Welcome, Guest. Please Login 3dfx Archive
 
  HomeHelpSearchLogin  
 
Pages: 1 ... 3 4 5 6 7 ... 23
Send Topic Print
3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700 (Read 11977 times)
trevormacro
God Member
*****
Offline


Voodoo5k6 Gold SE, Legend
& Mystical Edition

Posts: 622
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #60 - 26.06.13 at 19:10:15
 
According to 3dfx engineer Oscar need to put a sheet of copper between the rearside of the cooler and the rampage chip to make a very good pressure. All the pins of the chip will be in good contact with the golds contacts of the test socket.

nb. same idea as Loeschzwerg some days ago!

EDIT: But about the tickness size 0,5mm 1mm? I don't know.

...

I hope it will be the solution for the card... Sad

According to Oscar, the chip on my rampy now get hot while before no.
Possibly because the clamp helped and now test socket makes much better contact on the pcb.


We have a better contact on pcb now let's try to have a better contact on the other side with the chip.

@Loeschzwerg:

According to 3dfx engineer, if you look much closer, trace is fixed. No damage here.

...

Better RN112's pictures:

...

...
Back to top
« Last Edit: 26.06.13 at 20:15:34 by trevormacro »  
 
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #61 - 26.06.13 at 20:20:14
 
trevormacro wrote on 26.06.13 at 19:10:15:
According to 3dfx engineer Oscar need to put a sheet of copper between the rearside of the cooler and the rampage chip to make a very good pressure. All the pins of the chip will be in good contact with the golds contacts of the test socket.

nb. same idea as Loeschzwerg some days ago!


Yep, I hope this will help.

trevormacro wrote on 26.06.13 at 19:10:15:
@Loeschzwerg:

According to 3dfx engineer, if you look much closer, trace is fixed. No damage here.


Ok, thanks for the better picture. I also checked the AGP pinout for Pin A59 -> "Reserved"

According to the AGP 1.0 spec it does have no function but within AGP 2.0/3.0 it is used for an address strobe signal. This signal is used for clock synchronization between transmitter and receiver (board and card). The strobe signal is only used on speed setting 2x, 4x and 8x within spec 2.0/3.0.

So it shouldn't prevent the card from working even if trace would be damaged.

Edit: About the copper sheets... I would go for normal "copper tape". If one layer isn't enough you can add several more layers and find the right thickness by yourself. => something like that for example: http://www.amazon.com/Copper-Conductive-Adhesive-Thick-Length/dp/B009KB86BU/ref=...
Back to top
« Last Edit: 26.06.13 at 20:31:52 by Loeschzwerg »  
 
IP Logged
 
trevormacro
God Member
*****
Offline


Voodoo5k6 Gold SE, Legend
& Mystical Edition

Posts: 622
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #62 - 26.06.13 at 20:33:29
 
@Loeschzwerg:

Thanks about the agp 1.0 informations.

If Oscar doesn't have copper sheet he can use aluminium sheet (perhaps it's more easy to find in the kitchen)?

It can be really agressive for the chip or not dangerous?
Back to top
« Last Edit: 26.06.13 at 20:35:33 by trevormacro »  
 
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #63 - 26.06.13 at 20:53:56
 
Aluminum sheets or tape is fine also, but I wouldn't go for aluminum foil from the kitchen. It is very very thin and you will need more layers of it (and folding of aluminum foil sucks).

http://www.amazon.com/Duck-240225-1-88-Inch-50-Yard-Aluminum/dp/B0000DI82J/ref=s...

Just a few dollars like the copper tape too.
Back to top
 
 
IP Logged
 
Dexter
YaBB Newbies
*
Offline


If it ain't broken, why
fix it ?

Posts: 39
Romania
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #64 - 27.06.13 at 08:38:11
 
I've read all 5 pages and all I can say is Respect !
There's a noble 3dfx fan right there.

Now I don't want to barge in rudely on this sensitive electronical issue, but has anyone ever thought of the oven trick? Solely for the card, without the chip.

Can't it repair eventual "cold solderings" which are possible ?
Or is it too risky ?

...just saying.
Back to top
« Last Edit: 27.06.13 at 09:14:02 by Dexter »  

3Dfx. Proud to grow up playing with the pioneers.
georgeleus  
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #65 - 27.06.13 at 09:46:42
 
The oven trick mainly works for soldered BGA components. Beside this it can fix the connection from DIE to substrate.
Both isn't the case when it comes to the Rampage card.

So no, I wouldn't recommend this trick.
Back to top
 
 
IP Logged
 
gdonovan
YaBB Administrator
*****
Offline


Infernal Collector

Posts: 2361
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #66 - 27.06.13 at 11:07:00
 
Loeschzwerg wrote on 27.06.13 at 09:46:42:
The oven trick mainly works for soldered BGA components. Beside this it can fix the connection from DIE to substrate.
Both isn't the case when it comes to the Rampage card.

So no, I wouldn't recommend this trick.


I'm surprised that no one has suggested doing away with the socket yet. Just attach the chip to the board and be done with it.
Back to top
 

Rampage is alive!
 
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #67 - 27.06.13 at 11:14:00
 
If the "copper sheet"-trick doesn't work, soldering the Rampage to the board should be considered, yes Smiley

Back to top
 
 
IP Logged
 
osckhar
God Member
*****
Offline


www.3Dfx.es

Posts: 812
SPAIN
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #68 - 27.06.13 at 11:42:56
 
gdonovan wrote on 27.06.13 at 11:07:00:
Loeschzwerg wrote on 27.06.13 at 09:46:42:
The oven trick mainly works for soldered BGA components. Beside this it can fix the connection from DIE to substrate.
Both isn't the case when it comes to the Rampage card.

So no, I wouldn't recommend this trick.


I'm surprised that no one has suggested doing away with the socket yet. Just attach the chip to the board and be done with it.


LoL, yesterday I recommended it to Michel soldering the chip onto the card. He must think on that.

Regards,
Oscar.
Back to top
 
WWW  
IP Logged
 
trevormacro
God Member
*****
Offline


Voodoo5k6 Gold SE, Legend
& Mystical Edition

Posts: 622
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #69 - 27.06.13 at 12:01:16
 
@Loeschzwerg:

Oscar tried with double paper sheet. He already had problems to close the heatsink but card continues without working. Even used the clamp but the card does not want to boot. With double paper sheet heatsink is doing a lot of pression and even balls are more deteriorated. Did you think it's enough or he need to buy copper sheet?

Check rampy chip picture:

...

It looks bad but it works perfect though. The pression is deteriorating some balls.

@osckhar:

If you put the chip on the rampy and after that the card doesn't work, is it possible to come back to original state without damaged both parts (balls traces on pcb and pins balls for the chip)?

Soldering the chip will be done by you?
Back to top
« Last Edit: 27.06.13 at 12:13:13 by trevormacro »  
 
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #70 - 27.06.13 at 12:13:33
 
trevormacro wrote on 27.06.13 at 12:01:16:
If you put the chip on the rampy and after that the card doesn't work, is it possible to come back to original state without damaged both parts (balls traces on pcb and pins balls for the chip)?


It is not 100% possible to return to the original state. After soldering the chip to the PCB, the solder spots one the PCB are tinned forever. This mean if you unsolder the chip and mount back the socket it won't have contact to the gold surface it now has, but to a tinned surface => the conductivity isn't that good compared to gold.

Edit: Yes, I think the pressure was enough then Sad

Edit2: Btw. I would rework the missing solder balls on your 2nd Rampage chip. It mainly works cause the traces lead to other solder balls with the same trace too. But it might cause loading problem (ground area and voltage supply, as well as signaling).
Back to top
« Last Edit: 27.06.13 at 12:26:44 by Loeschzwerg »  
 
IP Logged
 
trevormacro
God Member
*****
Offline


Voodoo5k6 Gold SE, Legend
& Mystical Edition

Posts: 622
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #71 - 27.06.13 at 12:24:29
 
Loeschzwerg wrote on 27.06.13 at 12:13:33:
trevormacro wrote on 27.06.13 at 12:01:16:
If you put the chip on the rampy and after that the card doesn't work, is it possible to come back to original state without damaged both parts (balls traces on pcb and pins balls for the chip)?


It is not 100% possible to return to the original state. After soldering the chip to the PCB, the solder spots one the PCB are tinned forever. This mean if you unsolder the chip and mount back the socket it won't have contact to the gold surface it now has, but to a tinned surface => the conductivity isn't that good compared to gold.

Edit: Yes, I think the pressure was enough then Sad


When Oscar tried my test socket and my rampage chip, it worked on his card but on my card no.
Perhaps the problem come from another part?
Did you think it will be a good idea to solder the chip on my card?
Back to top
 
 
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #72 - 27.06.13 at 12:29:47
 
trevormacro wrote on 27.06.13 at 12:24:29:
When Oscar tried my test socket and my rampage chip, it worked on his card but on my card no.
Perhaps the problem come from another part?
Did you think it will be a good idea to solder the chip on my card?


Well it's still possible the socket doesn't have a good contact to the PCB. You mentioned the PCB is warped a little bit, right? Soldering might help here, but we can't say for sure. Still a problem on other components possible (vga bios, smd components, voltage supply).
Back to top
 
 
IP Logged
 
osckhar
God Member
*****
Offline


www.3Dfx.es

Posts: 812
SPAIN
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #73 - 27.06.13 at 12:31:23
 
trevormacro wrote on 27.06.13 at 12:01:16:
@osckhar:

If you put the chip on the rampy and after that the card doesn't work, is it possible to come back to original state without damaged both parts (balls traces on pcb and pins balls for the chip)?

Soldering the chip will be done by you?

Always there is a risk.  I cant tell you 100%  not risk but I have reemplaced several bga chip from 3Dfx cards without problems. I have done some reworks (vsa100 / tsops) for Gino and all perfect. No cards damaged.
Lately I had to fix a Daytona agp with problems in memory ddr and bga (reemplaced them) and card is perfect now.

About If is possible once soldered  the chip come back his original stated. Yes, it is possible.
Is possible reballing the chip with new balls? Yes, it is possible. Anyway I should build custom stencil from a standar stencil 40x40mm  for rampage chip.

Now the bga pads from rampage card are not leaded they are in his orginal copper stated.. Once I Solder the chip onto the card if I remove the chip pads willbe leaded but that is not bad since it can help makes better contact.

Yes, I would do all the process with my BGA reworks station.

Michel, is your choosing. You must think it and even talk with Hank about this. Always there is a risk and if something goes bad I dont want any responsability.

If you ask me what would I do? You know me and I would try but the main problem is if really the problem comes from there.

I know very difficult decission.

Regards,
Oscar.
Back to top
 
WWW  
IP Logged
 
Loeschzwerg
Senior Member
****
Offline


AAlchemy PSU maker :-)

Posts: 396
Germany -> Bavaria -> Augsburg
Gender: male
Re: 3dfx Specter 1000 AGP TEST SOCKET Rev.A1 4700
Reply #74 - 27.06.13 at 12:42:25
 
osckhar wrote on 27.06.13 at 12:31:23:
Now the bga pads from rampage card are not leaded they are in his orginal copper stated.. Once I Solder the chip onto the card if I remove the chip pads willbe leaded but that is not bad since it can help makes better contact


Conductivity goes down:
http://metals.about.com/od/properties/a/Electrical-Conductivity-In-Metals.htm (compare lead to copper)

But I assume it won't have an impact Smiley

Back to top
« Last Edit: 27.06.13 at 12:43:28 by Loeschzwerg »  
 
IP Logged
 
Pages: 1 ... 3 4 5 6 7 ... 23
Send Topic Print