Monday I received the pacakge from Michel. Rampy A0 is so beautifull.




05/12/2013

First I did is to know if 4 power rail are giving correct voltage.

Here all perfect.

C252 - 1.26
C276 - 2.3v
C210 - 1.9v
C232 - 3.2v

It is very good sign.

I was very curious why the test socket was so of modded, and why the the 2 PINS from test socket was removed since they are used for resgistration/alignment the test socket make perfect contact on the pads.
I removed the test socket and my surprise is the test socket was not good designed or REv A0 or Rev A0 had faills in the design. For such reason they had to remove both pins and make new holes for screws.


Here you can see how they tried to tin the pads for getting  a major zone of contact. There are in each corner there are 1-2 pads not tinned, and in the test socket in same postion are pins are removed. So I guess they used such holes without pins for alignment and the golden pad helped via microscope.



Checking voltage values


Here what I found


In the A0 design for some reason there are not traces connect Power Rail 1.9V to pads marked in white color but it should be connected.

I need to continue making more tested but I guess why 3Dfx built a new PCB A1 with only one week of difference.

Note- white pads are connected between them and I could do wrapping and connect in rear socket the power Rail 1.9V but not sure if it is a worth and is better to let the card as is.

ADDED 23/12/2013

Finally I had some time free and I worked on Ramapage chip with damaged balls.

• Before to reemplace damaged balls. Zone problem upper right corner

• Once reemplaced damaged balls.

Chip looks very well. This time I found the manner to reemplace balls without damage the green solder mask.

About If chip will work now- I am not very optimist  since seems all the chips decapped with open die had problems.

Not at home. I need to test it.


ADDED 15/01/2014

• I have been working again on Rampage A0. Finally I found the way to add 1.9V to pads in white. I did a bridge between 2 pads and now all the white line pad is connected 1.9V POWER RAIL. Now what I need is work in the bridge connection for reducing the tickness since it is little high and It will not help to make perfect contact between pad and socket.

• Here boards engineers already detected the BLUE POINT with 3.3V was not connected and they added the power rail across wrapping copper.

• I also detected another problem on RN101. It is in short but at the moment It is not important since it only would affect image displayed and not wake up the card.

Main problem now is if I will be to able all the pads make perfect contact with test socket. It will be hard. Checking the pads the boards engineer expanded the pads with tin  for trying to get a big zone of contact. Even so I have detected many pads are not in touch with test pin socket.


- Oscar.

trevormacro wrote on 05.12.13 at 15:15:26:

Your are crazy Michel!!! LoL

Well, let me I do some more test in PADS. There are more more pads need to get power rail. If rest of pads are correct and the only problem is present in WHITE PADS marked. It would explain because chip does not get HOT once motherboard is powering up. Possibly if I attach a wired to 1.9V power rail and connect it in only one VIA PAD in back of the card. Card could boot up. Is is my guess, but not sure, since I think 3Dfx engineer staff had done the same or simply they decided stop to work in A0 and go for new design A1.

I keep you informed.

- Oscar.

Someday, when I become really rich and buy back 3dfx corporation, I will invite Oscar to work with me 

If everyone who would want it would get a job at 3dfx it would be one of the biggest companies in notime. LOL.

Not saying everyone could ofcourse, what you guys do with those cards...I can't even dream of.

Maybe I'd do good in sales, or ict support.

osckhar wrote on 23.12.13 at 22:18:31:


Good chip decaped rework without remove the green mask.
Thank you

Oscar, can you put the Rampage chip under a scanner? Or take other die-photo's? I would love to see a close up of the chip, just like they did on CPU-World with CPU 'dies'.

Quote:


Hi Thandor,

Not stuff and not idea how I can take pics from Open Die with good quality.

Any tips?

- Oscar.

osckhar wrote on 26.12.13 at 16:43:00:
I don't have experience with open die photography but using a microscope will probably do the trick .

trevormacro wrote on 26.12.13 at 21:50:35:
Looks pretty good! It does not give away all the individual traces in the chip but the general idea is visible . Thanks!