PS3 Fault finding YLOD with the SYSCON - First steps and Error reporting

[tracedg0d]

Just wanted to chime in and report a successful YLOD repair while utilizing this diagnostic method. I was able to restore a COK-002 with a 0xa0403034 RSX error by re-flowing with a hot air station at 380C for 3 minutes. No telling how long it will last, but as of now it is working. I'll report back on how long it takes to die again. This is without a doubt a crucial step in diagnosing these YLOD systems properly!

 

[patricksouza472]

Just wanted to chime in and report a successful YLOD repair while utilizing this diagnostic method. I was able to restore a COK-002 with a 0xa0403034 RSX error by re-flowing with a hot air station at 380C for 3 minutes. No telling how long it will last, but as of now it is working. I'll report back on how long it takes to die again. This is without a doubt a crucial step in diagnosing these YLOD systems properly!

Congrats! Did you use a preheater or just hot air? Hope it survives for years to come

 

[patricksouza472]

@db260179 I have tested the caps on that +1.2V_MC2_VDDIO line and compared them to another known board, the caps on that line are just fine. Those caps between CELL and RSX seem to be shorted (can't compare anymore as I no longer have access to that other board and did not check it beforehand), also there are shorts on +1.5V_RSX_VDDIO. Should I just give up on this?

EDIT: Removed some tantalums from the RSX chip, now instant A0093004.. I guess CELL is really dead

EDIT 2: I'm stupid.. A0093004 means no power to RSX which is normal since I left it with zero caps.. Added tantalum caps back, and now I'm at square one. Tested +1.2V_MC2_VDDIO and there's 1.2V then 0V when YLOD comes up. +1.5V_RSX_VDDIO gives 0.09V and goes back to 0V on YLOD What else can I check? +3.3V_EVER is giving me 5V (weird)

EDIT 3: I *think* I'm getting somewhere.. Fuse F6302 is acting strange, it beeps for a split second and then it's resistance starts going up. Removed it, it's dead but board is still shorted

 

[tracedg0d]

Congrats! Did you use a preheater or just hot air? Hope it survives for years to come

Just a hot air station. I hope it will keep running for a while, I will probably need to de-lid the RSX and CELL to guarantee that for sure though. In Webman the Temps show the CPU at 67C and the RSX is at about 54C and the fans get pretty noisy after about 15 minutes. Only has 144 days of Power-on time, which is surprising because I honestly expected more.

On another note, I have revived another system, this time an A01, with the same approach. 3034 error in syscon, reflowed the RSX, and it booted right up. Oddly enough, there was some older 1001 errors for the CELL Tokins, so I re-flowed all 8 of them to be safe. Seems to also get noisy in a similar amount of time with similar temps as the E01 I repaired first.

 

[patricksouza472]

Just a hot air station. I hope it will keep running for a while, I will probably need to de-lid the RSX and CELL to guarantee that for sure though. In Webman the Temps show the CPU at 67C and the RSX is at about 54C and the fans get pretty noisy after about 15 minutes. Only has 144 days of Power-on time, which is surprising because I honestly expected more.

On another note, I have revived another system, this time an A01, with the same approach. 3034 error in syscon, reflowed the RSX, and it booted right up. Oddly enough, there was some older 1001 errors for the CELL Tokins, so I re-flowed all 8 of them to be safe. Seems to also get noisy in a similar amount of time with similar temps as the E01 I repaired first.

Nice, but if you had 1001 errors your Tokins are probably bad and must be replaced or else it will fail sooner rather than later. Still haven't lost hope on repairing my A01 with less than 100 days, waiting for the parts to arrive

 

[RIP-Felix]

Oddly enough, there was some older 1001 errors for the CELL Tokins, so I re-flowed all 8 of them to be safe. Seems to also get noisy in a similar amount of time with similar temps as the E01 I repaired first.

Could you please detail this further? I'm very interested in what you mean by, "get noisy in a similar amount of time with similar temps." Do you have an oscilloscope? If so, would you please post images of the noise? Also the lasterrorlog if you can get the 1001 error to trigger again. This is something I have been hoping to see for awhile now. @squeept and I are wanting to find the error codes associated with tokins, as confirmed with the oscilloscope images to prove they are bad. We also want to define the Vpp that makes them unstable and/or trigger an error/YLOD.

 

[patricksouza472]

Could you please detail this further? I'm very interested in what you mean by, "get noisy in a similar amount of time with similar temps." Do you have an oscilloscope? If so, would you please post images of the noise? Also the lasterrorlog if you can get the 1001 error to trigger again. This is something I have been hoping to see for awhile now. @squeept and I are wanting to find the error codes associated with tokins, as confirmed with the oscilloscope images to prove they are bad. We also want to define the Vpp that makes them unstable and/or trigger an error/YLOD.

I think that by "noisy" he was referring to the cooling fan.. Let's hope he can help providing more info on those errors

 

[tracedg0d]

Could you please detail this further? I'm very interested in what you mean by, "get noisy in a similar amount of time with similar temps." Do you have an oscilloscope? If so, would you please post images of the noise? Also the lasterrorlog if you can get the 1001 error to trigger again. This is something I have been hoping to see for awhile now. @squeept and I are wanting to find the error codes associated with tokins, as confirmed with the oscilloscope images to prove they are bad. We also want to define the Vpp that makes them unstable and/or trigger an error/YLOD.

Sorry, I think the way I typed that out was probably confusing. By noisy, I meant the cooling fans, haha! Unfortunately, I do not have an oscilloscope, I was really just referencing the OP's list of errors and what they could be, so I'm not even sure there was anything wrong with the TOKIN caps at all. According to the syscon logs, the last time the error was triggered was once in 2016, and every error that logged afterwards was the 3034 error, which I found pretty strange. I'll take another dive into syscon if/when it YLODs again.

 

[Pacorretaco]

Reflowing NEC tokins, now that's interesting.

Error 1001... I wonder if one can trigger that simply by removing the 12v prongs from the PSU. I should be able to check that soon.

3034 was your error and the heat made the error go away. Indeed this has been long documented.
As to why exactly... I'm afraid it will always be ambiguous.
Let's see how long it lasts. That ~150 day failure timeframe sounds familiar

Cheers

 

[RIP-Felix]

Any heat nearing reflow temperatures that close to the BGA is going to either warp the board such that a mechanical reconnection temporarily fixes a BGA defect, or it will cause a BGA defect. A 3034 is associated with BGA defects and requires a reball to fix. I have an A model with a single 3034 and no other identifiable problem. Tokins are fine, fuses are fine, no problems with voltages. Must be the BGA. I'm just working up the courage to attempt another reball. Already failed my last 2 attempts.

 

[Pacorretaco]

I have an A model with a single 3034 and no other identifiable problem. Tokins are fine, fuses are fine, no problems with voltages. Must be the BGA. I'm just working up the courage to attempt another reball. Already failed my last 2 attempts.

Not really ambiguous?
I hate to be that guy, but it "can" be the BGA.
Not "must", but "can".

I agree reball is certainly warranted and probably the best choice. But this is entering murky waters. Nothing is guaranteed. One can only be hopeful it's just a BGA defect, or even have a high level of confidence of it. But not sure sadly.

As you say when heat is involved, many variables change, but not just the solder balls.

 

[vyktormvmpay25]

Any heat nearing reflow temperatures that close to the BGA is going to either warp the board such that a mechanical reconnection temporarily fixes a BGA defect, or it will cause a BGA defect. A 3034 is associated with BGA defects and requires a reball to fix. I have an A model with a single 3034 and no other identifiable problem. Tokins are fine, fuses are fine, no problems with voltages. Must be the BGA. I'm just working up the courage to attempt another reball. Already failed my last 2 attempts.

What rework station do you have? Have you done any slim rsx? It is a long discussion and I can help with videos if you want, just let me know which model you want to record.

 

[RIP-Felix]

I Macgyvered an "e-bay special" from a cheap hot air station with a 45mm BGA nozzle, a 12" IR board heater, and a super cheap hand held temp meter with two K series theromocouples. It works, but I'm paying the price of learning what not to do.

Not really ambiguous?
I hate to be that guy, but it "can" be the BGA.
Not "must", but "can".

I agree reball is certainly warranted and probably the best choice. But this is entering murky waters. Nothing is guaranteed. One can only be hopeful it's just a BGA defect, or even have a high level of confidence of it. But not sure sadly.

As you say when heat is involved, many variables change, but not just the solder balls.

Process of elimination narrowed the possibility's to the BGA, but technically there could be something I missed. Believe me, I consider all the easier fixes first, and reball last.

 

[vyktormvmpay25]

It should work fine if board can reach 150 beside ic on top only with preheater, not sure how long will take, keep in mind this, then set as example if bottom preheater is set up for 280 with distance of 4cm should reach 120 beside top ic in 5 minutes on power of 2000 as example, starting top hot air set 180 (don't know distance from board to hot air nozzle may work with 3 cm). Should take 3 minutes to get 180, then set 225. This could take 2~3 minutes to get liquid stage. Most of time should watch with a camera when it starts to melt at 220 some they melt late totally at 225. May start testing on rsx without delid on some slims as they are not so fragile. From what I seen 60% is preheater work and 40 for top. Anyway found A0805fff is a fried cpu as I forgot to use aluminium foil on bottom cover. Cpu is more sensible than rsx. Another simple resistance measurement is that working board will have 3 ohms on cpu, mine is 7, took it out and same. So if work on cpu cove bottom middle with aluminium foil. This error I've found on most scrap boards where rsx was tested and worked well. After reball don't forget it there as can do short circuit.
A0805fff is common error for 2500 /3000/4000 models. Not sure for 2000. Have few boards in boxes over years, may research for 2000 with time.

 

[Pacorretaco]

I Macgyvered an "e-bay special" from a cheap hot air station with a 45mm BGA nozzle, a 12" IR board heater, and a super cheap hand held temp meter with two K series theromocouples. It works, but I'm paying the price of learning what not to do.

Process of elimination narrowed the possibility's to the BGA, but technically there could be something I missed. Believe me, I consider all the easier fixes first, and reball last.

I'm talking about more problems still with the RSX, beyond the 0.6mm solder balls.

Of course it's better if you can find a problem unrelated to RSX (not coincidence there's 170 pages chasing the tokins)

When reballing you are just hoping that everything else is OK, and only the BGA solder balls are causing the problem. Luckily it seems to actually be the case in many instances. But it's still enough of a gamble for guys like squeept to not want to offer repair services.
Sometimes reballing is just not enough.

This is why it's worth it to figure out the 65 or 40nm RSX swap.

 

[RIP-Felix]

When reballing you are just hoping that everything else is OK, and only the BGA solder balls are causing the problem. Luckily it seems to actually be the case in many instances. But it's still enough of a gamble for guys like squeept to not want to offer repair services.

Spoiler

I don't want to speak for him, but why offer repair services when you can buy and repair your own consoles with less headache and more upside? You get to control the condition of the consoles you buy. This improves the odds of fixing them. And if your @squeept you can sell them for $479 a pop. He puts in $100 overhead and a week of labor, but I imagine the upside is greater than any "charge per fix" business model would offer. First, no one is going to pay $300 for a reball. Second, people would try to have him fix their tantalum terror, heat gun special, or delid disaster. Then, they'll get all butt-hurt when he tells them "they killed it." Maybe he would repair 1/10 consoles doing that, the odds are worse. That business model only works on scale where you pay your employees minimum wage plus cancer. Buy, fix, sell makes more sense.

The 3034 error others have anecdotally noted tend to be BGA defects (It was in the documents with the SYSCON guide). PS3#4 just has this, it was sealed, and I checked everything else I can think of. There are a finite number of things it can be. Of course there are thing we can't fix, but I don't worry about those. You have to fight the battles you can win.

 

[squeept]

@RIP-Felix Yeah, no, that's pretty much it. I stopped most repair services almost 10 years ago. People would send in a Nintendo DS that doesn't power on... okay, I fix it and turn it on and find a screen is ruined. I let them know it will be $10 more, and then suddenly I'm a huge asshole scam artist that's ruining their kid's birthday or some bullshit. With a PS3 that I can't fix, I get accused of stealing their parts and sending them back my garbage. I don't like people much to begin with, so you can imagine how fun any of that is for me. So, yeah, buy then fix then sell only.

I do some local work for a few game stores and I do some microsoldering for local fix-it shops from time to time because we have a working, trusting relationship. I still take on some things like Sega CD model 1 systems because if I don't do it, someone else will ruin it beyond repair, and there aren't a lot of them left. I'll do a little "under the table" unwarrantied work for some people here if they agree to my insane terms. And I'll do some work for repeat customers as a favor. But for the most part, not a chance in hell. I can refer ya to some other techs, or I'll buy the heap of junk off of you outright or as a trade-in towards a finished system.

 

[Pacorretaco]

@RIP-Felix Yeah, no, that's pretty much it. I stopped most repair services almost 10 years ago. People would send in a Nintendo DS that doesn't power on... okay, I fix it and turn it on and find a screen is ruined. I let them know it will be $10 more, and then suddenly I'm a huge asshole scam artist that's ruining their kid's birthday or some bullshit. With a PS3 that I can't fix, I get accused of stealing their parts and sending them back my garbage. I don't like people much to begin with, so you can imagine how fun any of that is for me. So, yeah, buy then fix then sell only.

I do some local work for a few game stores and I do some microsoldering for local fix-it shops from time to time because we have a working, trusting relationship. I still take on some things like Sega CD model 1 systems because if I don't do it, someone else will ruin it beyond repair, and there aren't a lot of them left. I'll do a little "under the table" unwarrantied work for some people here if they agree to my insane terms. And I'll do some work for repeat customers as a favor. But for the most part, not a chance in hell. I can refer ya to some other techs, or I'll buy the heap of junk off of you outright or as a trade-in towards a finished system.

Yes, that makes a lot of sense.

So more or less, from your experience, how confident you are on being able to fix a 90nm system that hasn't been tampered with?

Or to put it differently, how worried are you about the chance of getting a bad RSX with problems beyond BGA defects?

Honestly this is what prevents me from getting into reballing myself. Granted, where I live people don't value these systems that much, either way. People are selling their systems "working" for well under 100€, closer to 50.
So it's mostly curiosity, and wanting to keep them out of the trash.

But if it's something more rare than what I had in my head... Maybe I could give it a chance. Right now my time isn't that valuable anyway. It's just that I was never that convinced.

 

[vyktormvmpay25]

Something perhaps ppl will need
http://s.go.ro/ax49drsu
Going up to slims they are similar, if anyone has more than those schematics please share them or ask me on pm to pay.

 

[squeept]

don't know if this is off topic or not so...

Spoiler

Yes, that makes a lot of sense.

So more or less, from your experience, how confident you are on being able to fix a 90nm system that hasn't been tampered with?

Or to put it differently, how worried are you about the chance of getting a bad RSX with problems beyond BGA defects?

I'm hoping the spreadsheet will do the talking for me in the future once it has enough results to be a little less anecdotal.

But prior to all the new diagnostics and information we have now, I'd say in general on a warranty sealed system (excluding disc drive problems) is very roughly 2/3 RSX, and 1/3 anything else. I'd guess once I narrow it down to a suspected RSX problem, warranty sealed are maybe a 65% win rate, and previously tampered with are like 25% or less.

Of course there's no way without x-ray to tell if I accidentally fixed a bump or temporarily revived a dead chip, but that's why I stress test them so long and include a generous warranty.