SM20 help, What do you think?
gtghm
New
Ok, I wanted to start a new thread because I thought it better for trouble shooting.
Here is what I did:
I re-soldered the cap again and now I'm almost back to zero. At least it doesn't go into a constant reboot mode.
I was inspired to try again by GHoosdum who mentioned that the board was fine when he took it out of his sisters machine.
Here is what its currently doing, maybe we can put our heads together and come up with the right soultion.
It wants to boot, I can tell. When I boot into the OS I get the black XP load screen, then it goes black but never quites get into the OS, it just stays at a blank black screen.
So I ran some memtest loops. I come up with some interesting results.
If I leave the CPU L1 and L2 cache enabled in the BIOS and run memtest full set of tests it locks consistantly on either test 4 or 5, both of which are cached. If I run the test with out cache meaning that it only loops tests 1,2, and 7 it will loop through with out any errors or lockups.
If I disable the cache in the BIOS and run the full set of tests it locks on the same tests, if I disable the cache option and it runs tests 1,2&7 it will loop fine, slowly, but it will loop.
If I slow the clock speed to 100Mhz I get the same results but slower, and if I boot it at XP 1600 I get the same results. Also I get the same sucesses and failures no matter which dimm slot I use.
The pattern I see is any time cache is used or tested it locks up.
According to the hardware monitor all voltages are normal and the CPU temp is around 30C. So its not an overheat problem.
I'm asking for all of the good PC doctors here to mull this because I'm sure that everyone that sent parts are sure or fairly sure that what they sent was working. I try to take every precaution when building rigs, but I am human and could make a mistake like everyone...
Not wishing to rial anyone or be accusing I was thinking that if I could out line my testing we might arrive a general consensis.
My thought is since there seems to be a direct corralation with the L1 and L2 cache and that its happend consistantly since I put the machine togther, that its the CPU not the memory or the board.
I can't test out my theroy tonight because I don't have a spare AMD CPU on hand... But either tomorrow or the next day I should be getting an AMD rig to look at that I might be able to use that CPU for trouble shooing in SM20 but I'm 100% positive that I will get the machine to look at.
In any case even though I don't have a CPU today/tonight to test I wonder what you guys think based on my testing?
I can try almost anything that you guys want except for swaping the CPU out right now. If I can get my hands on that other CPU I'll try it and post the results asap.
In the mean time I'm courious as to what you think.
"g"
Here is what I did:
I re-soldered the cap again and now I'm almost back to zero. At least it doesn't go into a constant reboot mode.
I was inspired to try again by GHoosdum who mentioned that the board was fine when he took it out of his sisters machine.
Here is what its currently doing, maybe we can put our heads together and come up with the right soultion.
It wants to boot, I can tell. When I boot into the OS I get the black XP load screen, then it goes black but never quites get into the OS, it just stays at a blank black screen.
So I ran some memtest loops. I come up with some interesting results.
If I leave the CPU L1 and L2 cache enabled in the BIOS and run memtest full set of tests it locks consistantly on either test 4 or 5, both of which are cached. If I run the test with out cache meaning that it only loops tests 1,2, and 7 it will loop through with out any errors or lockups.
If I disable the cache in the BIOS and run the full set of tests it locks on the same tests, if I disable the cache option and it runs tests 1,2&7 it will loop fine, slowly, but it will loop.
If I slow the clock speed to 100Mhz I get the same results but slower, and if I boot it at XP 1600 I get the same results. Also I get the same sucesses and failures no matter which dimm slot I use.
The pattern I see is any time cache is used or tested it locks up.
According to the hardware monitor all voltages are normal and the CPU temp is around 30C. So its not an overheat problem.
I'm asking for all of the good PC doctors here to mull this because I'm sure that everyone that sent parts are sure or fairly sure that what they sent was working. I try to take every precaution when building rigs, but I am human and could make a mistake like everyone...
Not wishing to rial anyone or be accusing I was thinking that if I could out line my testing we might arrive a general consensis.
My thought is since there seems to be a direct corralation with the L1 and L2 cache and that its happend consistantly since I put the machine togther, that its the CPU not the memory or the board.
I can't test out my theroy tonight because I don't have a spare AMD CPU on hand... But either tomorrow or the next day I should be getting an AMD rig to look at that I might be able to use that CPU for trouble shooing in SM20 but I'm 100% positive that I will get the machine to look at.
In any case even though I don't have a CPU today/tonight to test I wonder what you guys think based on my testing?
I can try almost anything that you guys want except for swaping the CPU out right now. If I can get my hands on that other CPU I'll try it and post the results asap.
In the mean time I'm courious as to what you think.
"g"
0
Comments
Another thought I had; when the other amd rig is in your hands you can drop that 1600 into it to see if it runs ok on that other rig.:)
I was able to loop tests 1,2, and 7 (the no cavhe tests) sucessfully for over 4 hours, so I really think that the CPU is the problem and not the MOB.
I was hopeing that some one else might read this and agree.
BTW, Before I can run your KG7 I'll need a Video card and a NIC too. I don't know if those are on the way yet or not.
Thanks,
"g"
Update, I am getting access to another AMD CPU today, I will try it and post the results later. Can we hold off until I try this other CPU?
"g"
--->//EDIT\ADD: Note, one other situation can cause a blank screen for aLONG time or infinitely.... For this to be true, ALL must be true:
Older monitor, misaimed slightly to LEFT, or misaimed monitor aiming too far to LEFT.
CD in CD-ROM drive.
BIOS set in boot order to boot from CD-ROM before booting from HD.
See. BIOSs, when checking for a CD boot sector, WILL show one little thing in upper left hand corner and nothing else.... A little underline cursor, flashing.... Misaimed monitor makes littel cursor show off left side of monitor, you never see it. Very dirty CD-ROM drive or scratched CD can make this seek take up to 15 min....
Solution in this case, easiest, is to first alter boot order so HD boots before CD-ROM, then if still no go, power down, and unplug IDE connector and power connector from CD-ROM drive.
John.
So send the new parts as soon as possible and I'll get her going again.
Mr K, I'll box up the ram and send it to you either tomorrow or the nextday.
Sucks, I really thought I could fix it.
"g"
I am truly sorry for letting the team down like this.
--Rob