Debugging spontaneous MS-DOS reboots: If the shoe fits

Article By : Robert Yankowitz

When workstations running MS-DOS begin spontaneously rebooting in a newsroom, the user's choice of footwear becomes an important clue to solving the issue.

Back in the early ‘90s, at the television station where I worked, we had a newsroom computer system that the reporters, anchors, and producers used to write and produce the newscasts. The system employed a pair of redundant servers and a newsroom full of PC workstations. The servers ran IBM’s OS/2 operating system, using the latest 90 MHz Pentium processors. (Ooh, that was fast!)

The workstations used 33 MHz 80386 CPUs running MS-DOS. No hard drives – just 3½” floppy disks. The floppy disks booted the computers to MS-DOS, and then ran a batch file that loaded the newsroom system’s executable program from the server. All the data, such as news stories and newscast rundowns, were stored on the server and everything connected over 10Base-T Ethernet running on a network hub (this was prior to network switches).

screen shot of a black and white MS-DOS command prompt

It all worked very well … that is, until one of our 11pm news anchors began reporting a problem.

Her workstation was occasionally and spontaneously rebooting, causing her to lose her unsaved work. We couldn’t find anything wrong with the workstation, so we replaced it. No good. We replaced her keyboard. Still no good. (No mouse change – there was no mouse; this was character mode MS-DOS.)

We replaced her network cable, changed to a different hub port, changed to a different homerun cable, and we even changed her monitor. Some of the swaps we did twice, but it still happened!

Do you have a memorable experience solving an engineering problem at work or in your spare time? Tell us your Tale.

Finally, one day she wrote a very polite email to the news director, complaining that she had lost an hour’s work the previous night because her machine had rebooted, and she had to stay well past midnight rewriting her story. “If this keeps up, I’m afraid I’m going to have to ask for a new machine.”

The news director called me and asked what was going on. I told him she apparently didn’t realize we had already given her two new machines. It didn’t seem to be a hardware problem. We decided to send one of my engineers to sit with her to see if he could figure out what was happening. I should have done that much earlier, as it took him less than five minutes to find the culprit.

/wp-content/uploads/sites/3/2021/02/tales-button.pngThe news anchor’s CPU tower was located on the floor under her desk, sitting on a short box. Back in those days, the CPU box had two buttons on the front – a power button and a reset button. The reset button connected directly to the NMI (non-maskable interrupt) pin on the microprocessor. Pressing the reset button produced an instantaneous hard reboot of the computer.

She liked to sit with her legs crossed while she typed. While sitting with her our engineer looked down, and said, “How often do you figure you hit the reset button with the tip of your pointed high-heel shoe?”

And that, was that.

This article was originally published on EDN.

Robert Yankowitz has been Chief Engineer at a television station in Boston, Massachusetts since 1999. Prior to that, he worked for 15 years at a station in an undisclosed city (to protect those with high-fashion footwear from embarrassment).

Related articles:

 

Expo booth:

New products & solutions, whitepaper downloads, reference designs, videos

Conference sessions:
  • Internet of Things (IoT)
  • Supply Chain
  • Automotive Electronics
  • Wave of Wireless
3 Rounds Lucky Draw:

Register, join the conference, and visit the booths for a chance to win great prizes.

Leave a comment