PS3 Frankenstein PHAT PS3: CECHA with 40nm RSX

lcferrum said:
BTW, note squeept's answer regarding state of my TOKINs. He says that he saw less noisy TOKINs causing problems. I think this indirectly confirms that Sony lowered PWRGOOD signal thresholds for both CELL and RSX VRMs on refurbished mobos.
Click to expand...

probing methods can vary. if you didn't use a ground spring, for example, you'll get a lot of coupled noise. Squeept usually just clips onto the gnd plane with a ground clip, aka an antenna.

What we're really worried about is the ripple (not noise). So if you can still make out and measure the ripple (use the cursor manually) then it doesn't really matter. I like to set my scope to 1x for better resolution and noise rejection. And I use a ground spring for a low impedance probing method. It cuts out most of the coupled and common mode noise, so I know what remains is really there. And it makes trusting my scopes automatic measurements easier.

Basically, measuring error needs to be accounted for before drawing conclusions.
 
RIP-Felix said:
probing methods can vary. if you didn't use a ground spring, for example, you'll get a lot of coupled noise. Squeept usually just clips onto the gnd plane with a ground clip, aka an antenna.

What we're really worried about is the ripple (not noise). So if you can still make out and measure the ripple (use the cursor manually) then it doesn't really matter. I like to set my scope to 1x for better resolution and noise rejection. And I use a ground spring for a low impedance probing method. It cuts out most of the coupled and common mode noise, so I know what remains is really there. And it makes trusting my scopes automatic measurements easier.

Basically, measuring error needs to be accounted for before drawing conclusions.
Click to expand...
I always use ground spring. And I was talking about ripple, not noise, sorry for my mistake. I you go back to TOKINs thread, you'll see that my CELL voltage has this sinusoidal shape periodically, that squeept identified as the sign of bad TOKINs. And the PS3 still works fine.
 
Summer is over, the boat is winterized, toys going into storage, and for me it's back to tackling some of these tabled projects such as the PS3 mods.

Since the spring I have successfully brought back to life a dead Atari Lynx (plus a BennVenn screen swap) and dead PC Engine Duo (with Voultar RGB mod), installed Tim Worthington's NESRGB on a front loader, along with many other console mods and repairs. But these are all extremely simple in comparison to the PS3.

I also got married earlier this month, my wedding band is made of pure Tantalum.

As mentioned in my SYSCON diagnostics here, I have (4) phat PS3's:
  1. CECHH - Came back to life after a NEC/TOKIN replacement, but failed again shortly thereafter.
  2. CECHL - Ceramic white import that has never been tampered with, no plans to open this up.
  3. CECHB - Grabbed YLOD'd on FB Marketplace, NEC/TOKIN replacement and is still working to this day. Ticking time bomb? Probably.
  4. CECHA - Grabbed YLOD'd on FB Marketplace, RSX delidding fail, first candidate for Frankenstein

I have on-hand (2) NOS 40nm RSX for the CECHA and CECHB, assuming that the B will eventually fail.

I also ordered (8) Tantalizer v0.2b per PS3 for a total of (24) plus a few spares to rework the inappropriately high ESR capacitors I originally used, and get the PS3s back into proper spec. Along with this, I also ordered (24) of Panasonic 2R5TPE470M7 capacitors per PS3 for a total of (72) plus a few spares.

So my planned scope of work is:
  1. Slowly pre-heat the CECHA main board up to temperature in a convection toaster oven, one I use for powder coating since I don't want to put a board that potentially off-gases dangerous chemicals into my cooking oven. Any suggestions on pre-heat step procedure and final temperature?
  2. The main board will have been wrapped in multiple layers of foil with exception of the exposed RSX. Once removed from the oven, I will use a handheld hot air rework to flow the RSX BGA and remove it with a BGA suction pen.
  3. I will then clean the main board BGA, and use the hot air rework to flow the 40nm RSX. The NOS 40nm RSX already comes balled.
  4. Allow the board to cool, and then complete the board modifications and botakompong's IC modchip installation per Artificial's YouTube video.

I am currently in the process of purchasing qty (10) of botakompong's IC modchip from Indonesia. I am located in the USA, and will offer the spares for resale.

After reading through this thread, it sounds like the jury is out on whether PS2 functionality is reliably maintained. There are still many unknowns and lots of speculation?

If anyone has any suggestions, please let me know. I'm still at least several weeks out from attempting this.
 
Last edited:
MuNaught said:
Summer is over, the boat is winterized, toys going into storage, and for me it's back to tackling some of these tabled projects such as the PS3 mods.

Since the spring I have successfully brought back to life a dead Atari Lynx (plus a BennVenn screen swap) and dead PC Engine Duo (with Voultar RGB mod), installed Tim Worthington's NESRGB on a front loader, along with many other console mods and repairs. But these are all extremely simple in comparison to the PS3.

I also got married earlier this month, my wedding band is made of pure Tantalum.

As mentioned in my SYSCON diagnostics here, I have (4) phat PS3's:
  1. CECHH - Came back to life after a NEC/TOKIN replacement, but failed again shortly thereafter.
  2. CECHL - Ceramic white import that has never been tampered with, no plans to open this up.
  3. CECHB - Grabbed YLOD'd on FB Marketplace, NEC/TOKIN replacement and is still working to this day. Ticking time bomb? Probably.
  4. CECHA - Grabbed YLOD'd on FB Marketplace, RSX delidding fail, first candidate for Frankenstein

I have on-hand (2) NOS 40nm RSX for the CECHA and CECHB, assuming that the B will eventually fail.

I also ordered (8) Tantalizer v0.2b per PS3 for a total of (24) plus a few spares to rework the inappropriately high ESR capacitors I originally used, and get the PS3s back into proper spec. Along with this, I also ordered (24) of Panasonic 2R5TPE470M7 capacitors per PS3 for a total of (72) plus a few spares.

So my planned scope of work is:
  1. Slowly pre-heat the CECHA main board up to temperature in a convection toaster oven, one I use for powder coating since I don't want to put a board that potentially off-gases dangerous chemicals into my cooking oven. Any suggestions on pre-heat step procedure and final temperature?
  2. The main board will have been wrapped in multiple layers of foil with exception of the exposed RSX. Once removed from the oven, I will use a handheld hot air rework to flow the RSX BGA and remove it with a BGA suction pen.
  3. I will then clean the main board BGA, and use the hot air rework to flow the 40nm RSX. The NOS 40nm RSX already comes balled.
  4. Allow the board to cool, and then complete the board modifications and botakompong's IC modchip installation per Artificial's YouTube video.

I am currently in the process of purchasing qty (10) of botakompong's IC modchip from Indonesia. I am located in the USA, and will offer the spares for resale.

After reading through this thread, it sounds like the jury is out on whether PS2 functionality is reliably maintained. There are still many unknowns and lots of speculation?

If anyone has any suggestions, please let me know. I'm still at least several weeks out from attempting this.
Click to expand...
I would start with that H you have there. Get a thermocouple or multimeter with temperature reader and check the preheater max temps, and do the same with your "top heater", your heatgun. You'll probably have problems, that's why is better start with pcbs that aren't that relevant. Cover the tantalums or remove them before doing the rework, and always check temperatures. For removing the preheater should be on 250ºC and top heater 220ºC, more or less. It depends of your equipments and ambient temperature. Use amtech flux. I wouldn't trust in that RSX you bought though, is better to get it from dead consoles.

An example of what you're trying to do:


Btw, if you care that much about that A and B (my favourite), and money isn't a problem. Better make them reball by a good reworker. Then add a few piggyback good tantalums next to the NECs.
 
250°C seems high for the pre-heater, while 220°C seems low for the hot air rework. Are those values reversed?

Why should I not trust the 40nm RSX I bought? They are brand new, old stock, never before installed. Have others found these to be dead on arrival?

I was planning to use Kester 951 no-clean flux, which I've had great success with.

As mentioned, the NEC/Tokins are already ripped out and long gone, which is why I'm using RIP-Felix's Tantalizer v0.2b with (24) Panasonic 2R5TPE470M7 capacitors per PS3. I plan to install those after the installation of the 40nm RSX such that they're not exposed to the heat.

I don't mind spending extra money to have the CECHA or B reballed, but I don't know of anyone reliable and trustworthy who is still doing that work. Any suggestions?
 
Yeah, that would be the max temperature for the pre heater, and for the top heater I'd say is the average melting temperature, but like I said, it depends of your equipment, ambient temperature and pcb, in your case you could use little bit on the top, but remember that the best way of extracting a processor is to get it in perfect condition, same as the pcb. For that you need to get a balance between the heaters and the no lead solder, in order to get the melting point the lowest you can, and prevent pcb deformation by using good supports for the pcb. In my eyes the most important aspect here is the pre heater, since you'll be using it to stabilize the whole work temperature.

If you're insterested or need more advices I suggest you to ask @vyktormvmpay25 on this thread: https://www.psx-place.com/threads/reball-ps3-cell-rsx.32376/

He's one of the best reballers in the forum.

I don't know if is safe to buy "brand new" RSXs from anywhere, since Sony never sold them. Or am I wrong?
 
Good resistance to 40nm out of board 2,7~3.2 ohms in vddc line. Personally I desolder on any ps3 at exactly 232 Celsius reading TC right beside ic. If I go with slow and long enough profile it may be 225.
ElGris is right we use only rsx from other boards even if we see "brand new ". Most part from gaming consoles aren't sold. If you find from big companies datasheet for one part it means is really sold for open market.
 
Last edited:
If you have the same puhi preheated I do, then I set it to around 180-200C when flowing. Set top heater to 250C and you may have to increase it to get to reflow Temps as measured by thermocouple. It took about 270C last time I did it, but thats with the IHS still attached, which I recommend. It takes longer but provides support and radiates the heat evenly. I think it might also help weigh the balls down, increasing adhesion...IDK, just speculation.

Practice on junk boards! And beware itll take all day.

2 hours dry time at 100C, not a minute less!!!
 
MuNaught said:
Since the spring I have successfully brought back to life a dead Atari Lynx (plus a BennVenn screen swap) and dead PC Engine Duo (with Voultar RGB mod), installed Tim Worthington's NESRGB on a front loader, along with many other console mods and repairs. But these are all extremely simple in comparison to the PS3.
Click to expand...
I know what you mean. After all this PS3 stuff, an N64 Digital install feels like a piece of cake! And PixelFX says its an expert install. I guess that means I'm an expert now!:victorious: Nothing like reballing a PS3 makes you feel like a noob all over again. If N64D is for xperts I'd say the frankenstein mod is for masochists.

Off topic:
I have been debating getting a TurboGraphx (or however it's spelled, feel free to revoke my gamercard). Is it worth looking into? I never owned one, and besides Bonk I don't even know what the game library is like...lol.
 
Last edited:
Hello ppl,

How is going this? is it possible to replace the rsx without the "orbis chip"?

I have 2 CECHA01 (and a rework/reballing "pro" station) and I'd love to replace their 90nm RSX BGAs.

@sandungas regards from eol
 
Hola capitan @quark :D
The orbis chip is the easyest solution by now, it works as a man in the middle to fake the communications in between syscon and RSX

We was speculating about how works that communications but is still unknown... lets say... the orbis chips knows about it (maybe not completly but at least partially because is exploiting them)... but we dont know what does the orbis chips exactly
We was also thinking that eventually we could find some way to "patch" the syscon firmware to send the correct identifyer for the new RSX but this is just a theory that requires some kind of "exploit" in syscon security... and we are far away to achieve this (not even to mention we dont really know what to patch because we dont understand the communications)
Are like 2 roads that worths to be investigated, but as far i know the research is freezed
 
I was wondering if @M4j0r has time and willing to explore one of these Frankenstein unit with modchip I should do one and send it to him. I have one cok002 as perfect candidate, just have to wait for one 40nm board /rsx . If there were better boards here was about to do it earlier this summer.
That one I have it here have dodgy WiFi port and if it needs to be opened may create problems with WiFi connection. I rather do another for developing and exploring.
 
Last edited:
sandungas said:
We was also thinking that eventually we could find some way to "patch" the syscon firmware to send the correct identifyer for the new RSX but this is just a theory that requires some kind of "exploit" in syscon security... and we are far away to achieve this (not even to mention we dont really know what to patch because we dont understand the communications)
Click to expand...

I know how it works but we simply can't patch it into the retail syscons using the available space we have.
It has three steps:
1) Setup (to select the right table, algorithm and misc. stuff)
2) Table Data Patching (apply patches to the training data) - RSX specific
3) Execute the training algorithm - RSX specific

Problems: Not enough patch space to just overwrite everything or just too many offsets to patch.
The Problem is really that we don't have enough offsets to patch, the difference between the algorithms isn't that big but we simply lack the ability to patch more than like 2 things because we also need to overwrite the table data (which also needs to fit into the patch) and need about 2 patches for the setup.
Sony also produced two variants of the 40nm chip which get automatically detected by Syscon and use different tables/algorithms. So instead of adding support for one chip we need to add two :/ (https://www.psdevwiki.com/ps3/Talk:Rambus_Registers).

vyktormvmpay25 said:
I was wondering if @M4j0r has time and willing to explore one of these Frankenstein unit with modchip I should do one and send it to him. I have one cok002 as perfect candidate, just have to wait for one 40nm board /rsx . If there were better boards here was about to do it earlier this summer.
Click to expand...
Thanks, but I guess that won't help.
 
You must log in or register to reply here.

Similar threads

Back
Top