VOGONS


First post, by pentiumspeed

User metadata
Rank l33t
Rank
l33t

I was using what I had from tin of parts to keep Latitude 3330 running. Had 16GB DDR3L but I needed them to test and work on Latitude 3340 notebooks bundle that friend found in the dump bin.

So I put in DDR3L 2x2GB for short time (elpida). Worked at first... then finally crashed on sleep and wouldn't boot properly, corrupting the 1TB SSD. I kept trying to reinstall Win 10 OS, always deleting the partitions and recreating but still unbootable. Started over with pair of Samsung 4GB modules DDR3L 1600 and OS finally booted properly after reinstall.

Just watch for unexplained problems. Always don't trust the memory, keep some on hand, always replacing all and toss the untrusted memory out if that resolves your problem.

I always buy OEM brands that made them: Samsung, Micron/Crucial, Hynix, Elpida and few others. I don't keep generics.

I was doing computer repair for decades and knew what to watch out for these kind of sneaky problems. Hard drive is easy to tell, slow, buzzing, 2 beat ticks etc or totally unresponsive, laggy.
I would pull the hard drive and read the SMART attributes using any utility my favorite is gsmartcontrol. Only few MB to download and is windows only and takes seconds to check. But I can check any storage drives using this: PATA, SATA and SSD does have smart reporting, does not matter where storage drives came out of: Game consoles, computers etc.

I must confirm and show this to customers that drive had failed.

Cheers,

Great Northern aka Canada.

Reply 1 of 11, by Errius

User metadata
Rank l33t
Rank
l33t

A faulty memory module can cause strange systemic problems that manifest anywhere. Overheating can also cause non-faulty memory modules to misbehave.

Is this too much voodoo?

Reply 2 of 11, by SodaSuccubus

User metadata
Rank Member
Rank
Member

I like to build gaming computers for friends and oddly, iv never really had much issues with bad memory these days.
Most of the time its just the memory throwing a fit with certain motherboards (Ryzen mobos in particular, can be VERY peculiar over what ram sticks they wanna play with).

Old computers though? hoo boy. Have i lost a few good days of gaming dealing with weird manifestations, only for it to be dead ram or cache.
I actually endedup throwng out a bunch of "dead" ram recently, out of my socket 7 build. Only to find out it was actually dead on-board l2 cache that was causing memory problems.

Reply 3 of 11, by PTherapist

User metadata
Rank Oldbie
Rank
Oldbie

I had a similar issue with DDR3 RAM years ago. A guy had bought a computer the previous year and it had been working absolutely fine, then suddenly it started having stability issues in Windows. I eventually diagnosed bad RAM, as it was failing memtest. Replaced his RAM and all was well. But the story doesn't end there -

I kept the RAM for about a year until I ended up putting together a cheap AM3 test PC. I still had that faulty RAM and so put it in this PC to test it out. It wouldn't POST at first with both sticks but I managed to get it to boot with just 1 and manually lowered the speed of the RAM in the BIOS. Then both sticks worked absolutely fine for years, until slowly some stability issues started to recoccur with Windows freezing at boot and random BSODs.

I'd turned the test PC into a Media Server by this point, so the dodgy RAM had to go, but I've actually still got it in my drawer somewhere.

Reply 4 of 11, by douglar

User metadata
Rank Oldbie
Rank
Oldbie

I had a similar problem years ago. Kept getting very clearly announced CDROM errors when installing the operating system. Swapped out drives, cables, controllers, disks. Still kept getting errors. Left memtest running out of frustration. 20 hours in, the errors started showing up. My best guess is that weak memory was causing the DMA transfers to fail.

Reply 5 of 11, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie

I have a process for any new hardware that enters my home for wayward computers: Clean it, replace fans, replace caps, run memtest. No CPU, motherboard, or RAM -- nothing -- goes into my inventory until it passes at least one round of memtest. 😀

Reply 6 of 11, by Errius

User metadata
Rank l33t
Rank
l33t
douglar wrote on 2020-03-10, 13:44:

I had a similar problem years ago. Kept getting very clearly announced CDROM errors when installing the operating system. Swapped out drives, cables, controllers, disks. Still kept getting errors. Left memtest running out of frustration. 20 hours in, the errors started showing up. My best guess is that weak memory was causing the DMA transfers to fail.

I had the exact same problem but it turned out to be overheating RAM. Removing 2 of the 4 memory modules to improve airflow (and the arrival of September) finally fixed the issue. CD-ROM drive was fine.

Is this too much voodoo?

Reply 7 of 11, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie

Hah, come to think of it.. I had that same issue with a Mac Pro G4 Cube. It had 1.5GB of RAM installed, one of the 512MB sticks was bad. Ran fine most of the time, but for whatever reason, when I tried to install OS X fresh from the CD, it would throw errors. I burned three or four copies of the install disc and thought my spindle of CDs was going bad before I gave up and installed from USB (1.1!), slowly, and still had the same issue. Then I discovered the diags disc, with a RAM test...

Reply 8 of 11, by Errius

User metadata
Rank l33t
Rank
l33t

I know that old iMac G4s won't read DVD-R disks. This is a different headache. I have two of these machines with the same problem. CD-Rs are OK though. You have to connect a FireWire drive to boot from a DVD-R.

Is this too much voodoo?

Reply 9 of 11, by halls_well

User metadata
Rank Newbie
Rank
Newbie
PTherapist wrote on 2020-03-10, 12:53:

I had a similar issue with DDR3 RAM years ago. A guy had bought a computer the previous year and it had been working absolutely fine, then suddenly it started having stability issues in Windows. I eventually diagnosed bad RAM, as it was failing memtest. Replaced his RAM and all was well. But the story doesn't end there -

I kept the RAM for about a year until I ended up putting together a cheap AM3 test PC. I still had that faulty RAM and so put it in this PC to test it out. It wouldn't POST at first with both sticks but I managed to get it to boot with just 1 and manually lowered the speed of the RAM in the BIOS. Then both sticks worked absolutely fine for years, until slowly some stability issues started to recoccur with Windows freezing at boot and random BSODs.

I'd turned the test PC into a Media Server by this point, so the dodgy RAM had to go, but I've actually still got it in my drawer somewhere.

If you want to be super cheap, you can keep running even with the bad RAM! GRUB lets you specify which areas of memory not to use, and so does BCD, but only on newer Windows 😁

I did it for a little while, I had 8GB and only had issues at the upper end of the memory, but the bad blocks started creeping downwards in address numbers so I ended up replacing it (this was DDR2 I picked up in ~2010 and it just went bad last year so I'll take it!)

Reply 10 of 11, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie
Errius wrote on 2020-03-10, 20:40:

I know that old iMac G4s won't read DVD-R disks. This is a different headache. I have two of these machines with the same problem. CD-Rs are OK though. You have to connect a FireWire drive to boot from a DVD-R.

I was using CD-Rs. I don't think any of the OSes I installed on it came on DVD. Er, wait, yeah Tiger did, but I think I had the original media at that point.

I had started buying boxed copies of OS X, and I was going to install 10.0 through 10.4 on the Cube, but not all of them had arrived yet, so I was installing some of them from CD-R. It read the discs fine, until it would suddenly choke and throw an error log showing that cpio or tar or something had bailed out.

At first I suspected the CD drive was failing, since the Cube was new (to me). After using the external USB drive, which also failed, I suspected the CD-Rs were going bad, but I could pop them in my Linux box and successfully MD5 the whole disc. Plus, the 10.2 install that actually finished without error crashed once, and generally felt a little unstable. So then I wasn't really sure what the issue was. I Googled the installer errors I was getting and someone suggested using the diag disc to test for hardware issues. That's when I narrowed it down to bad RAM.

Whatever it copied to memory during the install got corrupted and would fail to decompress or wouldn't pass CRC or whatever it was doing in the background. As soon as I took out the third (bad) stick of RAM, the rest of the OSes installed without a hitch.

Reply 11 of 11, by EvieSigma

User metadata
Rank Oldbie
Rank
Oldbie

I recently had an issue with a HP Pavilion laptop where the DDR3 RAM I bought would allow the computer to boot but it would blue-screen when it tried to load up Windows 10. Swapped it for different RAM and the problem went away.

RAM is weird sometimes...