PS3 FAT NEC/TOKIN Resistance Testing, need help interpreting values

[jkucukov]

I'm in process of diagnosing the yellow light of death symptom on my PS3 FAT console CECHG01, in process of getting the Arduino gadget to be able to access Syscon and be able to read fault codes. In the meantime, I'm trying to measure resistance of the NEC/Tokins while still attached to the board. I took the Ohm readings from each of the NEC/Tokins from front and back of the board (4 from the RSX and 4 from the CPU chips).

What I gathered in the attached pictures are values from the CPU NEC/Tokins are consistently measuring 1.6 Ohms while on the RSX they measure 3.4 Ohms except for two places where its open circuit (OL) no measurements. I understand that these are like capacitors and are hard to measure on the board. I hope somebody more experienced can tell me if what I'm seeing is a set of bad NEC/Tokins on the front side of the board or a bad RSX chip?

[stj]

[jkucukov]

Thank you for sharing, the values I was able to pull from looking at the shared diagram is as follows:
RSX_VDDIO = 85.2 Ohms
RSX_PLL = 5.78 kOhms
RSX_VDDC = 4.1 Ohms

Keeping my fingers crossed 😊

[stj]

values will be different for a "G" board because they use 65nm chips instead of 90nm

[jkucukov]

values will be different for a "G" board because they use 65nm chips instead of 90nm

I thought the board had a 90nm RSX chip per the psdevwiki site that showed all the SEM-001 boards had them, and tanks to you I realized from the back side that there are 5 rows & 6 columns of capacitors (30) behind the RSX chip which makes sense it to be a 65nm unlike the 90nm would have been 5 rows and 7 columns (35).

I'm completely guessing if that's a tell unless one was to delid the heat plate to confirm the die size:

CECHG01 CECHA01

[jkucukov]

I was completely wrong on in my previous post, the capacitor count on the back side of the mother board was of the CPU not RSX

[jkucukov]

I was able to pull internal Syscon error log both external and internal per the Psx Place tutorial from Rip-Felix.
It was very interesting process of wiring it all up and getting it all to read and work, took about 30 min of fiddling to get it to generate the output that I copied to the attached Notepad txt file.

External codes and internal codes both show 3034 indicating bad RSX chip.

The console has 100 days of operation and I also pulled data for the bringup and shutdown several times.

[stj]

it could be bad tokin's on the rsx

[jkucukov]

it could be bad tokin's on the rsx

I thought bad NEC/tokin with RSX had to have 1002 in the list, in my case I only see two codes 3034 and 4422.

[stj]

if the power isnt smooth the rsx cant run properly.
this is where a scope is handy

[jkucukov]

values will be different for a "G" board because they use 65nm chips instead of 90nm

I removed the IHS from the top of RSX, it measures a 90nm chip. So the resistance values being off on the chip potentially indicate a defective rsx

[stj]

possibly, normally i just use those charts to look for shorts.
i think you can get different readings based on the voltage your meter applies in resistance-mode too.

[jkucukov]

possibly, normally i just use those charts to look for shorts.
i think you can get different readings based on the voltage your meter applies in resistance-mode too.

Extremely funny thing is after deliding of the IHS from RSX chip I measured the values again several times and everything was the same except for one value that changed. RSX_PLL went from 5.78 kOhms to 1.47 kOhms. I couldn't believe what I saw so I tried seeing if the console would power on, guess what.. we have picture, I cant believe it.

[stj]

nice
now install EvilNat 4.92
THEN upgrade it to 4.92.2

dont try to clean-install 4.92.2 or you risk a brick!

[jkucukov]

nice
now install EvilNat 4.92
THEN upgrade it to 4.92.2

dont try to clean-install 4.92.2 or you risk a brick!

That's the 1st thing I did after I was sure the console was not going to YLOD again, I wanted to see what the CPU & RSX temps were out of the gate so I got the Webman too. The CPU I didn't want to delid just yet but the temps do go above 76C on the CPU from just browsing so I have to keep the fan speed up. The RSX temps are nice cool in the mid 50s, however I dont want to jinks anything knowing that all this is due to delid confirming cracked solder joints under the RSX chip.

Funny, reading online people are experiencing YLOD after performing Delid, in my case it managed to fix the issue god only knows for how long, not holding my breath lol!

[stj]

depends how they remove the plate i suppose.
if they get the thing too hot and pry on it they could lift the ram off it's balls!
i like to saw through the epoxy with a razor after applying a small amount of heat.
even then it's a bit of a risk but i'm not pulling on the ram.