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

[Pacorretaco]

Just to be clear, SONY didn't invent the mod chip that allows you to replace a 90nm RSX with a 65nm or 40nm. @botakompong's brother did. SONY doesn't need a mod chip to do this, they installed a newer model SYSCON chip with a FW update that allows it. We have attempted to replicate SONY's process and it didn't work. Then we found the "orbis" mod that has been allowing this for the last 5 years in Indonesia. It spoofs the SPI signal the SYSCON is expecting from a 90nm RSX. It's easier to install than trying to replicate SONY's method anyway.

SONY has been fine with allowing backwards compatible PS3 consoles die, consoles that could otherwise be saved with a 65nm or 40nm RSX. They have been siting on the procedure to do this since 2012 with no plans on ever doing it themselves again or releasing the instructions. There is no benefit to keeping it secret anymore, but they continue to. That's the SONY I know. I admire the engineers, but whoever made that decision deserves a block of coal for Christmas.

Hehehe, yesterday you were criticizing Sony for charging 128£ for this repair back in the day and now you are criticizing them because they no longer give the option.

Of course we can only thank botakompong and his late brother Kiaw.

They invented a workaround to do what Sony was already doing this back in the day. Not tokin swaps. RSX swaps.

 

[RIP-Felix]

This is why the "Orbis" Mod chip is a game changer. It allows you to install a 65nm or 40nm RSX which will not need to be replaced again.

Is this the part you are referring to @Pacorretaco? Perhaps that should read, "which will last much longer than a reball."

However, I think it's more likely the CPU fails before the RSX does again. And if the RSX outlasts the CPU, then technically I was correct. The RSX will not need to be replaced again, because the console is truely dead when the CPU dies. The CPU can't be replace with another easily.

 

[RIP-Felix]

Hehehe, yesterday you were criticizing Sony for charging 128£ for this repair back in the day and now you are criticizing them because they no longer give the option.

Of course we can only thank botakompong and his late brother Kiaw.

They invented a workaround to do what Sony was already doing this back in the day. Not tokin swaps. RSX swaps.

Man, you must have woke up on the wrong side of the Bed!

I didn't criticize them for their repair price. I said repairs are revenue, just like sales. Just because they loose money on console sales and repair, doesn't mean there is no incentive for them to reduce the BOM. All revenue sources add up, and any one of them is important in the larger picture. Which is why they do have an incentive to reduce reliability to a certain extent - to the extent that it doesn't initiate a recall or drive customers away. Building reliability in excess of that eats into profits. That it's a balance. And that everyone doing business does this. Reducing the cost of a single capacitor from 20 cents/unit to 10 cents/unit for millions of consoles will save the company ALOT of money (especially if the console is sold at a loss). Even if that cheaper capacitor reduces the console's lifespan by half, so long as the console lives beyond the warranty period, what do they care? The guarantee you agree to when you buy the console is that the console will last 1 year. If you believed it'll last longer, that's probably because you bought into their PR non-sense. Or it's because you know you will not use it as much as SONY's projects an average user will. Maybe you will get 3-5 years out of it, and that's good enough for you.

Let me be clear. If SONY still offered PS3 repair for $150 I'd gladly pay it! Especially since they started doing official 65nm and 40nm RSX swaps when the 90nm RSX's became unavailable. Chance are noadays, you would get a 40nm RSX every time SONY did a YLOD repair. That's definitely worth $150!

 

[Pacorretaco]

Is this the part you are referring to @Pacorretaco? Perhaps that should read, "which will last much longer than a reball."

However, I think it's more likely the CPU fails before the RSX does again. And if the RSX outlasts the CPU, then technically I was correct. The RSX will not need to be replaced again, because the console is truely dead when the CPU dies. The CPU can't be replace with another easily.

Yes, I'd probably say something like:
"which makes sourcing a working chip easier in case the original 90nm one is confirmed to be bad inside"

For all we know, it would also be possible for a reball to last more than a swap. If it was just the balls and the chip was OK, then replacing a good chip with a different good chip (and a black magic modchip) isn't necessarily that much better.

As far as the chip itself goes, there is still a bit of uncertainty because, just as before, whatever chip will come from another board somewhere. Which also has to be reworked again at least twice. This reduces those chances a bit as you say, and hope that the problem that this other board had was not another bad RSX internally. They are apparently more reliable than 90nm, but still can very much fail.

Unless Sony had made available new production of spare RSXs for this. Which they didn't. So we can criticize them for that.

Then of course there is the rest of the board. Changing the chip only changes the chip. I think you are right on that. Just as there is no single cause for 3034, there is no single cause for YLOD or other problems. We can be happy because we can now deal with most of them now, even the ones that seemed a dead end.

It's just that people here are desperate to jump to simple conclusions. In the same way that changing the tokins even if they are good is not the solution for all problems... Jumping to changing the RSX isn't either. Both address different things, but nothing else.
The first step is to make sure you need to take the step. Then take it.

Man, you must have woke up on the wrong side of the Bed!

Haha, yes Sorry if I sounded rude or something. I was talking more in general. At the end of the day, my point is simply be happy. Enjoy things while they last.

 

[ahcutemotor]

Hi Guys, sorry if this question had been asked before but I just couldn't find the correct answer.
Can the PS3 capacitors replacement be used with 470uF @ 2.0Volt Version?
I understands that almost all tutorials use 2.5V or 6.3V version.
It was because my seller sent the wrong capacitors to me and its rating is at 470uF @ 2.0V.

I also took note that newer PS3 such as Superslim model had this YLOD/Capacitor Issue resolved, but I noticed their SMD Capacitors were coded as 470 xxxxd. So what does that "d" means?

Example panasonic smd capacitor would have indicated "d" as 2.0v and "e" as 2.5v, however I cannot identify what is manufacturer of the capacitors at the CELL and RSX side. Can someone explain?

Thanks in advance!

 

[DeadEnd]

Hehehe, yes.
If you notice, that's what I was insinuating the whole time. There's something else going on with the RSX. Something nobody predicted. Not just the BGA defects or even the famous "bumps", the underfill... Or even the memory modules.

This is probably the real bad guy of this movie that is known to laugh every time heat is applied. Be it a fake tokin fix, a hairdryer trick, a towel trick, a "reflow"... Global warming... You name it.

But I didn't want to be too direct because people always read things the wrong way. It's an innate human desire for oversimplifying complex things. When in reality, it's just something that nobody can know 100%.

The reality is, that many things can cause a 3034/44XX RSX related error. There is still no single cause. And therefore it's dangerous to make assumptions.

This is why reballing is still the valid step forward. Because sometimes, it really is just the balls. And is the only way to isolate the problem further. Eliminate the possibility of problem under the chip vs on top of the chip. Perhaps more importantly, isolate what can be fixed vs what you cannot fix.

As I've said before, I wish I knew the real ratio of each type of failure, but I don't.
What I do know, is that many people including botakompong himself have reballed thousands of consoles and kept doing it because they know it's another type of failure that does happen.

This includes slim playstations which also get their share of RSX problems. Problems that again, often reballing can fix but not always. Remember the original 40nm CECHA official from SONY also had RSX problems.

If they invented the modchip is because they felt the need to solve a real problem that exists. But this doesn't mean that every single chip needs to be replaced.
When they still do so many reballings after inventing the modchip themselves, you can realize that reballing is neither stupid, nor a scam.

It's the correct step forward that shouldn't be skipped most cases. Of course different people have different priorities. Some are just little destroyers that just can't help it.

In any case we can be happy because thanks to them, (reballers at the end of the day) now even dead chips are no longer a Dead End.

I'm gonna have to disagree that reballing is a valid step forward, at least not on the BC consoles. First of all, botakompong himself has said that after reball it can last 3 to 6 months, MAYBE more (if you take care of the console and don't let it heat up too much). But the point still stands, like you said, whether it doesn't like too many heat cycles or has internal flaw inside the die, it is still a ticking time bomb. What is really the point of reballing now if you can simply solder a much more reliable and newer RSX? Yes, nothing lasts forever, but I'd much rather have a 65 or 40nm in there than keep reballing a 90nm that has been proven to have a tendency to fail internally. Again, if you have read what he was trying to tell us in the frankenstein thread, this type of internal short and separation in response to heat happens ONLY on the 90nm. So you won't have this type of fault on the newer revisions. Besides, are you going to rely on a service every 6 months? That's just playing the lottery, "Alright, I'll reball it and hope the issue was inside the balls". Aaannd by doing a reball on such a fragile chip, you are potentially sabotaging its lifetime and speeding up the same inevitable result (internal short that otherwise wouldn't happen on 65 or 40nm revisions).

Regarding the CECHA with RSX issues, well, we don't know if that case could have had other issues (could even be NecTokins). Didn't he say the console was in a pretty bad condition to begin with? We don't know much of its history. He didn't use syscon diagnostics on it or anything like that. So I wouldn't be jumping to conclusions just yet. Botakompong's method has existed for years, in his part of the world people are still using PS3s a lot. So I imagine he would have noticed if the swap was not such a long lasting solution either. I think, by telling us about how the issue only happens on

Btw, here is another good tip from his recent explanation, which means NecTokin is more often an issue on numbers 102 and 108, but rarely on 128.

+RSX CXD2971x, CXD2982x, CXD2991x: This type of RSX uses NEC TOKIN, if the needle on the analog meter shows movement as in the video, you can be sure that NEC TOKIN has a problem. Usually occurs on the NEC TOKIN which has the numbers 102 and 108 (Type CECHI,J,K,L and Slim DYN-001), on the number 128 (Type COK-SEM-DIA) there are also but rarely has a problem.

 

[Pacorretaco]

keep reballing

You see, the whole idea is that you only reball once. That's why it's the valid step in the troubleshooting. If and when it fails again (yes it can, but not always) that's when you swap the RSX. It used to be a dead end. Not anymore. We can be happy right? Meanwhile you can enjoy it working anyway. Guys like @squeept give 1 year warranty. That's because they don't think it's a scam. Otherwise they wouldn't risk their a$$ like that.

You want to skip reballing? As I said, different people have different priorities. You want to skip it, skip it. You don't care about the chip and don't want to give it the fair chance to work? Fair enough I guess.

I'm gonna have to disagree that reballing is a valid step forward

But that doesn't make reballing invalid, or a scam. By that logic, botakompong would be doing invalid things, or scams. I don't think he is doing either. Far from it.
Yet he has reballed more systems than he can keep track of. Including slims with RSX problems.

The point is, that not all RSX problems are an internal short inside the chip. He knows this, and we should too, instead of going crazy now and

Aaannd by doing a reball on such a fragile chip, you are potentially sabotaging its lifetime and speeding up the same inevitable result

So throw every chip to the trash instead to save time? That way it won't fail if it's already in the trash right? Oh but this includes every single 90nm RSX on existence then.

And sure... Maybe the internal short was solved in the 65 and 40nm revisions... But same as before, there are still more ways they can give 3034, and still more ways they can give RSX problems. Some have to do with the balls, others still not.
As the name suggests, reballing can only fix the balls. It's not magic, but it doesn't have to be invalid either.

Regarding the CECHA with RSX issues, well, we don't know if that case could have had other issues (could even be NecTokins). Didn't he say the console was in a pretty bad condition to begin with? We don't know much of its history

Ok hold on there. We do know it had graphical artifacting and he even tried a "reflow" on the new 40nm RSX and it even appeared to work. Not possible nec tokin. Clearly RSX problems even on 40nm.

I'm not arguing that this replacement is bad or it will last little time or anything. I'm the first one here to say it's amazing. If you look in this thread it was one of the first things I mentioned, long before it became real.
I'm just saying that not all cases need this. In the same way that not all systems need a capacitor replacement.
Not all "funny" people need a new brain.

This quick thinking can get very dangerous very fast. We are just talking of old machines, but I hope the day doesn't come where people are treated like this.

 

[DeadEnd]

You see, the whole idea is that you only reball once. That's why it's the valid step in the troubleshooting. If and when it fails again (yes it can, but not always) that's when you swap the RSX. It used to be a dead end. Not anymore. We can be happy right? Meanwhile you can enjoy it working anyway. Guys like @squeept give 1 year warranty. That's because they don't think it's a scam. Otherwise they wouldn't risk their a$$ like that.

You want to skip reballing? As I said, different people have different priorities. You want to skip it, skip it. You don't care about the chip and don't want to give it the fair chance to work? Fair enough I guess.

But that doesn't make reballing invalid, or a scam. By that logic, botakompong would be doing invalid things, or scams. I don't think he is doing either. Far from it.
Yet he has reballed more systems than he can keep track of. Including slims with RSX problems.

The point is, that not all RSX problems are an internal short inside the chip. He knows this, and we should too, instead of going crazy now and

So throw every chip to the trash instead to save time? That way it won't fail if it's already in the trash right? Oh but this includes every single 90nm RSX on existence then.

I detect a hint of bitterness in your text, as if I offended you personally by disagreeing. I think you don't need to take this so personally. I am not going to throw all your 90 nm chips into the trash bin

I never said reballing is a scam, I said I don't see the point in doing it on backwards compatible boards, it may be a valid option on other boards, of course. And I also never said ALL RSX problems have internal short, I was only talking about the CXD2971.

Let's rewind a bit, I did miss one point in Botakompong's text. I see now he has actually devised a way to know if it was indeed solder issue or not. It is all possible following his logic with the 'rules'. However, the internal short that can happen is actually a pretty big deal i'd say, so I'd rather just replace it. Plus he was telling us that after reballing it will last "a long time, 3-6 months maybe more". That doesn't sound like a long time to me to be honest...

And sure... Maybe the internal short was solved in the 65 and 40nm revisions... But same as before, there are still more ways they can give 3034, and still more ways they can give RSX problems. Some have to do with the balls, others still not.
As the name suggests, reballing can only fix the balls. It's not magic, but it doesn't have to be invalid either.

I never argued with that.

Ok hold on there. We do know it had graphical artifacting and he even tried a "reflow" on the new 40nm RSX and it even appeared to work. Not possible nec tokin. Clearly RSX problems even on 40nm.

I'm not arguing that this replacement is bad or it will last little time or anything. I'm the first one here to say it's amazing. If you look in this thread it was one of the first things I mentioned, long before it became real.
I'm just saying that not all cases need this. In the same way that not all systems need a capacitor replacement.
Not all "funny" people need a new brain.

This quick thinking can get very dangerous very fast. We are just talking of old machines, but I hope the day doesn't come where people are treated like this.

There is no quick thinking, I was simply implying to not jump to conclusions... Artifacting GPU, well, does that happen often on 40nm? I might have to reread the thread, but he was not the first owner of it, was he? He didn't get it straight from the service center. We don't know how the console was treated before it.

 

[RIP-Felix]

I'm gonna have to disagree that reballing is a valid step forward, at least not on the BC consoles. First of all, botakompong himself has said that after reball it can last 3 to 6 months, MAYBE more (if you take care of the console and don't let it heat up too much). But the point still stands, like you said, whether it doesn't like too many heat cycles or has internal flaw inside the die, it is still a ticking time bomb. What is really the point of reballing now if you can simply solder a much more reliable and newer RSX? Yes, nothing lasts forever, but I'd much rather have a 65 or 40nm in there than keep reballing a 90nm that has been proven to have a tendency to fail internally. Again, if you have read what he was trying to tell us in the frankenstein thread, this type of internal short and separation in response to heat happens ONLY on the 90nm. So you won't have this type of fault on the newer revisions. Besides, are you going to rely on a service every 6 months? That's just playing the lottery, "Alright, I'll reball it and hope the issue was inside the balls". Aaannd by doing a reball on such a fragile chip, you are potentially sabotaging its lifetime and speeding up the same inevitable result (internal short that otherwise wouldn't happen on 65 or 40nm revisions).

Regarding the CECHA with RSX issues, well, we don't know if that case could have had other issues (could even be NecTokins). Didn't he say the console was in a pretty bad condition to begin with? We don't know much of its history. He didn't use syscon diagnostics on it or anything like that. So I wouldn't be jumping to conclusions just yet. Botakompong's method has existed for years, in his part of the world people are still using PS3s a lot. So I imagine he would have noticed if the swap was not such a long lasting solution either. I think, by telling us about how the issue only happens on

Btw, here is another good tip from his recent explanation, which means NecTokin is more often an issue on numbers 102 and 108, but rarely on 128.

We still have to be objective and vet methods against the SYSCON and what we have come up with as a group together. When I hear claims like, "there's some flaw inside the 90nm RSX that degrades, which the 65nm and 40nm don't have," my QUESTION EVERYTHING meter rises. Yeah, what flaw? Just bcause the thing is complicated, doesn't mean there is some ethereal spirit haunting our RSX. Gremlins don't exist. If there's a flaw, there's an explanation.

Spoiler: about @botokompon's method

I'm not saying your were, but people do have a tendency to carve their hero's words in stone as if they're impervious to making mistakes.

Keep in mind that @botakompong's methods are qualitative not quantitative. He uses an analog multimeter and experience to come up with a working theory for diagnosing and fixing. While that's a great method and works for him, it's all intuition, gut feelings, and experience - not necessarily definitive diagnosis. The SYSCON codes and some combination of qualitative methods should be combined into a troubleshooting flowchart for a more quantitative diagnosis.

I would like to see what SYSCON codes pop up for each of his "rules." Also an oscilloscope measurement to characterize the voltage "kicks." There is more information to be gleaned from his method by getting sciencey with it. He has absolutely helpful insights to be sure, but difficult follow instructions, since much of it is lost in translation. I tried to sit down and follow it with the videos, but I couldn't tell which test points he was doing between tests and it's hard to tell from the pictures exactly where the test points are (C1, C2, etc.)

Qualitative analysis for your customer's benefit is one thing. Maybe you don't need to understand why it's bad, just replace and be done with it. That's a valid approach, it saves time and time is money. However, knowing why something fails is a worthy pursuit and may help us fix more consoles.

I think you have an unhealthy prejudice against these poor 90nm RSX's. As if they are born with a genetic defect that kills them before reaching adulthood. I challenge that assumption.

@squeept offers a 1 year waranty for his reballed consoles. Not all reballs are equal, just like how not all reflows are equal.
If kept cool and well maintained, who knows how long a well reballed 90nm RSX will last? It's often said a reflow only lasts about a year. If that's true then surely a reball should last longer than that! Perhaps the life of the console (which I define as when the CPU dies).

On the other hand, I have been giving thought to why the RSX is so much more susceptible to BGA defects than the CPU. I am aware that the CPU is considered to be stable. It runs even hotter than the RSX, but doesn't fail anywhere near the same rate, even with lead free balls. And @Pacorretaco makes a good point about other consoles like the latter models of XB360 and XBONE console not having an unacceptable rate of failure. Clearly BGA can be done in a way that's much less prone to failure (nowhere close to QFP using direct wire attached dies, or socketed LGA/PGA, but better than the RSX). I have 3 hypothesis'...

1. The RSX DIE is orientated at a 45-Degree (diamond) angle to the substrate. This might create odd twisting/shearing forces as the chip experiences thermal cycles. I'd like to see a simulation with a both orientations to see which one dies sooner. 0-degree (Square) like the CPU, or 45-degrees (Diamond) like it is.
2. The CPU has a hole in the motherboard directly below the DIE to allow the multileg MLCC bypass capacitors clearance and airflow. This Places the solder balls in a grid around the hole. As the hottest part of the die in the center of the substrate heats up, the substrate there can move up/down a little, acting like strain relief. Whereas the RSX has balls in the middle. When it's die heat up the substrate can't move and leverages a greater force against the balls on the outer edge.
3. The RAM chips being in the corners may also have a negative effect upon the BGA. Perhaps the heat they produce causes the corners to curl up more. Or maybe the balls/underfill adds ridigity to the corners and the twisting of #1 combine to flex the balls more in strange angles.

Again, simulation would be needed to model these effects.

 

[Pacorretaco]

Haha, sorry again if it sounded like that. I wasn't offended. I'm not a reballer, I'm not here as a business or anything. I'm just somebody with perhaps too much free time right now yes.
I'm sure the actual reballers are pretty used to this kind of talks anyway.

I never said reballing is a scam, I said I don't see the point in doing it on backwards compatible boards, it may be a valid option on other boards, of course. And I also never said ALL RSX problems have internal short, I was only talking about the CXD2971

Well not all RSX problems are an internal short on the CXD2971 either.
But look, I'm not even arguing with you. I agree there's obviously no point reballing an RSX that has an internal short. Just that not always is an internal short, and that's why reballing is the valid step.

Maybe it would even be possible to do some multimeter check thanks to this information before reballing. Assuming it's always the same pins. And we will have to thank botakompong again. Otherwise the actual reballing is still the way to find out.

 

[RIP-Felix]

Hi Guys, sorry if this question had been asked before but I just couldn't find the correct answer.
Can the PS3 capacitors replacement be used with 470uF @ 2.0Volt Version?
I understands that almost all tutorials use 2.5V or 6.3V version.
It was because my seller sent the wrong capacitors to me and its rating is at 470uF @ 2.0V.

I also took note that newer PS3 such as Superslim model had this YLOD/Capacitor Issue resolved, but I noticed their SMD Capacitors were coded as 470 xxxxd. So what does that "d" means?

Example panasonic smd capacitor would have indicated "d" as 2.0v and "e" as 2.5v, however I cannot identify what is manufacturer of the capacitors at the CELL and RSX side. Can someone explain?

Thanks in advance!

That voltage rating should be okay. IMO 6.3v is too high. If there is a surge or voltage spike the lower voltage rating will protect your CELL and RSX. I'd rather it burned out the capacitors than the GPU/CPU (kinda like a fuse). Operating voltage is 1-1.3v, so 2v is sufficient.

EDIT:
I realized that I only answered your question. It would be unethical if I didn't ask you why you bought the capacitors in the first place?

You should be starting Here. Only perform the tantalum replacement if you have an error 1002.

 

[Pacorretaco]

Plus he was telling us that after reballing it will last "a long time, 3-6 months maybe more". That doesn't sound like a long time to me to be honest...

Ok, you bring an interesting point. But Actually that's not what he was telling us. He's not even talking about reballing here. Actually he's talking precisely about one of the RSXs that have a confirmed internal short. He is saying that even one of these "terminal patient" RSX with intermittent short can last like 6 months or perhaps more after being "heated". Not necessarily reballed. He just says heated under 200c.
Which in reality I'd say it's impressive for something that's supposed to be doomed. That's why he says it's a long time. And I believe this. It is consistent with what some people claim that their fake nec tokin fix can last.

Artifacting GPU, well, does that happen often on 40nm? I might have to reread the thread, but he was not the first owner of it, was he? He didn't get it straight from the service center. We don't know how the console was treated before it.

Haha, yeah maybe they played football with the poor machine. I happen to have 2 of those with graphical artifacts after physical damage during shipment. But interestingly it seems that it was the RSX that was affected. Not the CPU for example or anything else besides the destroyed plastic case.
Sadly... Even without football these artifacts are not rare at all. Some eventually degrade into a GLOD. Not an intermittent short (which would always be a 3034 YLOD)
I was still talking about old fat models, but I've also seen slims and superslims like that yes.

 

[DeadEnd]

We still have to be objective and vet methods against the SYSCON and what we have come up with as a group together. When I hear claims like, "there's some flaw inside the 90nm RSX that degrades, which the 65nm and 40nm don't have," my QUESTION EVERYTHING meter rises. Yeah, what flaw? Just bcause the thing is complicated, doesn't mean there is some ethereal spirit haunting our RSX. Gremlins don't exist. If there's a flaw, there's an explanation.

Keep in mind that @botakompong's methods are qualitative not quantitative. He uses an analog multimeter and experience to come up with a working theory for diagnosing and fixing. While that's a great method and works for him, it's all intuition, gut feelings, and experience - not necessarily definitive diagnosis. The SYSCON codes and some combination of qualitative methods should be combined into a troubleshooting flowchart for a more quantitative diagnosis.

I would like to see what SYSCON codes pop up for each of his "rules." Also an oscilloscope measurement to characterize the voltage "kicks." There is more information to be gleaned from his method by getting sciencey with it. He has absolutely helpful insights to be sure, but difficult follow instructions, since much of it is lost in translation. I tried to sit down and follow it with the videos, but I couldn't tell which test points he was doing between tests and it's hard to tell from the pictures exactly where the test points are (C1, C2, etc.)

I think you have an unhealthy prejudice against these poor 90nm RSX's. As if they are born with a genetic defect that kills them before reaching adulthood. I challenge that assumption.

@squeept offers a 1 year waranty for his reballed consoles. Not all reballs are equal, just like how not all reflows are equal.
If kept cool and well maintained, who knows how long a well reballed 90nm RSX will last? It's often said a reflow only lasts about a year. If that's true then surely a reball should last longer than that! Perhaps the life of the console (which I define as when the CPU dies).

I see your point and I admire your dedication towards quantitative/definitive diagnosis. And true, he is not the almighty all-knowing playstation wizard either... I don't know if you noticed, but I rewrote one of his previous posts and tried to make it easier to read. I think I understood his last info quite well. I am also in touch with him on whatsapp so I can even ask about his thoughts on reballing...

In a way, I do have a prejudice. It just doesn't seem worth it playing the game of "how long will it last?" Of course, that's just my opinion. I am not forcing it on anybody else.

Ok, you bring an interesting point. But Actually that's not what he was telling us. He's not even talking about reballing here. Actually he's talking precisely about one of the RSXs that have a confirmed internal short. He is saying that even one of these "terminal patient" RSX with intermittent short can last like 6 months or perhaps more after being "heated". Not necessarily reballed. He just says heated under 200c.
Which in reality I'd say it's impressive for something that's supposed to be doomed. That's why he says it's a long time. And I believe this. It is consistent with what some people claim that their fake nec tokin fix can last.

Hmm, perhaps I misunderstood him. I think he was saying the opposite, that when there is a short which can be fixed by heat, then it needs to be replaced. But it's up to you if you want to keep heating it again and again instead of just replacing. And if it was never heated, then it will last a long time (I assumed he meant post reball, but maybe he just meant it can only last 3-6 months in general...) . I will try to clarify this with him.

Haha, yeah maybe they played football with the poor machine. I happen to have 2 of those with graphical artifacts after physical damage during shipment. But interestingly it seems that it was the RSX that was affected. Not the CPU for example or anything else besides the destroyed plastic case.
Sadly... Even without football these artifacts are not rare at all. Some eventually degrade into a GLOD. Not an intermittent short (which would always be a 3034 YLOD)
I was still talking about old fat models, but I've also seen slims like that yes.

Well, perhaps RSX from a superslim is the way to go then.

 

[Pacorretaco]

And true, he is not the almighty all-knowing playstation wizard either...

Sure... His brother was. He is just the next closest thing we have to the all-knowing wizard. Both total genius for sure.

I think he was saying the opposite, that when there is a short which can be fixed by heat, then it needs to be replaced.

Hmm yes and no. The way I interpret it, he's just talking about heating an intermittent short RSX in particular. Certainly not reballing. That's why he mentions 6months or more is a relatively long time.
So not quite the opposite either. Needs to be replaced but still can last months or years. But this is not about the balls.

In a way, I do have a prejudice. It just doesn't seem worth it playing the game of "how long will it last?" Of course, that's just my opinion. I am not forcing it on anybody else.

Well, we are just trying to say that it doesn't have to be always like that. 40nm doesn't necessarily last forever either.
For all we know, it would also be possible for a reball to last more than a swap. If it was just the balls and the chip was OK, then replacing a good chip with a different good chip (and a black magic modchip) isn't necessarily that much better either.

Well, perhaps RSX from a superslim is the way to go then.

Please no. This is what might have made me sound triggered because I see the problems coming with this kind of harmful thinking from a mile away. Don't turn a good thing into a bad thing... No need to become "racist" about the chips. There are working chips and not working chips. Making more arbitrary distinctions like these can only bring destruction. Be happy while it lasts. You also have several "ticking time bombs" inside your body. But you don't go searching for albino gorilla hearts to make your body "last forever". Luckily we are just talking about an old machine for which we have new treatments if it needs them.

I have edited my previous post. Just ask @vyktormvmpay25 about the slim AND superslim RSXs that have 3034 even after reball. They exist too. Maybe less, but the supposedly "good chips" can fail too.

 

[ahcutemotor]

Thanks @RIP-Felix for your excellent answer that I am truly looking for!
"Operating Voltage" (1-1.3v) are the key words here as I have been trying to find such technical data high and low on web.

I have helped my 2 of my friends to fix their old YLOD PS3 with "blunt assumption" capacitors replacement and it worked! Therefore I ordered another batch of 470uF / 2.5v capacitors (AlPol) but only to receive 2.0v version. I panicked thinking it will be a total waste and shipping back to China would be expensive. =/

Thank you for linking me to the syscon diagnostic, I will try it on next troubled PS3 if come across!

That voltage rating should be okay. IMO 6.3v is too high. If there is a surge or voltage spike the lower voltage rating will protect your CELL and RSX. I'd rather it burned out the capacitors than the GPU/CPU (kinda like a fuse). Operating voltage is 1-1.3v, so 2v is sufficient.

EDIT:
I realized that I only answered your question. It would be unethical if I didn't ask you why you bought the capacitors in the first place?

You should be starting Here. Only perform the tantalum replacement if you have an error 1002.

 

[RIP-Felix]

I should stop being cryptic and just say what I mean. I forget english is not everyone first language. What I mean is, I think you guys should consider the possibility the artifacting and "self healing" mechanism after the "heat test" are false positives due to thermalmechanical BGA reconnections and warping strain. Those are the key words I used reading through this thread. If I heard anyone say they used a heat gun or hair drier, or had artifacting, or it responded to the pressure test, I would place a "yes" into a "possible BGA defect" column.

He said, he hasn't used the SYSCON in diagnosing. That means he's not checking for 3034 before going about his qualitative troubleshooting process. 200C is just shy of lead-free melting point at 218C. So it warps the board as much as possible without actually reflowing. You will get a false positive if you're not ruling out 3034's first.

This is why we need him to find consoles that meet each step of his "rules" and get SYSCON codes. So we can identify if the "healing" process is nothing more than BGA false positive. Among other vetting purposes.

 

[RIP-Felix]

...I have helped my 2 of my friends to fix their old YLOD PS3 with "blunt assumption" capacitors replacement and it worked!...

Yeah, that's a common story.

May I trouble you for some information about the other console you have tried this on?

1. PS3 Model numbers (CECHA01, C03, G, L, H, 21xx slim models, etc?)
2. Model of Capacitors you used, or link to product page if you don't know.
3. Length of YLOD (eg, 2 seconds from PWR on to beeps)
   
4. Approximate date you first had success with the mods and how long they have been working.
5. Have your friends played with the PS3s much since they recieved it? Could you ask them to test it just to verify it's still working. Are they good enough friends to tell you if they stopped working?

I have been keeping a spreadsheet of everyone claiming to have attempted this mod and these details are more useful to me than 'I did the tantalum thing and it worked!' Unfortunately a lot of my DATA is skewed with exactly that kind of report and no update to let us know it's still working.

...but you, my friend, are not going to be one of them, right?


BTW: You may find these useful.
Schematics for A-G models can be found here.
SYSCON Script and Testpoints/diagnostic voltages

 

[Srassyfrass]

Just want to thank you guys for all your work. Going to get into reading the SYSCON myself. Got 20 broken PS3s myself and I hope to be of some help eventually

 

[ahcutemotor]

1. PS3 Model numbers (Working: CECHA06, CECHH06) (Tried but did not work: CECHE06)
2. Model of Capacitors you used, or link to product page if you don't know. (aliexpress: 470uF e96)
   https://a.aliexpress.com/_mqkbK3X
3. Length of YLOD (eg, 2 seconds from PWR on to beeps) (Working unit: around 2sec, Failed to repair unit: 0.5sec, as if something shorted)
   
4. Approximate date you first had success with the mods and how long they have been working. (Since May 2020, till now those 2 are still working.)
5. Have your friends played with the PS3s much since they recieved it? Could you ask them to test it just to verify it's still working. Are they good enough friends to tell you if they stopped working? (They usually play minimum 1 time per week due to current pandemic and all running GT5 and GT6 games.)

Btw, those Schematics and test points are very useful to me. As I have a few "dead" consoles which failed the capacitors replacement method.

Yeah, that's a common story.

May I trouble you for some information about the other console you have tried this on?

1. PS3 Model numbers (CECHA01, C03, G, L, H, 21xx slim models, etc?)
2. Model of Capacitors you used, or link to product page if you don't know.
3. Length of YLOD (eg, 2 seconds from PWR on to beeps)
   
4. Approximate date you first had success with the mods and how long they have been working.
5. Have your friends played with the PS3s much since they recieved it? Could you ask them to test it just to verify it's still working. Are they good enough friends to tell you if they stopped working?

I have been keeping a spreadsheet of everyone claiming to have attempted this mod and these details are more useful to me than 'I did the tantalum thing and it worked!' Unfortunately a lot of my DATA is skewed with exactly that kind of report and no update to let us know it's still working.

...but you, my friend, are not going to be one of them, right?


BTW: You may find these useful.
Schematics for A-G models can be found here.
SYSCON Script and Testpoints/diagnostic voltages

 

[DeadEnd]

Please no. This is what might have made me sound triggered because I see the problems coming with this kind of harmful thinking from a mile away. Don't turn a good thing into a bad thing... No need to become "racist" about the chips. There are working chips and not working chips. Making more arbitrary distinctions like these can only bring destruction. Be happy while it lasts. You also have several "ticking time bombs" inside your body. But you don't go searching for albino gorilla hearts to make your body "last forever". Luckily we are just talking about an old machine for which we have new treatments if it needs them.

I have edited my previous post. Just ask @vyktormvmpay25 about the slim AND superslim RSXs that have 3034 even after reball. They exist too. Maybe less, but the supposedly "good chips" can fail too.

I am going to still disagree with you, and I'm sorry if that triggers you. If it has been proven that the rates of failure in the original RSX are obviously higher than in others, then I am not going to be using that chip. Call it racist or prejudiced, but you are trying to equalize all of the chips, even though they are not all the same. Your logic is that just because some 40nm can fail too, then it somehow disqualifies the whole series of those chips. You easily swipe aside the fact that it fails far less as if it's insignificant, while it is actually a pretty big winning point right there. If you get 2 out of 10 slims with RSX failures but 8 out of 10 BC fats with RSX failures, that would be a great evidence to conclude that the original 90nm is a vulnerable part that would be better off replaced with another that is more durable. If you are replacing it with a potentially faulty 40nm, then yes, you are actually making the situation worse.

The comparison with human body is not at all appropriate. We are talking about a product, not a living thing. But let's do it your way, say you needed a heart transplant and they told you "Here we got several hearts, one has a tendency to give you a flat line of death sooner than others, but if you are careful, it might even last a few years, worst case you heat yourself up every now and then. Another is newer and more reliable, it needs a bit of extra work but it tends to fail less. The newest one in practice should be most reliable, but those also fail sometimes." Now go ahead and take the first one because "nothing lasts forever" anyways, right?

But back to the point, I am trying to figure out how to maximize reliability and minimize the risk of another YLOD . There are in theory several vulnerable parts , if we simply take statistics and see which ones are the most commoon/fastest to fail, we could, in theory, perform preventive maintenance by installing better parts instead of just waiting for another YLOD.