[comp.sys.dec.micro] Rainbow/LA50 Problems

ciaraldi@cs.rochester.edu (Mike Ciaraldi) (02/09/90)

I'm trying to help someone with a Rainbow who is having printer problems.
The system has worked fine for several years, but is now acting up.

Hardware:
	DEC Rainbow 100 (don't know if it's A or B)
	Dual floppy drive
	20 Meg hard drive (from some aftermarket company)
	DEC LA50 printer.

Software:
	CP/M 86/80 
		Application program "The Landlord"
		Microsoft BASIC
	MS-DOS 2.05
		Samna word processing.

The printer is in the default configuration:
4800 baud, 8 data bits, no parity, X-ON/X=OFF flow control.
The Rainbow is set to match.
--------------------------
OK, here's what's going on.  This system has been running for years
with no problem.  The user has both CP/M and DOS, because the programs
he uses are only available on one or the other.

About two weeks ago the LA50 printer started acting up.
When printing long reports (40 pages or so), it messes up after about 14 pages.
In the middle of a line it will switch into condensed printing,
then recover on the next line.  Or it will lose a few characters
from the line, then be OK for several more pages.
Sometimes it will start printing gibberish, then the "Fault" light will
come on and the printer shuts down.

When this problem started, the user tried other programs and even the
other operating system (MS-DOS) with similar results.  Again, all these
programs worked OK a few weeks ago.

My first thought was that flow control wasn't working,
so the printer buffer was overflowing.
I put an RS-232 breakout box on the printer port and watched.
After a few pages the data leaving the computer would stop, then
after a few seconds it would start again, and then repeat this pattern.
I couldn't see the flicker of light caused by an X-ON or X-OFF
(at 4800 baud it would only last 1/500 of a second), but 
this looked OK.

Then I wrote a test program in CP/M BASIC that printed a long line
to the printer and the screen, over and over.  About line 60 the screen
would pause for a few seconds, do a few more lines, pause again,
and so on.  Looking at the breakout box, these pauses matched when the 
data out of the computer stopped.  I let this run for several pages
and it all looked OK, so I concluded that X-ON/X-OFF was working.
Incidentally, I was sending just text, no 
control characters other than the carriage return-line feed.

A few minutes later I tried again, and this time I got down a
page or so on the printer, then the printer gave out a screech
(EEEEEEEEEEE sound), a line of lower-case e's were printed, and the
Fault light went on.

It turned out the user had another LA50 printer and another cable,
and these show the same problem.  He reports that he has swapped
the motherboard from another machine, and the problem persists.
I checked the output of the power supply at the floppy and hard drive
connectors, and my digital voltmeter showed 5.05 and 12.17 volts.
I couldn't easily get a probe onto the motherboard.  I also
have no way of measuring ripple.

I tried one more thing--I reduced the baud rate to 600, thinking this would
never overflow.  If anything, it fails more quickly!

I'm at a loss here, so I'm hoping someone has some suggestions on
things we can try.  I wonder if maybe the baud rate generator is drifting,
since I saw something like this once on an Apple 2C that had an
out-of-spec baud rate.

Thanks in advance.

Mike Ciaraldi
University of Rochester Computer Science Dept.
uucp:      ...rutgers!rochester!ciaraldi
internet:  ciaraldi@cs.rochester.edu