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

[marciolsf]

No worries. I learnt a lot from you. Mind the WD40 I only used on the IHS itself. I don't want to try it on PCB not to mention the RSX it could be corrosive. Contact cleaner maybe good?

Pretty sure wd40 is rather corrosive, and hard to fully remove as well (since its meant to lubricate, so it tends to stick around). goof-off would be a better solution.

 

[RIP-Felix]

I have used WD40 to cut the grease and grime on SNES/N64 cartridge slots in the past. It doesn't seem to corrode, but I do clean it off thoroughly with IPA afterwards, so nothing sits on the board. Works well.

If just using it on a q-tip and applying to the Resin coating on the RSX ram, it should be fine. Especially if removed quickly afterwards. I read ii's primarily composed of 9-14 chain Alkanes (hydrocarbons). It's a good de-greaser, water displacer, lubricant, and light rust remover. That's why I used it clean cartridge connectors. Sounds like it's somwhere between Gasoline and diesl fuel with a little mineral oil and xylene mixed in. Kerosine isn't a bad description. It may very well have some kerosene in it.

I wouldn't advise letting it sit on plastic parts or soak into resin microchips, but a short exposure to help knock off the thermal adhesive seems like a reasonable use to me. So long as you clean it off afterwards.

 

[Nosdio]

For really bad contacts you can use a soft sandpaper, lets say 2000 grit. It does a great job on the cartridge. If you don't like the idea simply use a white pencil eraser to rub the contacts. In all cases you can clean with IPA afterwards.
For the thermal glue one the ramchips you can use acetone to remove but not for prolonged periods of time. For the IHS there is no such issue.
Better not use WD-40 on electronics, i know it works some times but it is not the recommended product

 

[Chino Moreno]

Hi. Bought a cheap slim for parts and thought I'd revive it a few weeks ago. Unit is grossly infested with roaches. After I cleaned it up. It has Glod for more than a minute then shuts itself off. Getting ideas from reading past posts, I removed the hdmi chip but it still does the same thing. I am getting mostly a0902124 and a few a0902024. Read here that someone experienced the same but with an a0802124 but his ran after a minute and was solved with an hdmi chip replacement. So my question is, what does the A090 actually mean? Should I also remove the AV chip to see if there will be any difference? Btw wifi card looks shot as well, i just tried with the best of my abilities to resolder the pins and clean it with a microscope.

 

[RIP-Felix]

For really bad contacts you can use a soft sandpaper, lets say 2000 grit. It does a great job on the cartridge. If you don't like the idea simply use a white pencil eraser to rub the contacts. In all cases you can clean with IPA afterwards.
For the thermal glue one the ramchips you can use acetone to remove but not for prolonged periods of time. For the IHS there is no such issue.
Better not use WD-40 on electronics, i know it works some times but it is not the recommended product

So...don't do that! Sandpaper removes the hard gold coating that protects the the cart connector.

I start with IPA + Cotton Swab. Then WD-40 + Mr. Clean magic eraser. If I have to, Ammonia + Magic clean eraser. If still not working, I go back to Q-tips + Brasso/silver polish, but only on the stubborn pins.

I will admit to the sand paper travesty in my younger days. I feel so stupid now!

 

[RIP-Felix]

... So my question is, what does the A090 actually mean? Should I also remove the AV chip to see if there will be any difference?

Step number 90 is the power off state. There is an HDMI error being registered when the console is powering off. We've had a few of them reported recently, with mixed results from removing/replacing the HDMI encoder. Clearly there's more to the story than just the encoder chip.

The HDMI encoder is attached to the RSX VDDIO output. So RSX BGA problems can not be ruled out. It's common to see an HDMI error associated with the 3034/4xxx error we usually see as definitive evidence of RSX issues requiring a reflow/reball. But without those errors, a 2024/2124 may not be on the RSX itself. Where exactly it is, we don't not know yet. Would need to start lifting pins on a working console to trigger the error and track it down. No one wants to do that.

You're gonna have to troubleshoot.

 

[Nosdio]

So...don't do that! Sandpaper removes the hard gold coating that protects the the cart connector.

I start with IPA + Cotton Swab. Then WD-40 + Mr. Clean magic eraser. If I have to, Ammonia + Magic clean eraser. If still not working, I go back to Q-tips + Brasso/silver polish, but only on the stubborn pins.

I will admit to the sand paper travesty in my younger days. I feel so stupid now!

Brasso/Silver polish is a compound meaning it is abrasive*, i am only saying this because you disregard the sandpaper. If you use Brasso it is like using and sanpaper, a very fine one, but a sandpaper. Magic eraser it's an abrasive too. You can try all these in a clear plastic and see for yourself. There you should check the extend of the damage as a glimpse of what will happen to the contacts. To my understanding it is too little to say it is damaging because we are talking about seconds of "sanding" with either of these products and not like 15 minutes.

For contacts like pc memory or cartridges the white pencil eraser does a great job and it is borderline abrasive. It just removes the oxidation and patina lets say..

I think at the end, all DIY methods has to do with the person applying it. If someone does not have the skill to observe whether he will fix it or break it by applying to little or too much, then he better have someone else do it for him.

*In fact the proper way to use it is to let it dry in the surface (copper cooler) and then wipe it with a cloth.

 

[Chino Moreno]

Thank you, Felix. Yeah thats what I thought about the code 90 via the github guide. Looks like I'm missing a voltage rail given the poor state of the board. Just hoped there was something else easier to fix.

 

[RIP-Felix]

Brasso/Silver polish is a compound meaning it is abrasive*, i am only saying this because you disregard the sandpaper. If you use Brasso it is like using and sanpaper, a very fine one, but a sandpaper. Magic eraser it's an abrasive too. You can try all these in a clear plastic and see for yourself. There you should check the extend of the damage as a glimpse of what will happen to the contacts. To my understanding it is too little to say it is damaging because we are talking about seconds of "sanding" with either of these products and not like 15 minutes.

For contacts like pc memory or cartridges the white pencil eraser does a great job and it is borderline abrasive. It just removes the oxidation and patina lets say..

I think at the end, all DIY methods has to do with the person applying it. If someone does not have the skill to observe whether he will fix it or break it by applying to little or too much, then he better have someone else do it for him.

*In fact the proper way to use it is to let it dry in the surface (copper cooler) and then wipe it with a cloth.

Mr clean Magic Eraser is melamine foam. Melamine is an abrasive, but on much smaller scale than sandpaper. Yes, it'll scratch plastic. I couldn't find the hardness of Melamine listed anywhere, but VERY few things are harder than glass (silica). I think it's safe to assume an organic molecule such as melamine is softer than glass! So it shouldn't scratch fiberglass (FR4 substrate). and be gentiler on the hard gold fingers. It gets into microscopic pores of the material better too. So less abrasion is necessary to clean it. For those reasons, it won't damage the contacts as much as sandpaper.

Brasso is a 3-5% IPA, 5-10% Ammonia, 15-20% Oxalic Acid, and yes silica powder (fine sand). Ammonia removes oxides chemically. Oxalic Acid forms a complex with iron (removes rust). IPA helps clean/prep the surface so the chemicals come into contact with the surface oxidation and it evaporates. The evaporation concentrats the solution, increasing the chemical and abrasive action as you continue rubbing. Yes, silica power is the abrasive, but it's much finer than sand paper and wet. It's less evil than sandpaper and requires much less rubbing thanks to the chemicals rust removers.

BTW, you can make this yourself if you don't want silica in the solution! Store bought ammonia is only about 3-5%, but better than nothing. Zud cleanser contains Oxalic acid. IPA is easy to get. So you can easily make a non-abrasive alternative if you want to.

Like I said, it's a phased approach. Only escalate the damaging chems if really needed. And only on the pins that actually need it. Sandpaper is a scorched earth tactic!

 

[Nosdio]

Mr clean Magic Eraser is melamine foam. Melamine is an abrasive, but on much smaller scale than sandpaper. Yes, it'll scratch plastic. I couldn't find the hardness of Melamine listed anywhere, but VERY few things are harder than glass (silica). I think it's safe to assume an organic molecule such as melamine is softer than glass! So it shouldn't scratch fiberglass (FR4 substrate). and be gentiler on the hard gold fingers. It gets into microscopic pores of the material better too. So less abrasion is necessary to clean it. For those reasons, it won't damage the contacts as much as sandpaper.

Brasso is a 3-5% IPA, 5-10% Ammonia, 15-20% Oxalic Acid, and yes silica powder (fine sand). Ammonia removes oxides chemically. Oxalic Acid forms a complex with iron (removes rust). IPA helps clean/prep the surface so the chemicals come into contact with the surface oxidation and it evaporates. The evaporation concentrats the solution, increasing the chemical and abrasive action as you continue rubbing. Yes, silica power is the abrasive, but it's much finer than sand paper and wet. It's less evil than sandpaper and requires much less rubbing thanks to the chemicals rust removers.

BTW, you can make this yourself if you don't want silica in the solution! Store bought ammonia is only about 3-5%, but better than nothing. Zud cleanser contains Oxalic acid. IPA is easy to get. So you can easily make a non-abrasive alternative if you want to.

Like I said, it's a phased approach. Only escalate the damaging chems if really needed. And only on the pins that actually need it. Sandpaper is a scorched earth tactic!

Ok Felix, really it is not a matter of dispute for me and as far as my contacts i am clean
To tell you the truth i wish it could be a better way to remove the IHS of RSX than the scorched earth methods we are obliged to use..
Since you have made your reasearch, WD-40 is safe for electronics? I have seen guys spray whole motherboards (and with brake cleaner too) with the thing and if i am not mistaken it is whritten on the bottle that is not intended for electronics, but you may know better that is why i am asking.

 

[RIP-Felix]

WD-40 is the one chemical I can see a potential for bad results. It could work it's way inside small cracks and loosen things. It is a penetrant. That's why I use it for rust/oxidation, grease and grime. But only with supervision and I never leave it on the board! I don't use WD-40 for anything but contacts. I would never spray it wholesale on the board! And I clean it off THOROUGHLY afterwards. In fact I always clean the board thoroughly before final assembly anyway, just to be sure there are no residues left. Be it chemicals I used to clean, or crap from years of neglect.

I wouldn't reccomend brake cleaner, but an electronics contact cleaner works well and is safe to use on PCB's. I've sprayed it in the IO/USB ports and all the plactic parts attached to the PS3 MB. It doesn't hurt anything. I usually finish off spraying with 99% IPA.

 

[SkaziChris]

Well, the videos were helpful. I can see you have good equipment. Question: Do you have an Oscilloscope? If so, you can probe VDDC and measure the noise to confirm bad tokins. Set scope to 50mV/div and 1uS, trigger at 1v. Noise should be under 50mVpp.

I would start with the troubleshooting the VDDC filter on both the CPU and RSX. First compare the resistance and ESR of the first stage RC filter. It amplifies the attenuation of the second stage (tokins). So if the first stage is bad, the whole filter will be detuned. and cause the 1002 error. It's easier to replace a couple of SMD's. Compare the values to the other 3-4 in the filter. Here's what I mean...
View attachment 34886
...The resistanc and ESR of those SMD's should match (within reason). If not replace and retest, before removing the tokins.

Stuff you get to skip thanks to the SYSCON UART Diagnosis (Imagine how much time that would have wasted):

• ESR/resistance of the larger MLCC's next to the IOR VRM and the electrolytic caps. I'm sure they're fine or you'd get 1004 error and the RSX wouldn't start.
• Compare the SMD components next to the Inductors to a known good board (to the left of the inductors in the pic above. May be different on a VER-001 IDK). There are more on the back side.
• Check for voltage @ a few MOSFET in that section. The console wouldn't be booting this long or report a RSX OK PowerState if there were an issue here.

If the 1st stage filter is fine, then replace each tokin with 3x470uF Low ESR TaPol caps. I recommend Panasonic 2R5TPF470M6L or you can harvest them from dead motherboards. AlPol caps will work too, but they're theoretically not as good. You also need to bridge the + rails with at least 16AWG.

Hello Felix,

I bought a digital oscilloscope:
https://www.amazon.com/Hantek-HT6022BE20Mhz-Digital-Oscilloscope-Bandwidth/dp/B009H4AYII

Thanks for providing the location of VDDC filters on COK-001.
However, I cannot find such elements on VER-001.
Above Tokins, I only see some mosfets, fuses, and pmic.

Since there is no circuit diagram available for VER-001, any advice on where to probe for VDDC noise measurements?

Based on COK-001, the components are:
R6155
C6137

R6154
C6136

R6156
C6138

 

[sandungas]

For the thermal glue one the ramchips you can use acetone to remove but not for prolonged periods of time. For the IHS there is no such issue.
Better not use WD-40 on electronics, i know it works some times but it is not the recommended product

I use to say that one of the best tools to do that kind of job is the nails of our fingers
But obviouslly, thats only if you dont find any better tool at home, is very probable that the nails could get a bit damaged

To prevent leaving scratches in the surface we need to use a tool made of a material softer than the surface we are scratching (at the cost of some deformation/degradation in the tool edges)... but hard enought to remove the thermal glue

One of those used for hobby works to sculpt mud/ceramic, the ones that looks like a knife
In that photo are made of wood, but i have some made of plastic
And yeah... your could make it at home, starting with a stick of hard plastic + a file + sandpaper for the finish of the edges

 

[RIP-Felix]

Hello Felix,

I bought a digital oscilloscope:
https://www.amazon.com/Hantek-HT6022BE20Mhz-Digital-Oscilloscope-Bandwidth/dp/B009H4AYII

Thanks for providing the location of VDDC filters on COK-001.
However, I cannot find such elements on VER-001.
Above Tokins, I only see some mosfets, fuses, and pmic.

Since there is no circuit diagram available for VER-001, any advice on where to probe for VDDC noise measurements?

Based on COK-001, the components are:
R6155
C6137

R6154
C6136

R6156
C6138

Between the tokins and cpu or rsx is the vddc output of the filter. That's where you want to probe. The mosfets used on the BC models have more noise. Partly because they draw more power, partly because there are more switching VRM, and because they are different. I can see the pads where the first stage RC filter would be, but they didn't populate the pads. Evidently, the RLC filter was enough to attenuate the noise.

 

[SkaziChris]

Between the tokins and cpu or rsx is the vddc output of the filter. That's where you want to probe. The mosfets used on the BC models have more noise. Partly because they draw more power, partly because there are more switching VRM, and because they are different. I can see the pads where the first stage RC filter would be, but they didn't populate the pads. Evidently, the RLC filter was enough to attenuate the noise.

So, based on the link below, it would be the point (1) where I'd put the oscilloscope probe to measure VDDC, right?

https://translate.google.com/translate?sl=de&tl=en&u=https://www.trisaster.de/page/index.php?topic=484


I made following connections:

Then, in standby mode, I see this noise:

Then, I perform startup sequence from UART (auth, bringup, powerstats, shutdown).

 

[RIP-Felix]

So, based on the link below, it would be the point (1) where I'd put the oscilloscope probe to measure VDDC, right?

https://translate.google.com/translate?sl=de&tl=en&u=https://www.trisaster.de/page/index.php?topic=484


I made following connections:

Then, in standby mode, I see this noise:

Then, I perform startup sequence from UART (auth, bringup, powerstats, shutdown).

Yes, but you need need a low impedance path to ground or the noise floor will obscure the measurement. You should be using an oscilloscope probe with a ground spring.

I also like to trigger at a certain point after startup for consistent measurements. Kinda hard to explain without pictures, and I'm away from a computer. I'll try to explain later. Search YouTube for EEV Blog's how to measure ripple and noise video. There are other videos that show it too. IDK how that cheap scope will work, so you'll have to figure it out.

 

[SkaziChris]

Yes, but you need need a low impedance path to ground or the noise floor will obscure the measurement. You should be using an oscilloscope probe with a ground spring.

I also like to trigger at a certain point after startup for consistent measurements. Kinda hard to explain without pictures, and I'm away from a computer. I'll try to explain later. Search YouTube for EEV Blog's how to measure ripple and noise video. There are other videos that show it too. IDK how that cheap scope will work, so you'll have to figure it out.

Sure thing.

The probes are equipped with ground clips, and connected to ground.

The noise is waaay above 50mV...

I also made 1 measurement from upper metal case of the mobo and put the probe directly after NEC/Tokin capacitor...

The voltage noise is 131mV.

 

[RIP-Felix]

Ground clips are an antenna and most of the noise you are measuring there is coupled. The probe needs to be shielded and the ground needs to be as short as possible. A 50ohm coax can be used. Solder the center conductor to VDDC and the braided ground sheath to the ground plane as nearby as possible. That is the best way, but an oscope probe is good enough.

You're zoomed out too far. Here read this post and related. I show what you're looking for and go into details about the axis setup to see the noise. But like I said, you need a proper probe or it'll be obscured in the noise.

 

[Michael Crump]

I have a JSD-001 and have hooked up the RX/TX/GRD. I'd like to wire up the diagnostics wire now but I can't find a diagram of where it is located on a JSD-001 board.


https://imgur.com/a/gkLTBhL <- my system right now

 

[RIP-Felix]

Sure thing.

The probes are equipped with ground clips, and connected to ground.

The noise is waaay above 50mV...

I also made 1 measurement from upper metal case of the mobo and put the probe directly after NEC/Tokin capacitor...

The voltage noise is 131mV.

So here's the bad tokin wave forms you are looking for.

Spoiler: Bad RSX Tokin

Spoiler: Bad CPU Tokins

The sharp spikes is noise. The rectified sin wave is the ripple. We're mainly concerned with voltage ripple. That's what the tokins are tasked with attenuating (reducing). MLCC caps can be used to help reduce the noise spikes, but they aren't the major concern.

Spoiler: Good Tokins

CPU is in yellow and RSX is in Blue

Pay attention to the axis. Horizontal needs to be near 1-5us and vertical axis needs to be about 50mV/div. If you set the scope to AC coupling, you won't have to adjust the DC voltage offset to get the graph in view. Or you can start zoomed out and start triggering in on it.

The further in you are zoomed the more accurate the measurement. So the reason you had so much noise, is partly because of the ground clip and partly because you are zoomed out so far.