Because it hangs from the service dude's butt when not in use?

Because the dial pad is usually on the bottom (butt end) of the device.
But I like the first explanation better.

The modem now works as good as new -- equal to the st516v6 I had replaced it with. On our primary line, it (as well as the other two modems) is solid.

So today I put the known-excellent st516v6 modem onto the suspect line. And the same wanky line condition struck at about the same time again this morning (10:42am).

So, I think the three modems are all just fine (now).
And one of our two lines has an issue of some sort.

I'll ignore it for now, and deal with it in the New Year.

Cheers

About 20yrs ago all of our cad terminals used fiber to carry the analog RGB signals from the mainframe to the workstation CRT's. One fiber for each color and one for the kbd/tablet and terminal text.

Over several years a number of the fibers were broken eventually using up all the spares. Every one assumed it was the pigtails coming out of the drops that had been broken from mishandling.

It was a surprise when the OTDR showed that all the breaks were due to failed splices inside a closed junction box.

It was the first I had heard of or seen what a TDR could do. Seemed magical. The unit in use had a crt and a paper tape output. Even used on a good fiber the tech could spot where every splice was all the way back to the other end. He even showed how bending a fiber a little to sharply could be seen.

Those workstations were insanely priced $25k for a desk an rgb monitor, digitizing table, keyboard and a circuit card that mapped the terminal text on top of the graphic display. Today it's almost impossible to spend more than half that much on a high end PC workstation including all the compute power, which the old cadds workstation totally lacked.