VOGONS


First post, by Malik

User metadata
Rank l33t
Rank
l33t

Just a quick note on the memtests I did:

Motherboard : DFI CA61 Socket 370
CPU : Intel PIII 1GHz Coppermine processor
FSB @ 133MHz
Chipset : VIA Apollo 133 Pro

EDIT: Sticks used - RAM: 128MB PC133 & 256MB PC-133 sticks, and a PC-100 128MB stick (for testing) - tested individually, only one module running at a time.

Scenario : Windows ME started behaving strangely -

Error Message : "windows encountered an error accessing the system registry. Windows will restart and repair the system registry for you".

After consulting Microsoft's Knowledge Base, I tried checking the memory using the memtest. Fair enough, errors were detected.

But then, I tried using many sticks but all showed similar errors at the EXACT point of tests, beginning with Test 3.

Thinking something was wrong, and after taking into account my previous experiences with this particular motherboard, I tried reducing the bus speed to 100MHz.

Using the same sticks, at the same slots, MEMTEST passed all the tests SUCCESSFULLY.

Makes me wonder :

Either and/or :

1. My motherboard is faulty,
2. VIA's Apollo Pro 133 chipset is half baked.

I also wonder if slot problems will also bring out errors, even if the sticks themselves are clean.

And also, I wonder if we can feel snug and comfortable if MEMTEST passes all the test. Should we run the tests again and again to make sure the stick is not faulty? It's quite exhausting to run the memory tests.

Last edited by Malik on 2010-09-17, 15:11. Edited 2 times in total.

5476332566_7480a12517_t.jpgSB Dos Drivers

Reply 5 of 10, by sgt76

User metadata
Rank Oldbie
Rank
Oldbie

Having a couple of said VIA chipsets myself, I'm inclined to say implementation was half baked with the quality of these boards varying greatly. For instance my Matsonic MS7112C supports 133fsb bus speeds, but you can only run the memory asynchronously at 100mhz. Doing so at 133mhz will result in all sorts of wierdness.

On the other hand my AOpen AX33 with the same chipset not only runs at 150mhz bus speed rock solid, but can also crank any creaky old PC133 ram at 150mhz CAS2 Memtest stable....

Reply 6 of 10, by Tetrium

User metadata
Rank l33t++
Rank
l33t++

If you tested separate DIMM's in separate slots on the same board, I'm inclined to think the motherboard is toast.
I've seen this behavior once before (intel 815 board) except that it would see only 64MB of any 256MB DIMM I inserted.

If it were just 1 DIMM slot I could reason one of the slots might have been dirty or something.

Reply 7 of 10, by Malik

User metadata
Rank l33t
Rank
l33t
Tetrium wrote:

If you tested separate DIMM's in separate slots on the same board, I'm inclined to think the motherboard is toast.
I've seen this behavior once before (intel 815 board) except that it would see only 64MB of any 256MB DIMM I inserted.

If it were just 1 DIMM slot I could reason one of the slots might have been dirty or something.

There are two points I would like to point out:

1. MEMTEST showed the errors at exactly the same point with all the sticks at FSB 133MHz.

2. The are no more errors using MEMTEST when FSB is reduced to 100 MHz. Using the same sticks at the same slot.

I also later tried using TWO 128MB PC-100 sticks running concurrently at FSB 100MHz. No problems too.

Either THIS board is not stable at FSB 133 MHz, or the VIA Apollo 133 Pro chipset itself is at fault when it comes to the max FSB, again, with this board - as sgt76 had experienced the disparity.

There's also one more thing I've yet to try - the BIOS' CPU/PCI clock setting :

dsc00414o.jpg
It's currently running at default. Maybe I just have to specify it to run at 133 MHz, rather than leaving it at default.

5476332566_7480a12517_t.jpgSB Dos Drivers

Reply 8 of 10, by Mau1wurf1977

User metadata
Rank l33t++
Rank
l33t++

Have you tried lowering the timings when the Ram runs at 133?

Your issue does point towards a board / chipset issue rather than memory. However I found that usually I get line after line of red lines if a board / chipset can't handle a certain setting.

Reply 9 of 10, by sgt76

User metadata
Rank Oldbie
Rank
Oldbie

Malik, your DFI's bios looks the same as my Matsonic's! So maybe just let me share 2 more points about this if I may:

1) CPU/ PCI clock settings - From my experience this doesn't do anything when set from default to "4" - used to run the memory asynchronously using the "3" setting so that shitty PC66 and PC100 ram back in the day could still be used mainly for overclocking Celerons.

2) Your DFI should be able to run PC100 at "Turbo" timings and CAS2 which should be faster than PC133 with safe/ slow timings. This may be the best bet if you can't get it to run at PC133.

However, I'm myself puzzled as to why your DFI can't handle PC133 speeds. Should be able to as that was quite a good board back in the day- unlike my Matsonic which has always been budget stuff.

All being said and to be fair to this chipset, a motherboard with a properly implemented Appollo Pro chipset is a fearsome beast indeed and more than capable of competing with 440BX/ 815 chipsets. You get proper AGP dividers, ability to run ram asynchronously, up to 1.5gb ram support, etc. I nearly gave up on this chip set till I found the right board.

My Aopen AX33 is running a 933mhz cD0 at 1050mhz, 150mhz bus speed and 150mhz ram at 2-3-3-4 rock solid, so it might just be that the board manufacturers rather than VIA itself were to blame.

Reply 10 of 10, by Malik

User metadata
Rank l33t
Rank
l33t

Yes, I think it has to do more with this particular board. Maybe some bios updates (or downgrades) might help, to fit in this situation.

The only reason I started checking all this is because of that Windows ME registry problem. There are no problems in Windows 95 and MS-DOS 6.22. Looking back at certain dos cd based games, the random lockups might be even related to this problem, rather than downgrading the bios (which has resolved the problem anyway).

5476332566_7480a12517_t.jpgSB Dos Drivers