Aarrrgh! BSOD!

[justgary]

All -

I'm just venting here, so laugh at me if you want to...

I have had my mill on for over a week straight until last night, and had it on for over three weeks before that. I had four "Blue Screen of Death" errors yesterday while trying to mill a small part (a several hour job). The part is OK, but I got tired of re-zeroing and running the program only to have it cut air most of the time while it caught up to where it crashed before. I finally just shut it down and came in to watch the new ads (oops, I mean the Super Bowl).

I suspect it is a hard drive low temperature problem, but that is why I left the machine on during the cold weather. Also, I stuck a temperature probe on the drive after the fourth BSOD, and it was right at 75 degrees F (even though it was about 42 or so in the shop). I can play some games with it for now, but I may end up getting a solid-state drive for the machine.

By the way, most hard drives are not happy operating below 41 F. With it running in the cabinet under the deluxe stand, it stays much warmer. Of course, the problem could have started this summer when the temperatures inside the cabinet were much higher...

Aarrrgh!

- Just Gary

[keithorr]

why cut air? are you running canned cycles? I can usually find a G00 prior to the last known cut and restart from that line. Also, is this a dedicated PC or are you: running other software/connecting to the internet/not optimized for mach etc?

[justgary]

Good questions.

The task at hand is engraving lots of tiny letters on a small "coin", so finding the G0 that goes with the last full letter cut is like finding a needle in a haystack. I *could* do it, I suppose, but just starting over was the best option at the time (because I kept thinking that it would work). I just restarted it a few minutes ago and sped up to 300% feed rate, so I'll go check it every 20 minutes or so and slow it down when it's close. If it dies again, I'll make an effort to start in the middle somewhere the next time.

This is a Mach-optimized, dedicated machine running XP Pro. No internet, nothing fancy, but sometimes I do run a calculator, a text editor, and a feed/speed program. Since I have used it this way for over a year, I don't think it's the setup.

I strongly suspect that due to the size of the G-Code, winders is swapping memory during the run. I can't tell if it's true or not, but it is the only thing I have come up with so far. If the drive is having issues, maybe the swap file is getting corrupted.

Before I started tonight, I changed the swap file to another partition just for fun. Right now, I just want to finish this part (well, and make five more)...

Regards,

- Just Gary

[justgary]

Well, I got all the way through the run tonight without another BSOD. Was it the swap file, or the fact that it is almost 52 F tonight? It was 47 when I started, so it's definitely warming outside...

Since the swap file is on the same disk, I'm betting on temperature. Hopefully, I'll never know.

The best part is knowing that the new version of Mach doesn't freak out when Winders has a BSOD. The Tormach just seems to freeze in place, and the coolant pump and spindle both shut off (but you do lose position reference). It really could be a lot worse!

Regards,

- Just Gary

[philbur]

One of the good features in TurboCadCam is that when you highlight a particular line in the Cad drawing the corresponding line of G code is also highlighted.

Phil

Originally Posted by justgary The task at hand is engraving lots of tiny letters on a small "coin", so finding the G0 that goes with the last full letter cut is like finding a needle in a haystack.

[ger21]

BSOD are typically hardware related. You might want to check the memory.

[MichaelHenry]

Is there no chance of moving the mill to a warmer spot, say a basement?

Mike

[David Bord]

It can either be hardware or software that caused it.

Has anything been changed, HW or Software on the system between when the problem happened? I kinda doubt its the hard drive unless you a marginal one. Ive never heard that.. Ive worked for Intel for 16 years... and actively work in labs with hardware.

You can easily eliminate the hard drive issue. Just find or buy a new hard drive and use a program (free) called clonezilla to ghost the drive image on what might be the bad one to the new one. If you need help with this let me know...

You should also do a disk check. Goto a dos prompt. run C:\chkdsk /f It will say it can do it and ask if you want to run it at next reboot. Say yes. Watch it run and look for errors.

Memory is also good to check. If you have more than 1 DIMM than pull all of them except one and try it.. then try another DIMM.

Any other hardware that is not 100% needed remove for testing. That includes DVD drives, floppys, ad in cards etc.

Have you run a virus and malware scan on the system lately?

Let me know if you have questions.

David

[benji2505]

Gary,

we always cool down our shop floor over night to about 38 degrees and reheat it to 65 during the day. First start in the morning is the mill. Never had the BSOD with the original controller or any other PC in the place. I doubt it is the temperature.

How much humidity is in the compartment where you keep your PC? I remember there were some coolant issues with some Tormach stands. I used to have a water drop in my vehicle motor controller that shut down half of the cylinders when being moved hectically (quick left-reight steering), the problem was solved when I dried and sealed the Controller (just a stupid idea that crossed my mind).

Beside the memory check I would also check the connections between the boards and plugs internally. If you have a weak connection there it shuts down the machine as well (could be induced by the short and fast movements of the engraving job).

Good Luck.

Benji

[justgary]

philbur -

I use TurboCAD for drawing, but SprutCAM for toolpaths. Since I do both of those inside the house, I normally don't go back to either unless something is really wrong with a toolpath.

In hindsight, I could have started *anywhere* in the file and been better off than starting at the beginning. If I went too far, I could let it finish and then start a little earlier until it overlapped or got done. If I didn't go far enough, that's still better than the beginning. Oh, well.

ger21 -

Good point about the memory. That's why I started thinking hard drive, since the swap file could BSOD it, and all of the electronics like it cold. Also, the way Mach hijacks Winders' timing loop to run the controller could cause a BSOD with a bad connection at the parallel port.

MichaelHenry -

No chance. It is where it is. Here in Florida, most people do not have anything like a basement. In fact, until about 40 years ago, most people didn't even have garages. On the other hand, we generally only have a few weeks a year when cold is even a topic of discussion.

David Bord -

No changes to the mill or the PC running it. Its only link to the outside world is a thumb drive, and it is unlikely that a virus jumped onto it. I keep the host laptop scanned, both real time and periodic.

As luck may have it, I think I have two more of that model drive stacked up in the house. Neither of them are better off than the one running the mill (which is probably nearing the end of its life), but you are correct, I could find out more by cloning the drive. If I quit seeing the BSOD before I clone the drive, I won't learn anything by cloning it. So now I'm just waiting to see if it is a real problem or just my luck last Sunday. I didn't check to see how many RAM sticks are in it, but just reseating them could help. If the problem comes back, I'll keep digging.

Benji -

Hard drives do not want to work below about 41 F, and can have several different types of errors because of it. Of course, they warm quickly, and since I had mine running for a week I'm pretty sure is was well above that the whole time. I went out Monday evening and booted it cold at 47 F and didn't have a problem all night.

I don't have any coolant leaks into that compartment, but I can tell you that it is *always* humid here. Again, running the computer in the cabinet will lower the relative humidity considerably (inside the cabinet).

Sunday night after I shut down, I did press on a few connectors in the Tormach cabinet (and I wiggled the parallel connector in the back). That could very well have solved the problem. I'll have to wait and see if I get the BSOD again.

My goal is to make a full enclosure so that I can control humidity with a lamp inside. I mainly want to do that for rust control, but reliable electronics are always a plus.

I'll keep you guys posted on any further BSODs. Thanks for the help!

Regards,

- Just Gary

[Scott_M]

Just Gary

A simple boot time memory checker is available for free. Google Memtest 86+ , download the iso and burn it to a CD and restart with CD in drive.( With cd as first boot device ) It will take awhile but will check the memory . Any errors are unacceptable.


Scott

[kprice1658]

You can also tell alot by the last lines on the bsod. Something like:

Stop 0x00045549 - bla bla bla - this can point you to the general area that needs attention. Sometimes - its worthless, but worth looking at.

Our "SOP" for these problems are : Re-seat all memory chips, remove any other cards / peripherals and reboot. Run chkdsk, scan for virus's / trojans - also run any mfg's diagnosis software they provide - like Dell or HP does.

If you suspect the Hard drive, each mfg has utilities that you can download that will test them thoroughly.

I've fixed several of these by re-seating everything that would plug/unplug or by removing a bad memory chip.

Keep us posted....

Kevin