Frequent Crash of Windows XP

[RT]

Hi,

I am facing the problem very often especially at one specific
location...

As mentioned in various posts.. we also suspected the hardware and we
have changed the hardware (infact did it quite a few times). We have
now replaced the complete machine of different brand etc.

The same hardware is running in my office perfectly fine. I have put
the machine on the raw power and left it running for few days. It
works even when the machine is abnormally switch off (due to power
loss) multiple time in a day.

Where as when the same machine with same configuration is running in
the field it carshes very offten with the error messages similar to "C:
\Windows\System32\Config\System is missing or corrupt".

At times I am able to recover the disk from the crash.

I am basically looking for the answer that why does this machine
crashes, where as we do abnormal swtich off (without proper shutdown)
of the PCs in office / home, they don't crash. They why this machine.

Two configurations of the machine which we have tried are -
1. Via C7 Mother board with 1 GHz processor, 1 GB RAM, 4 GB Disk On
Module
2. Intel Borad with 1GHz processor, 1 GB RAM, 4 GB InnoDisk Disk On
Module (HDD), multiple serial ports (4 in number).

Operating System: Windows XP Professional SP2 without any hot fixes/
updates (earlier we were using Windows XP Embedded SP 2 with Feature
Pack 2007, we switched to XP Professional thinking we are missing
something in XP Embedded configuration)

We have tried different make of DOMs, suspecting something to do with
the quality of DOM. We have also tried switching of the "Write Cache
on Disk" option of Windows.

The application that we are running is basically polling multiple
devices on the periodic basis, every few milliseconds. It then writes
the data to SQL Express Database. Application is developed using MFC.

Any answer / tips are welcomed.

Thanks
Rajesh

 

[Bob I]

Perhaps you have a power voltage problem in the field?.

 

[Bob I]

Unless you can provide evidence of other environmental factors that
would explain the reason for the difference, sagging voltage is the only
thing that might explain the problem.

 

[Paul]

One major difference is also that these systems are in highly tough
environment (that is gas station).

There is no Air conditioning, the machine is kept in an Networking Rack
enclosure which has a fan. No one access the machine directly, it gets
switch on as soon as UPS is turned on every day in morning. In the
night when Gas station is closed, the machine is shutdown by pressing
the power switch once, this way windows initiate the shutdown
sequence.

I hope this information may help.

Best regards
Rajesh

Have you considered the type of UPS you are using ?

http://www.apcmedia.com/salestools/SADE-5TNM3Y_R5_EN.pdf

The simplest kind of UPS is the "Standby UPS" in Figure 1.
I have one of those, and one thing to note about those,
if there is a lot of variation in source AC line voltage, that
variation is passed right through the UPS unchanged. The
"standby" UPS is a "straight piece of wire" when not
running on battery. The Standby type UPS is also the
cheapest type of UPS you can buy.

It could be that your problems are related to power, and
could be corrected with a different kind of UPS.

The double conversion UPS is an AC to DC to AC conversion
box. By doing so, the output AC is virtually unaffected
by what is happening on the input. It means a fast
transient is much less likely to get through to the
computer and affect it. You might experiment with
something like that, to see if it makes a difference.
The double conversion UPS runs hotter than the standby
UPS, and some are equipped with a fan to cool them.

If you don't want to buy one, try renting one.

Also, you can get power analyzer equipment, that can
record how "clean" the power is, coming out of the UPS.
For example, the device here records "transients, sags, swells",
and perhaps if you rent a similar device, you could get
some idea how bad the power coming out of the UPS is.
A power utility may use devices (older models than this one),
to record problems in industrial power situations. Again,
you may be able to rent one.

http://www.powersight.com/products/inventory_details.cfm?id=60

Other possible problems are a strong source of
interference affecting the motherboard. Or a grounding
problem, with respect to the remote sensors or devices
the motherboard is connected to.

Since you have two systems (VIA C7 and Intel board), you
could run the systems in parallel at the gas station, and
experiment with the one not being used by the business.
Run them both off the same UPS. Don't connect the serial
ports, on the "experimental one", and see if it becomes
more stable. Or, connect a different kind of UPS to the
experimental one, and see if the symptoms improve
on that one.

To simulate a serial port, you might use one serial port
as a source of bytes, and send them to a second serial
port, all under the control of the same motherboard.
That way, there is no electrical interference (since
the cable could be very short, and connected between
serial ports). That way, you could continue to simulate
the application, while looking for the problem.

New_UPS ------- Test_computer Use one serial
| ^ port to test
| | another, to avoid
+--------+ ground problems.

HTH,
Paul

 

[Bob I]

I have appended your original post to this so as to keep all the
information together, please include previous if you reply. If you
verify that the power remains good the entire time, then the next
possible issue that you face is the problem of heat. If the system
temperature in the cabinet is too high then you are going to have
problems too. Can you document the temperature in the cabinet?

One major difference is also that these systems are in highly tough
environment (that is gas station).

There is no Air conditioning, the machine is kept in an Networking Rack
enclosure which has a fan. No one access the machine directly, it gets
switch on as soon as UPS is turned on every day in morning. In the
night when Gas station is closed, the machine is shutdown by pressing
the power switch once, this way windows initiate the shutdown
sequence.

I hope this information may help.

Best regards
Rajesh

Hi,

I am facing the problem very often especially at one specific
location...

As mentioned in various posts.. we also suspected the hardware and we
have changed the hardware (infact did it quite a few times). We have
now replaced the complete machine of different brand etc.

The same hardware is running in my office perfectly fine. I have put
the machine on the raw power and left it running for few days. It
works even when the machine is abnormally switch off (due to power
loss) multiple time in a day.

Where as when the same machine with same configuration is running in
the field it carshes very offten with the error messages similar to "C:
\Windows\System32\Config\System is missing or corrupt".

At times I am able to recover the disk from the crash.

I am basically looking for the answer that why does this machine
crashes, where as we do abnormal swtich off (without proper shutdown)
of the PCs in office / home, they don't crash. They why this machine.

Two configurations of the machine which we have tried are -
1. Via C7 Mother board with 1 GHz processor, 1 GB RAM, 4 GB Disk On
Module
2. Intel Borad with 1GHz processor, 1 GB RAM, 4 GB InnoDisk Disk On
Module (HDD), multiple serial ports (4 in number).

Operating System: Windows XP Professional SP2 without any hot fixes/
updates (earlier we were using Windows XP Embedded SP 2 with Feature
Pack 2007, we switched to XP Professional thinking we are missing
something in XP Embedded configuration)

We have tried different make of DOMs, suspecting something to do with
the quality of DOM. We have also tried switching of the "Write Cache
on Disk" option of Windows.

The application that we are running is basically polling multiple
devices on the periodic basis, every few milliseconds. It then writes
the data to SQL Express Database. Application is developed using MFC.

Any answer / tips are welcomed.

Thanks
Rajesh

 

[CNC]

I would inspect the placement of the system, incoming cables & routing. Any
chance or EMI or high frequency noise coming into the system? Are the
incoming cables routed with or close to electrical wiring or flourescent
lighting?