut downs like where the ps3 thinks it overheating when it not at all or some sort of temperature shut down or the power delivery section on the board may having an issue

 

[SiegHart93]

i had the same problem before where gran turismo 5 works fine but with some sort of stuttering then gran turismo 6 just shut down with 3 beeps flashing red light only played it for less than two minutes. i may say if something else may be causing shut downs like where the ps3 thinks it overheating when it not at all or some sort of temperature shut down or the power delivery section on the board may having an issue

that was happened to me with my CECHL01 it started yloding me with bioshock infinite and then with the last of us (a game that I could play normally before) but with the tokins replacement now it works flawless without any issues. Now my CECHE01 as i said before it was yloding me at start of boot and when it randomly allowed me to enter in XMB it ylods me with every game even the less demanding (yes even goat simulator xD) and then when i did the trick of the heat gun on the tokins , now it boots flawessly and i can play games that are not demanding because it still ylods with most demanding but that was a good sign it was tokins related problem and not gpu or cpu. Im trying to get some tantalums caps to fix it permanently. But im hopping to get in my hands a ylod CECHA01 and try to fix it with tantalums...

 

[wrx884]

I now have a spare lead reballed CECHE01 unit finished, so I can chop up my drive testing unit and not care if I botch it to smithereens. I'll be a few days since I need to Dremel some jigs so that I can show missing parts, keep it running cool, and play games all at the same time. I'm going to get some baseline images from my oscilloscope to show yet another way you can prove that the TOKIN's are fine. This time all you'll need is an o-scope, and you can leave the board completely intact.

Sincere question here no matter how much it sounds like I'm being a dick: I'm beginning to suspect that literally nothing will change some of your minds about this. If I manage to have it running Gran Turismo 6 and the Last of Us while missing any amount of TOKIN's..... you guys are just going to move the goalposts again, right? Like.... is there even any point to me trying this? I'd rather not waste a few hours setting it up and recording it if everyone is just going to close their eyes then plug their ears and hum.

do it man the ones that are more involved with these sorter repairs im sure u will get more respect for what ur doing so far, regardless u will still get the nay sayers but meh at the end of the day if we get to the same result to show what the issue is thats all that matters.

 

[Naked_Snake1995]

Its been a long time, ive read a few articles about on how to diagnose the NEC/TOKINs, unfortunately it needs an Oscilloscope to detect any anomalies in the Positive Via, and Grounded for any Voltage Ripples, the uF value can be the same as factory,but if the NEC gives a Voltage Ripple,the PS3 enters in a shutdown sequence, hence the YLOD, the top layer replacement still applies, not all NECs need to be removed,only in extreme circumstances, only the Top ones do heat a lot, the bottom ones are cooled by the Fan itself, so there degradation is particularly none, so for us who cant afford an Oscilloscope, the best way is to remove the top ones only, leave the ones one the Chips as they are.

My CECHC its still working fine as we speak, already closing to the 7th month since the repair, and no anomalies detected, tortured with GT6 and TLoUs, since ive sold my Slim, it gave me more reason to torture this one

Happy repairs

 

[Maroon Storm]

Its been a long time, ive read a few articles about on how to diagnose the NEC/TOKINs, unfortunately it needs an Oscilloscope to detect any anomalies in the Positive Via, and Grounded for any Voltage Ripples, the uF value can be the same as factory,but if the NEC gives a Voltage Ripple,the PS3 enters in a shutdown sequence, hence the YLOD, the top layer replacement still applies, not all NECs need to be removed,only in extreme circumstances, only the Top ones do heat a lot, the bottom ones are cooled by the Fan itself, so there degradation is particularly none, so for us who cant afford an Oscilloscope, the best way is to remove the top ones only, leave the ones one the Chips as they are.

My CECHC its still working fine as we speak, already closing to the 7th month since the repair, and no anomalies detected, tortured with GT6 and TLoUs, since ive sold my Slim, it gave me more reason to torture this one

Happy repairs

Can I have a question to you? The tokins were damaged on your COK-002 (those discoloration like I captured)? Front of me mine (I was lucky to get this one, because I bought some phat in bulk this one was shining through between them and I was just hoping this one is going to be at least a half compatible and turned out it is, happiness) I ask you because I want to but I can't provide a good quality images but this ones tantalums don't look burnt (I popped up the top of them, where I see the little parts under the CELL and the BD driver silver cable quick slot is placed).
Nevermind, the coin is dropped, that is not evidence if you see burn on them. I've just realized what you are saying because my C doesn't have those marks. Instead of a question.

Is this normal around those capacitors those black smudges? I've tried cleaning them off, but it was just a half success. I'm afraid of some cooking occurred on this unit. I have not tried starting it.

 

[Naked_Snake1995]

Can I have a question to you? The tokins were damaged on your COK-002 (those discoloration like I captured)? Front of me mine (I was lucky to get this one, because I bought some phat in bulk this one was shining through between them and I was just hoping this one is going to be at least a half compatible and turned out it is, happiness) I ask you because I want to but I can't provide a good quality images but this ones tantalums don't look burnt (I popped up the top of them, where I see the little parts under the CELL and the BD driver silver cable quick slot is placed).
Nevermind, the coin is dropped, that is not evidence if you see burn on them. I've just realized what you are saying because my C doesn't have those marks. Instead of a question.

Is this normal around those capacitors those black smudges? I've tried cleaning them off, but it was just a half success. I'm afraid of some cooking occurred on this unit. I have not tried starting it.

If i recall correctly, the NECs did have a brown like burnt descoloration inside them once you take them apart, at least every COK-002 Board i did take apart had the same brown burnt descoloration, and these were sealed YLOD units, might be related to the differences i mentioned on my posts, about the active and passive cooling, i suspect that the NECs that are passive cooled, excede they´re operation temperature, since upstairs you have the PSU and the EMI Sheild to add as a furnace inside, while the bottom ones are active cooled by the fan itself, so they are safe from degradation.

Now for the black color pattern around the NECs via the Ground points and etc, ive never saw such a thing, perhaps someone added Flux, and tried to reflow it, usualy that black could be dry unclean flux residue, although with alcohol and a q-tip it should come off fairly easy.

 

[Edwired]

I would use brake cleaner spray or electronic spray to clean the entire board if the board was dirty or sticky

 

[wrx884]

on the A and C units ive done so far ive not seen any burn marks like this on the tokins caps ive opened, ive popped the caps off the majority just to see this because it was discussed earlier when squeept showed some results but i have checked this 1st b4 i tackled them and so far none of these were revived with this method so far. I have however seen the discolouring on a couple H,K and L models so far, i suspect these just must run hotter on that side of the board ur seeing. I even have a G model that has the rainbow colouring (suggests over heating spots) around the power rails just b4 the tokins but this too never saw any results.

@Maroon Storm the black brown u can see in the vias in ur pics are in fact flux residue so that unit has either been reflown or reballed at some point, more than likely just a reflow. u can see it all directly under the GPU and CPU so who ever did it never cleaned the mother board correctly. just use IPA 100% (or close to) and an old toothbrush if u want to clean it up.

 

[Maroon Storm]

On the previous page I include some images of my 200xxx model have discoloration on the bottom tokens, but on this C model tokens don't have any on the bottom side, I haven't destroyed any on the top side tokens of the units.

I was trying using 95% IPA with q-tips to clean the black stuff (I also assumed that is flux residue) but as I sad it wasn't a success with simple q-tips, maybe with a toothbrush could help I am going to try it later. I never used electronic spray before maybe it is a good idea.

 

[wrx884]

On the previous page I include some images of my 200xxx model have discoloration on the bottom tokens, but on this C model tokens don't have any on the bottom side, I haven't destroyed any on the top side tokens of the units.

I was trying using 95% IPA with q-tips to clean the black stuff (I also assumed that is flux residue) but as I sad it wasn't a success with simple q-tips, maybe with a toothbrush could help I am going to try it later. I never used electronic spray before maybe it is a good idea.

For real stubborn baked on flux u can use a mild thinner or even acetone just make sure u use IPA after so get rid of any traces.

 

[squeept]

I've been dicking around all day trying to come up with a good way to run this thing without setting it on fire while also having access to show that components are missing, and I just can't come up with anything that isn't a gigantic pain in the ass. So, once I get started, you'll just have to trust that I didn't add extra capacitors somewhere else to fake results. Just accuse me of that now if you're going to so I don't waste my time documenting all this. Tonight is probably just these reference images, might get one or two ripped off before I get tired and pass out.

Anyway, here's the baseline results. Console is a CECHE01 on the most recent official firmware. All parts/peripherals are connected. Controller is not plugged in, only wireless. This console was YLOD, reballed to fix, checked capacitance when GPU was lifted. All good.
Has been turned on for about 50 hours total since being fixed. Runs GT6 and The Last of Us with no issues.

GPU sitting idle. Average 1.28V. Voltage appears to mostly stay between 20 millivolts either way with the occasional slightly larger spike. Vpp says 50, but it's almost entirely contained within 40.

GPU running GT6. Slightly higher max voltage (looks like a lot more, but I was messing with some adjustments and just move the channel up a little - just check the auto measurements). Loading the GPU does appear to cause more fuzz. Voltage mostly between 25 millivolts either way. (50 mvpp if ignoring the auto). Anomalous large spikes are also larger.

CPU sitting idle. Almost perfect 1.2 volts. Fuzz appears the same - 20 millivolts either way (changed the scope to 100mv / division from 50 mv / div). The spikes, however, are much larger and more frequent. Almost 100mv either way.

CPU running GT6. Interesting result here: no change at all.

If your voltages look like these on any backwards compatible model: congratulations, your TOKINs are fine! (just like 99.9% of them!)

 

[squeept]

Alrighty, ripped off one RSX TOKIN. Results are as expected (for me, anyway... you guys were probably expecting a different kind of crash at the end of the video).

Here's a video of it running GT6 like a trooper:

Things are getting fuzzier, though. Here's the GPU sitting idle. Note that I'm now on 100mv/div. Looks like the fuzz is almost entirely contained by around an 80mv window. Medium spikes are getting more frequent. The big anomaly spikes are getting really big, around 150mv in either direction.

GPU while running GT6. Looks almost the same except the mid range spikes got a little more frequent and pronounced.

Anyway, everything still works fine. So even if your voltages look this bad - missing 25% of the RSX TOKIN capacitance - you're still just dandy. This is taking forever to document and upload everything, so I might be done for tonight. Stay tuned.

 

[squeept]

One off RSX, one off CELL. Running GT6 just fine again:

More interesting results: CPU idling about the same as before.

But there WAS a difference under load this time. Still about a 40mv window, but the spikes are getting over 150mv each direction

Again, still running totally fine. So even if your voltages look this bad, again... you're golden. I'm tired, going to bed. Tune in tomorrow for the exciting conclusion that's going to be totally ignored!

 

[squeept]

2 RSX TOKINs removed, 1 CELL TOKIN removed, still running GT6 perfectly:

GPU idling. Curious results again. No change from 1 to 2. Looks cleaner if anything.

GPU running GT6. Things are really starting to get hairy under load, though. Still seems to be about an 80mv window, but the transient spikes are WAY more pronounced and frequent. About 100mv in either direction. It's getting to the point that if I zoomed out to 1V/div, it won't actually look like a straight clean line.

Still running perfectly fine. These chips don't appear to be particularly choosy about how clean their power is. So, now we've lost 50% GPU capacitance and are still running GT6 without issues. Even if your voltages look this disgusting, you're still in the clear!

Imma go grab some breakfast, then I'll get the next CPU TOKIN off. I'm stopping there since we know from my previous set of videos that it will shut down as soon as it tries to display video once I take the third GPU cap off.

 

[Edwired]

i reckon the nec tokin act like a filter cap before the gpu/cpu as the more nec tokin comes off the board more spikes is appearing on the o-scope so must be something else is trigggering the ylod

 

[squeept]

Yes, they're high speed, low ESR filter caps to filter transient voltages from switching power supplies. The larger spikes you're seeing (much clearer on the CPU images) are each time it switches. The results here are exactly what one would expect: the noise gets bigger the more capacitance we remove. That's not why I'm showing the images, though. I wanted to provide a reference set of healthy, fully functional TOKIN performance images for others to compare their systems against to figure out if there is any problem with the TOKINs. I verified the set of reference images were consistent against another CECHE01 and a CECHA01 that are fully functional after reballs, surviving stress testing and GT6 / TLOU play. We now have a simple, quick method to prove your TOKINs are fine using only one piece of common equipment. The rest of the images are because we want to know just how bad things can get before there are any issues.

Here is 2 TOKINs each removed from the RSX and CELL. 50% total TOKIN capacitance gone. Running GT6 fine again:

CELL idling with 2 missing TOKIN. Pretty ugly, but still working fine. Note that I'm at 200mv/div now. Appears contained within an 80mv window, but the spikes are massive. Over 200mv each direction.

CELL running GT6. I did not get these backwards or change the amplitude. It got better under load this time. Guessing at this result is above my pay grade.

And with 50% of the total TOKIN capacitance pining for the fjords, everything still runs perfectly.