I doubt that this motherboard killed your SIMMs. Test with another PSU, preferably more or less fresh ATX one via adapter.
I must be some kind of standard: the anonymous gangbanger of the 21st century.
wrote:You can't test whether it's a fake, you can just test different settings to determine its true capabilities, regardless of what […]
You can't test whether it's a fake, you can just test different settings to determine its true capabilities, regardless of what is claimed by the inscription.
Given the photograph I'm pretty sure the inscription is fake:
- lousy print quality
- non-existent brand name similar to existing one
- nonsensical chip codeOne interesting point: this is a parity SIMM (or at least it pretends to be). Do you have parity on or off in BIOS on the boards? And which So5 board are we talking about here?
It is some intel socket 5 motherboard and it doesn't have parity option in bios. I don't know the model. It seems to detect memory automatically.
The memory modules currently on the motherboard in the picture pass memtest86+ without errors.
and the fake memory modules give hundreds of errors in memtest86+
Both the fake memory modules and the modules currently on the motherboard in the picture are detected by bios as EDO.
BANK 0 SIMM DETECTED EDO MODE
I can't test memory modules individually on the socket 5 board because there is 2 banks of 2 slots so it doesn't boot with just 1 module.
wrote:I doubt that this motherboard killed your SIMMs. Test with another PSU, preferably more or less fresh ATX one via adapter.
Unfortunately I don't have that kind of adapter. This PSU works with other motherboards and has stable voltages. 12.28V and 5.08V when measured from a molex connector when under load.
Not sure if there is any other way for me to make sure the psu good.
wrote:It is some intel socket 5 motherboard and it doesn't have parity option in bios. I don't know the model.
Always a good idea to known what you are playing with. That's an Intel Advanced/ZP board. Very solid and reliable, but not many options.
Here's the manual:
http://tulrich.com/tectrixvr/intel_pba-638995_manual.pdf
Parity generation and detection is
NOT supported.
It seems to detect memory automatically.
Nope, these are SIMMs, not DIMMs, there's no such thing as SPD. It can detect EDO vs FP (which it is doing) but outside of that it's your responsibility as user to supply the correctly-specced SIMMs for your config. To quote the manual again:
For external CPU speeds of less
than 60 Mhz (used with 75, 90 and 120 Mhz processors) memory timing requires 70 ns fast page devices or, for higher
performance, 70 ns EDO DRAM. For external CPU speeds of 66 Mhz (used with 100 Mhz processors) you must use 60 nS
EDO DRAM, but 70 nS fast page DRAM may still be used.
The memory modules currently on the motherboard in the picture pass memtest86+ without errors. and the fake memory modules give […]
The memory modules currently on the motherboard in the picture pass memtest86+ without errors.
and the fake memory modules give hundreds of errors in memtest86+
Both the fake memory modules and the modules currently on the motherboard in the picture are detected by bios as EDO.
BANK 0 SIMM DETECTED EDO MODE
I can't test memory modules individually on the socket 5 board because there is 2 banks of 2 slots so it doesn't boot with just 1 module.
The i430FX chipset was a spectacular performer by 1995 standards, but it managed that by being as inflexible as hell, so no, no way to test individually there. Your 486 board has no such issues as its memory bus is only 32b wide anyway. In your OP you said that there the one started failing, then the other too.
I'm beginning to wonder whether this isn't just ESD damage... Do you have any idea where these SIMMs came from before you had them and how likely it was the previous owner treated them well? Did you handle them (remove from board etc) yourself before the trouble started?
No idea how previous owner treated them. I was given the memory in small plastic bag that was not an antistatic bag and in the original picture I saw they were installed on the motherboard. I installed them with ESD wrist strap and everything worked fine for at least an hour while I was setting up config.sys/autoexec.bat and doing benchmarks and rebooting it several times during that time. Then there was the first sign of trouble when I got memory test fail during reboot and later I also got error himem.sys has detected unreliable memory at address... and I started booting with just one module at the time. So I didn't touch them when the trouble started.
I did just set up the 486 motherboard again for testing. First time it went through post memory test and himem.sys memory test without any problems, but after that I immediately got those errors again whenever I rebooted it. I didn't find any option in the bios settings about parity.
The pentium motherboard has 100Mhz pentium cpu, so I think either 60ns or 70ns should work then. There was no any kind of markings or stickers on it that would have told me what model the motherboard is, so I really appreciate you finding the manual.
I did just put this sacrificial memory module on the 486 motherboard. May it rest in piece.
Right after I turned it on, I did memtest86+ and it passed it. Now I know if it stops working, it is the motherboard that is breaking the modules.
wrote:I did just put this sacrificial memory module on the 486 motherboard. Right after I turned it on, I did memtest86+ and it passed it. Now I know if it stops working, it is the motherboard that is breaking the modules.
It might rather be a problem of the modules being bad. I had a series of 4 16MB sticks (that supposedly came out of some Compaq machine) that would only give RAM beeps in ANY motherboard, and recently I got another 4 sticks that are EXACTLY the same that work perfectly fine in my Biostar MB-8500TVX-A Rev2.3 board.
Definitely try some new sticks.
"Enter at your own peril, past the bolted door..."
Main PC: i5 3470, GB B75M-D3H, 16GB RAM, 2x1TB
98SE : P3 650, Soyo SY-6BA+IV, 384MB RAM, 80GB
wrote:No idea how previous owner treated them. I was given the memory in small plastic bag that was not an antistatic bag and in the original picture I saw they were installed on the motherboard. I installed them with ESD wrist strap and everything worked fine for at least an hour while I was setting up config.sys/autoexec.bat and doing benchmarks and rebooting it several times during that time. Then there was the first sign of trouble when I got memory test fail during reboot and later I also got error himem.sys has detected unreliable memory at address... and I started booting with just one module at the time. So I didn't touch them when the trouble started.
The trouble with ESD is that it can manifest itself (long) after the damage is done. Sounds like you took admirable precautions, but whoever put the SIMMs in a regular plastic bag didn't. Not that a bag in itself will kill stuff, but I sort of get an image of someone wearing rubber-soled sneakers on a synthetic carpet with polyester sleeves (and hell, maybe a woollen jumper just for good measure) dumping them in there...
I generally can't be bothered with wrist straps, but ensure I'm wearing cotton, linen and/or leather (NO wool or synthetics), with conducting soles. Also I make sure to touch the case of any system I'm working on before adding or removing components. Oh, and components rest on paper, wood or antistatic bags. Not clean-room safe, but at least as good as wherever else the old parts have been.
I did just set up the 486 motherboard again for testing. First time it went through post memory test and himem.sys memory test without any problems, but after that I immediately got those errors again whenever I rebooted it. I didn't find any option in the bios settings about parity.
But the 486 did work fine with the 'sacrificial' SIMM? I'm 99.99% sure that board isn't the problem.
The pentium motherboard has 100Mhz pentium cpu, so I think either 60ns or 70ns should work then.
Nope, see the bit from the manual I quoted. A 100MHz Pentium runs at 1.5x66MHz, and 66MHz FSB specifically requires 60ns EDO. 70ns is only OK if it's FP. Simple test: set the board to run at 1.5x60MHz (underclock the CPU to 90MHz) and re-test with the suspect SIMMs. If they're good 70ns modules, they should work reliably then. If not (as I fear) you can drop it down to 1.5x50MHz (i.e. CPU at 75MHz), but as your 486 was already having trouble at something lower than that (33MHz-50MHz) I doubt that would make much difference...
There was no any kind of markings or stickers on it that would have told me what model the motherboard is, so I really appreciate you finding the manual.
You could probably google some or other part number, but for Intel OEM boards (which this clearly is), I generally just take note of the overall layout (number of slots, position of chipset & socket) and then nose around TH'99. Once I spot a likely match, I google up a good picture to double-check. That's how I figured out this one was an Advanced/ZP.
I am starting to wonder if the 50ns ram was actually real because the pc seems much slower with that one 8Mb 70ns stick that I did put there as sacrificial ram. Doom benchmark was around 2000 realticks when I first benchmarked the system and with that ram I am getting around 2400 realticks.
I will still wait a while before I will put any of the 60ns edo modules I have to make sure.
It does not seem like the motherboard is breaking that module. I also realised that it does not matter what ram I put on the motherboard as long as the memory settings in bios are set to fastest. I get the same results in benchmarks no matter if the motherboard has 60ns edo or 70ns fpm ram.
