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

[marciolsf]

Would those models be better ? Lowest ESR I could get...
https://www.mouser.ca/ProductDetail/KEMET/T528Z337M2R5ATE005?qs=/ha2pyFadugxZFgON0xyPFjNbM9lt0ZzZGOvaYNWuro=
https://www.mouser.ca/ProductDetail/Panasonic/ETPE330MA9GB?qs=sGAEpiMZZMtZ1n0r9vR22YIZno/SogTnwtMlR7acw6%2B%2BLmUPgvcHog==

Thanks!

There was a discussion a few pages back (only a few!) with @RIP-Felix and a few others about a target value for total both voltage and ESR. The original thinking was that total capacitance didn't matter so much as long as it was above the original, but I think we're moving towards keeping capacitance and ESR as close to the original spec as possible.

 

[RIP-Felix]

PS3 #3 - Part 1: Good Working Tokins​

@squeept Those look right, but I'm starting to think we need to zoom out to ~50uS/Div to see anything useful. No need to hold up the flip on your console, I got my own working unit now.

I just disassembled a new CECHA-01 (PS3#3). This one is working fine (no YLOD), except it doesn't read discs (so they said, I didn't even test...lol). Anyway, I wanted a stock working one to compare my measurements with the Tantalum capacitors (theoretically optimized capacitance and ESR). Here is what I came up with:

Spoiler

WORKING STOCK (PS3#3):

You can see this one doesn't YLOD. The yellow CH1 (or is it green, I'm color blind) is the CPU and the RSX is the blue CH2. The last plateau is when it finishes boot and idles stable from there on. Here is the YLOD PS3#2 again for comparison:

TANTALUM REPLACEMENT, STILL YLOD (PS3#2):

What's interesting is that the CPU follows the same pattern except for that last plateau before the YLOD (PS3#2). It goes to a High voltage state instead of dropping like it does for the working console (PS3#3).

The RSX's voltage doesn't fluctuate as much on the working PS3. If you look closely it has the same four states, but they are less pronounced. The scale is the same, so it's not that. I don't know what it means, if anything, but it's interesting. Speaking of interesting signals...Remember, before I zoomed in on the last plateau region? This is what it looks like on PS3#2, just before the YLOD:

TANTALUM REPLACEMENT, STILL YLOD (PS3#2):

Here is the same location on the working PS3#3...
WORKING STOCK (PS3#3):

It looks like the tokins have more noise, but the vertical axis is zoomed in a bit more on this image. They look the same as far as I can tell. So this is normal behavior. I would expect the NEC/TOKINs and Tantalums to look different if there was an appreciable difference. Notice that this is the only region that I can get both a working console and a YLOD console to shoe comparable regions. So I would try to locate this spot and base your before/after conclusions on it.

Also, notice the time scale is set to 50uS. I think this is how far you need to be zoomed out to catch switching signals from the voltage regulators. Remember that their switching frequency is between 300KHz-1MHz, so if you zoom into 1uS/Div (1MHz/Div) it's really too close to see anything useful. That's what I'm thinking now.

Now that I have a console that Idles, I can stop using the single trigger capture feature and see the voltages in real-time. This allowed me to find some interesting features of the idling RSX. Not that this isn't possible with the YLOD console because it doesn't get this far into the boot process and these never show up. The voltages stabilize here as well, at least during idle:

WORKING STOCK (PS3#3):

The RSX looks almost like a square wave. It reminds me of the sync pulse in analog video, but this is way less pronounced and irregular on the low state. I thought, "maybe it's the clock," but it's not constant. It fluctuates. Maybe this is switching noise from the VRM, but the time scale is way too far out for that to make sense. Each pulse is like 7ms long, that's a long time for a processor. It's deliberate, like some kind of communication or sync pulse. Honestly, I don't know what it is, but again it's something of note I haven't seen before.

Continued on Part 2 here...

 

[squidfox]

This might not yield any usable results, but can anyone (@squeept , @RIP-Felix ) do some FFT and diffing for me?

One measurement that could be of interest is to see what the FFT of the signal on the input leg to the caps look like, and how the same signal on the output leg looks like.

Another one that could be of interest is to see the difference between input and output legs. Probe 1 on cap in, probe 2 on cap out and do a subtraction, that way the only thing that is left is noise. (And then if the raw noise data can be captured do some FFT post processing on a PC to see the spectrum of the noise)

Maybe this makes no sense as the singals are not periodic but maybe it is worth a try.

 

[squidfox]

I ordered some "original" NECs from Aliexpress, and when all of my soldering equipment arrives I will try to solder them on and see how it goes. Too bad there is always a tool missing. Now that I have a soldering iron, hot air station and a scope I really want to dip my toes into BGA stuff, but the proper machines cost way too much, and the 200 USD junk that I ordered was good for nothing

 

[demetris]

Would those models be better ? Lowest ESR I could get...
https://www.mouser.ca/ProductDetail...=/ha2pyFadugxZFgON0xyPFjNbM9lt0ZzZGOvaYNWuro=
https://www.mouser.ca/ProductDetail...r9vR22YIZno/SogTnwtMlR7acw6%2B%2BLmUPgvcHog==

Thanks!

I have the Panasonic ones and they work fine.
I don't know kemet and i wouldn't buy a kemet even if it was 1/2 the price

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk

 

[SeraphTC]

Forgive me if I'm posting this in the wrong place, but does anyone in the UK offer a PS3 delid/thermal paste replacement/cleaning/nec tokin replacement service?

I bought a used PS3 60GB fat from ebay in january. For various reasons I didn't have a lot of time to play around with it, although I did have it running for a few hours to verify that it was working at the time. I was tidying my office and relocating things this week, turned it on, and got a YLOD.

It had supposedly been refurbed before I bought it, with the thermal paste replaced (although the tokins hadn't been touched). The seller seemed ok (I did a lot of research at the time and fired a lot of questions at them before I purchased), but they have since removed their account from ebay, so it's possible I was scammed, or I just got unlucky.

I'm not worried about the money I spent, I'd just really like a working system, and I neither have the tools nor the experience to perform the job myself.

 

[demetris]

Smells like a reflowed junk to me. I would be worried i throw good money away. No one would touching this with confidence.

Στάλθηκε από το POCOPHONE F1 μου χρησιμοποιώντας Tapatalk

 

[SeraphTC]

It's possible.

I had a long conversation with the seller, who insisted it was not reflowed, and that the paste had been replaced over and under the IHS. Based on my research at the time (I did a lot) he knew what he was talking about.

Ofc knowing what he was talking about doesn't mean that he did what he had claimed.

Whether I buy another PS3 or stick with this one, I would still be looking for someone trustworthy to perform a full refurb including tokins.

 

[RIP-Felix]

This might not yield any usable results, but can anyone (@squeept , @RIP-Felix )
One measurement that could be of interest is to see what the FFT of the signal on the input leg to the caps look like, and how the same signal on the output leg looks like.

I tried FFT, but I need to read up more using it properly. I messed with it a bit, but setting up the options really changes the outcome, so I don't know exactly how I should set it up yet. The signal I latched onto only appears after bootup. It's always there during system Idle, but disappears once you initiate the power off sequence. I don't know what it is, but I don't think it's noise.
EDIT: I let the scope auto latch on the signal and it picked it up pretty good. Then I looked up some videos on FFT and came up with these. I tried a variety of settings to try and get the peaks, but I'm still not sure of the significance yet.

Spoiler

Another one that could be of interest is to see the difference between input and output legs. Probe 1 on cap in, probe 2 on cap out and do a subtraction, that way the only thing that is left is noise. (And then if the raw noise data can be captured do some FFT post processing on a PC to see the spectrum of the noise).

I won't be able to get this for the CPU, @squeept has a better setup for getting that. My test bench doesn't break the power supply out, so it's in the way of the output side of the CPU. I can only get at the input side through a small hole I cut in the RF shield. So here's the RSX.

Spoiler

RSX with OG NEC/TOKINs. CH1 (Yellow) is output side of tokin (+/GND) and CH2 is Input side:

Difference noise appears to be about 30mV P-P, but this is using hires acquire mode and using normal makes it look noisier. I'm not sure if there's a correct way to do it, but this way looks cleaner.

 

[squeept]

Last one is back together, and the next one isn't a backwards compatible. I can dick around with another A01 in about a week if it arrives on time.

It's been almost 20 years, and just thinking about diff eq or fourier transforms will still give me nightmares for a week. At the risk of triggering my PTSD, what is that going to tell us in the context of what we're working on here? I really don't want to refresh my memory with google and end up drooling in the corner.

 

[squidfox]

I tried FFT, but I need to read up more using it properly. I messed with it a bit, but setting up the options really changes the outcome, so I don't know exactly how I should set it up yet. The signal I latched onto only appears after bootup. It's always there during system Idle, but disappears once you initiate the power off sequence. I don't know what it is, but I don't think it's noise.
EDIT: I let the scope auto latch on the signal and it picked it up pretty good. Then I looked up some videos on FFT and came up with these. I tried a variety of settings to try and get the peaks, but I'm still not sure of the significance yet.

Spoiler

View attachment 28576 View attachment 28577 View attachment 28571 View attachment 28572 View attachment 28573 View attachment 28574 View attachment 28575

I won't be able to get this for the CPU, @squeept has a better setup for getting that. My test bench doesn't break the power supply out, so it's in the way of the output side of the CPU. I can only get at the input side through a small hole I cut in the RF shield. So here's the RSX.

Spoiler

RSX with OG NEC/TOKINs. CH1 (Yellow) is output side of tokin (+/GND) and CH2 is Input side:
View attachment 28570 View attachment 28568 View attachment 28569
Difference noise appears to be about 30mV P-P, but this is using hires acquire mode and using normal makes it look noisier. I'm not sure if there's a correct way to do it, but this way looks cleaner.

Thank you for these measurements! Was the FFT only done on the input legs? Too bad these are taken while in idle mode as that signal is not that interesting in the context of YLOD, but what is interesting is that I thought the caps would remove much more noise than this. Can you also do a capture on a YLOD console for comparison (of the startup signal, not the idle one)? Even with 30 mV VPP removed the signal is still looking noisy as hell (guess that is why there are so many caps under the CPU/GPU). Maybe the chips are cycling between different modes of operation for power saving measures?

Last one is back together, and the next one isn't a backwards compatible. I can dick around with another A01 in about a week if it arrives on time.

It's been almost 20 years, and just thinking about diff eq or fourier transforms will still give me nightmares for a week. At the risk of triggering my PTSD, what is that going to tell us in the context of what we're working on here? I really don't want to refresh my memory with google and end up drooling in the corner.

What I was looking for was the filtering characteristics of the cap on a known good console and on one that has YLOD most likely due to the caps. AFAIK you don't have to do the FFT yourself, the scope should do it.

What I wanted to capture is not a picture of the signal on a working and YLOD console, but the how the noise looks like on one that is good and one that is bad. Since the input signal can vary slightly between consoles, the noise profile should mostly stay the same so it would be easier to see if a console has a YLOD due to the caps maybe?

Don't know it yet, this is all speculation on my side, maybe these measurements are worth nothing, but maybe it helps someone.

 

[RIP-Felix]

It's been almost 20 years, and just thinking about diff eq or fourier transforms will still give me nightmares for a week. At the risk of triggering my PTSD, what is that going to tell us...

Luckily I was able to avoid differential equations! In this case subtracting the output noise from the input noise on the tokin would allow us to see the effect the capacitor is having. How much is it reducing it, basically. FFT is useful if you can latch onto a signal. It shows the spectrum of signals contributing to the waveform. So you get a peak for each sin wave that contributes to a square wave, for example. In our case it "might" reveal what frequencies of the noise. That could be useful to design a bypass cap array that would more effectively target them. I don't think it's necessary when the noise is less than 50mV p-p. My tantalum cap array appeared to be working as good as the OG tokins, so I don't think there's any point, but that console doesn't boot far enough to see the idle signal. If the reball goes well, then I'll be able to compare them.

Thank you for these measurements! Was the FFT only done on the input legs? Too bad these are taken while in idle mode as that signal is not that interesting in the context of YLOD, but what is interesting is that I thought the caps would remove much more noise than this. Can you also do a capture on a YLOD console for comparison (of the startup signal, not the idle one)? Even with 30 mV VPP removed the signal is still looking noisy as hell (guess that is why there are so many caps under the CPU/GPU). Maybe the chips are cycling between different modes of operation for power saving measures?



What I was looking for was the filtering characteristics of the cap on a known good console and on one that has YLOD most likely due to the caps. AFAIK you don't have to do the FFT yourself, the scope should do it.

What I wanted to capture is not a picture of the signal on a working and YLOD console, but the how the noise looks like on one that is good and one that is bad. Since the input signal can vary slightly between consoles, the noise profile should mostly stay the same so it would be easier to see if a console has a YLOD due to the caps maybe?

Don't know it yet, this is all speculation on my side, maybe these measurements are worth nothing, but maybe it helps someone.

CH1 is output side, CH2 is input. FFT works on signals, I'll mess around with some YLOD consoles tonight. Last night I received PS3#4 (CECHA01). This one was described as untested, but I bought it because it had the warranty sticker inplace. So this console has never been opened. It does have a YLOD. So this gives me a pristine YLOD with it's OG tokins. So now I have a working console with tokins, YLOD with tokins, and YLOD with tantalum caps. This should make capturing images of what's happening with these caps easier.

Note that those scope images are zoomed in to 10mV/Div! The noise may look big, but it's not really. It's also worth noting that I now have 3 YLOD consoles. 2 of them were not permanently fixed with tantalum! Now we'll see if the tokins on a YLOD look the same as a working console, and a YLOD console with tantalum caps. If they all look the same, then there's no evidence to suggest that the bypass caps are the problem - at least with these 3 consoles. But then, 3 consoles in a row with good tokins is significant, since the narrative (myth?!) is that the YLOD is caused more often by the tokins. So far my data doesn't support that conclusion, but let's see what the tokins on this YLOD look like.

Stay tuned...

 

[squeept]

Gotcha. Diff Eq almost made me switch majors to business. Maybe I need to go "back to school" this winter with my old circuits textbook while I'm hiding from covid until spring. I was a CpE way more on the embedded systems and programming side. I've been an overqualified repair man from day one, so I drank most of that knowledge away years ago.

Definitely unnecessary from the diagnosis standpoint, the failed caps look so significantly different. Although it may be useful on non-BC models since they appear to be more picky about noise issues, and it looks like they either need NOS TOKIN or a homebrew, well designed solution beyond just tantalum to ensure they work right 100% of the time.

I also expect that since we're pushing the limits of these scopes already, the results from math operations like fft are going to be.... sad. So I don't know how much help this will be without a much more expensive scope or a spectrum analyzer.

I'll update the "good" images in my signature with the new cleaner images once I get the next one on the bench.

 

[RIP-Felix]

Definitely unnecessary from the diagnosis standpoint, the failed caps look so significantly different. Although it may be useful on non-BC models since they appear to be more picky about noise issues, and it looks like they either need NOS TOKIN or a homebrew, well designed solution beyond just tantalum to ensure they work right 100% of the time.

Well, I'm only interested in BC consoles, soo...I'm not going down that rabbit hole to hell!

...I also expect that since we're pushing the limits of these scopes already, the results from math operations like fft are going to be.... sad. So I don't know how much help this will be without a much more expensive scope or a spectrum analyzer.

+1

I'll update the "good" images in my signature with the new cleaner images once I get the next one on the bench.

I looked into this more and what HiRes mode does is average the extra samples instead of throwing them away. So on normal mode your scope is sampling at some rate (Sa/s), but depending on your horizontal resolution setting it doesn't display every data point on screen. It may only display 1 out of every ten, or whatever. In HiRes mode it averages those samples to display one. Since there are transient noise events that might get captured in one sample at (say 100mV), but the other nine are closer to the noise floor (say 30mV), then HiRes mode will average that sample out and the resulting image looks closer to the noise floor.

If the point is to measure noise, use normal mode. Hires mode will artificially reduce it. If the point is to try and visualize certain patterns that could be hiding in the noise, HiRes mode is a good way to see it. This is the take away.

 

[RIP-Felix]

PS3 #3 - Update
(...continued from part 1 here.)​

I wanted to post some backstory for posterity. It's important to record the state the console was in when I received it, so I can better tell the story later. And it's best to do so when the memory is fresh.

PS3 #3 was described as "turns on, doesn't read disks" Upon receiving I did test to see if it would turn on, and it does no problem. However, I was primarily interested in getting Oscilloscope measurements from it's NEC/TOKINs. I did not test to see if it outputs video or reads disks...lol. During the teardown I made these notes:

Spoiler

1. The manufacturer warranty sticker (tamper evident) was missing, indicating that someone may have opened the console before. The rubber foot was torn.
2. The torx security screw was in place. Often people loose the metal retention clip that this security torx screws into. It was also in place.
   
3. When most people reassemble their PS3, they tend to only use the 7 long screws. The Card Reader is also attached from under the top shell with 2 of the small Machine screws that the Bluetooth/WiFi module and Power supply are attached with. These 2 screws are recessed in a hole and are hard to get back in without dropping them and having to open the case back up to retrieve them. The card reader is clipped into place, so most people leave them off. However, these were fastened correctly where they belong.
4. The top shell opened fine. The clip wasn't broken. Also the small plastic hinges that the top shell pivots on were all unbroken. These are often broke off!
5. See picture below:
   
6. I disassembled the BluRay drive and it was reseated correctly, no game inside. The only non-stock bit was that the RF shielding was missing, indicating it was a replacement. Hopefully this doesn't mean the daughter board is different than the original. If that's the case the reason the disks wont read is because the BluRay drive isn't married to this motherboard. I'll have to remarry them.
   
7. All the screws were accounted for and had that tight "crack" when you first start unscrewing them. It's indicative of having never being unscrewed since assembly.
8. Interestingly the NEC/TOKINs on the bottom side of the motherboard did not have that thermal paper cover/tape??? Either someone had been inside the case and did not replace this tape, or it was never installed in the first place. This could indicate that this was a SONY refurbished unit, but I didn't see any sticker to indicate it. Game Stop refurbished units at least have a sticker with a date and service tech initials indicating when the console was repaired and by whom. I have a certified premium refirbished CECHA01 that does anyway. It has been working fine for the 10 years or so I've owned it (though I modified it a bunch over the years...some of which I regret). Anyway, I don't know if that thermal tape stuff was always on A01's when they came off the line. Maybe SONY stopped using it at some point during production of the A models, or this was a SONY refurbished and they had stopped using it by them. But I've never opened a SONY refurbished unit and don't know if they use a stickier indicating when it was last worked on, as is usual industry practice. If they do and that Thermal tape should be there, then this must be a well done repair by some 3rd party who knew what they were doing!
9. Besides the usual dust bunnies, the motherboard was cleaned well of flux residue. If someone performed a reball, there is no evidence of it due to left over flux. It was very well cleaned off if they did.
10. There was a bit of foam still stuck to the MB RF Shiled, indicating the 2 hadn't been separated before.
    
11. Thermal compound was your average cheap white crap over applied such that it globed over the sides of the chips. If this was a professional service doing repairs in bulk, this is the mixture of professionalism and cheap crap I would expect to see.
12. During my oscilloscope measurements I noticed that the fan was ramping up through it's profile. I know that the SYSCON temperature profile is insanely generous toward allowing the chips to get hot before intruding on your quiet. So even hearing the fan ramp up once was my cue to end the test and shutdown. I knew I would need to delid, everyone does!!! After deliding (my best job yet), it's clear that they had never been delidded.
    
    This doesn't surprise me, as professional repair shops that offered reballs at scale, back in the day, often did not remove the Heat Spreaders. Usually the reason for the YLOD was overheating and just reballing without replacing the TIC between the heat spreader and die will just lead to another YLOD. In the picture above you can see that the CPU has lost all of it's TC and the RSX has fared better, but was in need of some love. This is probably why this console is still alive and doesn't have a YLOD - The RSX was still able to sink it's heat some. That's a good sign that this console may last.
    
    
    
    That CPU though...OUCH! That's the reason for the fan ramping up, for sure!
13. I applied new MX4 and reassembled back to the test bench stage. It just went through a burn in cycle. I let it idle until the fan ramped up to the first increase, then shutdown. Before, within a minute of turning it on, it would start ramping up until it was quite loud within five minutes. What's interesting is that touching the power supply revealed very little heat - meaning heat from the CPU die was causing the fan to ramp up, but wasn't transferring to the Power supply which sits directly above it. If the heat was being transferred from the die to the heatsink and RF shielding, then it would also have heated the power supply. After new paste, it took about 10 minutes for the fan to ramp up to the first step increase and the whole console was toasty. This means that the heat is being evenly spread across the whole console, sinking as designed. The fact it was so hot is a good thing because it means the heat is getting off the die and BGA! As for the fan profile, that's total SYSCON crap and needs a jailbreak to fix, but at least the heat is sinking and the fan can do it's job. I just turned it back on and will repeat the burn in a few more times just to relieve stress and flatten the TC.

I still need to test it out with a TV to see if I get a picture. Then I'll test the bluray drive to see if it doesn't read disks, like described. I have working BR drives, so I can swap the daughter board in and see if that's it. I hope I don't need to remarry the BR daughter board to the MB. If the BR drive was replaced along with the daughter board, then I'll have no choice. Lastly, I need to jailbreak to get control of the fan profile. Anyway that's the story so far with PS3#3. This one I'm using as control for measurements, so it's good to document the console's history as best I can in case it makes a difference later.

Continued here...

 

[squeept]

The tape isn't always there, I'm positive of that. I believe the drive shielding isn't always present, either. I try to buy original warranty sealed consoles as often as possible. I've never actually made a mental note of that shielding in correlation to a warranty sticker, but I'd guess the fact that I've seen so many without it would mean it's normal. I don't think I've seen any stickers from Sony warranty repairs except on the drive. Sounds like someone was actually thwarted by the security torx and immediately gave up without a fight.

 

[RIP-Felix]

I think you're right. I compared the RF shielding on the BluRay of another A01 to this one, and it has the pads for those grounding pins (those things that look like the contact switches for the PWR & Eject sensors), but they were not populated. Whereas they are on BR drives that don't have the RF shielding. This must be a cost down measure. And if this was a cost down A model, then it makes sense that the tape on the TOKINs was gone too.

Upon further testing, the console does output video and seems to work fine, with the exception that it doesn't read disks. So I got exactly what was described. I switched the BR daughter board over to the BR drive of PS3#2 which YLOD during a PS2 game (with the game still in there BTW). Then I tested it out on PS3#3. It read the disk just fine! I played that PS2 game for about an hour, then tested a PS3 title for another 30 minutes. The fan only kicked into the second ramp level. So I closed up the console and it's now my primary rig.

Combined with other evidence I would conclude this was a lucky find. That elusive case where a laser went bad on the BR drive, no YLOD yet, and where someone was actually thwarted by the security screw! Actually that makes sense, because the BluRay drive still took and ejected disks. So the person didn't have to tear it open to retrieve their game (something I've seen done quite brutally before). That laser probably saved this console's life...lol! Delid, new TC, a new laser, and this thing should be good to go for years to come. I had to pay a little more for it, but I'd say it's a well spent $75!

 

[RIP-Felix]

PS3 #4: Teardown & Delid​

PS3 #4 was described as "from an estate auction and I do not have the wires to test. Is sold As-Is for parts or repair." Upon receiving I tested it and it has a fast YLOD. This one was $80 shipped. I wasn't so lucky with this one! I will note that the listing was a bit decieving, I could tell that certain angles were missing in the pictures. People do this to hide things and try to get a better price. However, I took a chance that the seller may not be lieing when they said, "I do not have the wires to test." So there was a remote chance the thing could be be working. Confidence bias, I know, but that's not the real reason I bought it. In one of the pictures the warranty sticker could be seen!

Spoiler

Yeah, it's out of focus and really only the outline can be seen, but it's a clean line - as if it hadn't been peeled. I was taking the chance that it meant the sticker was in place. After receiving I can see why the seller didn't have a picture from all the angles:

Well that could explain the YLOD. If it fell and jolted the chips, that certainly can do it. But this console has another YLOD inducing problem...here's a hint:

Now it was time to break the warranty seal, and see what we have inside:

Security screw clip is in place. I taped it in so I don't loose it. This is how you do it people, please take note!

These are the card reader screws (left) that people often do not replace when reassembling.

Here is the first open. Look at all that dust! I'm afraid it only gets worse as I dig deeper.

Note that the ribbon cable is taped in place and unbent. All the plastic hinges are intact. Even the curl in the slack for the WiFi cable is is present. That's a HUGE clue the console has never been opened. I noticed this on PS3#3 as well. SO now I know it wasn't opened either. Cool!

Also, here is the BluRay drive daughter board. It doesn't have the RF shield either, so that's defiantly not always there.

My hypothesis is that the cost down A model doesn't have the RF shield or NEC/TOKIN tape on the backside of the motherboard. So we'll see when I get down to it...

"Why is my fan so loud? What causes a console to overheat?", you ask?

DUST!!! SO MUCH DUST!!! I had to take this thing outside to blow it all off. No way that crap is getting in my house and in my lungs. Oh, and it wasn't all dust. There were cockroach carapaces too! Disgusting!

Here is the heatsink foam sticking to the motherboard RF shielding. If you see this, there a good chance your console hasn't been opened before.

Even the HS spring screws had dust bunnies!

Here's SONY's application of thermal grease.

Here's the Delid...

The RSX has lost some contact but not all. The CPU is pretty bad too. Yeah, this is only proving my point. No matter what, you're chips need a delid! The paste is gone, in need of love.

And now let's see if my hyopthesis about the cost down A models without BR drive RF shield also lack the TOKIN tape...

Yup!

Something worth noting is that this console has a mixture of Black silicone thermal pads and white pads. I've seen all black/grey ones, all white, and sometimes red pads. 2 unopened console have had the black, grey, and white ones. I'm not sure if the red ones are stock, but I can confirm that White, Grey and black ones are.

I cleaned the board up and will begin testing on it over the next few weeks. I need to solder the SYSCON serial conductors, then get an error log off of it. Then I'll get some scope images off it. We'll see how they compare to the same info I got off of PS3#2 (YLOD that I performed the Tantalum caps "fix" on - which didn't fix it BTW). That'll be in the next update for PS3#4. I suspect I'll get to this before I perform a reball on PS3#2. I'm still waiting on a BGA square nozzle for my Hot air rework station. It's still on the boat from china, so it could take awhile.

Continued here...

 

[Kleon1876]

I have a cechh01 which had ylod. I took it apart and replaced one nec tokin on the non RSX side with 4 capacitors. I tried it and still ylod. I delid the gpu and cpu carefully without scratching any of the chips. Reapplied mx4 and turned it in. I got GLOD. I tried booting into safe mode, got the beeps etc still nothing on screen. I removed the hdd and still nothing. The controller keeps blinking. I get the hdd light blinking upon boot but then it stops blinking. I took the capacitors i put on off the board off and tried it again. Same thing. Glod. My guess is I'm not soldering the capacitors correctly on the rail or I'm shorting something somewhere. Or maybe I messed up the repasting after delid. Any suggestions? Has anyone removed a nec tokin and tried booting. Do you get ylod or glod?Thanks in advance.

 

[Danook28]

My ps3 fat model cechc04... Wich nec tokin must use...