Cleaning out Ram?? WOW!

[SC68Cal]

Hey everyone, I've been having some siginifcant problems with my Titanium Powerbook, where I literally could count on programs crashing, along with the entire operating system.

Originally I blamed it on bad ram and hard drive problems. However, I think i've been thinking on too high a level. I gave up and tried cleaning out the RAM pins and the slots on the motherboard with "QD Contact Cleaner" by CRC, which is a great electrical cleaning solvent that dries very fast. I took out the ram and sprayed the solvent onto the chips, then wiped them with a single QTIP to clean off any extra solvent. I then took the Q-Tip, sprayed some solvent on it and then used it to wipe the slots that the ram goes into.

Booted up my computer, Holy COW! I can already tell a difference in response time.

I'm hoping this will end the kernel panics and system instability, since most of my problems were linked to bad memory addresses. :crosses fingers:

 

[topgunn]

Call me a pessimist but I think you just wasted two Q-tips.

 

[SC68Cal]

I actually only used one

I found the suggestion from here, so at least another person has has similar problems

http://blog.polaris-digital.com/archives/97

 

[mad jew]

I'd be interested to hear the outcome. I guess it's a possibility that dirt on the end of the RAM chips was short circuiting the controller, resulting in your failures. However, I'm inclined to agree with topgunn at this stage.

 

[Makosuke]

Application crashes or kernel panics, I could imagine in *rare* cases being caused by corroded contacts.

But "I can already tell a difference in response time."? Sorry, I wouldn't believe it unless I saw timed benchmarks. In fact, I still wouldn't believe it even then. If the RAM is corrupting data, it should cause outright crashes unless it's ECC RAM (which this isn't), not degraded performance.

I admit, I'm not an expert, but that sounds like placebo at work to me.

 

[topgunn]

I admit, I'm not an expert, but that sounds like placebo at work to me.

That is exactly what I was thinking.

Perhaps the memory was not seated correctly and having removed it and replaced it solved the problem?

 

[SC68Cal]

This has crossed my mind. I know it sounds really suspect about the response time, but I swear the boot time was significantly shortened.

I've spent all day seating and reseating the ram (and switching out modules for other ones) only this time around cleaning out the contacts did I actually get any sort of result (percieved or actual, we'll see with some more use of the laptop)

 

[mad jew]

Boot time isn't really related to RAM though. As long as it works now though, it's all cool.

 

[Angrist]

I had horrible instability problems on my PowerMac a couple of months ago, turned out to be RAM related.

Applications would crash, kernels would panic, filesystems would corrupt. When I ran memtest, it failed miserably. So in an effort to track down the bad dimm, I pulled and tested each stick. Every single one passed. Somehow the act of removing and re-seating the RAM fixed the issue.

I didn't benchmark anything, or notice and speed differences, BUT nothing has crashed since.

 

[generik]

I think it shouldn't be necessary too, considering how fast those components are clocked, a "loose connection here or there" won't just result in your computer being slower, it'd just *poof* and do wonderful things to a portion of total memory chips on the module, resulting in a night of Kickboxing with my Mac

 

[KingYaba]

Maybe something was loose and you fitted it in properly?

 

[SC68Cal]

Well. The honeymoon is over. Still having kernel panics. (sighs)

It's almost like this TiBook doesn't want any other ram sticks except the ones that came from the factory. Of the original two that I bought from OWC, one has already been sent back to be replaced, and this one that I originally kept because it passed Rembr is now looking like it needs to be sent back too.

I think if the second stick of ram that OWC sends me still makes me crash, I'm going to either ask for a store credit or an outright refund.

 

[Makosuke]

Can't say I'm surprised. Anyway, if it definitely works with factory RAM and not with the OWC stuff, that would be odd if there wasn't something actually wrong with the RAM, since OWC sells stuff that should definitely work in the model they okay it for.

You mentioned Rembr, not something I'm familiar with, but have you tried running Memtest ( http://www.memtestosx.org/ ) to see if it comes up with errors?

Just to offer a similar situation, I recently upgraded the RAM on an Intel mini and ran into a similar issue; worked perfectly with the factory RAM, would spontaneously reboot with the two new DMS sticks in it, but only when it heated up. And ironically, it was fine with either one of the DMS sticks alone or paired with the factory sticks, and Memtest never found a thing wrong. I eventually exchanged both sticks for a new pair, and that worked perfectly.

So... it's possible to get flakey RAM that checks out fine but misbehaves in some particular situation, even from a reliable vendor. If you're lucky, that's your problem.

 

[SC68Cal]

OWC has been very accomodating to my issues, and after calling a rep and explaining my issues and concerns, and a few apologies for causing them so much trouble over the ram, they reiterated that their policy for ram is no-questions-asked replacement and that if I have an issue I should call them.


Rembr is a GUI front end for Memtest. I know that it's not as powerful as memtest, since memtest runs in single user mode, but I've been occupied recently and Rembr was the quick and dirty way to check the memory.

 

[California]

That is so weird. I have never had a ram problem on any Tibook and I have owned...
one 400mhz
two 500s
one 550
one 667
one 1ghz

BUT I did have "kernal panics" on the 1ghz that had a bad logic board.
Had nothing to do with the ram.

What model Ti is it?