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

[Fallen777]

Interesting theory. I really thought the bad ones were under the board. I most likely will be replacing the top side as well but first id like to play around with webman fan control

I am running rebug cfw with webman, so i am wondering if anyone knows of a setting available where i can have the system start up in syscon, then after say 1 or 2 minutes, switch over to my webman fan control settings?

did you test the NECs first with multimeter before replacing.
From what @ElGris says the NECs were barely alive before have now died and needs to be replaced But you should test both NECs and Tantulums with multimeter first before getting practical

 

[wrx884]

I believe you removed the good NECs and left the bad ones. Why? You applied heat to remove them in first place, now we come back to my previous comment. Now those caps aren't warm, so you'll need to replace all of them.
If you wanna try something, remove all the NECs and don't put any tantalum there, and see if the console boots.

If u read the description of what was done im not so sure this is the case if anything it only means bad ones were left there not the fact good ones were removed.

did you test the NECs first with multimeter before replacing.
From what @ElGris says the NECs were barely alive before have now died and needs to be replaced But you should test both NECs and Tantulums with multimeter first before getting practical

Yes all Tant caps were tested and verified individually with an ESR tester b4 the work was done. The caps were already in question as they didn't work previously on other consoles but there were a couple members that had success with this type so that verified they do work.

 

[Fallen777]

If u read the description of what was done im not so sure this is the case if anything it only means bad ones were left there not the fact good ones were removed.



Yes all Tant caps were tested and verified individually with an ESR tester b4 the work was done. The caps were already in question as they didn't work previously on other consoles but there were a couple members that had success with this type so that verified they do work.

Apart from tantulums did you test the NECs on your ps3's motherborad.
In case you din't nows a good time to test them

 

[marciolsf]

I know my O-Scope is a different cheap brand but to get my good readings from the tokins i just had it set at 1uS/Div and 50mV/Div as advised. Also make sure its set on AC and not DC as we want to see the pulses it sends, not DC current. @squeept assuming this is correct to?

I get the same results when set to AC... When do you get your readings, though? Anytime you plug the probes, or during boot?

 

[wrx884]

Apart from tantulums did you test the NECs on your ps3's motherborad.
In case you din't nows a good time to test them

Unless they are removed completely and intact no u cant test them while still on board, @SeanRanklin only wanted the top side done so it was left at that once it was confirmed no shut downs during TTT2 and ran GT6 fine. The cheap O-scope i have, i dont have enough images to say they are good or bad yet as they sit on board but when the console was in my possession and tested for 3-4 days i never had these shutdowns, it ran perfectly fine until after it was posted back.

I get the same results when set to AC... When do you get your readings, though? Anytime you plug the probes, or during boot?

During boot u only have that short window when the console turns on briefly so u have to try capture the images in that time. If u can get hold of a working console with original tokins id suggest probing that console so u get the working images ur after and document them for urself.

 

[Fallen777]

Unless they are removed completely and intact no u cant test them while still on board, @SeanRanklin only wanted the top side done so it was left at that once it was confirmed no shut downs during TTT2 and ran GT6 fine. The cheap O-scope i have, i dont have enough images to say they are good or bad yet as they sit on board but when the console was in my possession and tested for 3-4 days i never had these shutdowns, it ran perfectly fine until after it was posted back.



During boot u only have that short window when the console turns on briefly so u have to try capture the images in that time. If u can get hold of a working console with original tokins id suggest probing that console so u get the working images ur after and document them for urself.

Aren't NECs unreliable now a days after all those years of their manufacture date.And should be replaced all at once with reliable replacements.

 

[squeept]

I know my O-Scope is a different cheap brand but to get my good readings from the tokins i just had it set at 1uS/Div and 50mV/Div as advised. Also make sure its set on AC and not DC as we want to see the pulses it sends, not DC current. @squeept assuming this is correct to?

at what time did you get those readings?

I just checked and I have mine set to DC coupling, which will just show you the whole signal. I checked to make sure, and all AC coupling did was center the same signal at zero volts. I think the signal is small and ugly enough that it doesn't matter.

As WRX said, you just have to have fast fingers on the memory buttons to catch it in the 2 to 3 seconds before it shuts back down.

 

[squeept]

If I'm not mistaken, from CECHA to CECHE they all use the same 12V 32A PSU as they both have 90nm RSX and 90nm CELL, so they should be drawing around the same current, except for the PS2 hardware inside CECHA and CECHB that should make them draw a little bit more in some cases I guess. Also the card reader could make a little difference but I don't think that it would be that noticeable. Except CECHB they all seem to hace card readers.

Unsurprisingly, popped the 10A fuse in my Fluke. $5 each, geez. This little experiment will wait for another day.

 

[squeept]

At this point with this A01 thats been clearly dropped

When they're dropped so hard the case is mangled, there are often torn traces under the RSX. For whatever reason, the CELL doesn't rip traces.

 

[squeept]

Hey guys.

I've replaced all but one tokins for each chip with 3x470uf 10V tantalums, and after seeing that it's still dead, added two more per chip, still didn't work.Then i noticed that many many capacitors on the chip side of the board are shorted or have low ohm continuity. Photo shows tantalums that are beeping with multi meter.

I wasn't the first one to work on this board, in fact, i found quite a few anomalies so i guess it's safe to presume the main chip connections have been bridged with a heat gun and the only fix is a reball, right?
If anyone has any ideas, i'd love to hear them, because i'm out of options on this PS3.
Oh and i get a 3 second YLOD, no different from when i got it with no tokins replaced.

Many of the caps on the board are going to appear shorted. It's normal.

 

[squeept]

Pretty much, but it's really the other way around -- heat is by-product of voltage. There is the general idea that capacitors can be "restored" by being heated, but I'm not too familiar with the principles to speak of it.

Why after heating the mobo on that zone, miraculously, the console boots again? This was a problem even for experts in rework jobs, since they believed this was only a bga problem, and not anything else. So, after the repair job was done, the console came back again to the repair shop 'cause the bgas cracked again, but they didn't, and that WASN'T the issue in the first place. This was the reason why many of these shop didn't want to make these kind of jobs on these consoles. But thanks to this, many others stole money from people by
only doing a worthless reflow that will last weeks, months..

Regular electrolytics can be restored temporarily with heat. This is one of the main things that caused me to pull my hair out for 30 some pages of this thread early on. Nobody has presented any non-anecdotal evidence or documentation anywhere of this kind of polymer capacitor being temporarily revived with heat. They have a self-healing mechanism that is DRIVEN by heat, which is where I think this misinformation is coming from. When areas of the polymer start to fail, their resistance increases. The increase triggers a spiral of more heat and more resistance until the failed areas are literally burned out of existence. Those are the burn marks you see when you pop the tops off. They burn up microscopic amounts at a time. The surrounding areas are completely unaffected, and the cap keeps on chugging like nothing happened, just with a tiny amount less capacitance. This is normal and expected during operation.

Now, it should be obvious why external heat triggering this reaction is not what is happening. It would need to reach the temperature required to literally BURN the substrate, and since you'd be doing it to the entire cap, you'd first melt the entire casing, and second, if you heated the entire cap to the temperature required to cause it to self-heal.... you'd cause the ENTIRE polymer to burn up.

On the 3 cases of verified bad TOKIN caps I have seen, I tried heating them while connected to the scope and observed no change whatsoever in the output. It is well documented, however, that heat can temporarily revive a dead chipset. A few degrees is all that's needed to warp the board enough to make mechanical contact in the case of a cracked BGA connection as well. This is why I keep saying the heat test is useless. Even if we work on the assumption that the caps can be revived with heat, it still hasn't narrowed anything down.

Next time I get a console that responds to heat, I'll try to remember to change out the caps and hopefully get a video of it still responding to heat before I reball it.

 

[marciolsf]

It seems that my little scope can't reliable display anything lower than 500mV, anything higher and it'll just draw a line at the very top of the display, even when I set the baseline at the bottom of the display. I can set it interval to 1uS, however, and get some readings... Sadly, I can't freeze frame at that speed.

Now, just for the sake of comparing things apples to apples, I removed all the tantalums from the Cell (both top and bottom), but left all the tantalums from the rsx in place. With that done, I compared scope readings between the Cell and the RSX, and they were interesting. See the video below

The first beep you hear is when I'm testing the RSX (which, again, has all the caps in). All I get with it is... a flat line. I mean, you can see it go up, and then down, but there's no visible wave.

Right after the RSX test, I test the Cell (remember, no caps) and you get some very jagged patterns fly by. So, here's my theory -- my cap installation is perfect, and my problem is elsewhere, not cap related at all. It's very likely that my lack of sufficient resolution is hiding any kind of "good patterns" out of the RSX reading... But I did a few readings on the Cell early on, when I removed the first 2 caps, and the change on the wave was immediately noticeable (it's too bad I didn't save those, but I'll record them when I put the caps back on).

So now I'm really curious about why the RSX reading is flat. Is it really just my lack of resolution, or something else? I'm in a bit of a denial, I don't necessarily think it's a BGA issue (for a variety of reasons).

I'm also really tempted to put all the Cell caps back on and then enable the tristate, just to see if it boots. I don't know that it will, but hey, science! If it does (big if), then it might prove that the problem is in the southbridge somewhere (since the tristate acts as a disconnect between syscon and southbridge -- I might be getting this slightly wrong), or maybe the firmware is corrupted (which is something that the psxdev wiki hinted at). Anyone around here experienced with hardware flashing can speak to this idea?

 

[squeept]

my little scope

There are a few weird issues that can lead to an oscilloscope not showing correct waveforms when they're pushed to their limits. I'm afraid I'm terrible at troubleshooting anything unless I have my hands on it in front of me.

I'll stop assaulting everyone to buy one for now, at least until I get my hands on another verified set of bad caps. I'll buy both of those models and maybe one or two more to test out then, since I can just sell them on eBay afterwards and get most of my money back.

 

[wrx884]

I'm also really tempted to put all the Cell caps back on and then enable the tristate, just to see if it boots. I don't know that it will, but hey, science! If it does (big if), then it might prove that the problem is in the southbridge somewhere (since the tristate acts as a disconnect between syscon and southbridge -- I might be getting this slightly wrong), or maybe the firmware is corrupted (which is something that the psxdev wiki hinted at). Anyone around here experienced with hardware flashing can speak to this idea?

It should boot, I cant remember what board ur working on but the tristate to ground to rule out corrupt firmware only works on NOR consoles from memory? and not NAND, i could be wrong tho im not full bottled on this hopefully others can chime in here? It is however a good test to do since most consoles run CFW these days now so u should always have this on ur list "to Test".

I just checked and I have mine set to DC coupling, which will just show you the whole signal. I checked to make sure, and all AC coupling did was center the same signal at zero volts. I think the signal is small and ugly enough that it doesn't matter.

As WRX said, you just have to have fast fingers on the memory buttons to catch it in the 2 to 3 seconds before it shuts back down.

I'll change mine back to DC and get some new images just for references, last time i probed with DC was on a board that was using the tant caps installed and the signal was like Marciolsf's so i thought the tants signals just weren't able to show such a clean signal like the Tokins can, for some reason on AC the signal i could capture appeared better but as u said it probably shouldn't really matter.

 

[joe musashi]

i have never seen here any one said that he made his ps3 work so good after he replace his neces with tantalum have someone here succeeded

 

[SeanRanklin]

i have never seen here any one said that he made his ps3 work so good after he replace his neces with tantalum have someone here succeeded

Really?

 

[SeanRanklin]

maybe you can disable webman mod and reboot. with console completely boot make a note of stock syscon temperatures than use webman mod

can confirm that disabling the fan in order to heat it quicker didn't do much. The shutdowns pretty much happen within the first 30 seconds now upon a cold boot. After that, it stays powered on for hours with no problems.

To anyone reading this, what is generally the cause of YLOD only at cold boots, but after that no shutdowns for hours and hours of use? What's my prognosis? How long can it keep working in this stage?

 

[pinky]

can confirm that disabling the fan in order to heat it quicker didn't do much. The shutdowns pretty much happen within the first 30 seconds now upon a cold boot. After that, it stays powered on for hours with no problems.

To anyone reading this, what is generally the cause of YLOD only at cold boots, but after that no shutdowns for hours and hours of use? What's my prognosis? How long can it keep working in this stage?

possibly the gradual heating of the system is causing the ylod. it's cold at first, but then it gets very hot. if the same temp is maintained, it may be less inclined to ylod. I think that's the reason the ps4 continues to cool the system even after shutdown (fans are still on).

it was suggested to turn the fans off on the 360 after replacing thermal paste. it's supposed to allow the paste to cook evenly, but I've never confirmed if that's what's happening. I unplugged the fan for a couple minutes to let it 2 ring (overheating) the first time I did it. it didn't stop the pos xenon from eventually dying though. I repaired it a couple times, but at that point, it wasn't worth it anymore. those things are ticking time bombs.

 

[evaldas256]

Many of the caps on the board are going to appear shorted. It's normal.

Oh, then i have no clue what else to do. Would cleaning the board with IPA possibly help?

 

[SeanRanklin]

possibly the gradual heating of the system is causing the ylod. it's cold at first, but then it gets very hot. if the same temp is maintained, it may be less inclined to ylod. I think that's the reason the ps4 continues to cool the system even after shutdown (fans are still on)..

That's... an interesting theory indeed. It might actually explain why my most recent cold boot YLOD happened quicker than ever before when I had syscon enabled.

I'll try setting the fans at a manual speed of 40% before I turn it off next and test the console boot up next day that way. If I survive the YLOD, your theory is correct!