01-04-2018 10:28 AM
8+ years ago I bought Crucial memory (two 1 GB boards) to replace two smaller memory boards in my PC.
This worked just fine up until a couple of months ago when I noticed that the BIOS claimed there was now only 1536 MB present ; before it showed the full 2 GB.
Windows XP and HWiNFO32 confirm that there is only 1.5 GB present now.
However, the Crucial scanner shows that slots 1 and 2 are still occupied by 1 GB boards.
All those years nothing has been changed in the BIOS, which still displays all the memory timing settings I made upon inserting the Crucial memory boards.
I do not overclock, and the /maxmem option in the boot.ini is not present.
Still using Windows XP 32-bit.
I read the obvious reasons that can cause the system to declare less memory than is actually present, and executed the obvious remedies, but to no avail.
Removed slot 2 board : it says I only have 512 MB.
Put the removed slot 2 board into slot 1, with no board in slot 2 : still says I only have 512 MB.
Tested each board separately and both simultaneously with Windows Memory Diagnostic - extended test suit : no errors were reported.
So, I'm pretty sure there's nothing wrong with the Crucial memory boards.
What else can cause this, besides the obvious reasons mentioned in the knowledgebase ?
01-04-2018 10:46 AM
@petoetje Thanks for contacting us today. We would be happy to help you troubleshoot this memory further. What is the make and model of the computer? This will help us look into all possibilities of what the problem may be.
01-04-2018 11:14 AM
The PC is entirely self-built (January 2002) : motherboard MSI K7T266 Pro2 Series MS-6380 2.0, AMD Athlon XP 1500+ 32-bit processor.
Quite a few repairs have been made and replacement components added in the past 16 years, all done by myself.
Every change has been meticulously documented, so I can tell you that the BIOS settings are still the way they were just after I added the Crucial memory replacements 8+ years ago (I photographed the BIOS settings then) - two pieces of Corsair DDR 1 GB/PC2700U-25331-Z/DDR333/184-pin/Unbuffered/Non-ECC/DIMM 128MX6/CL2.5 desktop memory with type code CT12864Z335.16TFY.
I suspect something must be failing on the mobo?
01-10-2018 08:46 AM
Might a failing CMOS battery have something to do with the erratic amount of memory modules that go missing? Today half of it went offline, the PC stalled when I tried a restart, and the LED indicators on the mobo indicated the memory detection test had failed. After a cold start minutes later, it worked again and now shows the full memory capacity is available to Windows, but only half of it is shown in HWinfo32?!
The 3 Volt lithium CR2032 CMOS battery in my PC has never been replaced and is 16+ years old, but I do not see any of the other quirks attributed to a depleted battery : the clock isn't losing time, none of the BIOS settings changes (except the amount of available memory that is), and the battery voltage still hovers around 3 Volt?! People on the Net report it lasts between 3-5 years, so I will change the battery ASAP.
01-10-2018 11:09 AM
@petoetje It could potentially be the battery. Please replace and let us know if that solves the issue or not.
01-12-2018 04:18 AM
Yesterday unplugged the PC from the mains, removed the old cell battery from the mobo, left it for more than 1 hour to clear the CMOS RAM, inserted a new battery, restarted the PC, re-entered the BIOS settings anew (no overclocking), and restarted the PC.
Made no difference as to the amount of available memory shown in the BIOS, Windows, and HWinfo32 - still missing 512 GB.
Today, after restart the available memory had dropped even further, to 1 GB instead of 2!
Switched the memory modules, and the BIOS and Windows say 2 GB are available now, but only one memory module is seen in HWinfo32.
CPU-Z shows weird specs for memory module in slot 1, slot 2 memory looks normal (to me). HWinfo2 only shows module in slot 2.
CPU-Z screen dumps are down below.
Before the module switch, slot 1 looked OK, but slot 2 showed the weird specs.
So, is there something wrong with one of the Crucial memory modules after all?!
Next step will be to replace BOTH modules with cheap Chinese ones.
Before I give up.
01-15-2018 04:09 PM
@petoetje Please contact us directly using the info in my signature. We would be happy to help you troublshoot and replace the memory modules if necessary. They are under warranty.
01-16-2018 03:50 AM
January 1st I ordered replacement memory modules from China with the same specs as the Crucial ones I ordered back in November 2009.
I will try these Chinese modules first, to make sure the Crucial ones are faulty.
If the Chinese modules work I'll leave it at that ; 2 GB only cost EUR 8,16 including shipping, and I'm pretty sure sending the Crucial items back (to Britain, where they were bought 11/2009) as a tracked insured parcel alone will cost more than that - the reason why I ordered the new replacements from China. They too boast "Limited lifetime warranty" but that will certainly be a joke.
If they display the same problem as the Crucial ones, then there's something wrong on the mobo.
If the Chinese modules do not work at all then I'll reconsider your offer.
If the replacements last for one more year I will be happy with that ; the then 17-year old PC will be due for replacement next year.
01-28-2018 11:12 AM
My disappearing RAM problem got finally resolved; the cheap Chinese replacements (Hynix chips) work fine for a week now, so the Crucial ones are definitely faulty.
The SPD content on the Crucial module with weird (corrupted) specs shows "CRC error" in Thaiphoon Burner.
What I also learned is not to trust programs claiming to read SPD data; HWinfo32, SPDtool, CPU-Z (especially), and Thaiphoon Burner disagreed on one or more settings - I had to use the official JEDEC standard to make sense.
Just out of curiosity, how can SPD EEPROMS be messed up on quality memory such as the Crucial/Micron ones?
And could it be that only the SPD content contains errors but that the Micron RAM chips are still good?
02-10-2018 06:41 PM
That is a crazy failure! I know it is very rare for Crucial memory to fail and to have two sticks fail in such a way is amazing (frustrating & disappointing for you obviously). I wonder if perhaps you might be starting to have a power supply or motherboard failure which might have somehow damaged your memory.
It is amazing inexpensive no name Chinese memory actually works. I hope it will prove to be stable & reliable in the long run. I would advise running Memtest86 or Memtest86+ on it along with the Prime95 stress test option. I long ago vowed never to use anything but memory from proven name brand vendors such as Crucial. While some companies may use the same memory chips as Crucial, the support circuitry on the DIMMs may be subpar and cause problems. This always seemed to be the case for us years ago when we purchased lower priced memory.
You should contact Crucial Support to see if they will replace your memory under warranty. I believe Crucial has a lifetime warranty on their memory. Due to the testing you have done and proving Chinese memory of all things is working in your system, I think you will be able to get them replaced.
BTW, I forget if you mentioned it in your prior posts, but did you reset the memory settings in your motherboard's System Settings? Try Auto detect and did you try to manaully customize the settings to reflect the memory you have without trying to overclock or increase performance?