Power Loss With Surge Protector

Dell have disabled the Windows RAC task.
We'll see what happens now

 

Thats didn't work.
Back on the phone

I'm going to try running on OpenSuse Live for a few hours to see if get any reboots.

 

You did not mention the ethernet on power line. Every wire (including where various devices get power) can be significant for inane reasons. That ethernet connection is a potential suspect (but should be made irrelevant by galvanic isolation provided by the Dell). I do not know enough about the internal design of an ethernet over power device to completely eliminate from the suspect list. It should not cause a problem. But I can think of rare electrical anomalies where that ethernet device can cause a system crash.

I meant to say 30 degree C (not F) room. 30+ degree room combined with a running hardware diagnostic has located (and therefore permanently solved) many a strange computer failure. In fact, I always test my every hardware in a high 30 C room with comprehensive diagnostics so that failures do not happen later. Any computer must run everything perfectly happy in a 40 degree C room - which locates defects that may cause crashes months or years later. Temperature is a powerful diagnostic especially to isolate or provoke intermittent failures.

For test purposes, it is advisable to connect monitor, computer, and Ethernet on power device to a common power strip. To eliminate (and therefore locate) a defective household wiring problem. If that 'single point ground' test makes failures disappear, then a problem is traceable to household wiring or a combination of wiring mistakes combined with another appliance 'defect'. A problem that an outlet tester typically would not locate. One of those inane problems that makes life interesting. Do it just to eliminate that strange possible suspect or to find a suspect made obvious by how power (especially safety ground) is provided. Power can be perfectly fine. But in combination with something otherwise perfetly normal inside the machine, therefore causes this intermittent.

Try powering everything from a common 3 quid power strip (that only has a circuit breaker).

BTW, I was serious - very serious - about the voltage measurements even one hour later reporting a voltage problem. If voltages are causing a sudden failure, then a 'defective' safety margin is obvious both before and after power off happens. But your load and your numbers says your margins (at least at that time of day) are more that sufficient. Power supply is out of it (but not power supply controller).

To power off a machine suddenly, only functions that talk to the power supply controller can do that. Most all other failures result (worse case) in a BSOD message. I was hoping that numbers and error message might happen again. Hopefully it might point to a HAL level software that can order the power supply controller to power off.

Appreciate that this software may have been completely rewritten recently (or may be in the near future) because of how CPU hardware security is changing again. (another reason why the service tag number may be helpful).

What is the service tag number? Important for details that might zero in on the suspect.

Just thinking while typing - measure voltage of the gray wire that goes from power supply to motherboard. This 'Power Good' signal should be well above 2.4 volts (less than 5) to have a sufficient and stable margin. Just a long shot that might explain why the power supply controller suddenly removes power.

If not, if BSOD numbers are not appearing, and if failure still happens with everything on a common power strip, then the only remaining suspect is software. Which is why a service tag number can be helpful.

Well, leaving the diagnostics running overnight every night might be useful. It sudden shutdown is due to hardware (ie an intermittent front panel power switch), then it would also happen when running diagnostics. If sudden shutdown is due to Windows, then it would never happen when diagnostics are looping constantly overnight.

We are almost at the point where using the System Recovery disk may be necessary to completely wipe the disk clean and to reload the HAL level software - to replace software that talks to the power supply controller.

Hopefully you are following these various thought lines in order to appreciate what actually happens inside. Your power off is somehow related to the power supply controller. Inputs for that include BIOS and Windows softwares, power switch, Power Good wire, some signal wires to various peripherals (ie that make 'Wake Up' functions work), and some otherwise 'should not be' current flows (ie created by anomalies on the safety ground or via that Ethernet on Power device).

 

That still makes absolutely no sense. What does an hour here or there have to do with anything. AC power cycles through 0 volts two times every second. An hour later may as well be 10 years later or a minute later (unless you are in the middle of a thunderstorm) or a cheap camping generator or the like. An anomaly now does not mean you had one an hour ago, or that you will have one an hour from now. Anomalies typically last a few cycles at most - usually a 1/2 cycle, as a spike, or a dip.

Your comment suggests it takes an hour for power to stabilize after an anomaly. The power grid does not work that way, nor do power supplies.

 

All my deviced are now powered from the same power strip / surge protector. I downloaded OpenSuse Live CD and ran this for several hours without any problem. I rebooted into windows and within 2 minutes it lost power and rebooted. I know this does mean much but it is coincidence? This was done with the room at 30C~.

Windows 7 has been re-installed, no issues, yet.

I am going to install ShadowProtect 4 and install drivers & apps one at a time so I can identify any offending apps/drivers.

 

Maybe there was a conflict with Windows APM.
Maybe OpenSuse Power Management features are better integrated.

ACPI is a hardware BIOSlike device for power management.
APM is a driver in the Windows kernel that runs when ACPI is not available.

Personally, I think a 500Watt PSU is right for your system.

 

As I said in my previous post I have re-installed windows 7 which is software/driver-less. I am not getting any reboots for now but the system is freezing instead at random points, just like the reboots were.

I have asked for a new machine because this is bollocks.

 

Which is why running the Dell comprehensive hardware diagnostics overnight provides hard facts that even Dell cannot ignore.

All Windows have a HAL (that includes the ACPI, etc). These 'drivers' adapt a standard Windows to each unique motherboard and chipset. And to new security hardware implemented in newer CPU chips. Etc.

Quite possible that the HAL software for your hardware has a bug. That would be obvious if other software (ie Dell's comprehensive hardware diagnostics) loop repeatedly every night - without crashes. Then you have a fact that Dell cannot dispute. Then Dell would not replace your machine with a refurnished duplicate. Might replace it with a machine using a different chipset.

 

How many times would you suggest running the hardware test?

 

Infinite. Just let it keep running. In the morning, learn if it powered off.

Next night, do same. Collect a statistical fact. More loops means a stronger conclusion.

 

Damn this doesn't power off in OpenSuse :/

Hypothetically speaking; If my computer is failing because of a Windows & Chipset compatibility issue, would this affect the windows power management or is this seperate? Also, is that something that could be fixed with a driver update (future release)?

 

The HAL is the interface between standard Windows and each unique chipset. The power controller is part of that chipset - another of the unique hardware devices that HAL (Hardware Abstraction Layer) interfaces to.

Power management is just a tiny part of HAL - and talks to the power supply controller. What controls the power 'system' and can suddenly power off a computer? Power controller.