• Please review our updated Terms and Rules here

Checksum error on a Toshiba T1200

nintenloup

Member
Joined
Apr 3, 2017
Messages
28
Location
Québec
Hello !

One month ago, I bought a T1200, it was working perfectly. It's only problem was that the HDD gets stuck after some time, but nothing a little swing didn't resolved. A week ago, I booted the computer and it showed me : A checksum error occurred in the configuration RAM

It asks me to select if the computer is either a 2 floppy or 1 floppy + 1 HDD system, so I choose the option with HDD ( I also tried the other one) and then the computer reboots and shows the same errors and option.

At first I was just thinking that one of the 2 internal batteries that went flat, so I got a rechargeable Ni-Mh battery from a cordless phone. I got it today and plugged it in, but to no success, the computer is still telling me checksum error and reboots to the same error after I choose an option.

The other internal battery still is over 4v and the computer booted without it anyway when I was cleaning/testing it when I got it, so that battery shouldn't be the problem. I read the manual multiple time on minuszerodegrees but it doesn't say anything useful about that. I hope it's not bad ram that's causing this as everything is SMD in there. I dismanteled and disconnected everything too just to be sure it wasn't a bad contact.

Thank you in advance.
 
I was helping him in the #vc IRC chat and yes the battery is getting a charge voltage when the computer is powered on.
Are you absolutely positive it will not let you boot from a floppy disk? The inability to use the test/setup floppy really complicates the diagnostics.
 
I was helping him in the #vc IRC chat and yes the battery is getting a charge voltage when the computer is powered on.
Are you absolutely positive it will not let you boot from a floppy disk? The inability to use the test/setup floppy really complicates the diagnostics.
Yes I am, the computer doesn't try to access the floppy drive, it just reboots and ask me the same question again and again.
 
It's so weird. There's not a lot of computers that lose the CMOS defaults on a reboot unless the power management completely cycled the board.
Likewise I've not seen a CMOS just up and die like that. Are you sure you are not looking at trace damage?
 
It's so weird. There's not a lot of computers that lose the CMOS defaults on a reboot unless the power management completely cycled the board.
Likewise I've not seen a CMOS just up and die like that. Are you sure you are not looking at trace damage?
There's a place on the PCB where there's a thick plastic band on the PCB close to the ram, it's foggy and I can't really see the state of the traces. I'll get some googone and check if there's any damage under it.
 
Update : I've took the plastic band out and the traces under it look as perfect as the rest of the board. I also touched all the chips and they're all cold.

I'm really stuck as to what to do there.
 
Hi nintenloup,

Here are two ideas that could maybe help:

According to maintenance manual, jumper PJ17 is used to configure if it is a model with 2FDD or FDD + HDD. Maybe play with it, if it helps. I would also suggest you to play without BIOS battery.

You can also try to see what output you get from parallel port. If I am not wrong, it should show the binary error code in parallel port, but you will need to have some kind of port checker. You can easily build it yourself...it is just diode in serie with for example 470 Ohm resistor on each on the data lines.

Cheers,
Branko
 
As promised on Reddit I have been digging through my archive this weekend but have not found anything useful yet.

I'm with Branko on checking the jumpers and using a parallel port adapter to read the fault codes (reference maintenance manual p2-12)
In addition to that, check the voltages on the different connectors are within spec (see maintenance manual p2-8).

I did find my disassembly of the BIOS ROM and have located the subroutines triggering this error but I'm not sure what is being checked that triggers it. It seems to write the result of your choice to F000:0214h and F000:0215h but that segment just doesn't make sense to me.
Still digging through the assembly though.

By the way, as far as I'm aware these machines do not have a CMOS nor a CMOS/BIOS battery. The 2 internal batteries are for the SRAM used for the suspend function and for keeping the RTC powered (timekeeping).
Please ignore my total brainfart above, the RTC is what contains the CMOS configuration data.
It could be the RTC chip is damaged so new configuration values can't be written to it.
Wouldn't know how that would happen other than a wrong (too high voltage) battery though.
 
Last edited:
Something came to mind late last night.
On Reddit you mentioned you replaced the RTC battery (by which I assume you mean the brown 2-cell one) with a "3.6V cordless phone NI-Mh battery".
That might be too high a voltage for the RTC, the original battery is a 2.4v 50mAh Ni-Cd battery (as per this picture).

Also see this page for a reference of the batteries and cap values.
Note that the "SUB" battery is incorrectly named CMOS battery, it is not used for CMOS but for the suspend function to the on-board SRAM chip.
 
Was talking to nintenloup tonight as everyone else in the IRC is currently on the way to VCF and he was tracing the path of the battery and saw it runs to a BS108 switching MOSFET and a (P)ST531 low-power battery controller. His quick check found no voltage at the Source pin of the BS108.

What I'm suspecting is that this circuit switches the CMOS power from the 5v rail when the power is on and the CMOS battery is charging to the CMOS battery when the main power is turned off. In theory if power never makes it to the CMOS it would never retain anything on a warm boot. I've seen similar on other devices but usually it was a flat or missing CMOS battery.
It would be really nice if I knew batter hw this circuit worked....
 
Alright, I have some updates with this.

I finally found the RTC chip not thanks to the manual. It get's 2.4v when the computer is off and 5v when on. I used my logic probe on the RP5C01 and it seems to be accessed and written to when powered on and reseted after I press enter respectively. The RTC battery circuit seems to be working correctly if I understand this remotely correctly.

It doesn't makes me less lost though, I dunno what's (not) happening in there as I don't have a scope and can only rely on what the flickering LEDs on the logic probe does. Would replacing the RP5C01 be the best try I can do with the information I have ? As this chip has 26x4bits ram it could be that some of it died and is the reason why the checksum error is triggered ?
 
It's so weird. There's not a lot of computers that lose the CMOS defaults on a reboot unless the power management completely cycled the board.
Likewise I've not seen a CMOS just up and die like that. Are you sure you are not looking at trace damage?
There are many. Ask me how I know... :ROFLMAO: Often, the CMOS is powered by the battery only, even when the system has power. I ran into that issue so often I can't even count. You won't always notice it is this way if the defaults match the actual hardware, but if the system e.g. defaults to a 5.25" DD disk drive while you have a 3.5" HD drive installed, you can't even boot from floppy - as in his case.
 
Have you tried with parallel port adapter?
I'm not sure what the parallel port POST reader would do. To me it seems like the CMOS logic passes POST, but the CMOS will not remember anything written to it, which is a very bizarre fault that the computer itself isn't detecting because it's not smart enough to remember that it just initialized the CMOS and the contents should be valid.
 
Hello !

One month ago, I bought a T1200, it was working perfectly. It's only problem was that the HDD gets stuck after some time, but nothing a little swing didn't resolved. A week ago, I booted the computer and it showed me : A checksum error occurred in the configuration RAM

It asks me to select if the computer is either a 2 floppy or 1 floppy + 1 HDD system, so I choose the option with HDD ( I also tried the other one) and then the computer reboots and shows the same errors and option.

At first I was just thinking that one of the 2 internal batteries that went flat, so I got a rechargeable Ni-Mh battery from a cordless phone. I got it today and plugged it in, but to no success, the computer is still telling me checksum error and reboots to the same error after I choose an option.

The other internal battery still is over 4v and the computer booted without it anyway when I was cleaning/testing it when I got it, so that battery shouldn't be the problem. I read the manual multiple time on minuszerodegrees but it doesn't say anything useful about that. I hope it's not bad ram that's causing this as everything is SMD in there. I dismanteled and disconnected everything too just to be sure it wasn't a bad contact.

Thank you in advance.
Hi, did you ever solve this? I have a T1200 that does exactly the same thing. I've ultrasonically cleaned all the boards, but haven't recapped it yet. Is there a way to reset the CMOS via the Parallel port diags?
 
Hi, did you ever solve this? I have a T1200 that does exactly the same thing. I've ultrasonically cleaned all the boards, but haven't recapped it yet. Is there a way to reset the CMOS via the Parallel port diags?
Sadly, no, got stuff happening in my life that's preventing me from going much further.
 
Back
Top