Crucial PC2-8500 Ballistix Tracer Heatspreader off center ?

[eva2000]

Well I bought a few sets of 2x 1GB Crucial Ballistix Tracer PC2-8500 modules and started to test my second set. But for some reason in the second set, I couldn't install one of the modules in my dimm slot

So I examined the heatspreader carefully, comparing it to the other module in the set and look what I found. The problematic module's heat spreader (bottom module in image below) was installed off centre to the left as you can see the below image. No wonder it couldn't fit the dimm slot!



Question is should I RMA what could be a perfectly good set of Micron D9xxx modules, or should I try to correct the off centre heatspreader myself ? RMA back to USA would of cost AUD$60 (US$50) via USPS EMS/ECI courier from Australia!

I've never taken the heatspreader off any DDR2 module before for fear of ripping the ICs off the PCB. But decided to try and correct the problem myself with a hair dryer to heat up the heatspreader and my finger nails (lucky and hadn't clipped the nails yet as they were very useful :heh: ).

It took bloody ages with several bouts of hair dryer heating up to very very slowly loosen the TIM under the heatspread! I did one side at a time until finally they came off

Sexy bare naked Micron D9xxx module next to it's clothed partner



Close up.. At first, can't seen any clearly indentifying markings to suggest whether it's Micron D9GMH or D9GKX. But then if look very carefully you can see this module is Micron D9GMH

[eva2000]

Looks like the bare naked module is fine.. phew!

System

• Intel Core 2 Duo E6750 ES - L710A438 G0
• Corsair Nautilus 500 3/8" ID
• Asus P5K Deluxe 0311 bios
• 128MB Gainward FX5200 PCI
• 1x 1GB Crucial PC2-8500 Ballistix Tracer single naked module from dual channel kit #2 in black dimm slot
• 74GB WD Raptor 8MB on SATA 2 port
• Pioneer DVR-08 Burner
• Sony FDD
• 1KW PCP&C SLI psu
• WinXP Pro SP2

Prime95 v25.2 Blend Load test


Post load idle


(AI Suite cpu temp sensor gone a bit wacky although coretemp reports the right temps)

It's late now so will test the second kit as a pair tomorrow

[stephen23]

I got mine open b4, but not dare to fully open it as I have rip off one of the D9 chips from my G.Skill HZ b4!

[erwinz]

nice EVA2000.. now you see the nakedness.. I doubt that you will remove all you heatspreader now..

[eva2000]

I got mine open b4, but not dare to fully open it as I have rip off one of the D9 chips from my G.Skill HZ b4!

yeah it's scary removing the heatspreaders with knowledge you could rip off the ICs if you're not careful

nice EVA2000.. now you see the nakedness.. I doubt that you will remove all you heatspreader now..

yeah i doubt i will... no choice on this module - it just wouldn't fit into the dimm slot LOL

[Techie]

May i ask where you found these modules? I can't find the tracer version anywhere (i live in europe )

[stephen23]

May i ask where you found these modules? I can't find the tracer version anywhere (i live in europe )

many online shops and ebay will help!

[Techie]

many online shops and ebay will help!

I did find many online shops, but all in US. They don't ship to europe.

[eva2000]

try http://www.priceguidenetwork.com/

[MotF Bane]

I always thought that it was better to freeze the modules, so that the adhesive would give way.

Eva, how were you able to tell those are D9GMH chips?

[aussie-revhead]

Eva, how were you able to tell those are D9GMH chips?

Perhaps its under the Ballistix logo .

Eva that looks scary , has it affected clocks at all - higher or lower than expected ? Or about the same ?

[eva2000]

I always thought that it was better to freeze the modules, so that the adhesive would give way.

Eva, how were you able to tell those are D9GMH chips?

i don't think freezing would be good for DDR2 as the ICs joints to the PCB aren't that strong ??? I read up lots of DDR2 heatspreader removal before gather enough balls to do this and hair dryer method was more common.

D9GMH stamp is barely visible under the crucial text on ICs if you look very very closely

Perhaps its under the Ballistix logo .

Eva that looks scary , has it affected clocks at all - higher or lower than expected ? Or about the same ?

too early to tell... give me a few days...

When i started testing this second kit in a pair for dual channel I noticed another QA issue with the other module.. one side of the heatspreader's TIM wasn't making much contact with the ICs.

Tested the heatspreader module at default DDR2-1066Mhz 5-5-5-15 at 2.1v first to make sure it worked. And you guessed it, stripped it bare naked as well

As you can see the TIM is partially lifted off the ICs.. I tried holding it together to see if it will stick but it lifts off again by itself. So bye bye heatspreader



In the flesh Micron D9GMH based 2x1GB Crucial Ballistix Tracer PC2-8500 dual channel kit

[MotF Bane]

D9GMH stamp is barely visible under the crucial text on ICs if you look very very closely

Oh. Thanks, I see it now.

[eva2000]

System

• Intel Core 2 Duo E6750 ES - L710A438 G0
• Corsair Nautilus 500 3/8" ID
• Asus P5K Deluxe 0311 bios
• 128MB Gainward FX5200 PCI
• 2x 1GB Crucial PC2-8500 Ballistix Tracer naked modules - dual channel kit #2 in black dimm slot
• 74GB WD Raptor 8MB on SATA 2 port
• Pioneer DVR-08 Burner
• Sony FDD
• 1KW PCP&C SLI psu
• WinXP Pro SP2

Looks like bare naked Micron D9GMH from Crucial Ballistix Tracer PC2-8500 are running more than fine @625mhz 5-5-5-15 3-30-3-3-8 at 2.1v vdimm!





PC2-10000 5-5-5-15 at 2.1v vdimm

Temp probe reading off ICs for Super Pi 32M run read idle = 28-32C and load = 32-35C with 92x38mm Sunon 120cfm fan on medium speed rheobus.

[StealthyFish]

i don't think freezing would be good for DDR2 as the ICs joints to the PCB aren't that strong ??? I read up lots of DDR2 heatspreader removal before gather enough balls to do this and hair dryer method was more common.

D9GMH stamp is barely visible under the crucial text on ICs if you look very very closely



too early to tell... give me a few days...

When i started testing this second kit in a pair for dual channel I noticed another QA issue with the other module.. one side of the heatspreader's TIM wasn't making much contact with the ICs.

Tested the heatspreader module at default DDR2-1066Mhz 5-5-5-15 at 2.1v first to make sure it worked. And you guessed it, stripped it bare naked as well

As you can see the TIM is partially lifted off the ICs.. I tried holding it together to see if it will stick but it lifts off again by itself. So bye bye heatspreader



In the flesh Micron D9GMH based 2x1GB Crucial Ballistix Tracer PC2-8500 dual channel kit

whoa, do you know how to make that RAM cooling solution with the spacers holding up a fan? I usually just jam a fan between my scythe and the IDE//power supply cables to let it hold the fan, but it doesn't look very pro . That looks like a good solution

[eva2000]

take away chop sticks sawn off

[erwinz]

that is one ram cooling..

that is what you get for a memory addict..

very nice eva2000

here are other ghetto ram coolers..
http://www.xtremesystems.org/forums/...d.php?t=148783

[eva2000]

Yeah it works well and rheobus fan controller allows me to tune the fan speed for quietness or best performance

[eva2000]

Still testing this 2nd kit which bare naked now, but I inspected the 3rd and final kit I have here and again another module in the pair isn't making full TIM contact with ICs but no where as bad as this 2nd kit's one module pictured here.

It seems one side of the module's LED controller (middle square chip) is slightly taller in height than the other side's middle square chip meaning the HS lifts slightly off the ICs in the middle on either side of the LED controller chip. The heatspreader on that side is clearly at a slight angle like the picture here but not as bad.

Solution: No stripping of heatspreader necessary ?

Instead of stripping this module's heatspreader going to try another method, grab one of those strong paper clips, 2 pieces of cardboard and trusty hair dryer.

Place one piece of cardboard on either side of heatspreader, use the paper clip to clamp both sides of the heatspreader using the cardboard as protective layer to prevent scratching on the heatspreader. Now using hair dryer to heat up the heatspreader while leaving the heatspreader clamped down to making contact with ICs.

I'm hoping this heat and clamp method will allow the heatspreader to bend back inward towards the ICs to allow better contact

End Verdict:
The clamp and heat method seems to have worked, the heatspreader's TIM has made firm and full contact on the ICs and seems to have stayed for now. Guess have to wait until I get around to testing the 3rd and final kit

[StealthyFish]

Still testing this 2nd kit which bare naked now, but I inspected the 3rd and final kit I have here and again another module in the pair isn't making full TIM contact with ICs but no where as bad as this 2nd kit's one module pictured here.

It seems one side of the module's LED controller (middle square chip) is slightly taller in height than the other side's middle square chip meaning the HS lifts slightly off the ICs in the middle on either side of the LED controller chip. The heatspreader on that side is clearly at a slight angle like the picture here but not as bad.

Solution: No stripping of heatspreader necessary ?

Instead of stripping this module's heatspreader going to try another method, grab one of those strong paper clips, 2 pieces of cardboard and trusty hair dryer.

Place one piece of cardboard on either side of heatspreader, use the paper clip to clamp both sides of the heatspreader using the cardboard as protective layer to prevent scratching on the heatspreader. Now using hair dryer to heat up the heatspreader while leaving the heatspreader clamped down to making contact with ICs.

I'm hoping this heat and clamp method will allow the heatspreader to bend back inward towards the ICs to allow better contact

End Verdict:
The clamp and heat method seems to have worked, the heatspreader's TIM has made firm and full contact on the ICs and seems to have stayed for now. Guess have to wait until I get around to testing the 3rd and final kit

hmm.. I'm hoping my tracer don't get here all imperfect like your kits.. lol. That or you're pretty particular about everything. 4 sticks of tracer... 4 times more likely to end up with some sort of deformity

[eva2000]

Yeah hope yours are better in terms of QA

Some 4-4-4-12 tests with my bare naked Micron D9GMH from Crucial Ballistix Tracer PC2-8500. Seems my first and second kit (bare naked) both have a slightly higher 4-4-4-12 max then my non-tracer Ballistix PC2-8500 modules. 570mhz 4-4-4-12 at 2.45v vs 555-565mhz 4-4-4-12 at 2.45v (tested with 3-30-3-3-8 subtimings).

Memtest86+ v1.70 looped test #5 @570mhz 4-4-4-12 3-30-3-3-8 at 2.45v the bare naked modules with 92x38mm 120cfm fan on rheobus set ~60% fan speed was only registering around 34-36C temps

[MotF Bane]

Why haven't you put the heatspreaders back on?

[[XC] DragonOrta]

I removed my heatspreaders on my Corsair 5400UL by heating them with a hairdryer and slipping a very thin knife between the ICs and the heatspreaders. It worked great for me.

[ozzimark]

Seems my first and second kit (bare naked) both have a slightly higher 4-4-4-12 max then my non-tracer Ballistix PC2-8500 modules. 570mhz 4-4-4-12 at 2.45v vs 555-565mhz 4-4-4-12 at 2.45v (tested with 3-30-3-3-8 subtimings).

come on Eva, you can do better than that

with 2.5v set in the BIOS with my PC2-8500 Tracers:

[Don_Dan]

I did find many online shops, but all in US. They don't ship to europe.

Have you tried buying from Crucial's own online shop?
http://www.crucial.com/eu/