Memory fault XT computer

[Robin4]

After installing some parity ram to the motherboard (which was a good idea to do) my system is now halting on something..

On screen it says: Ram Parity Chk 1 CS:IP,DS,ES= C800:0750, 0000, 0BA5

I think he`s noted it because there is some ram bad i guess.. But iam not familiar off puzzle where this bad ram chip would be..

Can someone with more experience help me out.. Because still conflicting on the ram failure isnt the best thing to want.

 

[Chuck(G)]

That CS:IP puts you squarely in the ROM of the hard disk adapter; judging from the DS value, it's during initialization, ES puts it in somewhere in the 48K range, so the first bank of RAM--possibly/b]. More likely, it's a fault in your hard disk controller. Can you boot without the hard drive card inserted--and do you get the message?

 

[H-A-L-9000]

Once again:

The address reported by the parity error handler is not very likely where the memory problem actually happened. The error handler can only tell the address of the instruction executed when the error occurred. This instruction can reference to a very different address.
Also we have a prefetching CPU which means even if the instruction fetch caused the problem the address may be some bytes off.

In this case it was certainly not the instruction fetch because there is no on board parity checking happening for ROMs. The error message is no hint.

Did you install parity RAM but left the related regular RAM sockets unpopulated? Might be a problem...

The harddisk BIOS' business to access RAM is:
- BDA for parameters
- at the position to read or write harddisk data
- the stack
- maybe EBDA (private parameter space at the top of RAM)


If you can get the computer to boot by some means try a memory checker tool. If not, remove memory and add back banks one-by-one and check each time if it works. Or you can try the Supersoft ROM.

 

[Chuck(G)]

Once again:

The address reported by the parity error handler is not very likely where the memory problem actually happened. The error handler can only tell the address of the instruction executed when the error occurred. This instruction can reference to a very different address.
Also we have a prefetching CPU which means even if the instruction fetch caused the problem the address may be some bytes off.

Permit me to expand on this.

The CS:IP indicates the location of the instruction that produced the error (C800:0750). In this case, the address lies within the ROM at segment C800, which is usually part of the hard disk adapter's BIOS extension ROM. The ROM isn't parity-checked, so you know it's not that. That leaves the possibility that the error occurred within 64KB of the XT zero segment (DS register) or within absolute locations 0BA50-1BA4F (hex) in memory.

It could also have occurred within the stack segment (not dispalyed) memory range. It could have also occurred as a DMA access (the XT hard disk controller iuses DMA for data trasnfer).

And finally, it could be none of the above--the message could be the result of a wild program jump.

What we do know for certain is that the system was executing an instruction in the hard disk BIOS when the interrupt hit or was entered. Which is why I said to pull the hard disk controller and see if you can get the thing to happen again. It might be a problem in the HDC.

 

[Robin4]

That CS:IP puts you squarely in the ROM of the hard disk adapter; judging from the DS value, it's during initialization, ES puts it in somewhere in the 48K range, so the first bank of RAM--possibly/b]. More likely, it's a fault in your hard disk controller. Can you boot without the hard drive card inserted--and do you get the message?

I had two controllers in it.

The first was a WDXT-150 controller card with an XT 40MB harddisk to it at adress C800 (in rom)
The second was an Seagate ST11R *rll* controller with a NEC D5126 20MB harddisk (i think previous owner formatted it in rll i guess because now its sees 30MB capacity. D800 (in rom)

Yes the system just can boot.. But then there is no message. (only happens when the XT IDE harddisk is installed) Maybe it fails because the harddisk is almost worn out.. (makes high noise (looks like a coil) (or because the head is never been parked)
I dont have the parkide tool for it..

I found the parking tool here:

http://www.filewatcher.com/m/parkide.exe.1024-0.html

confirms its working!