Discussion in 'Tutorials & Guides' started by Naked_Snake1995, Jul 15, 2019.
Hi again..the tantalium caps should be 470 pico farad or micro farad?
UPDATE: Just got a warranty return from a customer after almost 6 months, popped it open and found it was one of the very few that I actually replaced the caps on. Somewhere in the last few pages, I said I've been using 2x 26 gauge stranded wires per chip and it was fine. So, uhhh.... turns out that was a lie. CPU side wires burnt to a crisp. Don't know if it was some kind of unexpected power surge, or if they slowly degrade over time, but 2x 26 gauge is definitely NOT sufficient!
Wouldn't a single core, pretty thick wire act as a copper trace in some way? Don't know the math to propperly prove that but it sounds reasonable to me, like the shunt wires that multimeters use. I guess it could be difficult to bend and solder but it may work.
Edit: when doing my ATX mod I had a 12V wire burn to ashes because the second and third ones weren't making propper contact. The CELL and RSX use lower voltages and higher current so this is not a good comparison but it was a 13 AWG wire handling at max 13A, it probably started burning at lower current tho.
I was going off of this chart with a guess at 2 amps max per chip (which I see now was obviously way too low), and then doubled it for good measure: https://www.powerstream.com/Wire_Size.htm
However, I was using the "chassis" rating, when I guess I should be using the "power transmission" rating since there is so much added heat from external sources already. So, I'd be more than comfortable switching to 2x 18 gauge per chip. Between the 3 consoles with bad caps I've found, and the handful that I've ripped the caps off of for various experiments, there are are least half a dozen of mine floating around that I haven't had back for warranty repairs with 2x 26.
Thanks! I'm getting them ordered. What do you think of the location for ground? some quick testing shows continuity between the outer edge of the board and the negative poles of the caps, so I should be getting the same scope results on either one...
Either would be fine, but I'd go for the copper on the edge with no shielding on so there's less chance of touching anything else. At 50mv, just you touching the shielding can affect the readings a little.
@squeept model of the problematic console?
Also, I had a question for you. Have 2 fatties with HDMI problems (shorts), and the consoles are in GLOD condition. That means those RSX are dead, or should I see those shorts? It's wierd since the consoles should be shutting down. And btw, could you boot a console with only the CELL, or that's not possible?
@Workz_777 Sir, you have quite a hand for electronics https://i.imgur.com/UnvmIHS.jpg
I would expect each chip to draw more than 2A, knowing that they work below 2,5V and that the 12V lane from the PSU is designed to deliver up to 32A in CECHA's and B's, so in the worst case both chips could use 384W, I know other components use the 12V as well, but CELL and RSX are the most demanding, I don't know the max voltage this chips scale up to, but let's wrongly assume that they use 2V, in that case they could be drawing 192A at the same time (obviously this is not the case) but they draw huge ammounts of current, so 2A seems too little to me.
I would go for 2x13AWG wires per chip if I was looking at chassis wiring amps, or 2x9AWG wires per chip if looking at power transmission amperage and still falling below the (theorical and for sure exaggerated) 192A.
I've never bridged the cap contacts because I never replaced all the NECs so I have no experience on this, so you try and tell us what ends up working for you
CECHE01. Where are you seeing shorts? I honestly don't pay any attention to GLOD versus YLOD and the times before showing up / shutting down. There are an unholy amount of BGA pads, bumps, internal connections on each chip that there are myriad ways for any BGA/chipset issue to manifest.
Yeah, I screwed up putting stuff in to the little calculator I was using, using the internal resistance of the chips as the starting point instead of the wattage. Time to go back to school. I only purposely get CECHA01 and CECHE01 in here, so next time I have to do it, I'll just run the jumpers through a multimeter and see what the real world tells us. Since CECHA01 sucks the most power, that will tell us what to use on all of them.
I haven't put this CECHE01 back together yet. Max overall power draw is different, but I can't see why the individual draws of just the RSX and CELL wouldn't be the same as a CECHA01. Anyone wanna ring in there if there's something I'm not thinking of? I'll just do it on this one then, should get to it this evening.
If I'm not mistaken, from CECHA to CECHE they all use the same 12V 32A PSU as they both have 90nm RSX and 90nm CELL, so they should be drawing around the same current, except for the PS2 hardware inside CECHA and CECHB that should make them draw a little bit more in some cases I guess. Also the card reader could make a little difference but I don't think that it would be that noticeable. Except CECHB they all seem to hace card readers.
@squeept on a "L" I had a burned Panasonic chip, removed it but the GLOD is still there, and there's a short at the other side of the mobo, can't remember the IC/capacitor. But I didn't try seeing if it is a NEC problem, since I don't believe it's a bga one. The other one, a "H" it has the HDMI IC already removed, and a short on that area, again, with GLOD.
My doubt is if those electrical issues had to do with the RSX not giving any kind of signal through the A/V. And the funny thing is that the other day I managed to get another "H" functioning by removing a couple of NECs, that were producing a YLOD (even with 8 tantalums on the other side). I removed them, and got GLOD, added tantalums, and got signal. So, I'm gonna be testing if those GLOD of mine are a NECs problem, since I couldn't do anything else.
About your CECHE, I have a few BCs and those run perfectly fine with ATX PSU wires, 2 per processor. I don't think it's only a matter of wire sizes, the working temperature in there is important too.
So i did some minor poking around on the A01 I got from a ebay a couple weeks back. I havent done a reflow yet, but i'll be doing that after I post this.
As I said before the outer shell appears to have been dropped and i thought maybe a component got knocked loose. I did try a new power supply but that didn't work.
I checked the fuses and they all have continuity so its definitely not a fuse issue.
Nothing on the board itself appears to have actually been damaged or knocked loose, just the shell itself which can thankfully be replaced.
When doing a proper reflow its best to use some actual flux correct?
I'm guessing all my previous readings were me not making proper contact with my alligator clips. While I wait for my new probes to arrive, I poke around until I'm fairly sure I'm making contact, and here's the only thing I can get out of it
Basically, it'll be flat, then it'll shoot up to about .7V, then it'll drop back down... ???
If I set any other lower ranges, all you see is the line shoot up to the top of the screen, and then back down. Configuring for 1v/div at 1.2s is the only time I can get anything "readable", and that's during boot time (you can hear the beep go off). Even when I set it to 250nS, all is flat until I boot. So @squeept (and sorry for pinging you so much over the last few days), at what time did you get those readings?
I know my O-Scope is a different cheap brand but to get my good readings from the tokins i just had it set at 1uS/Div and 50mV/Div as advised. Also make sure its set on AC and not DC as we want to see the pulses it sends, not DC current. @squeept assuming this is correct to?
superslims use tantulums right would it be possible for them to go bad i have a lot of burn marks on my MB PQX 001
Having something like glod to
So I did a reflow or what I assume is a reflow. Got my heat gun up to about 400c on both the CELL and RSX as well as the NEC/Tokins and still instant YLOD.
I tried swapping out power supplies thinking that was the issue, it was not.
Not sure what else to do to try and revive it at this point. All fuses checked out ok for continuity.
I'm not confident about simply replacing the caps and wasting time doing that if that doesn't work.
I checked over the motherboard, this makes it 3 times, and theres nothing I can find that seems off or destroyed.
So either I can give up and sell it to someone who will take it off my hands, or I may use this to practice delidding as i did find a possible tool that will make it easier with little to no risk of scratching the substrate.
At this point with this A01 thats been clearly dropped, it took the wind out of my sails
When u did the reflow were u able to lightly nudge the BGA's? if u can and it acts as if its floating then yes u at least got the solder to flow at melting point. If it didnt nudge and float then the reflow wouldn't have done anything. Takes about 7 minutes to reflow properly.
I highly doubt u were able to reflow both BGA's at the same time with a heat gun (unless u did them separately??) the area is too large to keep it at a reflow temp.
hello friend i forgot to tall what is 2nd best tantalum if i haven't find panasonic tantalum
Separate names with a comma.