I didnt see it suggested but it would be nice if before you throw it in the freezer, test it again to make sure its dead. If it is degraded and still posts, note operating volts and stability then freeze and compare.
Printable View
I didnt see it suggested but it would be nice if before you throw it in the freezer, test it again to make sure its dead. If it is degraded and still posts, note operating volts and stability then freeze and compare.
at least stick it in and warm it up first
Any particular reason why a rat? :p: I'm hoping theres a story behind it, its one of the more unusual choices for an avatar and unusual normally means theres a story :DQuote:
Originally Posted by Vatos_locos
sorry about the OT.
But unless the chips are your baby super overclocking chips, why not just RMA them since most of them have lifetime warrenty?
Because the replacement won't be D9s ;)
:down: And there is no logical explanation - at least for me.Quote:
Originally Posted by Oj101
I killed mine with 3,2v - I'll try that "tweak".
The seriously high clocking, tight timing RAM used "Micron D9" ICs, normally D9GMH or D9GKX. Those insane little ICs haven't been made for well over a year now, so your replacements will probably be PSC or similar - not all that marvellous for benching.
PC2-8800 Gold XTC
5.0-6-6-18 @ 2.2 ~ 2.4V
Micron D9GMH (B6-3) (only if produced before 2008)
Just bought another set - no D9's any more :( (And sold it even faster)
Now you understand the logic :up:
:rofl:Quote:
Originally Posted by Oj101
then i guess im lucky cuz i have all these to pick from
:rofl:
All but the Patriots i believe are D9's. I understand what you guys are saying, i keep 3 benching pairs ready. :rolleyes:
http://i125.photobucket.com/albums/p...a/IMG_0892.jpg
:wasntme:
Still id think a working stick is better then a non working stick no? :p:
I dont know about you guys, but usually i do throw my kit on freezer after heavy bench session :S just after i shut down pc, ill just throw them there, no logic here, maybe trying to keep them alive little longer. doing this since BH-5 days
Wee it worked with my good old Corsair bh5
no logic in it you were verry lucky you did not have the reverse affectQuote:
Originally Posted by KeZzZu
i only throw in any electronic part it when they dont work or stop workin not just ram
ram likes cold when runnin its the connections that need the cold when broken
im curious how long these revived sticks will work though...
would be annoying if they need to be thrown into the freezer every other week :lol:
Mine's still happily shunting away :)
First attempt - no luck. I put them back to freezer and will forget their existance for some time. Mayby I'll have more luck next time.
My Team D9GMH dies on a fairly regular basis (i.e. when ever I stress them at over 2.35v for a protracted period of time and then encounter and error or shut off the system). To all tests, one or both DIMMs appear dead. No system will post with the RAM, and clearing CMOS, pulling the battery...nothing.
I wrap them in aluminum foil and put them in the oven at just under 400F for 20 minutes...it's resurrected them three times in the last 12 months.
I figure the rapid cooling they go through (from about 70C to 25-30C) when an a stress test locks up, or is canceled, might be breaking one or more of the solder balls holding the ICs to the PCB and that the heating in the oven is just enough to reflow the solder and get them working again. Or the solder connections themselves could be more prone to electromigration than the ICs, and partially reflowing the solder removes the traces of electromigration damage.
This freezer idea is interesting, though I cannot for the life of me come up for a plausible reason why it should work.
I was about to ask if you're "cooking" mad then I remembered that this is XtremeSystems, not XtremelyAverageSystems
400F is 204*C, that's quite hot, although the foil would somewhat shield it. Good to hear it has worked for you though.
The reason I used foil was to hold everything firmly in place, so when and if the solder did become weak, components didn't just start falling off the dimms. I also did not want the direct radiation from the heating element heating the dimms, or the dimm itself would have gotten a lot hotter than 400F before the air in the oven was that temp.
400F/204C is not any warmer than what they were exposed to during manufacture, and since the dimms were not powered on there was little risk to the ICs. Still, I would obviously not store them long term at such a temp.
They do smell like burnt plastic now, as the thermal pad team uses is some rubberised gapfiller that apparently has a lower melting/ignition point than any other part of the ram.
I put them in the freezer and forgot their existence ... turns out, that they are GKXs so Vatos_locos, no RMA for me ;) Anyway, I don't know what their exact problem was 'cause the only thing I did, was suddenly feeding them with 2.6V ... Normally they shouldn't die, even if GKXs do get really hot ... Well, in conclusion, they are now working like a charm :) :woot: ... I have them here, on the P5K Deluxe with an E6600, running memtest without the faintest problem :yepp: :)Quote:
Originally Posted by George_o/c
Thanks Oj101 for the tip :up: These Hyper-Xes are not the greatest modules I have, but still worth being alive :) I was planning on using them for 3D benching, 'cause I'm afraid of murdering my "killer" ballistix :D
i just threw my pair of crucials 667 at 3-3-3-8 2.2, they did 640 with something between 2.32-2.38(can`t remember).
they we`re getting me errors at 535 with 5-5-5-15 at 2.36, so i had no other choice :))
we`ll see in a day or so
Woohoo my Balistix are back :D
Gona try if this works on GFX's too.
i pulled out my fatbody kit and forgot about testing them since i've been putting my patriot power chips through the ringer - any one else try this trick on an old fat body d9 kit???
It works as well:clap:Quote:
Originally Posted by ieatrawbacos