PS3 (Research/Experimental) - NEC/TOKIN Capacitors Replacement - YLOD

[RIP-Felix]

It's also important to think about time. @squeept made a good point about the fact something is true today does not mean it will be be in a year's time. Things are always changing. We don't know if there were more BGA problems early, then the tokins started going bad. How may are dead tokins vs defective BGA etc. That ratio is likely to change over time. Especially as mor and more consoles get reballed (which is by far the most commonly believed solution for the YLOD). If they don't replace the tokins, they could be the next thing to fail.

This thread certainly has merit. Proadlizers are no longer in production. And as time goes on the equation may flip. More tokins than BGA and no new old stock tokins left to buy. So we'll nee tantalum caps to fix consoles. We just need to be sure it's the caps before we start, otherwise a false positive could cause us to assume the repair worked, sell the console and have a bad day when the customer ruins our 100% seller feedback rating e-bay.

 

[Revak3115]

This reminds me btw, do you normally care to check the "usage time" which is displayed on WebMan settings?
I'm not sure how relevant or even accurate this may be, but the longest I've seen on a 90nm RSX is 250 days on a H model, still working with the warranty seal intact (and Cell needing delid, which may be significant).

Coincidence or not, about all 5 C model consoles that I could check, had begun giving problems at around the 150d mark.

Maybe good thing to check on those consoles that come back, checking before and after. It should show how much they were actually used before coming back.

(Btw This is just some recent underdeveloped thought of mine but, could it be possible that the 90nm RSX present on H models is somehow revised or something? After all, they were the last ones to incorporate it.
But of course If I'm not mistaken it's not like H models aren't prone to YLOD. They very much still are, so faint hopes there)

i have had CECHA/B/E units with less then 50 days give artifacts and BGA issues, i currently found a CECHE01 with 2 days total and that seems to be fine, while a system with 1,200 days is fine also. it really depends on how well it was taken care of. i also do agree with you on 150d mark, i have some systems that needed New NECs around that time usually. i have a few A units that have around 1,200 days up. its very strange. maybe some units are just manufactured defective

 

[RIP-Felix]

... i have a few A units that have around 1,200 days up. its very strange. maybe some units are just manufactured defective

Totally depends on use. You have to remember that Blu-ray players were about $1000 then. The PS3 was a bargain by comparison and that's what a lot of people only used theirs for.

When it released I couldn't afford one. I waited a couple years for prices to come down and finally bought a slim. Later I got an A model used and gave my slim to my dad. I did at least change the thermal past and drill holes above the fan so it would run cooler before giving it to him (the slim). It's been running like a champ ever since (~ 10 years with daily use). But he only uses it as an MP3 player and DVD/Blu-Ray player. He never plays games on it. So the RSX doesn't get stressed much. It must have a crazy number of days on it.

 

[Alexander RU]

Hello, I am from Russia, I write through a translator and all responsibility for spelling lies with the translator. I have a Sony Playstation 3 CECHH08 PAL, when I receive it, I assume it is already heated with a building dryer. One day it just didn't turn on. Error, as later, YLOD. I did not have sony architecture, i repair notebooks and TVs. before, then I read 153 pages and decided to replace nec tokin. I have all the technique for replacing chips and balls. Having disassembled the nec tokin, I checked their performance. 6 nec tokin (4 for rsx and 2 for a cell) turned out to be from 900 to 1200 uF, 2 nec tokin from CELL on the reverse side were 2700 and 2900 uF. Read the whole topic, why is there such a difference in capacity? Why is Cell such a large capacitor? Replacing them with tantalum capacitors in
configurations 4x470 and 6x330 there will be an error, they will not correspond to the nominal value. I mean, don't disassemble nektok until you are sure they are not working. I put all the tokens in place, then I want to add nc 559 or rma 223 flux under the graphics chip and pump heat on balls, then rinse with isopropyl alcohol. I have a completely faulty rsx, because after 20 attempts to turn on the playstation, the CELL heats up, which indicates its performance. I'll write about the results.

 

[Pacorretaco]

i have had CECHA/B/E units with less then 50 days give artifacts and BGA issues, i currently found a CECHE01 with 2 days total and that seems to be fine, while a system with 1,200 days is fine also. it really depends on how well it was taken care of. i also do agree with you on 150d mark, i have some systems that needed New NECs around that time usually. i have a few A units that have around 1,200 days up. its very strange. maybe some units are just manufactured defective

Wow, and here I thought 250 days for a 90nm RSX was pretty good.
But 1200 days is a lot.
If you think about it, that would mean it's been on for over 3 years straight. Or... 8 hours a day for 10 years. Imagine at least 6 hours a day every day since launch day 15 years ago... Now imagine after the first 100 days the fan screaming for a CELL delid. Not even talking about the rest of the components in a 200W+ console (even idle in the XMB).
How did the machine even look inside?

Hahaha, I'm not saying it's imposible but... just a bit Hard to believe. (Or maybe the ps3 loving owner was also deaf? Hmm)
I mentioned not being sure of how reliable this webman figure is in the first place. On some instances, I remember it displaying strange numbers too, like 1536 days and other head scratching data.
After reboot these numbers showed more reasonable figures though.

 

[Revak3115]

Wow, and here I thought 250 days for a 90nm RSX was pretty good.
But 1200 days is a lot.
If you think about it, that would mean it's been on for over 3 years straight. Or... 8 hours a day for 10 years. Imagine at least 6 hours a day every day since launch day 15 years ago... Now imagine after the first 100 days the fan screaming for a CELL delid. Not even talking about the rest of the components in a 200W+ console (even idle in the XMB).
How did the machine even look inside?

Hahaha, I'm not saying it's imposible but... just a bit Hard to believe. (Or maybe the ps3 loving owner was also deaf? Hmm)
I mentioned not being sure of how reliable this webman figure is in the first place. On some instances, I remember it displaying strange numbers too, like 1536 days and other head scratching data.
After reboot these numbers showed more reasonable figures though.

whenever i boot those systems again ill look at webMAN once more to see if the data is different

 

[truemaster]

hello im looking to replace the nec/tokins from my cechc ps3. i read that for all phats every nec/tokin is rated at 2.5v 1200mf. that meens 3x470mf of tantalums for each nec/tokin total 24. im looking to buy from mouser. but what about esr value? can somebody point me to corect set? panasonic brand preferably. thanks in advance

 

[Yugonibblit]

hello im looking to replace the nec/tokins from my cechc ps3. i read that for all phats every nec/tokin is rated at 2.5v 1200mf. that meens 3x470mf of tantalums for each nec/tokin total 24. im looking to buy from mouser. but what about esr value? can somebody point me to corect set? panasonic brand preferably. thanks in advance

I would go with the lowest esr value

 

[RIP-Felix]

There are 3 factors you need to consider:

1. Are you sure you need to replace the NEC/TOKINs
   
2. Capacitance (total of your array)
   
3. ESR (Total of your array)

• See if your YLOD goes away when you add pressure to the RSX. If it does, you most likely have a BGA defect. If you have an oscilloscope or access to one, you can directly measure the waveform on the TOKIN +/GND to see if the tokins are bad. Once you have ruled out BGA defects and confirmed a bad tokin you can be assured this fix is what you need.
  
• Capacitance: NEC/TOKIN array was 4800uF total. This number is not arbitrary. It is ideal and calibrated for maximum efficiency. If you choose 3 X 470uF X 4 pads, your array will = 5640uF. That's not ideal, your filter will perform worse, but it's not super far over, so that should still be okay.
  
• ESR: Each NEC/TOKIN array was 0.375 mOhms. It is best not to exceed this. Choose the set with lowest ESR possible. To equal the token array, 0.375 mOhms is the number to shoot for. So, 0.375 X 12 caps = 4.5Ohms. In this case the combined ESR = ESR of an individual cap / #of caps in the array. It is easier to find lower ESR caps if you reduce the capacitance. Also you can't calculate ESR this way if you combine non-identical caps, such as when you leave a tokin in place (bad idea, it greatly increases combined ESR).

 

[truemaster]

yes i have warm the tokins and the console boots now. so according to what you say i need rsx top pcb side 3-3 tandalums cell top 3-3 tandalums (single tandalum cap in farad 470) botom side cell 2-3 rsx 2-3 to match the original capatitance more accurate correct?

 

[RIP-Felix]

yes i have warm the tokins and the console boots now...

That test suggests there are BGA defects that mechanically reconnect through thermal warping. It tells you nothing about the health of your filter caps.

I think you may have been mislead. Before you continue on perhaps you should tell us why you want to perform this fix.

1. Is there a reason you don't just buy another PS3?
   
2. Do you have the tools and experience needed to do this?
   
3. When does the YLOD occur?
   
4. Does the fan ramp up to jet engine mode?
   
5. What other work has been done on the console? Is it factory sealed?

...so according to what you say i need rsx top pcb side 3-3 tandalums cell top 3-3 tandalums (single tandalum cap in farad 470) botom side cell 2-3 rsx 2-3 to match the original capatitance more accurate correct?

Your previous idea overshoots the ideal 4800uF Capacitance. Yes, using fewer caps gets you closer to 4800uF, but lowers the required ESR that each cap would have to be in order to match the NEC/TOKIN array they are replacing:

Spoiler: Repeating Myself...again!

Combined ESR = ESR of an individual cap / #of caps in the array , where all capacitors are identical.

What you want to do is search mouser or digikey for TaPol caps:

• Filter by the lowest ESR.
  
• Try to find caps that will fit. 4.3mm Height caps are too tall, they'll interfere with the RF shielding on the backside of the board. 3.5mm is perfect. 7.3mm length will require you to scratch off solder mask to install them. 2.8mm length will allow you to solder directly onto the rails. I don't remember if it's been mentioned before, but those holes you see on the motherboard above and under the TOKINs (Vin/Vout and GND) are thermal VIAs. In addition to connecting the layers of the board electrically, they are critical for conducting heat too. Filling them with solder reduces their performance. This is why I prefer 3.5mm Length caps.
• I recommend Panasonic caps. Avoid the yellow ones!
  
• Once you have narrowed the list by the size, ESR, and manufacturer you want. Then calculate how many you would need to achieve 0.375mOhms Combined ESR.
• Then calculate the total capacitance. Adjust the above to achieve ~4800uF.

For Example:

470uF 2.5v 5mOhm TaPol
Combined ESR = 5.0mOhms / ? capacitors
5.0mOhms / 0.375mOhm = ~14caps
14 X 470uF caps = 6580uF (this is greater than 5760uF. It's too high.)

270uF 2.5v 6mOhms TaPol
Combined ESR = 6.0mOhms / ? capacitors
6.0mOhms / 0.375mOhm = 16 caps
16 X 270uF caps = 4320uF (this is low, not advised)
18 X 270uF caps = 4860uF (this is perfect)
21x 270uF caps = 5670uF (this probably fine, it's within the +20% threshold)

So the problem with the 470uF TaPol caps was there aren't any with an ESR low enough to get the total capacitance within 20% of 4800uF and maintain a combined ESR below 0.375mOhms. At least not in the size we wanted from the manufacturer we wanted. I had to go down to 270uF to accomplish this. I have previously posted all of this. You're lucky I felt generous enought to spoon feed this to you instead of slapping your wrist for not reading the thread!

I confirmed with an oscilloscope the noise rejection for 18x 270uF TaPol caps was better than the NEC/TOKINs, but I couldn't thoroughly test, as PS3#2 still had a YLOD on boot! It needed a reball, there was nothing wrong with the tokins to begin with.

There is wiggle room for the upper limit for capacitance, but I wouldn't sacrifice ESR. While 330uF & 470uF capacitors don't allow you to get capacitance down to 4800uF and still maintain 0.375mOhms Combined ESR, they may still work, but the filter's sweet spot is at 4800uF.

 

[truemaster]

i have warm the nec-tokin area with no more than 100c without hot the rsx or cell.its a cechc motherboard otherwise i woudnt bother. no i replace thermal paste every 2 years, i had delid rsx and do the rubber trick on cell. the fan never goes crazy. i have good solder skills. but from the way i read from you i lack other knowledge. thanks for helping me. i would definetly go for panasonic brand that is what i had in mind but please tell me what you recomend regarding capacity and how many per rail. im looking to replace them all

 

[Jaxuk]

Hello, i buy not open YLOD PS3 80 GB - CECHM03 - but i see this ? What is it and can this make YLOD on this PS3 ? I just buy 470uf 6.3 Transistors, to change Nec,Tokens... I need to change this or it doesn't matter for YLOD?

 

[Pacorretaco]

i have warm the nec-tokin area with no more than 100c without hot the rsx or cell.its a cechc motherboard otherwise i woudnt bother. no i replace thermal paste every 2 years, i had delid rsx and do the rubber trick on cell. the fan never goes crazy. i have good solder skills. but from the way i read from you i lack other knowledge. thanks for helping me. i would definetly go for panasonic brand that is what i had in mind but please tell me what you recomend regarding capacity and how many per rail. im looking to replace them all

It's very difficult to apply heat to the nec Tokin area and not heat the rsx or the cell. Among other things, because it's pretty much the same area.
Sadly also, I don't think these machines are failing precisely because they are running hot.

You'll have to decide yourself how to balance the number and the parameters of the capacitors together with the availability for you, price you are willing to pay, risks you are willing to take etc.

The parameters as @RIP-Felix kindly pointed out are to ideally match the original 4800uF and the ESR to be as low as possible. It's still not very clear how much you can actually deviate from this and still call it acceptable. But the closer the better.
Generally speaking, the more capacitors you put in parallel the lower the ESR will be.
But all in all, we know tantalum capacitors are not going to be an exact 1:1 match for the original NEC. So you will always be aiming for something that is "close enough", sadly.


Now if you only care about doing it the best way possible, then you most likely are already doing it wrong because you could do further diagnosis first. You may do the best NEC/TOKIN replacement in the world, and maybe your original ones were ok in the first place.
You just have to balance your own priorities.

Cheers

 

[RIP-Felix]

i have warm the nec-tokin area with no more than 100c without hot the rsx or cell...no i replace thermal paste every 2 years, i had delid rsx and do the rubber trick on cell. the fan never goes crazy.

It doesn't matter if it was only 100C or less. The reason your console turned on is because the motherboard flexed from thermal warping. This was enough to mechanically reconnect the failed connection. These defects are only a few microns apart. Any tiny change could be enough to cause the console to boot. The warped board will eventually relax and cause the defect to disconnect again. Only a reball will fix that.

It's good you have been diligent about replacing your thermal paste often. Brownie points for deliding! But that only puts off a BGA defect, it won't prevent it. Nothing can...

its a cechc motherboard otherwise i woudnt bother.

So if I understand you right, you want to get it working because it's a backwards compatible (hardware/software hybrid) model, not because it's sentimental and contains all your hard earned saves? Right? I'm mean that's a good enough reason for me, but I'm not you.

Spoiler: Beating a Dead Horse

It's a PITA to remove the tokins and install Tantalums. If you scour e-bay for a few months you may find a deal. I got lucky a couple months back and found an A model a with just a bad laser for under $100(US). I had a spare BD so it cost me nothing to fix. I just got a sealed working A model yesterday for $160. It even came with a broken sixaxis controller. Probably the one it came with. Considering these were $600 new, I'd say anything under half price is pretty good. Now, I'm in the US and we have a larger market than most. It may not be the same for you...IDK. You'll have to weigh the risk/reward accordingly. Tantalum isn't cheap. So the mod will cost you $30-50 in caps alone. Add to that any supplies you don't already have (solder, braid, flux, conductors, hot air, tweezers, & etc). You'll need to decide if it's worth trying to fix it. That's a personal decision.

Assuming you do decide to forge ahead, be advised you're entering "experimental" territory! There is a very likely chance the mod will fail, or if it works initially, then it may fail later for any number of reasons - not the least of which is you misdiagnosed the problem. Because of your history with hot air resolving the YLOD, it sound to me like you have a BGA defect that would require a reball to fix. It would take an oscilloscope to rule out the tokins and the SYSCON error codes to rule out something easier to fix (like southbrige, VRM, MOSFETs, HDMI encoder, and other common causes of the YLOD). I'm assuming you are just hoping the tantalums were your fix. If you thought this was an easy 1 fix fit's every YLOD situation, I'm sorry to tell you it's not.

I STRONGLY ADVISE YOU READ THIS ENTIRE THREAD BEFORE MAKING AN ATTEMPT.

1. To save you the time and money you would have wasted
   
2. To save this C model from being destroyed.

However, it's your property to do with as you please. If you want to print a picture of my avatar, tape it to the console and light it on fire, you have every right to do so (at least you would in the US). So if you say the words, "I still want to try it and see," I'll gladly help you waste your money.

i have good solder skills. but from the way i read from you i lack other knowledge. thanks for helping me. i would definetly go for panasonic brand that is what i had in mind but please tell me what you recommend regarding capacity and how many per rail. im looking to replace them all

Well, yeah. If you are going to replace any, you should replace all of them. Mixing caps of different capacitance can negatively affect the ESR. If I thought it would help your cause, I would recommend 18x (RSX) + 18x (CELL) of the caps I linked and would try to equally spread them out so that there isn't a current bottle neck forming on any one pad. I would also recommend plenty of Jumpers and a thorough cleaning to remove all flux residues.

Eww, now I feel like an accomplice.

 

[RIP-Felix]

Hello, i buy not open YLOD PS3 80 GB - CECHM03 - but i see this ? What is it and can this make YLOD on this PS3 ? I just buy 470uf 6.3 Transistors, to change Nec,Tokens... I need to change this or it doesn't matter for YLOD?

You said "not open". Did you mean it was sealed??? That sure looks like it's been tampered with. Looks to me like someone used WAY too much thermal compound (Not SONY's white love juices either). Yeah, sony uses alot, but this guy got super happy! You can see it bridging some components. If that is arctic silver or another conductive TIC, it could explain the scorch marks on the board. Looks like some components are scorched. That inductor coil on the bottom there looks bad. The cap in the bottom right looks suspect too. Even those top right tokins look a bit discolored.

To answer your question. Yes, that massacre can absolutely explain your YLOD. I think you're looking at more than just bad caps to fix this one. I would start by checking the SYSCON error messages to see how may there are. I suspect many! Then replacing the bad electrolytic. An ESR meter can diagnose them. I'd replace that inductor too. Then your going to have to start checking every tiny MLCC capacitor in the PWR circuit from 12v mains toward the chips. Check the VRM and MOSFETs for proper voltages (VDDR, VDDIO, Vref & etc). Then check Switching DC-DC converters for proper VDDC. I think troubleshooting that board would be a PITA. I don't envy you.

DIA-001 huh? More like DOA-001. I probably wouldn't even bother.

 

[traplio24]

Hi everybody.... i replaced all nec tokin under logic board but now i have no signal... i have to do jump wire of tantalum?

 

[Jaxuk]

You said "not open". Did you mean it was sealed??? That sure looks like it's been tampered with. Looks to me like someone used WAY too much thermal compound (Not SONY's white love juices either). Yeah, sony uses alot, but this guy got super happy! You can see it bridging some components. If that is arctic silver or another conductive TIC, it could explain the scorch marks on the board. Looks like some components are scorched. That inductor coil on the bottom there looks bad. The cap in the bottom right looks suspect too. Even those top right tokins look a bit discolored.

To answer your question. Yes, that massacre can absolutely explain your YLOD. I think you're looking at more than just bad caps to fix this one. I would start by checking the SYSCON error messages to see how may there are. I suspect many! Then replacing the bad electrolytic. An ESR meter can diagnose them. I'd replace that inductor too. Then your going to have to start checking every tiny MLCC capacitor in the PWR circuit from 12v mains toward the chips. Check the VRM and MOSFETs for proper voltages (VDDR, VDDIO, Vref & etc). Then check Switching DC-DC converters for proper VDDC. I think troubleshooting that board would be a PITA. I don't envy you.

DIA-001 huh? More like DOA-001. I probably wouldn't even bother.

Where can find diagram, to find RX,TX,GRN for this Motherboard, to read Syscon ? Thanks for the fast reply

 

[RIP-Felix]

Hi everybody.... i replaced all nec tokin under logic board but now i have no signal... i have to do jump wire of tantalum?

It sounds like you replace all the TOKINs on Side B of the Motherboard. If so the tokins on the other side are acting as the bridge. So you don't "need" jumpers. You want them for long term stability though. Let's focus on your lack of signal first...

Please provide more information. These details will help us:

1. What model PS3 are you trying to fix?
2. What was wrong with it to begin with (before you installed tantalum)? Please describe the fault and how the console behaves during startup. Does the fan ramp up to aloud level? Did it try to spin, then the console shuts down? If you had a YLOD how long was it? Did it occur in less than a few seconds after boot (Instant), or was it delayed a few minutes? Or was it only when playing intensive games? Did you see any graphical artifacts?
3. After replacing the tokins with tantalums, how does it behave? When you say "no signal," what do you mean? A YLOD? Or does the console boot, but no video?
4. Also can you provide a picture of your work,describe how you removed the tokins and installed the tantalum. Did you use any hot air to assist you or just a soldering iron?

 

[RIP-Felix]

Where can find diagram, to find RX,TX,GRN for this Motherboard, to read Syscon ? Thanks for the fast reply

Try asking in the SYSCON thread. I only work with A models, but I think I saw that model in there.