No, that rumor has been busted. Tokins and polymer caps do not heal capacitance with heat. That only applies to aluminum electrolytic and MLCC. It is possable that the array of bypass MLCC's are responable for attenuating the noise below a threshold that would cause the error, and that it will come back soon when the effect wears off.
Another potential explanation is the ESR of the feedback compensation loop for the PWM controller. As the console ages, solder deforms, components loosen in specification, the ESR/ESL of the loop increases and throws off the attenuation. This is why you must use low esr caps to replace the tokins (if they need replaced that is). The higher the ESR, the higher the voltage ripple.
Hello guys. My cechb ps3 is started giving ylod again. I was fixed it with changing 2 nec tokins on rsx side. It was giving ylod randomly, now its not even turning on. I think i need to change another nec tokin on rsx side. Cuz it shows rsx related error codes from syscon. What do you think?
Well I tried to re-flow the RSX chip with my electric frying pan and 700W (really doubt this rating) China heat gun and all I wound up doing was popcorning the chip so it is toast now. Oh well at least I tried.
Hello guys. My cechb ps3 is started giving ylod again. I was fixed it with changing 2 nec tokins on rsx side. It was giving ylod randomly, now its not even turning on. I think i need to change another nec tokin on rsx side. Cuz it shows rsx related error codes from syscon. What do you think?
OMG HELL NO
i just realized that. Im so sorry. Unit was working perfectly yesterday wtf. It has no overheating issue. I keep it cool, why the hell cold solder problem appeared on it there is no reason ( it was my favorite ps3 f*vk
Tried my best to clean up and solder the syscon wires then re-assembled it. Then suddenly it won't turn on and became a very short YLOD. Error code shows 2110(Clock Generator Error IC5001). Some earlier code shows 1001 and 1002...
Made a note about this on the SYSCON wiki. However, could you please post the errorcodes for this? I want to see the step number for the 2110 specifically.
I'm operating on the idea that subsystem voltages are one of the first things the SYSCON does in the Power On Sequence. If F6001 pops (caused by whatever, in your case C6010) then it cuts most of the peripheral voltages. USB, HDD, 5v_Misc, 3.3v_Misc, even some of the SB, CPU, and GPU voltages. The reason I think it complains about IC5001 is because clock generators are one of the most important functions for coordination. So it probably checks that clock signals are available first, perhaps even before continuing onto voltages.
So to test that hypothesis, went on the hunt for the step number of simalar 2110s caused by blown F6001. It was 02...
OMG HELL NO
i just realized that. Im so sorry. Unit was working perfectly yesterday wtf. It has no overheating issue. I keep it cool, why the hell cold solder problem appeared on it there is no reason ( it was my favorite ps3 f*vk
Yeah, it sucks when it happens. We all had that experience.
Sometimes it can go from random YLOD to full blown overnight. BGA deformation can make a weak bond that is precarious in the months leading up to it's final crack. Thing is that once it's on and warmed up the board warps enough to keep it running. And that strain can even keep it connected for a few hours between sessions. But after sitting and cooling the strain relaxes and the crack forms, preventing the console from turning on until it's heated up. This is why, people often get the console to boot after multiple attempts, after which it is stable until cooling down. That can be one way to retrieve saves and stuff.
Anyway, there is a pressure test to confirm a BGA or bumb defect. If you press down on the RSX while attempting to turn it on, it might boot. If that works, you definitely have a BGA or bump defect. But the opposit isn't true. So if it doesn't work, that doesn't mean you don't have a defect. It's kinda a mute point now though, those error codes are pretty definitive.
Yeah, it sucks when it happens. We all had that experience.
Sometimes it can go from random YLOD to full blown overnight. BGA deformation can make a weak bond that is precarious in the months leading up to it's final crack. Thing is that once it's on and warmed up the board warps enough to keep it running. And that strain can even keep it connected for a few hours between sessions. But after sitting and cooling the strain relaxes and the crack forms, preventing the console from turning on until it's heated up. This is why, people often get the console to boot after multiple attempts, after which it is stable until cooling down. That can be one way to retrieve saves and stuff.
Anyway, there is a pressure test to confirm a BGA or bumb defect. If you press down on the RSX while attempting to turn it on, it might boot. If that works, you definitely have a BGA or bump defect. But the opposit isn't true. So if it doesn't work, that doesn't mean you don't have a defect. It's kinda a mute point now though, those error codes are pretty definitive.
yeah it was litteraly like that. Randomly giving ylod when playing gran turismo 5. Now its completely dead. Also it was working after 1-2 ylod. Like you said. Thank you for technical explanation. I tried putting pressure on rsx and cell but didnt effect anything. Also tried coin trick but no results.
Thanks, that indeed took me a while to figure out. Here's my work log:
F6001 fuse blown. IC5001 or related fuse/resistor/cap problems. Need to check for short before resolder a new fuse. (IC6017 1.7V_MISC input was short to GND with resistance 28Ohm. Q6307 and Q6303 has resistance of 28Ohm and it's main voltage regulators output to 1.7V_MISC, all of them might be the cause of the blown fuse. <-- This was validated with #13 board but looks like they should have resistance of 28Ohm.) Replaced F6001 from #13. 12V line still has resistence about 100Ohm, turned on, blown F6001 again. Now both #12 and #13 has F6001 open. Tested resistance and do the comparison, there's reason to believe it's either Q6002, or C6019, C6020, C6023 are faulty. Either case, I need to wait for hotair rework station to arrive to begin further test. Update: turns out C6019 is short while Q6002 maybe okay. After replacing a new F6001, it bringups to a 1001 error, then a second try with 1200(BE thermal) because I didn't put fans on. Assembled and got no errors.
Made a note about this on the SYSCON wiki. However, could you please post the errorcodes for this? I want to see the step number for the 2110 specifically.
I'm operating on the idea that subsystem voltages are one of the first things the SYSCON does in the Power On Sequence. If F6001 pops (caused by whatever, in your case C6010) then it cuts most of the peripheral voltages. USB, HDD, 5v_Misc, 3.3v_Misc, even some of the SB, CPU, and GPU voltages. The reason I think it complains about IC5001 is because clock generators are one of the most important functions for coordination. So it probably checks that clock signals are available first, perhaps even before continuing onto voltages.
So to test that hypothesis, went on the hunt for the step number of simalar 2110s caused by blown F6001. It was 02...
No, that rumor has been busted. Tokins and polymer caps do not heal capacitance with heat. That only applies to aluminum electrolytic and MLCC. It is possable that the array of bypass MLCC's are responable for attenuating the noise below a threshold that would cause the error, and that it will come back soon when the effect wears off.
Another potential explanation is the ESR of the feedback compensation loop for the PWM controller. As the console ages, solder deforms, components loosen in specification, the ESR/ESL of the loop increases and throws off the attenuation. This is why you must use low esr caps to replace the tokins (if they need replaced that is). The higher the ESR, the higher the voltage ripple.
Out of curiosity, this is what I found on a GLOD, potentially video memory problem board. Is this a so called RSX bump? I don't know if it's clear enough on the picture but if you look at it under a light, you'll see the area around the circle has a round bulged surface.
Out of curiosity, this is what I found on a GLOD, potentially video memory problem board. Is this a so called RSX bump? I don't know if it's clear enough on the picture but if you look at it under a light, you'll see the area around the circle has a round bulged surface.
Well I tried to re-flow the RSX chip with my electric frying pan and 700W (really doubt this rating) China heat gun and all I wound up doing was popcorning the chip so it is toast now. Oh well at least I tried.
Well, the way I look at it is this. You can't break what's already broken. I guess it's a Frankenstein candidate now.
Yeah, trying builds skill and confidence. So don't be afraid to try new things. You can see the need for better equipment now, can't you. I just wish competent reballing equipment wasn't so expensive. But I guess if it were cheap and accessable, the YLOD wouldn't be notorious.
@RIP-Felix Yah I did not expect it to work. I have started working on my third system (I got around 20 of these from that friend I mentioned) and it is a Slim system with DYN-001 motherboard and I think the RSX may also be bad as it also has the blueish color to it and this time both chips have had there heat spreaders removed and the RSX RAM chips did not have any thermal paste on them and there was flux around the RSX as well. The odd thing is that it stays powered on (wit a green light) when I turn it on but there is no video out (this test was with out the hard drive or disk drive connected so that may be normal?) and the error codes seem to indicate a issue with the AV or HDMI chips. I have attached the error log to this post just in case you are interested. Also the built in fan test functions as well but I don't think this really shows anything. Is there a way on this motherboard to enter internal mode or is that not possible due to there not being a diag pin to ground?
Great thanks! The good tokins looks really pretty clean.
I do have a question though. During live probing yesterday, and from your oscilloscope image too, I see the voltage of both Tokins are around 1.2V - 1.3V. So if the colour you mentioned is right, the RSX VDDC gets 1.3V and BE VDDC gets 1.2V. This is different from what service manual says. Shouldn't it be RSX VDDC=1.2V and BE VDDC=1.0V? I got around 1.25V-1.3V on BE VDDC yesterday on a(another) short YLOD DIA-002 board, the signal looked clean but I was wondering if the voltage was right.
Great thanks! The good tokins looks really pretty clean.
I do have a question though. During live probing yesterday, and from your oscilloscope image too, I see the voltage of both Tokins are around 1.2V - 1.3V. So if the colour you mentioned is right, the RSX VDDC gets 1.3V and BE VDDC gets 1.2V. This is different from what service manual says. Shouldn't it be RSX VDDC=1.2V and BE VDDC=1.0V? I got around 1.25V-1.3V on BE VDDC yesterday on a(another) short YLOD DIA-002 board, the signal looked clean but I was wondering if the voltage was right.
The nominal voltages stated in the manual were for for new chips. The VID is selected by the CPU/RSX on the fly based on load and other factors. The controller outputs what the processor demands, which can get larger with age and use. 1.2v to 1.3v is what you'll typically see.
Can I just say, this post completely flew past me when I first read it! After countless hours of watching videos and reading the datasheets of the PWM controller and DC/DC converter this is finally beginning to make sense. The 1002 error was correctly characterized way before I did it, and more completely. Props to you guys!
I think I have an explanation for why SONY has 1002 labeled "RSX VRAM Power Fail." I think they meant to type "RSX VRM Power Fail." I mean, that makes A LOT more sense! VRM is more encompassing than "no drive signal," which is a symptom not the cause. So it would cover a broader spectrum of potential causes. And they have this section on the service manual labeled "VRM for RSX (2Phases)." I'm pretty sure that's just a typo.
Under Voltage Lockout (UVLO) circuit:
Consists of the Enable and V input to the controller. Enable is controlled by the SYSCON directly. The controller will refuse to work if it's input voltage is not sufficient. Now these controllers are powered directly off 12v_Main which is supplied by the PSU. There isn't much conditioning before it gets there. So a bad PSU could trigger an UVLO. I suspect this is the cause of some 1001 errors.
Both PWRGD and Drive need to be high to enable the buck converters that power the PS3's processors. With that in mind, lets dig back into those oscilloscope measurements.
First thing I notice is that the Drive pin is low when it should be high (blue = Bad RSX tokins | WHT = Good Tokins):
The controller's drive signal tells the buck converters to switch on and power the processor. It does so through an AND Gate, but only when both PWRGD and Drive signals are present!
This gives us a new thing to test. If Drive and PWRGD are high on the input side of that AND gate (IC6107/IC6204 for CPU/RSX VRM respectively), then output should be high! Check it has all the correct voltages. Replace if not.
PWRGD is also interesting (Blue = Bad Tokins | WHT = Good Tokins)...
PWRGD (AKA reset) notifies the system when voltage falls out of regulation. It has protection built in that prevents false triggering due to voltage spikes/noise from transient loads. So PWRGD "shouldn't" go low without cause.
From watching this video...
...Power Good is usually sent 100-500ms after power is switched on. Someone will need to double check this reasoning, but I gather the SYSCON is timing how long it takes for the PWRGD signal to arrive after the Power Control line is enabled. When the SYSCON enables the Power Control Line, enable signals travel across the board to many subsystem controllers, initializing main system voltages, VRM, etc. There is a proper amount of time that needs to elapse before the SYSCON begins to worry. If the time between the Power control line enable signal and Power Good signals coming back to the SYSCON from across the board are too short or long, it get's mad, takes control of the situation, and waits ominously for your to figure out how to resolve the issue.
By way of analogy:
Jeff Goldbloom = Enable Signal
Infant = Hardware issue
The Trailer = A busy Controller trying to resolve the issue...
Mommy = SYSCON
I think the time delay is needed to account for Softstart (SS = Pin6), which limits inrush currents. It's set using C6208 (0.1uF) according to the equations presented in the Introduction to Buck Converter Features video. Also, it just takes a certain amount of time for a chip to power up and report power good. That's the rise time you see in the power good signal.
Without a proper delay the SYSCON assumes something is wrong, triggers an error code, and shuts off the console. @DeadEnd found that out when he injected 3.3V into PWRGD (power good) of IC6103 (CPU Buck controller). It generated error 20 1001 and 20 3010. So, I think we can now conclude definitively 1001 is related to the CPU VRM, and that 3010 can be interpreted as "PWRGD signal time too short," since it would have been a 0ms delay if powered externally.
I wonder if @DeadEnd could have used an RC circuit with a 100ms time constant, long enough to delay the rise time by the required amount. He would just need to tie it to a switched power line so it got voltage when the system powered on, then the RC would delay the rise time and perhaps create the delay necessary to trick the SYSCON that PWRGD was present.
Speculation:
That jagged startup to the PWRGD signal has higher amplitude (bigger/wider peaks) in the console with bad tokins. I wonder if that is used to gauge the extent of the ripple. Say it's higher than half the final 3.3v PWRGD signal, maybe that over the threshold for triggering a 1002. Say it's more than 75%, then maybe it's over the threshold for triggering 3004. Of course it could be 25% and 50% respectively, IDK. It's just a hypothesis, but it makes sense there would be some mechanism for generating a standard voltage ripple/noise pulse during the check. I'm wondering if that's what those spikes are before PWRGD goes high.
( it was my favorite ps3 f*vk
I had an error 2110 and in my case PS6001 was bad.
