Bad engineering from Sapphire: Radeon 9600 the heatsink damages the open die beneath

  • Thread starter Thread starter Jopo75
  • Start date Start date
J

Jopo75

Yeah, ad subject says, my heatsink has damaged the die of the Ati chip
beneath .
See www.digit-life.com/articles2/radeon/pcolor-1.html for a visual
explaination of my accident (see the pics at the middle-bottom of that
page).
My card is running fine but I'm wondering why Ati has licensed such a sloppy
manufacturer to produce that bad engineered card : on the 9600 the heatsink
is too way big and too heavy to being placed on the chip via two ultrasmall
plastic spacers (like those on old mobos).
Take nVidia for example: all nVidia chips are virtually mechanically
unbreakable (like all Pentiums IV), not quite like Ati's: infact nVidia's
haven't the open die touching the heatsink.
Remember how many Amd Athlons got permanently damaged because the open die
touches directly the heatsink??
Bad engineering.
 
Just curious. How did you know the cooler damaged your chip?

If it's running fine, you had to pry it off to see. If so, perhaps you were
the one to damage it.

Draw any conclusions you want about Sapphire, but they are ATIs biggest
partner and make more of the cards ATI sells under its own brand name than
anybody else.

I've had three of them, and had the coolers on and off all of them with no
such problem. They cut corners to keep costs down, but they perform well.

That's my definition of good engineering. You're certainly entitled to a
different opinion.
 
Just curious. How did you know the cooler damaged your chip?
Simple: touching the heatsink I've heard a sound similar to the sound that
you make when you rub a piece of glass against sand.
Have you seen how BIG is the heatsink on the Sapphire Radeon 9600??
And that big heatsink simply cannot be placed firmly against the die using
only two small piece of plastic.

If it's running fine, you had to pry it off to see. If so, perhaps you were
the one to damage it.
Click to expand...
Yes, but only touching once the heatsink (who hasn't touched once the
heatsink, and by saying "touched" I don't mean that I've taken the videocard
BY the heatsink) doesn't automatically imply that I've ruined the chip.
If so it means that the card is mechanically bad engineered.

Draw any conclusions you want about Sapphire, but they are ATIs biggest
partner and make more of the cards ATI sells under its own brand name than
anybody else.
Click to expand...
Frankly I don't care how big or not a brand is: I care for quality and
robustness.
I've had three of them, and had the coolers on and off all of them with no
such problem. They cut corners to keep costs down, but they perform well.
Click to expand...
Well you're lucky. Probably the Sapphires you own have the heatsink glued to
the chip via a thermoconductive epoxy (in that case rub against the chip is
avoided).
That's my definition of good engineering. You're certainly entitled to a
different opinion.
Click to expand...
That's not a definition of good engineering: Pentiums IV are GOOD
engineered, their type of heatsink (the way they are mounted and the quality
of construction) are good engineered, nVidias are good engineered.
Old proper Atis are good engineered (infact I'm very happy of my Ati Radeon
7500).

Tell me one thing: why Ati stopped to make the entire video board and relied
to 3rd part manufacturing only for Europe?
 
Jopo75 said:
Simple: touching the heatsink I've heard a sound similar to the sound that
you make when you rub a piece of glass against sand.
Have you seen how BIG is the heatsink on the Sapphire Radeon 9600??
And that big heatsink simply cannot be placed firmly against the die using
only two small piece of plastic.


Yes, but only touching once the heatsink (who hasn't touched once the
heatsink, and by saying "touched" I don't mean that I've taken the videocard
BY the heatsink) doesn't automatically imply that I've ruined the chip.
If so it means that the card is mechanically bad engineered.


Frankly I don't care how big or not a brand is: I care for quality and
robustness.
well.
Well you're lucky. Probably the Sapphires you own have the heatsink glued to
the chip via a thermoconductive epoxy (in that case rub against the chip is
That's not a definition of good engineering: Pentiums IV are GOOD
engineered, their type of heatsink (the way they are mounted and the quality
of construction) are good engineered, nVidias are good engineered.
Old proper Atis are good engineered (infact I'm very happy of my Ati Radeon
7500).

Tell me one thing: why Ati stopped to make the entire video board and relied
to 3rd part manufacturing only for Europe?
Click to expand...

I don't think ATI ever manufactured ANY of the Radeon series sold under the
ATI brand name. Those have all be done by Asian companies. They are all made
to ATI design and spec, and even the OEM cards sold by Sapphire and others
have to use a minimum reference design.

Sorry you feel you got a bad one, QA slips happen. But I think you're
reaching to conclude that Sapphire is second-rate because you discovered a
chipped core when you pried the cooler off.

The fact that no one in this group has agreed with you in the multiple
threads you've started should tell you something.

Feel free to buy another brand next time. I'll feel free to recommend
Sapphire cards based on my personal experience with several of them and the
good vibes from friends I've suggested them to.

Opinions are like snowflakes, there are no two exactly alike.
 
Uhhh.. do you actually know what a die is? What you're refering to is
actually the ceramic substrait that the die is inside of (acutally in
this case it appears the die is attached to the substrait below and
there is a ceramic cap on top of it). Typically the ceramic can be
significantly damaged and the chip will still function. But after
reviewing the pictures that you show, it doesn't appear to be cause by
vibrational friction as you are proposing. What it appears is that
someone lifted on side of the heat sink diagonally across the chip's
ceramic causing the corner to either chip off or be scraped off. If
it was vibrational damage you would see more consistant wearing of the
ATI silk screen on top of the ceramic; but this is not shown in your
photos.

Conclusion. You did the damage yourself; it does not look like what
you claimed (bad heat sink design / connection).
 
On 11 Nov 2003 21:57:40 -0800
Uhhh.. do you actually know what a die is? What you're refering to is
actually the ceramic substrait that the die is inside of (acutally in
this case it appears the die is attached to the substrait below and
there is a ceramic cap on top of it). Typically the ceramic can be
significantly damaged and the chip will still function.
Click to expand...

With many contemporary chips the the die is exposed, with the pure
silicon base on which the circuits were grown pointing up. The reason
this is done is that that approach gives the lowest thermal resistance
in the cooling solution. The Athlons are the best known chips
constructed in this manner, and cracking the die was a fairly common
installation error until AMD started putting some bumpers on the chip
to ensure that the heat sink went on level--occasionally someone still
manages to crack the die. The ATIs appear to be similar in construction
to the current generation Athlons, which means that that little black
rectangle in the middle of the substrate _is_ the die. Further, the
substrate is often a fiberglass PCB with current designs.

However some of the immense heat sinks used with Athlons are secured by
a spring-clip to the skinny little tabs on the socket--if the mass of
the heat sink is not a problem with Athlons, which it generally is
not, then the relatively tiny one used on ATI boards is most assuredly
not going to be a problem when secured using the drilled-through
mounting holes in the circuit board.

Intel and nvidia put a metal heat spreader on top of the die that serves
to protect it from hamfisted installers, however it also adds another
layer of thermal resistance to the cooling. AMD has started doing the
same thing with their 64-bit chips.
 
On 21 Nov 2003 23:45:27 -0800
<BIG BIG SNIP>

Yeah, I understand that 100%, but I was refering to this case. If you
look at the pictures, the die is not exposed. But the way he was
talking about it, you would think that it is. First thing I did was
to look at his pictures and you can quickly determine the die is not
exposed.
Click to expand...

Gee, that sure looks like an exposed die to me. What leads you to
believe that it is not?
 
With many contemporary chips the the die is exposed, with the pure
silicon base on which the circuits were grown pointing up.
Click to expand...

<BIG BIG SNIP>

Yeah, I understand that 100%, but I was refering to this case. If you
look at the pictures, the die is not exposed. But the way he was
talking about it, you would think that it is. First thing I did was
to look at his pictures and you can quickly determine the die is not
exposed.

As far as I know AMD is the most popular manufacturer using the
exposed die. It does tend to open a whole new can of worms when it
comes to the general public (I wonder how many were cracked during
installation and returned as defective). The bumpers help, but
doesn't emlimate the issues.

Dave
 
Back
Top