YLOD, Theories, and an X-Ray Inspection

[alonehacker]

So, I've spent the last couple of weeks trying to investigate why the YLOD happens to the PS3. I've been to many forums, read multiple threads, and really wanted to get to where the actual root cause was.

Now I've read many theories regarding solder opens, solder cracks, voiding, and solder bridging on the RSX. I wanted to see which one of these theories were true, and I wanted to see if the YLOD issue was something the folks over at SONY could fix during manufacturing.

So, I decided to buy a YLOD PS3 (80 gig, CECHK01) to test some of these theories, and because I wanted to X-Ray the motherboard. FYI, in the manufacturing process, the X-ray is used to inspect BGA components, like the RSX to find defects. If there is net access to the chip, boundary scan could also be run on the part through a JTAG TAP to ensure further reliability.

With the heatsink on top of the RSX, I wasn't able to get a good look look at the RSX through the X-Ray. The heatsink would shadow the RSX's pins and would make it look dark through the X-Ray. So, I decided to take off the heatsink on top of the RSX. The results are attached below.

You could see significant amount of pins had solder voids within each ball. While solder voids aren't necessarily defects, the voiding that was witnessed within each solder ball was less than 25%, which most manufacturers pass as acceptable.

There was some evidence of insufficient reflow in a few pins, and you could identify this with the "jagged" perimeter of the solder ball. Again, this isn't necessarily bad, as the pin could still be soldered and deemed acceptable.

Also, in the other picture, you could see some real small solder cracks, which could be characterized as a defect.

Unfortunately, even after the X-Ray inspection, what I found was quite inconclusive. BGA components are difficult to inspect and I don't think SONY could've caught many defects even with a trained X-Ray operator. YLOD's are most likely caused by insufficient reflow on the pins, solder cracks, or some other reason. I see no evidence of bridging or shorting between any of the pins.

Let me know what your thoughts are and if you all would like more pictures!

[patmorita]

Nice job man, and very interesting pics. idk maybe is something you can not watch by x ray. By the way how did you take the pics?? are you working in hospital or something?

Watching carefully you can see how in some spots the dots are like melted with other dots. maybe try take some shots of a perfectly working console(a new one). Remember too that this issue is common in pc micros with the same soldering technique so i doubt the cause of ylod is elsewhere.-

[damox]

Let me know what your thoughts are and if you all would like more pictures!

Awesome pictures.

I always like looking at motherboards x-ray scans.

Would love to see all your pictures. Do you have any complete boards?

[alonehacker]

Nice job man, and very interesting pics. idk maybe is something you can not watch by x ray. By the way how did you take the pics?? are you working in hospital or something?-

I work in a manufacturing facility where we have an X-Ray inspection machine for circuit boards. Given the volumes, I'm assuming not all boards that are manufactured by SONY are inspected with the X-Ray, they're very likely only sampled. In other words, there can be a lot of opportunities for escapes and defects to reach the end user.

Would love to see all your pictures. Do you have any complete boards?

By complete boards, do you mean an X-Ray picture of the complete board? Or do you mean a working board? I do have another 40 gig and 60 gig PS3 lying around, although I'm not sure if I want to take it apart to X-Ray. Also, I'm not sure if inspecting that through the X-Ray will show me any major differences either.

[Mark Webber]

The old lead based solder was soft, the new lead free is ridgid. It's not so much a case of bad solder, but the board flex caused by heat and the balls eventually let go.
I have removed failed graphics chips from laptops (infrared) and the traces are still attached to the BGA, but broken off the PCB.
This is really common in graphics cards and graphics chips in laptops too. A lot of heat over a small section of board.

The best way is to not use the lead free solder in areas where it will cause problems. People cry "environmental murder" at the mention of using lead. But really, how much better is it for the planet when the device is barely a year old and in landfill?

[damox]

By complete boards, do you mean an X-Ray picture of the complete board?

Yes, X-Ray picture of the complete board.

[alonehacker]

Yes, X-Ray picture of the complete board.

Yeah, I could definitely take an X-Ray image of the complete board. Hopefully sometime this weekend I'll do it!

[ttz]

nice job!

[HKKory]

YLOD doesn't always mean it's the CPU, or GPU. It signals a general hardware failure. I have a few consoles, that ylod because the ram chips are bubbling. If only it could do a error check like xbox 360

[alonehacker]

YLOD doesn't always mean it's the CPU, or GPU. It signals a general hardware failure. I have a few consoles, that ylod because the ram chips are bubbling. If only it could do a error check like xbox 360

I agree that it's not always the CPU or GPU. And yes, it would be nice if there was some sort of test point we could find which would tell us more about YLODs. Which RAM chips are you talking about that bubble?