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

[squeept]

This means the capacitance of a 1200uF NEC varies from 960uF to 1440uF. I think those tantalum caps have 10% of tolerance too. Anyway If the circuit works with 1880uF (4x 470uF) instead of 1440uF max means it just can handle it but this extra capacity may put your device under pressure?

https://www.electronics-tutorials.ws/rc/rc_1.html
https://www.electronics-tutorials.ws/rc/rc_3.html

Changing either capacitance or voltage rating will alter multiple important aspects of the circuit. Time constant, inrush current, frequency cutoff, etc. Ideally, you want to use exactly the specs that were originally there. But again - staying within reason, I'm betting you won't notice any real difference (FOR THIS SPECIFIC APPLICATION). I'd bet a dollar that IF there is any noticeable difference, then 6.3V 1200uF will be uglier than 2.5V 2400uF because it's going to screw with the current sensing more.

I'll toss a few different ratings on my next Mouser order to show you on the oscilloscope.

 

[128x.k]

Hello!

My name is 128x it is nice to meet you all! This is my first time joining a forum and also my first post! I just finished reading all 75 pages on this thread for the past few hours and I must say WOW! What an amazing journey, all the amazing information and tests being shared actually has made me so happy (not to mention some of the drama). You guys have inspired me to join in. Thank you to everyone in this thread especially @Naked_Snake1995 & @sandungas any many others. Would love to be a part of the community and help others as well as myself.

Ok so here is what I have to share. I have a CECHA01 that I installed cfw on as well as re-applied thermal paste (arctic silver). I delid the heatsinks on both the cell and the rsx using tool similar to the ihs buster. I purchased the tool at an art store that looked like something I saw in another thread on this site (no damage to substrate or traces, practiced on a dead board) and re applied thermal paste under the ihs'. Temps have been in the 49-57c on both chips while in xmb ps1, ps2 and ps3 games, with 40% fan, auto at 60c and everything has been running cool. I used the console for a little under a year, not too too often and no problem then one day when I got YLOD when I was deleting a .iso file through ftp (now I just delete using multiman).

I then revived the system by taking it apart and heating up the NEC caps for about 30 seconds each using a 200c heatgun (only one i had access to). Which is far less time than what I have seen suggested here (3 minutes and 30 seconds). Either way VOILA! it ran nice again for about 2 months. I knew this was a temporary fix though from some prior reading.

Then yesterday night I reallly ran the system for a long time id say 9 hours straight. Playing ps3 titles watching movies on multiman and finally 2-3 hours in to playing the psn title "Sonic Adventure DX", not a taxing game id assume, constantly checking the temps throughout playing, nothing out of the ordinary... On the 3rd hour of playing it froze for a second then black screen with 3 beeps and continuing red flashing until I clicked the power button. Then I got the same YLOD as I did before, 2-3 seconds start up green light fan kicks in, then 3 beeps amber light, red flashing.

Then I read this entire forum and while doing so decided to take the NEC caps off my dead practice board to see if there were any signs of damage and if some places on the board were worse than others, because I would like to keep one set of NEC caps so I do not have to solder a bridge. I used the same tool I used to delid my CELL and did not use any heat. I have attached my findings. I thought it was very interesting that on my COK-002 board the NEC caps looked veryyy rough. Notice the burn marks and the dark grey discoloration. When I purchased this COK-002 ps3 it was not powering on and a game shop in my city had tried to fix it to no avail, so they sold it to me for $20 as I needed it for the screws used to tighten the shield to the board/ secure the fans (those strip so easy!) So I do not know how this unit was taken care of but I can see the RSX NEC caps on both sides have visible damage and the CELL top caps have some damage but the CELL NEC bottom caps had almost no visible damage! .

Though I have another CECHA01 COK-01 dead board donor from a friend, which had scratched traces from poor delid job so I knew it was dead for good. (such a shame) But I delid the NEC Caps on the top of that board (see example) and they looked better but still little browning on the corner, maybe this unit was never used that much in the past, any-who thought it was interesting. Not sure what any of it means just thought I would share. (Note both pictures are of dead boards not the one I want to fix, I will post those pics when I get all the parts and equipment)

I am now ready to purchase what I need to replace the old NEC caps, Soldering kit (fine tip), Flux, Solder, wick, New Tantalum Caps, and Kapton Tape.

So my questions are:
1. Does anyone have a soldering iron kit brand they recommend for this job ?
2. Is there a specific solder brand and flux brand that people recommend for this job?
3. Should I replace all NEC caps and just get the thick gauge wire to make bridge? or leave a set on bottom, I see mixed answers so just wondering your thoughts on both. If so what wire should I use and can I get from anywhere on my other 2 dead ps3s?
4. These are the caps I am thinking of getting from what I have read here, just want to confirm these are good for CECHA01 NEC replacement. https://www.newark.com/panasonic/2r5tpe470mc/tantalum-capacitor-470uf-2-5v/dp/98W0352 if not please link ones that are best for job please

Thank you so much look forward to hearing everyone's responses! Very excited to do this project and share the process with everyone here to see if it works/ how my soldering job on the new caps will be once I order and receive everything.

Sorry if this post is a handful I know it is very long

 

[sandungas]

NEC/TOKIn capacitors have 20% tolerance. Refer to this data sheet for more information

https://www.bulcomp-eng.com/datasheet/Tokin 0E907.pdf

This means the capacitance of a 1200uF NEC varies from 960uF to 1440uF. I think those tantalum caps have 10% of tolerance too. Anyway If the circuit works with 1880uF (4x 470uF) instead of 1440uF max means it just can handle it but this extra capacity may put your device under pressure?

About the voltage, it is common to use caps with two times bigger in voltage than the voltage of the line. I think the voltages of running CELL&RSX is about 1.2v? (correct me if I'm wrong) and sony used 2.5volt NECs. Surely we can use higher voltage caps but there is restriction in space as higher voltage caps are bigger caps.

Just for curiosity sake, in one of the tests @squeept removed half of the tokins from a motherboard that was working normally, and the motherboard continued working normally after the removal (but displaying lot of innaccuracies in the osciloscope meassurements)

The moral of the story is the total capacitance of the factory tokins is not the reason why this PS3's stops working
In other words... we could reduce it... and we could increase it too
The usual thing made by engineers at factory is to exceed the real requirements of the machine, because his performance decreases with time
As example for a PSU... if the machine needs 200w they uses a PSU of 400w because after 8 years of use the PSU is going to have an output of 300w or so
First they does a very accurate meassurement of the real requirements, and after that they adds a +30% to that value for a safety margin

With the tokins is pretty much the same, the first PS3 models have 4800uF for RSX and 4800uF for CELL
But if we take the test made by @squeept as valid then we can say that the real requirement is around 2400uF for RSX and 2400uF for CELL
Personally i think that test was too much extreme though (i doubt the engineers kept a 50% safety margin) but it should be something around that

 

[Ayu]

Hi! Has anyone tried using regular cylindrical aluminum electrolytic caps or is there a risk to it?

SMD Tantalums are pretty hard to get and (super) expensive in my area and the only ones locally available are non-SMD 50-100uF ones (too low lol)

Since tantalum caps are a type of electrolytic caps, I was wondering if other electrolytics would work just as well? Barring space concerns. If the space is too tight and if heat is an issue, I could wire up an external board with the caps mounted there and place them somewhere else inside the PS3. (Should also make replacement easier in the future.) I don't know the current flowing through so I'd probably go for the thickest gauge wire or have multiple wires in case I burn those out.

Would this be a plausible idea? I have a bunch of cylindrical aluminum 25v-40v 470uF caps I use for repairing my LCDs but I could easily buy smaller 6.3v uF ones or a full 6.3v 1200uF altogether for every NEC/TOKIN cap. I could also try to get solid-state (polymer) electrolytic caps instead (the kind used in motherboards nowadays).

If its not a deadly endeavor to try with my YLOD PS3, I would like to know before trying. Thanks a lot!

 

[sandungas]

Hi! Has anyone tried using regular cylindrical aluminum electrolytic caps or is there a risk to it?

SMD Tantalums are pretty hard to get and (super) expensive in my area and the only ones locally available are non-SMD 50-100uF ones (too low lol)

Since tantalum caps are a type of electrolytic caps, I was wondering if other electrolytics would work just as well? Barring space concerns. If the space is too tight and if heat is an issue, I could wire up an external board with the caps mounted there and place them somewhere else inside the PS3. (Should also make replacement easier in the future.) I don't know the current flowing through so I'd probably go for the thickest gauge wire or have multiple wires in case I burn those out.

Would this be a plausible idea? I have a bunch of cylindrical aluminum 25v-40v 470uF caps I use for repairing my LCDs but I could easily buy smaller 6.3v uF ones or a full 6.3v 1200uF altogether for every NEC/TOKIN cap. I could also try to get solid-state (polymer) electrolytic caps instead (the kind used in motherboards nowadays).

If its not a deadly endeavor to try with my YLOD PS3, I would like to know before trying. Thanks a lot!

Yes, take a review at the previous pages of the thread, there is someone that posted a photo of a motherboard with a "forest" of cilyndrical electrolitical capacitors soldered in vertical like trees, and he reported success

My mention about soldering them in vertical "like a forest" is not a critic though, i like ghetto repairs eventually is just it was funny
The first obvious problem is that there is no room for them... remember, you need to asemble the interference metal shields later as a sandwich
A solution could be to crop holes in the metal shields, but this is a bit counterproductive, because that area of the motherboard generates heat because the voltage regulators block next to the tokins, and have thermal pads to transfer that heat to the metal shields.. so... by cropping the metal shields over there most probably you are decreasing the colling of that area next to the tokins

And additionally there is an argument that im not so sure yet because nobody confirmet it clearly yet
The point is the electrolitic capcitors are going to overheat (a lot ?, how many ?) because they have a bigger internal resistance to the pass of the electrons
In other words... some of the current that passes throught the electrolitic capacitors is going to be converted to heat... and this cummulation of heat could destroy the capacitor itself

In my oppinion it worths the try... the worst thing that could happen is the capacitor is going to "puff"

And i think it really worths the try to use electrolitic capacitors for diagnose purposes
You know... instead of soldering lot of tantalums trying to do a good solder job.... just take a dirty electrolitic capacitor (with a huge capacitance), and solder it next to a tokin... and try to turn ON the PS3 and see what happens

 

[sandungas]

Btw... in electrolitic capacitors the biggest capacitance (exceeding the capacitance value from factory) is going to cause less heat. Just an example with numbers...

In the old PS3 fat models we have 4800uF for CELL
If we replace that by a single electrolitic capacitor of 4800uF is going to generate heat=X
But if we replace it by a single electrolitic capacitor of 10000uF is going to generate something around heat=X/2

Spoiler

This numbers i used are just orientative, i dont really know how is calculated that... the noob aproach i did was to duplicate the capacitance with the hope this could halve the amount of heat generated

In simple words... you can reduce the heat generated by an electrolitic capacitor by increasing his capacity

 

[Ayu]

Btw... in electrolitic capacitors the biggest capacitance (exceeding the capacitance value from factory) is going to cause less heat. Just an example with numbers...

In the old PS3 fat models we have 4800uF for CELL
If we replace that by a single electrolitic capacitor of 4800uF is going to generate heat=X
But if we replace it by a single electrolitic capacitor of 10000uF is going to generate something around heat=X/2

Spoiler

This numbers i used are just orientative, i dont really know how is calculated that... the noob aproach i did was to duplicate the capacitance with the hope this could halve the amount of heat generated

In simple words... you can reduce the heat generated by an electrolitic capacitor by increasing his capacity

I'm at least familiar with that Though with bad-cap LCDs I increase voltage instead while retaining capacitance or just slightly increase it. I find they last longer this way. I once changed a bad 37v 470uF cap with a 1200uF one --> the LCD won't display anything lol. A 50v 470uF solved it though. Another board I changed a 200-ish uF one with my 470uF and the display has some very slight snow in it but kept it that way anyway.

I'm worried that if I increase the capacitance too much, the chip might get problems from getting delayed power.
Would running 3x-4x 470uF caps be better than a single 1200uF because of lower ESR? Or did I get that wrong? (My electronics is very rusty). I remember being recommended to use multiple caps instead of one big cap in a DIY LM7805 PSP charger I made (a long time ago) because this also filters the line cleaner.

Yes, take a review at the previous pages of the thread, there is someone that posted a photo of a motherboard with a "forest" of cilyndrical electrolitical capacitors soldered in vertical like trees, and he reported success

Hmm I think I missed that. Only saw the guy asking to do the single 1200uF/2400uF one. And another guy very early on (page 20-ish?). If you can remember around which page number it is, that would be a nice sight Knowing someone did it already gives me hope though.

And i think it really worths the try to use electrolitic capacitors for diagnose purposes
You know... instead of soldering lot of tantalums trying to do a good solder job.... just take a dirty electrolitic capacitor (with a huge capacitance), and solder it next to a tokin... and try to turn ON the PS3 and see what happens

That's a great idea actually. I'll buy 8x 6.3v 1200uF cylindrical electrolytic on my next visit to the electronics store.
This PS3... I had it delidded and reflowed where it got revived to the point it stayed alive for a few seconds. The technician I went to said it was a no-go and refunded me. After getting home, I opened it up and increased tension by the HSF by adding wooden popsicle sticks (super ghetto I know) where the ps3 worked for a couple of months playing Tekken and MGS. Expelled air was very hot and the PS3 never went jet-mode. The system then went YLOD again while watching a DVD though. Thought I needed to tighten the HSF again maybe to increase contact or maybe the PS2 chip died.

I'm afraid I may have over-tightened the HSF out of frustration but soldering 8 caps would be worth it to see if I didn't! I'll report back if it does work and try to make my own "forest" eh :P I figure if the external board version works (if I do try it) I can make these boards in advance for my friends with YLOD PS3s.

BTW, thanks for the response No one at youtube or reddit is telling me anything

 

[sandungas]

Hello!

My name is 128x it is nice to meet you all! This is my first time joining a forum and also my first post! I just finished reading all 75 pages on this thread for the past few hours and I must say WOW! What an amazing journey, all the amazing information and tests being shared actually has made me so happy (not to mention some of the drama). You guys have inspired me to join in. Thank you to everyone in this thread especially @Naked_Snake1995 & @sandungas any many others. Would love to be a part of the community and help others as well as myself.

Ok so here is what I have to share. I have a CECHA01 that I installed cfw on as well as re-applied thermal paste (arctic silver). I delid the heatsinks on both the cell and the rsx using tool similar to the ihs buster. I purchased the tool at an art store that looked like something I saw in another thread on this site (no damage to substrate or traces, practiced on a dead board) and re applied thermal paste under the ihs'. Temps have been in the 49-57c on both chips while in xmb ps1, ps2 and ps3 games, with 40% fan, auto at 60c and everything has been running cool. I used the console for a little under a year, not too too often and no problem then one day when I got YLOD when I was deleting a .iso file through ftp (now I just delete using multiman).

I then revived the system by taking it apart and heating up the NEC caps for about 30 seconds each using a 200c heatgun (only one i had access to). Which is far less time than what I have seen suggested here (3 minutes and 30 seconds). Either way VOILA! it ran nice again for about 2 months. I knew this was a temporary fix though from some prior reading.

Then yesterday night I reallly ran the system for a long time id say 9 hours straight. Playing ps3 titles watching movies on multiman and finally 2-3 hours in to playing the psn title "Sonic Adventure DX", not a taxing game id assume, constantly checking the temps throughout playing, nothing out of the ordinary... On the 3rd hour of playing it froze for a second then black screen with 3 beeps and continuing red flashing until I clicked the power button. Then I got the same YLOD as I did before, 2-3 seconds start up green light fan kicks in, then 3 beeps amber light, red flashing.

Then I read this entire forum and while doing so decided to take the NEC caps off my dead practice board to see if there were any signs of damage and if some places on the board were worse than others, because I would like to keep one set of NEC caps so I do not have to solder a bridge. I used the same tool I used to delid my CELL and did not use any heat. I have attached my findings. I thought it was very interesting that on my COK-002 board the NEC caps looked veryyy rough. Notice the burn marks and the dark grey discoloration. When I purchased this COK-002 ps3 it was not powering on and a game shop in my city had tried to fix it to no avail, so they sold it to me for $20 as I needed it for the screws used to tighten the shield to the board/ secure the fans (those strip so easy!) So I do not know how this unit was taken care of but I can see the RSX NEC caps on both sides have visible damage and the CELL top caps have some damage but the CELL NEC bottom caps had almost no visible damage! .

Though I have another CECHA01 COK-01 dead board donor from a friend, which had scratched traces from poor delid job so I knew it was dead for good. (such a shame) But I delid the NEC Caps on the top of that board (see example) and they looked better but still little browning on the corner, maybe this unit was never used that much in the past, any-who thought it was interesting. Not sure what any of it means just thought I would share. (Note both pictures are of dead boards not the one I want to fix, I will post those pics when I get all the parts and equipment)

I am now ready to purchase what I need to replace the old NEC caps, Soldering kit (fine tip), Flux, Solder, wick, New Tantalum Caps, and Kapton Tape.

So my questions are:
1. Does anyone have a soldering iron kit brand they recommend for this job ?
2. Is there a specific solder brand and flux brand that people recommend for this job?
3. Should I replace all NEC caps and just get the thick gauge wire to make bridge? or leave a set on bottom, I see mixed answers so just wondering your thoughts on both. If so what wire should I use and can I get from anywhere on my other 2 dead ps3s?
4. These are the caps I am thinking of getting from what I have read here, just want to confirm these are good for CECHA01 NEC replacement. https://www.newark.com/panasonic/2r5tpe470mc/tantalum-capacitor-470uf-2-5v/dp/98W0352 if not please link ones that are best for job please

Thank you so much look forward to hearing everyone's responses! Very excited to do this project and share the process with everyone here to see if it works/ how my soldering job on the new caps will be once I order and receive everything.

Sorry if this post is a handful I know it is very long View attachment 24347 View attachment 24348

Welcome to the forum, nice first post
The tokins in your photos doesnt looks so bad, with electronics is hard to tell, sometimes is posible to identify the faulty components just with a visual inspection but other times the components looks fine visually but is "hiding" a problem and is faulty
From the photos ive seen here my personal oppinion is having a bit of brown color in the corners is something normal. But when appears several colors (different tones of brown, black, and even blue) is very damaged
The point is that metal sheets probably have some kind of electro-chemical-thermal treatments... if we could cut one at half and take a look at it with a microscope probably the "core" is made of a metal, and then it have several external "layers" of other kind of particles/substances

When it overheats too much all that electro-chemical treatments is decomposed is different substances, some burns and becomes smoke, others could make a shortcircuit, etc...

The decission of replacing only a few tokins or all them, depends mostly of the amount of usage time cummulated by the motherboard, webman can display this info... but if the motherboard doesnt turns ON is hard to know :/

I use to consider the cosnoles that are suffering this problem with the tokins all them have a huge amount of hours of usage... and this is why i use to suggest to replace all the tokins
Using examples with cars... is like if you change the tyres of a car and you start driving it (non stop) for months/years, there is going to be a point where 1 of the tyres fails, and the best you can do at that point is to change all the tyres... because you know... the others are about to fail too

 

[pure3d2]

Just for curiosity sake, in one of the tests @squeept removed half of the tokins from a motherboard that was working normally, and the motherboard continued working normally after the removal (but displaying lot of innaccuracies in the osciloscope meassurements)

The moral of the story is the total capacitance of the factory tokins is not the reason why this PS3's stops working
In other words... we could reduce it... and we could increase it too
The usual thing made by engineers at factory is to exceed the real requirements of the machine, because his performance decreases with time
As example for a PSU... if the machine needs 200w they uses a PSU of 400w because after 8 years of use the PSU is going to have an output of 300w or so
First they does a very accurate meassurement of the real requirements, and after that they adds a +30% to that value for a safety margin

With the tokins is pretty much the same, the first PS3 models have 4800uF for RSX and 4800uF for CELL
But if we take the test made by @squeept as valid then we can say that the real requirement is around 2400uF for RSX and 2400uF for CELL
Personally i think that test was too much extreme though (i doubt the engineers kept a 50% safety margin) but it should be something around that

Do you recall what sort of tests he ran? Did he simply power up the board with the "bad" TOKINs removed? Did he run any sort of stress testing (demanding games like GT6 or LastOfUs)? I recall someone mentioned that when the TOKINs go bad, they also act as some type of internal short, which is why removing them "fixes" the console.

 

[Edwired]

Did i just see like a hairline crack in one of the nec tokin on one of the pictures?

 

[sandungas]

Do you recall what sort of tests he ran? Did he simply power up the board with the "bad" TOKINs removed? Did he run any sort of stress testing (demanding games like GT6 or LastOfUs)? I recall someone mentioned that when the TOKINs go bad, they also act as some type of internal short, which is why removing them "fixes" the console.

The tokins was not bad, he removed them for science in a motherboard that was working fine

If i remember right he ran GT6 following the tests other people was suggesting in this thread (in a special track that did seem to be problematic), and as far i remember the PS3 continued working fine
But i guess he is going to see the mention and he could tell better than me

Since that point the theory we have (im including myself becaue i agree with him) is most of the times in this PS3's that are having this problem, the cause is not the total capacitance... but some kind of internal problem in the tokin that creates either interferences, or random resistance, or like shortcircuits, dunno
It looks simply by removing the tokin that is causing this "garbage" in the power line the PS3 should work

But as i mentioned before, if you have 1 tokin with this kind of weird problems it means the other tokins are about to have the same problem soon, so for a serious repair is better to replace all them
Otherway, for diagnose purposes or dirty/unexpensive fixes... yeah do whatever you consider needed even if it looks dirty

 

[squeept]

Do you recall what sort of tests he ran? Did he simply power up the board with the "bad" TOKINs removed? Did he run any sort of stress testing (demanding games like GT6 or LastOfUs)? I recall someone mentioned that when the TOKINs go bad, they also act as some type of internal short, which is why removing them "fixes" the console.

Played GT6 and TLOU just fine missing half. Missing 3 out of 4 would start to boot, show an image for a split second, then YLOD. Missing all and just shorting with a wire would YLOD instantly.

I could be wrong on this, but I think anyone fixing anything by using electrolytic caps is only working because they removed the bad TOKIN (or more pressure from reassembly ). Is there any video or reports of a system working using ONLY electrolytics?

As far as I can tell on an extremely limited sample size, the internal "cosmetic" condition of the caps does not correlate to whether or not they are causing problems. One bad set I tore open was burnt to a crisp, one bad set looked almost mint.

At best, heating them up tells you nothing. At worst, it tells you there is a BGA/GPU issue. It did nothing on the few boards that I had verified bad caps on. No change in functioning, no change on the scope.

 

[sandungas]

As far as I can tell on an extremely limited sample size, the internal "cosmetic" condition of the caps does not correlate to whether or not they are causing problems. One bad set I tore open was burnt to a crisp, one bad set looked almost mint.

But note in some photos it can be seen the corners are like "delaminating" and it looks like was composed by different materials... and everyone of that materials became a different substance

I think thats the point where we can say for sure is damaged only with a visual inspection, lol. I only saw a photo where i noticed this though
In most of the other photos is doubtful



Edit:
I cant find that post with the photo damnit, this thread is becoming huge, lol
All i remember is i replyed to that posts saying something like... "nice photo of a damaged tokin"
But i cant find the photo neither my post, sorry :/
That photo called my atention up to a point where i took it, and opened it in a image editor to take a better look at it at 400% zoom or so
And the damaged materials was very notable... are not just a cummulation or rust... the corner had a deformation like a "bump" and it was like if the particles of whatever was that material started decomposing into different substance/s

 

[ElGris]

@squeept Nice to know you will able to test the capacitors performance with a scope. Don't know if you're gonna find something that tell us which is the best setup for both processors, but will be really interesting indeed. Are you gonna try tantalums, both 330/470 in different numbers over the mobo? I think the brand of the capacitors really matters, since mines (bought from Aliexpress) have a not so real value when doing measurements with the multimeter.

@Ayu I don't recommend you in any ways using electrolytics. They're big and warm a lot, you won't be able to mount the PS3 and you can also provoque a short at any moment when trying assembling everything back. I strongly recommend you to buy some tantalums from China (or anywhere cheaper than that), the wait will be granted with a great performance and without having to worry about important problems later.

Example of what you aimming:

Better try buying them from other countries, since China has this little biohazard problem that's mutating while I'm writing this.

@128x.k

Those caps are nice, similar to the originals on 3xxx PS3s (maybe the same?). About the soldering iron, use any soldering station, most of the people use them. But in my case I only use a common chinese one that I have since many years ago, and is enough. I can't remove the NECs enterily like some of you, though, I have to destroy them by using a heat gun :/
Use any common flux, and wire from an old ATX PSU if you have one. If you have good pics (using some good lens) show us how are those damaged CELLs. Maybe you can still have a shot.

Important note: Check every tantalum before installing, values and if it is showing any value at all. A fat with all tantalums died again with YLOD, had to remove every tantalum and test them. I realized one of them was "open" in my multimeter, but this isn't a short, so
why the PS3 was shorting at all? Misteries from China.

@sandungas Sorry pal, Sony enginners are morons. If it wasn't for them, this whole thread wouldn't even exist lol. Why you even write "S@ny" hahaha.

 

[sandungas]

@sandungas Sorry pal, Sony enginners are morons. If it wasn't for them, this whole thread wouldn't even exist lol. Why you even write "S@ny" hahaha.

Well, the decission of using the tokins instead of standard tantalum capacitors was something that increased the manufacturing costs
So they was trying to have a robust design, but the tokins betrayed them

Is hard to know if they was aware of the problems of the tokins

 

[ElGris]

But they used them until the 2xxx models, very negligent one their part. The only good thing about all of this, is that this console can have a second chance, and that's all that matters

 

[sandungas]

But they used them until the 2xxx models, very negligent one their part. The only good thing about all of this, is that this console can have a second chance, and that's all that matters

And they could have continued using the tokins, i mean... instead of 4 tokins "per processor" and at the time they was upgrading the processors (CELL/RSX) they could use 2 tokins instead of 4.... and later only 1 tokin for the latest superslims

Instead of that they dropped the tokins, i guess at that point they realized it was not worthy and it was better to use the (cheaper) tantalum capacitors
Probably they regrets of using the tokins at all in the first PS3 models, at this point it looks it could be better to replace them by an army of tantalum capacitors, lol

We cant be sure though...

 

[pure3d2]

@ElGris I checked the RSX SMDs for shorts, found zero. I put everything back together and did the video reset and it came up, but the video was glitched on the HDMI selection screen (could hear sound, but video didn't update between Yes/No): https://i.imgur.com/9VCZX8g.jpg

@squeept, should I try reflowing the RSX or is it toast and needs a replacement via reball?

 

[ElGris]

@sandungas there're a lot of questionable aspects on Sony's decisions. Since they are still using those wifi/bt modules or HDMI ICs on PS4 knowing that they're pretty bad and can leave you without PS4 at any moment. Everything is cheap here, that's the bad thing about consoles, but you're right they tried to improve reability through the years, while they were reducing processors sizes.

@pure3d2 Then maybe that RSX is dead, sadly. Don't you have an A/V cable to confirm that?

 

[squeept]

@squeept Nice to know you will able to test the capacitors performance with a scope. Don't know if you're gonna find something that tell us which is the best setup for both processors, but will be really interesting indeed. Are you gonna try tantalums, both 330/470 in different numbers over the mobo? I think the brand of the capacitors really matters, since mines (bought from Aliexpress) have a not so real value when doing measurements with the multimeter.

Just gonna toss some random stuff in the cart. A few different combinations of voltage, capacitance, and ESR. I'll try some of the electrolytics I have here when I'm dicking around then, too.

Sony enginners are morons. If it wasn't for them, this whole thread wouldn't even exist lol.

Eh, from my vantage point, they picked a suitable part, especially without hindsight from 15 some years later. I still don't believe they're that big of a problem. Imma need those cheap oscilloscopes to flood this thread to change my mind.

should I try reflowing the RSX or is it toast and needs a replacement via reball?

Still not really a definitive way to see if it's a dying chip or a bad connection. Does it change if you position a weight or a shim on the GPU in a jig somehow? Otherwise, without some expensive/exotic equipment, you just have to inspect after lifting the chip and hope there's obvious signs of a BGA issue. I stopped reflowing a few years ago, the results got worse and worse as years went by. My guess is that they work for a pretty long time even with cracks forming and present, allowing oxidation and dirt to build up that makes reflowing them reliably these days next to impossible. And with the first model of PS4 coming with oxidized pads straight from the factory, I decided that reflowing in general is a poor practice. Just assume shit needs cleaned up underneath, and I want to see the evidence of a defect before I'm happy.

There is not and I don't believe there ever was a source for new, working replacement GPU chips. Anything you're going to find to buy is an untested pull from a scrap board, and the seller is hoping that you'll think you screwed up the installation and won't bother to fight to return it for a refund after it doesn't work. This is why I got so excited in the "Frankenstein 40nm CECHA" thread - known working chips to transplant from cheaper models.