NoahHatePolitics2000
Forum Noob
alright thanks
Shawn Shakir
Member
I believe you will need to use a copper shim between the RSX die and the IHS or it wont make proper contact. Otherwise, it should work fine.
RIP-Felix
Senior Member
The IHS from a 90nm will work fine. You'll need the added height to mate with the HS from the BC model. However, I have been told that the SS RSX is slighly thinner than the 40nm RSX from a slim. Meaning you may need a copper shim to get them to mate properly.
The training will be possible and the rest of the process is the same. So that not an issue. You may just need to shim the processor to get it to make solid contact. Keep it in mind, but it's not a deal breaker.
The training will be possible and the rest of the process is the same. So that not an issue. You may just need to shim the processor to get it to make solid contact. Keep it in mind, but it's not a deal breaker.
RIP-Felix
Senior Member
Sniper!
NoahHatePolitics2000
Forum Noob
I've never de lidded before and more concerned about me screwing that part up than the re-balling part even. Do you have any advice for doing that and adding the lid to the 40? Also is there any place I could purchase a copper shim for my RSX?
Shawn Shakir
Member
You can get a bunch of them from Amazon. Ive been told razer blade in the corners when rsx is off board and it should pop off. Haven't gotten around to trying this method myself. I was doing the pry after reballing it on board, but broke ram chips on my last 40nm reball, so won't be doing that no more!
NoahHatePolitics2000
Forum Noob
Well, as i'm really not that experienced with de-lidding and re-lidding is there a special kind of paste that is used to put the lid back on since the lids are usually wanted to kind of firmly stick on top? Also would it be reccomended to remove the cell microprocessor heat spreader to repaste it's connection with the heat spread and if so are there any good methods to do that without damaging the cell?
Shawn Shakir
Member
Tried razor blade on 40nm RSX today, slighly slide in on each corner between RAM chips and she comes off easily. I don't use any paste to glue it back down, just thermal paste on each ram chip + die and haven't had any issues thus far. For cell i use a painters knife with board at 100C, screwed down to a JIG, but everyone has their own method when it comes to delidding in general. I'd test on a few dead boards to figure out what works best for you
redcyclone
Member
Hello, I'm about to try this modification on a CECHA01. I've already removed the 90nm RSX successfully, but now I have some questions:
- I accidentally ripped off 3 pads from the board while cleaning the old solder, but I saw on the schematic that these pads are not connected to anything on the board, so it's OK. My question is: is it safe to have the solder ball for this missing pad on the RSX even though there won't be the pad for it to solder to? Wouldn't this possibly cause a bridge to adjacent balls?
- How exactly should I align the RSX? I know there are the four squares on the corners for reference, but should they be covered completely by the chip? I'm really afraid of misaligning it and having to do it all over again
- Is it safe to solder the RSX without delidding it? I'm worried that the weight of the IHS might press it when soldering and cause bridges
- Also, am I susceptible to the VRAM solder "bleed" when soldering it? The balls are leaded, not lead-free, and the replacement RSX is 65nm. I'll bake the chip a bit before trying to solder it to remove moisture
iComputer7
Forum Noob
Hi,
Just finished a 40nm swap on a CECHA and I'm getting an instant YLOD with error code A0093004. I have done the resistor mod (remove R2054, remove R2153 then scrape mask near R2054 and solder diagonally, remove R2001, leave R2002 alone) and have written the proper training data into the syscon.
This console booted before I swapped the GPU and was able to make it into safe mode. The display had severe artifacts and trying to boot the system and complete the initial setup caused it to freeze after the time setting screen.
I am certain that I did not knock off any passives in the process and did not touch anything other than the GPU and the syscon's test points. I have measured resistance to ground on a few points and they all checked out. I forgot to write them down so I'll have to check again and update my post. The actual solder on the GPU fully melted and I can see (with a magnifying glass) that the balls along the outer perimeter have attached to the motherboard.
I feel like this has happened to multiple boards that I've done BGA work to. Are there any components that can't survive this process? I did need to go quite hot, up to 290C on the top with hot air and my T-8280 board preheater set to 265C. No popcorning seems to have occurred and I didn't hear any pops or crackles while heating it.
Just finished a 40nm swap on a CECHA and I'm getting an instant YLOD with error code A0093004. I have done the resistor mod (remove R2054, remove R2153 then scrape mask near R2054 and solder diagonally, remove R2001, leave R2002 alone) and have written the proper training data into the syscon.
This console booted before I swapped the GPU and was able to make it into safe mode. The display had severe artifacts and trying to boot the system and complete the initial setup caused it to freeze after the time setting screen.
I am certain that I did not knock off any passives in the process and did not touch anything other than the GPU and the syscon's test points. I have measured resistance to ground on a few points and they all checked out. I forgot to write them down so I'll have to check again and update my post. The actual solder on the GPU fully melted and I can see (with a magnifying glass) that the balls along the outer perimeter have attached to the motherboard.
I feel like this has happened to multiple boards that I've done BGA work to. Are there any components that can't survive this process? I did need to go quite hot, up to 290C on the top with hot air and my T-8280 board preheater set to 265C. No popcorning seems to have occurred and I didn't hear any pops or crackles while heating it.
db260179
Member
A0093004 = RSX_POW_FAIL poweroff state (Potential NEC tokins issue and VCC)
Have you replaced the nec tokins?, its possible there is a short on the VCC line?
From experience, especially underneath of the board the smc,caps etc will just die, short or pop off - i now cover them with aluminum tape to keep them alive.
db260179
Member
Make sure the four edges have their smd resistors still on, otherwise your RSX will just sink and merge its balls. They are very tiny, but keep the rsx upright.
To keep bleeding away, just pre bake the board for a few hours to get the moisture out - i do this all of the time as havent had popping since.
Alignment is tricky, just tilt the RSX on its side and align up as best you can - usually before i remove an RSX chip i mark the corners so that the new one can be placed a bit better.
redcyclone
Member
I didn't even know the SMDs were there before, and it looks like I lost them. Only their pads are there. Could I replace them maybe with cellphone caps? Do you know what size they might be?
Thanks
iComputer7
Forum Noob
I haven't replaced the tokins. Last I checked before the first powerup, RSX_VDDC had about 4 to 6 ohms to ground. I haven't noticed any missing SMDs but I need to check with my microscope to make sure. It's entirely possible that's my issue.
db260179
Member
OK, if you haven't replaced the nec tokins, if you have an oscillscope you can see if there a ripple - rip felix has done some great analysis on this and pics - its very noticable when the tokins are below spec - i did a board recently, tokins read 989uf, and on my oscilliscope showed the effect of bad filtering - comparing before and after change.
So before removing them make a note of the current 20s output from the oscilliscope. Then compare with new caps.
If you dont have a oscilliscope, i highly recomend you get one, they are cheap now and can get one for 90$ or less.
Last edited:
db260179
Member
Its quick to lose them when taking off the original RSX, if you still have the old rsx chip just take the ihs off and you will see some tiny caps/resistors on the chip use those, its just to rest the rsx on the four corners.
Example - https://www.psdevwiki.com/ps3/File:RSX_90nm_without_IHS.jpg
iComputer7
Forum Noob
I took some screenshots on my oscilloscope. It doesn't look like bad filtering to me but I can't be sure. Probe was on RSX_VDDC.
Also I took a look at the board under my microscope and found a couple passives (R2131 and R2058) by the RSX that are missing. With this board it's difficult to tell if something was not populated from the factory or if I knocked it off. R2131 seems important so I'll replace it. R2058 is marked as XX on the schematic so I think that means it wasn't there from the factory?
db260179
Member
Interesting, read this page - https://www.psx-place.com/threads/f...nd-error-reporting.30100/page-101#post-316228
And check if the input drive pins are correct - checking expected voltages.
As its a runtime error, it could be the caps are degraded and not holding a charge long enough. I would just change them then go from there.
iComputer7
Forum Noob
As much as I don't want to do it, it's worth a try. Gonna order some PS3 tantalizers since I'm already getting some stuff made by OSH Park.
db260179
Member
It can be a daunting process, but just take your time - rewards will come later.
If you can use a preheater makes getting the tokins off easier.
From experience fixing ps3 boards, the heating process is important and not to rush it - pre baking it will help alot.
Good luck!
Similar threads
-
-
PS3 PS3 CECHA RSX Swap (40nm) – JP Limited – SYSCON Patch, VDDC Mod – Need Detailed Help- Started by jw0710
- Replies: 4
-
-
