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

[vyktormvmpay25]

If board is in standby you will get to test at least 3v3 for Tx pin.
Usually a short hdmi ic isn't detected in 5v line . Even dead AV ic isn't detected.
If you manage to connect as "comX SW"
It should ask for "auth"?
Or at least you should get Auth1 response invalid (swap rx Tx)
Can't remember if 3v3 voltage regulator is same for syscon and WiFi board or have one individual.

 

[Pacorretaco]

Here is something i never seen before. This is my 2000a slim which i delided. On cold boot, GLOD for about half a minute, then beep and shutdown. Subsequent attempts, 2-3 second ylod. After 10+ tries, console works no problem. I turn off, turn on everything fine, no crashes, no shutdowns. Leave for the night. Cycle repeats. So today, instead of booting up cold (from overnight being off), i decided to hook it up to the syscon GUI and use the execute command bringup. instead of glod, it ylod on first try. subsequent attempt, it booted, console beeped...then beeped again...booted. Dumbed the codes and this is what i got....
===================================
ERR 00: FFFFFFFF Answer length
ERR 01: FFFFFFFF Checksum
ERR 02: 00000000 A0801201 0B4D061A
ERR 03: 00000000 A0801201 0B4D0588
ERR 04: 00000000 FFFFFFFF FFFFFFFF
ERR 05: 00000000 FFFFFFFF FFFFFFFF
ERR 06: 00000000 FFFFFFFF FFFFFFFF
ERR 07: 00000000 FFFFFFFF FFFFFFFF
ERR 08: 00000000 FFFFFFFF FFFFFFFF
ERR 09: 00000000 FFFFFFFF FFFFFFFF
ERR 10: 00000000 FFFFFFFF FFFFFFFF
ERR 11: 00000000 FFFFFFFF FFFFFFFF
ERR 12: 00000000 FFFFFFFF FFFFFFFF
ERR 13: 00000000 FFFFFFFF FFFFFFFF
ERR 14: 00000000 FFFFFFFF FFFFFFFF
ERR 15: 00000000 FFFFFFFF FFFFFFFF
ERR 16: 00000000 FFFFFFFF FFFFFFFF
ERR 17: 00000000 FFFFFFFF FFFFFFFF
ERR 18: 00000000 FFFFFFFF FFFFFFFF
ERR 19: 00000000 FFFFFFFF FFFFFFFF

thoughts? I think this whole cold boot might have something to do with the checksum since prior to delid etc i dumped the codes and hooked it up to uart. Not sure, if it messed the syscon up and is now stuck in this weird glod first, ylod, and finally boot after 20 attempts. Also its not the power supply, cuz i have a 2000b with a wokring powersupply which i used on the "A" model and same thing. I used th "A" model suspect power supply and on the "B" model and worked on first try. I thought NECS..but my "A" model only has necs on the RSX. Thoughts?

Have you tried checking and fixing the checksum?

Auth in manually. Use eepcsum command to see if its wrong. Then write the "should be" value in. You can use the tutorials to figure it out or we can walk you through it.

eepcsum doesn't really work like that in Sherwood. It's a bit different from Mullion.
You can't exactly fix it just like that. eepcsum will not tell you what address is broken or not. Only the calculated value that "should be" yes. Which is only half of the puzzle.

The address needs to be found by other means and then compare its value to the output of eepcsum. Address appears to be

7fe

I went over this here for example:
https://www.psx-place.com/threads/modifying-syscon-fan-settings-better-than-webman.34626

Also if he can't use SW mode with python, maybe he is using wrong old script?
He could still use the execute.exe AND debug.exe to do the same thing... If somehow he likes the utility more.
You can type whatever command you want with execute.exe while you have debug.exe open to see the outputs.

But...
In his case, he doesn't have to do anything. If his checksum were a problem, the machine would beep red immediately and not turn on at all. No response from power button.

His machine was working perfectly fine for years. Then he opened to delid or whatever, and now strange problems...

Try to disassemble again carefully and put together exactly like it was before? Maybe something is shorting due to improper assembly or something stupid like that.
Otherwise the problem could be delid damage, perhaps precarious contact under CPU, or NOR flash who knows, from flexing the board anything can happen.

But no need to overcomplicate things. It's funny to have this kind of talk here. Looks like it's unrelated but this is exactly the spirit of this whole thread hehehe.

Good luck

 

[vyktormvmpay25]

Yes a bad flash won't affect syscon, it won't start debugging mod but still syscon should be accessed fine unless 5v line or 3v3 line faulty. I can access boards on syscon even without cpu /rsx or both out of board. It is a hardware issue in power line. Unless he writed something to syscon and will cause that. I always use old script.py drag to cmd.

 

[Pacorretaco]

Yes a bad flash won't affect syscon, it won't start debugging mod but still syscon should be accessed fine unless 5v line or 3v3 line faulty. I can access boards on syscon even without cpu /rsx or both out of board. It is a hardware issue in power line. Unless he writed something to syscon and will cause that. I always use old script.py drag to cmd.

Hm, but he has no problems accessing the SYSCON. He can get errlog without problems.
But the problem is he's having random CPU/ check stop errors.

Or maybe I read something wrong?

Actually I have an untouched COK 001 board suspicious of SYSCON problems. No red light. 5v Ok. 3.3v OK but 1.8v is 3+v! Resistance there is about 2kohms instead of rising to about 50k

I guess I should go to Syscon thread for that. I'm not blaming the poor tokins here too.

 

[truemaster]

for the no standby led try this.

i had to do that for a friend he has cok-002 board
i had to take this component from a rlod dead board and replace it look for the marked 5 pin compomend near syscon

 

[vyktormvmpay25]

for the no standby led try this.

i had to do that for a friend he has cok-002 board
i had to take this component from a rlod dead board and replace it

Not that video is for 5v and return of 3v3 to power supply. Here is another problem.

 

[Pacorretaco]

for the no standby led try this.

i had to do that for a friend he has cok-002 board
i had to take this component from a rlod dead board and replace it look for the marked 5 pin compomend near syscon

Aha, so you confirmed that replacing ic6009 (LCARKI) solved the problem completely?

I was very suspicious it was the case, but I was lazy to test it.
Soon I will update the respective thread then.
https://www.psx-place.com/threads/not-dead-c-ps3-nlod.31921/
The quick workaround was good enough but eventually I'll do the proper thing too.

Thanks anyway although now I was talking about a different more severe issue. I posted more in Syscon thread.

 

[truemaster]

yes that was the fix. although it im not sure if its always the cause for consoles that wont go even on stand by. but it wont hurt to try bridging psu pin 1 and 5 will tell if this is the issue

 

[squeept]

Just got back around to that CECHE01. I'm sure this won't come as much of a surprise, but I swapped out the caps and it's still dead. Same errors, except 4401 is now 4411. Verified the RSX and CELL signals are now perfect on the scope. In to the drying oven for another reball.

These ones always confuse me. How did it get 2 separate issues that would prevent it from working at all? The CPU caps were totally dead. The RSX issue is not intermittent. Doesn't the first of those problems stop you from putting more hours on it to develop the second problem? Anyway, back to work.

 

[camperotactico]

This past week I found a local repair shop that offers reballing work for as low as 20 euros, which is awesome.

Also, I bought another CECHE04. This console has already been opened before as there was a lot of dried flux everywhere and the GPU lid was lifted. I replaced all the caps but that didn't get to revive the console. While I wait for the adapter to read the SYSCON and plan my next move, I wanted to share the current state of the boot process with you:

The green LED stays for a couple of seconds, it is not an instant YLOD and the fan gets to spin at full speed for a brief period. The first time I booted it up after the cap replacement, the console played a total 6 beeps, in 2 bursts of three. Why did this console beep 6 times and not just 3?

 

[squeept]

repair shop that offers reballing work for as low as 20 euros

There is no way a qualified technician with the right equipment can stay in business with those prices, it's definitely a scammer. They're going to heatgun your shit to oblivion, or they're just doing some stupid washer/shim fix.

Their total is what I would consider my "shop fee" for a reball - the cost of consumables, and the general wear and tear on my tools and machinery. It's the break even point without even considering the book rate on labor. That price is not a great deal, it's a giant red flag.

 

[RIP-Felix]

I second this ^

Just got back around to that CECHE01. I'm sure this won't come as much of a surprise, but I swapped out the caps and it's still dead. Same errors, except 4401 is now 4411. Verified the RSX and CELL signals are now perfect on the scope. In to the drying oven for another reball.

These ones always confuse me. How did it get 2 separate issues that would prevent it from working at all? The CPU caps were totally dead. The RSX issue is not intermittent. Doesn't the first of those problems stop you from putting more hours on it to develop the second problem? Anyway, back to work.

My guess is the BGA was precarious before you started messing with the Tokins. Depending on how you removed/installed...what, NOS?...the strain was too much and it finally gave way. What's the becount? If it's on time is in the 100-150+ day range, it makes sense.

Even with PS3#7 that you reballed with leaded solder I popped a BGA with all those tantalum tests. And I was just using an Iron! I had to reflow and it's been reliable ever since, but BGA is a POS technology IMO!

 

[vyktormvmpay25]

This past week I found a local repair shop that offers reballing work for as low as 20 euros, which is awesome.

Also, I bought another CECHE04. This console has already been opened before as there was a lot of dried flux everywhere and the GPU lid was lifted. I replaced all the caps but that didn't get to revive the console. While I wait for the adapter to read the SYSCON and plan my next move, I wanted to share the current state of the boot process with you:

The green LED stays for a couple of seconds, it is not an instant YLOD and the fan gets to spin at full speed for a brief period. The first time I booted it up after the cap replacement, the console played a total 6 beeps, in 2 bursts of three. Why did this console beep 6 times and not just 3?

Can you check with your local repair shop, if they can provide just a single photo with serial of board that have already ic's desoldered . This is what I do when friends recommend me for this service from many years. Just photos as reference I've attached.
As @squeept said is impossible even with workshop not so crowded as I have because I avoid to take to much income work.

Spoiler

Kind of a real proof.

 

[Kleon1876]

eepcsum doesn't really work like that in Sherwood. It's a bit different from Mullion.
You can't exactly fix it just like that. eepcsum will not tell you what address is broken or not. Only the calculated value that "should be" yes. Which is only half of the puzzle.

The address needs to be found by other means and then compare its value to the output of eepcsum. Address appears to be

7fe

I went over this here for example:
https://www.psx-place.com/threads/modifying-syscon-fan-settings-better-than-webman.34626

Also if he can't use SW mode with python, maybe he is using wrong old script?
He could still use the execute.exe AND debug.exe to do the same thing... If somehow he likes the utility more.
You can type whatever command you want with execute.exe while you have debug.exe open to see the outputs.

But...
In his case, he doesn't have to do anything. If his checksum were a problem, the machine would beep red immediately and not turn on at all. No response from power button.

His machine was working perfectly fine for years. Then he opened to delid or whatever, and now strange problems...

Try to disassemble again carefully and put together exactly like it was before? Maybe something is shorting due to improper assembly or something stupid like that.
Otherwise the problem could be delid damage, perhaps precarious contact under CPU, or NOR flash who knows, from flexing the board anything can happen.

But no need to overcomplicate things. It's funny to have this kind of talk here. Looks like it's unrelated but this is exactly the spirit of this whole thread hehehe.

Good luck

I'm not sure about python and old script. Could be the issue.I always get glod for 30 seconds follow by a bunch of ylods followed by working no problem. Every time on cold boot. So confusing.
I'll post it on YouTube

 

[RIP-Felix]

I'm not sure about python and old script. Could be the issue.I always get glod for 30 seconds follow by a bunch of ylods followed by working no problem. Every time on cold boot. So confusing.
I'll post it on YouTube

I don't trust this cold boot thing. I always suspect a soldering issue if the temperature can affect the YLOD in any way.

 

[Kleon1876]

Here is the video. Haven't turned it on since last night (10 hours ago). First try ylod. Thoughts?

 

[truemaster]

that looks like cold solder joint. the heat generated from those tryes simply make the cold joint to reconect. now if its cpu or gpu is a trivia

 

[Kleon1876]

that looks like cold solder joint. the heat generated from those tryes simply make the cold joint to reconect. now if its cpu or gpu is a trivia

What does cold joint mean exactly? Like the solder balls under the Chip? If that's the case how come it happens every time I leave it off. If it reconnects after several tries...shouldn't it stay connected?

 

[RIP-Felix]

Here is the video. Haven't turned it on since last night (10 hours ago). First try ylod. Thoughts?

Yes, that's what I call a warm start YLOD (Only starts when warm). The old trick was to use a hair drier to warm the console first. Initially people thought that meant the tokins were going bad, but that's a myth. Heat doesn't affect this kind of capacitor. Instead it's mor likely that the heat generated over multiple startup attempts finally warps the board enough to mechanically reconnect a BGA crack. They are microscopic, micrometers apart. It doesn't take much.

To confirm if it is a BGA you could try the pressure test. Press on the cpu while booting cold, to see if it starts. If so you need a reball/reflow, which isn't a 100% fix as @Pacorretaco always corrects me on. If not, it doesn't rule out a BGA/bump defect. This is a + test result = reball/reflow, - result means nothing kind of test.

What gets me is the error log. This doesn't have the usual 3034/4xxx error I would expect to be associated with a BGA/Bump defect. This console might be a candidate for @botakompong's method of troubleshooting.

Above is the long way of saying, "The hell if I know!"

 

[truemaster]

mm dont know about tokins but electrolitic capacitors will perk up if lightly heatet up, can you try another psu? is psu isnt give the correct power ylod can occur.