PS3 Frankenstein PHAT PS3: CECHA with 40nm RSX

UIMD said:
Thanks for reply.
I check resistance on R2054. Without any resistors - really 100 ohm.

But about R2153 is not yet clear. It seems to belong to the /RS_TRSTNEG signal. On DIA-002 it has been removed. I think after checking the RSX (soldering) it's worth trying with and without it.

As i see, R2153 removed\not soldered. But do I need to do this...?
**DIA-002 comes wqith 65 nm RSX from factory, as I remember.

In BE, I'm pretty sure the soldering there is good.
Since the console started up, after a spot warm-up of the RSX 90 nm(warmed ic die).
Therefore, I will first check the soldering of the RSX that I soldered. Perhaps I really made a mistake somewhere, or did not notice something. (e.g. bad pads on PCB)
I see three scenarios:
1. BGA problem (RSX)
2. BGA Problem (BE), although I don't want to believe it.
3. Dead RSX, this can also be.

*Additionally, I'll look at the resistances on the rest of the RSX supply lines, as far as it's realistic without a schematic. (for SEM-001 schematic I have, it seems, but not 100%)
Click to expand...

They both work without RS_TRSTNEG and 100 ohms for cg reset.
3034 is a data line miscommunication between cell rsx. Reball again rsx should work if you are confident cell is good soldered. If board look straight /flat then you will fix it.
Oh and if you reball again rsx and get again same 3034 then cell was all this time your problem.
I somehow got very confident when I reball one ic and if rsx did not work, then will cell reball always working when this type of miscommunication errors happen.
 
Last edited:
vyktormvmpay25 said:
They both work without RS_TRSTNEG and 100 ohms for cg reset.
3034 is a data line miscommunication between cell rsx. Reball again rsx should work if you are confident cell is good soldered. If board look straight /flat then you will fix it.
Click to expand...

Vic do you think there is a way to straighten the board, like maybe soaking the board in deionised water, then screwing the board down to a PS3 bga jig, then reflowing the whole board (attached to the jig) in an oven at 85c - 90c for 48 hours?
 
Workz_777 said:
Vic do you think there is a way to straighten the board, like maybe soaking the board in deionised water, then screwing the board down to a PS3 bga jig, then reflowing the whole board (attached to the jig) in an oven at 85c - 90c for 48 hours?
Click to expand...
Yes if both ic are out of board cell and rsx. Or you don't have to get it in to water and dry as board naturally becomes flat on jig when both out. But if deviation is to big and cannot be corrected by desolder both then I don't lose time with it if it doesn't naturally go flat to each spacers on jig.
 
vyktormvmpay25 said:
Yes if both ic are out of board cell and rsx. Or you don't have to get it in to water and dry as board naturally becomes flat on jig when both out. But if deviation is to big and cannot be corrected by desolder both then I don't lose time with it if it doesn't naturally go flat to each spacers on jig.
Click to expand...

Thanks Vic, most appreciated :encouragement:
 
So how did the story end:
I re-soldered the RSX and it seems that the problem was solved.
I removed the chip and immediately noticed two bad contacts. Just in the direction of BE.

Strange, another RSX was soldered on the first try.

So I didn't check well. What can I say, it happens.
Cell BE was soldered really well. Didn't have to do anything with it.

vyktormvmpay25, thanks for your assistance, it really helped and supported. Board saved. :)
To put RSX 65 nm in DIA-001, you need to remove R2153 and patch syscon.
And that's it. DIA-002 is ready. Bon appetit. =)

P.S. In fact, I was lucky that the RSX chip from the donor board was fully functional. Usually it's exactly the opposite.
And then, perhaps, an exciting adventure awaits me to repair the superslim version (NPX-001, 4000 series) with a crooked board and a darkened CELL. But that is another story.
 
Last edited:
that would be interesting. but keep in mind cell is maried to syscon bdrom and nand-nor the bd can be remaried through cfw. syscon can have the eeprom patched. but 45nm cell is tied to nor never seen a 45nm cpu and be nand based bc consoles are nand based nor and nand are not pin compatible. and i dont know if the nor can be readed by external programmer and if any strange modification needed to be done in software before flashed back to nand. long story short too much hastle. and cpu rarely dies
 
In addition to this problem, we have two other predominant factors:

the Syscon, which in the COK 00X models is soldered by BGA. Example https://www.psdevwiki.com/ps3/CXR713120-201GB

On 45nm models and most 65nm models, Syscon SMD is used, and CELL BE 45nm models are only available on DYN 001 models. Example: https://www.psdevwiki.com/ps3/SW2-301 .

And as @sandungas already explained a while ago (if I'm not mistaken it was in this topic) Syscon cannot be fully re-recorded, we can't do a Remarry as is done with Renesas.

And we also remember that BC models are NAND, two 128mb chips each, keeping all the system data on them, while in the other models it is a 16mb NOR chip, leaving only vital data on the NOR and the rest on the HDD. Be circumvented by means of software to create a manager that allows you to join the data inside the NOR along with the other remaining data that is not on the HDD, but I think that maybe it's something like this difficult to replicate.

in addition to all these problems, we have the cost that would be extremely high and risky, almost not worth it.
 
_XBrD4shDOE_808_ said:
In addition to this problem, we have two other predominant factors:

the Syscon, which in the COK 00X models is soldered by BGA. Example https://www.psdevwiki.com/ps3/CXR713120-201GB

On 45nm models and most 65nm models, Syscon SMD is used, and CELL BE 45nm models are only available on DYN 001 models. Example: https://www.psdevwiki.com/ps3/SW2-301 .

And as @sandungas already explained a while ago (if I'm not mistaken it was in this topic) Syscon cannot be fully re-recorded, we can't do a Remarry as is done with Renesas.

And we also remember that BC models are NAND, two 128mb chips each, keeping all the system data on them, while in the other models it is a 16mb NOR chip, leaving only vital data on the NOR and the rest on the HDD. Be circumvented by means of software to create a manager that allows you to join the data inside the NOR along with the other remaining data that is not on the HDD, but I think that maybe it's something like this difficult to replicate.

in addition to all these problems, we have the cost that would be extremely high and risky, almost not worth it.
Click to expand...
In some way the syscon is ike the BIOS of a PC motherboard, the "base" syscon firmware contains software support for a specific list of component revisions.. and thats the part of the syscon EEPROM that is not upgradeable
The 45nm CELL was introduced after the sherwood syscons (with pins), i dont remember if there is some page in wiki where could be seen this, but i guess mullion syscons (with BGA) doesnt supports the 45nm CELL because is newer
Maybe m4j0r could cook one of that black magical patches to achieve it, but there are other 2 problems

As mentioned above, the flash contains a couple of regions (bootldr and metldr) that are tied to CELL, because are encrypted with a key unque for each CELL unit... we cant rebuild that files, so to do a CELL "transplant" is needed to take that files with it to the new flash

And the last and biggest problem is the BGA pad layout for CELL is different as far i remember... it seems with RSX they was doing an effort to keep the same BGA pad layout as most time as posible (and they almost completed the whole PS3 familly with the same RSX layout, the only exception is when the RSX was reduced to 28nm), so there are only 2 RSX pad layouts in total
But with CELL there are several pad layouts (3 or 4, dont remember right now, i wrote an accurate list some weeks ago in this same thread as far i remember)
 
sandungas said:
In some way the syscon is ike the BIOS of a PC motherboard, the "base" syscon firmware contains software support for a specific list of component revisions.. and thats the part of the syscon EEPROM that is not upgradeable
The 45nm CELL was introduced after the sherwood syscons (with pins), i dont remember if there is some page in wiki where could be seen this, but i guess mullion syscons (with BGA) doesnt supports the 45nm CELL because is newer
Maybe m4j0r could cook one of that black magical patches to achieve it, but there are other 2 problems

As mentioned above, the flash contains a couple of regions (bootldr and metldr) that are tied to CELL, because are encrypted with a key unque for each CELL unit... we cant rebuild that files, so to do a CELL "transplant" is needed to take that files with it to the new flash

And the last and biggest problem is the BGA pad layout for CELL is different as far i remember... it seems with RSX they was doing an effort to keep the same BGA pad layout as most time as posible (and they almost completed the whole PS3 familly with the same RSX layout, the only exception is when the RSX was reduced to 28nm), so there are only 2 RSX pad layouts in total
But with CELL there are several pad layouts (3 or 4, dont remember right now, i wrote an accurate list some weeks ago in this same thread as far i remember)
Click to expand...
maybe there is a way for this to work, but it is very specific and we would need dev's to test it.

the only PS3 model with NAND and with CELL different from 90nm are the CECHG, which have a 65nm CELL, are NAND, use BGA Syscon (as far as I remember) and use the same Renesas as the BC.

if we are lucky enough that the pads are the same, maybe a PS3 CECHA01 Frankenstein 2.0 with CELL 65nm and RSX 40nm is possible, but it would take several patches to recognize the firmwares and other problems, I think @M4j0r can help us with that ;)


in addition, CECHG are models that, for the most part, are found in Europe, since most of the CECHG's sold were from region 04, and region 01 was the CECHH, at least as far as I know (I only see CECHG04 here in Brazil, it rarely appears a CECHG01).
 
I found my post with the list of CELL pad layouts
sandungas said:
...the best way to see how different they are just with text is if i tell the number of pads missing at the hole at center this way, the hole are the numbers inside the brackets... im substracting them to calculate the total number of pads

-CELL 90nm 41x41 - (19x19) = 1681 - 361 = 1320 pads
-CELL 65nm 41x41 - (15x17) = 1681 - 255 = 1426 pads
-CELL 45nm 41x41 - (17x17) = 1681 - 289 = 1392 pads
-CELL 45nm 42x42 - (22x18) = 1764 - 396 = 1368 pads

The motherboards that was designed for a 90nm CELL (like the COK-001) only have 1320 pads
I dont know if the pinouts of the pads at the surrounding matches in between the ones using a 41x41 array, but are physically incompatibles because the "hole" at the center
Click to expand...
 
sandungas said:
I found my post with the list of CELL pad layouts
Click to expand...
so frankenstein 2.0 project went down the drain without even getting off the ground:rolleyes: but I figured they weren't compatible, maybe in the number of pads or what each pad does.

a doubt, why companies choose to use SMD chips instead of BGA, as cocrreu with Syncon? decrease mechanical work, cost savings?
I don't know if I'll know the answer either, no problem, of course, it's just out of curiosity:D
 
sandungas said:
I found my post with the list of CELL pad layouts
Click to expand...
Mate Well that Not Meant by Any way that they are not pin compatible, the 72 Pins Cell B.E 45nm hás more than the 90nm model could bê only for Support Soldering purposes or maybe Even Things that are used only on the Ps3 Slim Board, the Cok 001 would Not need those pins
 
Victor Hugo Alvarez said:
Mate Well that Not Meant by Any way that they are not pin compatible, the 72 Pins Cell B.E 45nm hás more than the 90nm model could bê only for Support Soldering purposes or maybe Even Things that are used only on the Ps3 Slim Board, the Cok 001 would Not need those pins
Click to expand...
You can even ignore the fact that the pads are different, but we still have the other components that must be replaced together with the CELL BE, Syscon, flash chip and Renesas, which are different.
for a possible exchange of a 90nm CPU for a smaller lithograph, just for a 65nm one and a CECHG, this ignoring the incompatibility of pads, not only from not being together at the time of soldering, but from, for example, pad 27B on the MB provides 1.2v voltage, on the CPU pad 27B is a GND pad or 1.5v pad or who knows about data transfer, besides that SEM 001 boards are relatively difficult to find even more with the 'CPU kit' working, and changing to a 45nm one is almost impossible, it would take a lot of work, it was tried to replace the RSX a few times before this case was reported, but it was not successful because it was not recognized, now imagine this with four different chips and that are the main chips on the motherboard, it is very unlikely that it is right.
 
_XBrD4shDOE_808_ said:
maybe there is a way for this to work, but it is very specific and we would need dev's to test it.

the only PS3 model with NAND and with CELL different from 90nm are the CECHG, which have a 65nm CELL, are NAND, use BGA Syscon (as far as I remember) and use the same Renesas as the BC.

if we are lucky enough that the pads are the same, maybe a PS3 CECHA01 Frankenstein 2.0 with CELL 65nm and RSX 40nm is possible, but it would take several patches to recognize the firmwares and other problems, I think @M4j0r can help us with that ;)


in addition, CECHG are models that, for the most part, are found in Europe, since most of the CECHG's sold were from region 04, and region 01 was the CECHH, at least as far as I know (I only see CECHG04 here in Brazil, it rarely appears a CECHG01).
Click to expand...
the Cell B.E 65nm is a too Hoto being working at 80c Just in the xmb from the Models Cech L or H, the 45nm is much better ideia to Install on the ps3s fat
 
_XBrD4shDOE_808_ said:
You can even ignore the fact that the pads are different, but we still have the other components that must be replaced together with the CELL BE, Syscon, flash chip and Renesas, which are different.
for a possible exchange of a 90nm CPU for a smaller lithograph, just for a 65nm one and a CECHG, this ignoring the incompatibility of pads, not only from not being together at the time of soldering, but from, for example, pad 27B on the MB provides 1.2v voltage, on the CPU pad 27B is a GND pad or 1.5v pad or who knows about data transfer, besides that SEM 001 boards are relatively difficult to find even more with the 'CPU kit' working, and changing to a 45nm one is almost impossible, it would take a lot of work, it was tried to replace the RSX a few times before this case was reported, but it was not successful because it was not recognized, now imagine this with four different chips and that are the main chips on the motherboard, it is very unlikely that it is right.
Click to expand...
OOHHHHHHHH OK so the voltages are opose You sure? ? IF voltages are like that then its game over for Cell Frank
 
Last edited by a moderator:
vyktormvmpay25 said:
What? How did you find out? Can you please share with us!
Click to expand...
Hello im Sorry I Just mean that I know that the Cell B.e 45nm has a few more pads but that dont mean they're not pin compatible, but mate I think May be better we really forget about Cell Swap or People will Start Destroing 90nm Cell just to Replace with the 45nm one, if we want the 45nm Cell we can just Install an Motherboard of an super Slim inside an Cecha00 Case só we Will have an Fat PS3 with 45nm Cell and even 28Nm rsx, the motherboard just sits perfectly inside and we could Install massive heatsinks from the fat 360 aluminium cpu heatsink on its Cell and Rsx and for me that modding is looking supee epic right now, the Cok-001 has nothing special beyond the Rtx-001, Neither Special neither I want nobody damaging Ps3 Cell cpus neither it's 90, 65 or 45, also all PS3s can play PS2 and via Emulation is Só much better cause via Emulation it Utilizes 7 Spes and the 2 PPE Threads to Emulate the ps2 Full Hardware all acording to Dev Wiki Ps3 and ps2 Emulation is Super computacional intensive, if anyone could unlock the Ps2 Net Emu Locked Frame rate we could Finaly see 120 Fps and beyond by Ps3 Cell and Finaly showing all the Ps3 Cell Processor Power
 
You must log in or register to reply here.

Similar threads

Back
Top