PS3 Frankenstein PHAT PS3: CECHA with 40nm RSX

[botakompong]

On the other hand, after replacing it, it reads 33K (in circuit) now. So I'm wondering if it's actually a 47K resistor after all I didn't think to read it while I had it off the board). I know reading resistance in circuit is frowned upon, but in truth I forgot. There wouldn't be any question had I used R6016 like Paco suggested.

Hallo friend..
the 47k resistor that you have taken is correct, the 33k size value that you measure with the installed position is due to the influence of resistance in other components

Okay, I just replaced R2002 with that resistor I measured (in circuit) was 47K and I removed R2001. Nothing changed at all. No YLOD or any change in behavior. It still plays PS1/3 games, but fails to load PS2. Dang, I was hoping that was it.

R2001 and R2002 removed, put 47k in place of R2002, this was done specifically for cok-001. to avoid GLOD problems on console

 

[botakompong]

That greainy picture is hard to tell if R2001 is actually unpopulated or not

right, it's empty, there's no component in R2001.
you can also put 47k into R2001 but leave R2002 empty

And @botakompong's instructions have a confusing translation that let my primitive ape brain be lazy

I'm sorry, I'll explain in more detail in the future....lol

 

[DeadEnd]

I will try to order some 20 or 60gb versions from Japan to do more tests on Cok-001 later this year. Haha, sorry Felix, I also had the same issue with finding a 47k resistor until I asked @botakompong about its locations on slim boards . I used to use a spare GTX7950 graphics card with some leaked schematics to source different resistors. I forgot that not everybody is in possessions of spare boards and parts. I've been too busy with work, so apologies for not providing more info.

Based on this new info, I will edit my video as well. Although, I bet not many have attempted this swap yet.

I still believe we need somebody to do another swap on a cok-001 before coming to any definitive conclusions. Me and Vyktor have been trying official method with a syscon swap, and unfortunately both got the Flex IO_ID error. So at the moment modchip is our only best option.

 

[botakompong]

I haven't been able to open the shop (it's been more than 1 month), while all the equipment and motherboards are in the shop, later when I open I will make sure the cok-001 has a problem or not for the PS2 games.
I also don't want to give input that I haven't tried, because it will only add unclear work.
if someone asks why don't you know cok-001 is having problems with ps2 games? because we are not focused on ps2 games, in our country we don't play/rarely play ps2 games using the console ps3, so for the past few years there have been no questions related to ps2 problems, for that I apologize because I just found out about this problem, for the solution I will definitely look for it soon, I have to refresh my memory from the last few years about cok-001/002, I hope you can be patient..., tq

 

[RIP-Felix]

Yeah, no worries guys! I just wore myself out mentally today, so if I came off crass, I apologize for it. I just wanted you keep you apprised of the situation while I struggle to figure out what's wrong with this console. I'm probably just being impatient, but there were some nuggets of discovery gleaned!

I certainly can't rule out the PS2 HW. I have yet to attempt troubleshooting it. I hate troubleshooting! And I still need to modify a PSU so I can power the board with it off to the side. I have an Extra ZSS, so that's a good candidate! I also need to find a HS solution so I can test voltages with the console powered, without it overheating! @vyktormvmpay25 are those slim HS you are using?

I must admit, I have been avoiding troubleshooting like the...I was going to say plague, but COVID seems more apt. Again it's my ape brain being lazy!

R2001 and R2002 removed, put 47k in place of R2002, this was done specifically for cok-001. to avoid GLOD problems on console

Could you reword that last bit. I don't understand. Are you saying that a COK-001 woud GLOD if you didn't do that? Because I didn't modify them until today and is was running just fine with the stock 49.9Ohm resistors. In fact I notice nothing different by removing them and replacing R2002 w/47K???

I could try moving it over to R2001. In the schematic, it is part of a low pass RC filter. Could that have something to do with it?

I'm also curious to hear what you think about the voltage divider? Is it useful to populate both pads but with different values to drive the voltage High or low? I'd like to know why we're leaving it empty like the mod calls for.

 

[RIP-Felix]

Me and Vyktor have been trying official method with a syscon swap, and unfortunately both got the Flex IO_ID error. So at the moment modchip is our only best option.

I forget, was that different than what @squeept was getting?

Can you try the 47k R2002 mod to see if it changes your error? Now that we know it's part of the official method too, we may as well have one less monkey throwing wrenches (is that an american saying - Monkey Wrench? We sure have alot of idioms. I apologize if they confuses anyone).

 

[vyktormvmpay25]

Felix those HS are removed from 2500 rsx. But is a minute test. Won't not stand more than that, fan is key off cooling, I don't test only image comes on screen, not more than that.

 

[botakompong]

Could you reword that last bit. I don't understand. Are you saying that a COK-001 woud GLOD if you didn't do that?

Right, there was a GLOD for some of the cok-001 that we installed, at that time we used cxd2991 and cxd530x, I forget the exact cause of the GLOD happening in cxd2991 or cxd530x or both can happen, which is clear with a solution changing 47k can eliminate GLOD, so we apply until now.
the memory is blurry now because we did it maybe 5 years ago...

nb: strangely for the type cok-002 we never had a GLOD problem (installation of cxd2991/cxd530x), if you pay attention to R2001 and R2002 there are no resistors, as if they were left open, I will find out later, there is a difference in resistors (R2001&R2002) on the cok-001 and cok-002

 

[DeadEnd]

Could you reword that last bit. I don't understand. Are you saying that a COK-001 woud GLOD if you didn't do that? Because I didn't modify them until today and is was running just fine with the stock 49.9Ohm resistors. In fact I notice nothing different by removing them and replacing R2002 w/47K???.

I think it applies more for COK-002 , if I remember correctly, Vyktor once tried the modchip without a 47k resistor and it didn't work.

 

[vyktormvmpay25]

Yes got 3034 4002 on that time for purpose test.

 

[DeadEnd]

I forget, was that different than what @squeept was getting?

Can you try the 47k R2002 mod to see if it changes your error? Now that we know it's part of the official method too, we may as well have one less monkey throwing wrenches (is that an american saying - Monkey Wrench? We sure have alot of idioms. I apologize if they confuses anyone).

Squeept may have flashed the wrong EEPROM. My board seems to be busted as I got other errors related to temperature sensors 2030, 2031, 2033.

As you remember, we took the syscon 302GB from a dia-002 and flashed it with the patched EPPROM, then soldered it back to COK-002. We've tried different patches, but no luck. Vyktor did the same with a 301GB from DIA-001, but got the same errors as me. However, the board will start with a newly installed syscon and the modchip. The resistors were changed correctly in all tests.

The preliminary conclusion is that we cannot use the 301/302GB syscons for a 65nm RSX. The only way is to follow Sony's steps seen on the refurbished models and use either 304GB or CXR713F120A .

We were working on this behind the scenes so to speak (there is no point in flooding this thread every time a test fails).

And do continue using the idioms, I'm taking notes to sound more fluent on the next video

 

[Pacorretaco]

As you remember, we took the syscon 302GB from a dia-002 and flashed it with the patched EPPROM, then soldered it back to COK-002. We've tried different patches, but no luck. Vyktor did the same with a 301GB from DIA-001, but got the same errors as me. However, the board will start with a newly installed syscon and the modchip. The resistors were changed correctly in all tests.

The preliminary conclusion is that we cannot use the 301/302GB syscons for a 65nm RSX. The only way is to follow Sony's steps seen on the refurbished models and use either 304GB or CXR713F120A .

We were working on this behind the scenes so to speak (there is no point in flooding this thread every time a test fails).

Hmm, so then what I was saying was actually correct huh?

I think this is the same brick wall that squeept was hitting.

If my hypothesis is, let's say there is something wrong with this method and 4002 means some kind of "unsupported RSX" error.
What if you try to solder a modchip there right now, see what happens?
If it works it would confirm this and save many headaches.

Kind of bad news, but hey this is also progress.
If modchip indeed solves the 4002 RSX FLEXIO_ID error, then that's it.
I guess no more need to wonder about resistors, soldering, or other random hardware mysteries then. Now we know.

I guess there's nothing wrong with "flooding" this thread though. Fail or success, it's still progress.

 

[Pacorretaco]

As it turns out, my cecha01 is a refurb, I think. It died of ylod one day fairly early in its lifetime... I want to say it was around 2009. The circumstances were odd too -- I was playing one late afternoon, then turned it off. I turned it back on about 2 mins later to check on something, then turned it off. I then accidentally bumped the home button and bam! Ylod. I contacted Sony and sent it back for repairs, and about a month later I had it back. It worked fine until it died again late 2019 (in much the same way - power off, then power one a few mins later to be greeted by a ylod). It also played ps2 games just fine.

I guess the long story short is -- why did Sony started doing the mod? My bet is that they were low on supply, so it would have to have been some time into the slims lifetime. @DoublesAdvocate, did you acquire your moded ps3 from a 3rd party? Or was it your own refurb?

By the way are you saying that you also have a COK 001 board with a retrofit RSX?

And this board now has YLOD?

That's very interesting, but of course what we want is the details hehe.

Pictures of the board/ rsx area / syscon area...
And what kind of YLOD is it? / What does the syscon say? Typical 3034/4xxx?

Cheers

 

[RIP-Felix]

The resistors were changed correctly in all tests.

Not unless you knew that the R2001/R2002 mod applied to both COK-001 and 2! Am I an idiot, and that was clear all this time? Because to my knowledge neither of you touched those resistors, leaving the voltage divider alone on COK-001 boards because we thought that mod only applied to COK-002 where those pads are unpopulated.

We were working on this behind the scenes so to speak (there is no point in flooding this thread every time a test fails).

I'd prefer everything be public. You never know what little nugget of information sparks the inspiration that finally solves this, or stays a mystery because it was a closed conversation.

 

[RIP-Felix]

...the 47k resistor that you have taken is correct, the 33k size value that you measure with the installed position is due to the influence of resistance in other components

Confirmed!

The 47K R6016 @Pacorretaco found measured 23K in situ and 46.85K after removal. I retested the 47K I took from the KTE-001 board and installed on R2002. In situ it read 23K (not 33K like I said previously). After removal it read 46.9K ish. So yes, the circuit influences the reading quite a bit. This is good to know, because I came across other resistors that read near 23K while probing. It's possible they were actually 47K. In fact that might be why I missed R6016. I gave up looking at the schematic and went probing, forgetting that doesn't work - LIKE A NOOB!

You can use resistance to compare a working part to a non-working one, the resistance won't match. That's a valid diagnostic tool, but you can't accurately measure the resistance value unless you remove it first. Not with a DVM. Lesson learned...again! Me ape [scratches noggin..then pokes live wire!]

Moving on...

PS3#8 - Mini Update:

right, it's empty, there's no component in R2001.
you can also put 47k into R2001 but leave R2002 empty...

This morning I moved the 47K to R2001 leaving 2002's pad unpopulated. The opposite of what I tried yesterday. The result is the same. No change in behavior. PS2 is still not loading, the console is otherwise unaffected. If my issue is still related to this voltage divider (GS_CLKI), we're now back to my original 3 scenarios.

But I'm starting to think my problem is unrelated. The SONY refirb had a "Black" resistor on R2002. It's value is unknown, but if the in circuit reading is 23K, then we know it's 47K. I already tried that and that didn't solve my issue. I'll have to move on to...Oh God I'm going to say it... troubleshooting. You know how much I'm looking forward to that!

@DoublesAdvocate (or someone with a SONY refurbished 40nm COK-001) would you please measure the in circuit resistance of R2002 & R2054, as well as confirm the removal of R2001 and R2153? And a nice close-up picture would be much appreciated!

 

[RIP-Felix]

Populating R2002 with a 47K resistor then would have what effect?Does it pull voltage High? In effect saying, "yeah I'm on." I'm struggling with that part. Also, COK-001 boards are more complicated and my brain hurts from staring at the schematics. If "feel" like I'm onto something, but have worn myself out trying to put together the above. So I'll pick it back up later.

Here's where we are. I still need to figure out what effect populating R2002 has in COK-002. What's the point? We know it's what the mod requires, but why? And I still need to figure out what effect doing the same would have on a COK-001.

Maybe some of this makes sense to someone out there, but for now I'm fried!

I just circled back to this today with a sleep refreshed mind...

The math(s) is(are) below, but basically it's the same as the scenarios I mentioned before. Any value of resistor on R2002 will have the same effect. It doesn't even need to be a 47KΩ, you can leave the 49.9Ω resistor in place and just remove the other one. The resistance of air is so high a number that it drives the voltage to zero. And if you remove R2002 instead, it's 1.5v.

The only thing that 47K changes is the attenuation of that low pass RC filter. But like I said before, I don't think it matters. It certainly doesn't in the case of the mod, since it drives voltage to 0v. If R2001 is missing then the crossover frequency is 0Hz! No voltage means no signals to filter. It's that simple.

Spoiler: Here's the math(s):

Spoiler: Voltage Divider Diagram

• Vin = 1.5v_RSX_RC_VDDA.
• R1 = R2001 = (SONY Refirb = 20 TΩ, the resistance of 1mm of air)
  
• R2 = R2002 (Stock = 47KΩ )
  
• Vout = RSX Pin AM39 (GS_CLKI)

R2002 is replaced with 47KΩ, R2001 is removed:
Vout = 47,000Ω / (20,000,000,000,000Ω + 47,000Ω) X 1.5v
Vout = 0.000000003v! ≈ 0V

R2001 is replaced with 47KΩ, R2002 is removed:
Vout = 20,000,000,000,000Ω/ (47,000Ω +20,000,000,000,000Ω) X 1.5v
Vout = 1.499999996v ≈ 1.5V

If you use the stock 49.9Ω resistor:

R2001 is removed, R2002 is stock:
Vout = 49.9Ω / (20,000,000,000,000Ω + 49.9Ω) X 1.5v
Vout = 0.00000000000374v! ≈ 0V

R2002 is Removed, R2001 is stock:
Vout = 20,000,000,000,000Ω/ (47,000Ω +20,000,000,000,000Ω) X 1.5v
Vout = 1.5V (even my TI84+ rounds it to 1.5! That's how close it is.)

The Math(s) suggest it doesn't matter which way you go. If you have a COK-001, it would certainly be easier to just remove R2001 and not have to worry about locating and micro-soldering on a 47K. Except that's not what SONY did!

Spoiler: SONY refirb COK-001 w/40nm RSX

So the question is, why did SONY even bother to replace R2002 with a black resistor? What else is affected by this voltage divider? Because I can't see anything else that would matter!

Spoiler: COK-001 Schematic (GS_CLKI)

 

[marciolsf]

By the way are you saying that you also have a COK 001 board with a retrofit RSX?

And this board now has YLOD?

That's very interesting, but of course what we want is the details hehe.

Pictures of the board/ rsx area / syscon area...
And what kind of YLOD is it? / What does the syscon say? Typical 3034/4xxx?

Cheers

thank you! yes, that is what I'm wondering. i don't know what the original error was (I yanked the tokins before we introduced the syscon scripts in the thread), but other than when I broke something, I've only ever gotten a 3034, and even when I fixed what I broke, I still got a 3034. Im pretty sure it would have been a 3034 as well, at first.

As far as the Sony repair, what exactly am I looking for? I can dig up the ol' cecha01 and see what I got in there, I just need to know the cues.

 

[botakompong]

right, it's empty, there's no component in R2001.
you can also put 47k into R2001 but leave R2002 empty

Sorry I corrected what I said, my old brain remembers R2001 & R2002 are resistors connected in parallel, and that's WRONG, the 47k resistor is only installed in R2002..., Tq RIP-Felik with his math formula reminded me.

 

[DeadEnd]

Felix,I think you lost me with the resistors. Cok-002 never had R2001 or R2002 resistors populated, so we did use a 47k in the R2002 spot all the time. This is what it said in the instruction. The only confusion we bumped into was with the Cok-001.

I am still focused on trying to revive the board with a 302GB syscon. You'll never guess, but turned out the thermal errors 2030, 2031 and 2033 I was getting were caused by a dodgy button board of all things. After reconnecting it once more, the ps3 stopped beeping and its light almost faded. The moment I replaced it , it returned back to 3034 just like in Vyktor's test.

I could use a little advice though, can anybody confirm how important are the capacitors on top of RSX? Can one missing cap cause a 3034 error? Because I had some doubts of whether I should take that tiny missing capacitor seriously or not.

 

[Perpetual_Noobness]

I'm just hung up on the fact that you got this model from THE Japan. Never entered my mind that Japan needed a "the" header. Lol.