[rescue] Trying to get Cadmus 9200 running (serial console trouble)

[Mouse]

>> Is the garbage consistent?  That is, if you type (say) x, do you
>> always get the same garbage back?
> I retried it. it's less consistent than I remembered.

> Aaa at a@aa at aa@aaaaaaAaA at AAaAa@aAcssssbscsbbsbbsbssbd``DD at d@`@d`d at d@ddD at D@`@`dDDDddfDfDDfDdDfffDf
> ``qqqqqqqqqaaq`qqqqa`a`aqqqqq`qaq`qvwwwwwvwwwwwgwvwwwvwwwwwADeDe at eAe@eDDaDdaeEeaeaDae`eDED

> that's
> aaaaaaaa....sssssssss.....dddddddd...
> qqqq.....wwwww....eeeeeeeeeeee...

Curious.

a = 01100001
A = 01000001
@ = 01000000

s = 01110011
b = 01100010
c = 01100011

d = 01100100
@ = 01000000
D = 01000100
` = 01100000
f = 01100110

q = 01110001
` = 01100000
a = 01100001

w = 01110111
v = 01110110
g = 01100111

e = 01100101
A = 01000001
@ = 01000000
D = 01000100
` = 01100000
E = 01000101

My first guess was that whatever drives the serial-line oscillator has
drifted out of spec and the two ends are, thus, running at slightly
different speeds, but that's not very consistent with the observed
errors.

What it _is_ consistent with, though, is the 1-state driver partially
failing, or the voltage source driving it having gone weak, so that 1
bits turn into 0 bits.  I see only one error converting a 0 bit to a 1
bit, that being d turning into f (unless you actually typed
aaaaa...sssss...ddddd...fffff...), but lots of 1s turning into 0s.

Another possibility, prompted by Dave's mention of differential serial,
is that perhaps something has broken the ground connection (pin 7 on
the DB-25 end); I've occasionally had that produce very oddly corrupted
serial data.

> DB25  DB9
> 2      3
> 3      2
> 4      7
> 5      8
> 6      6
> 7      5
> 8      1
> 20     4
> 22     9

(Minor nit: the 9-pin end is almost certainly a DE-9; I've never even
seen a DB-9, a DB shell with only 9 pin positions.)  This does, as you
say, appear to be a relatively standard RS232-to-PC serial cable; all
the signals match up with what my notes say are usual for PCs.

> Cadmus docs about RS232 cables denote DB9 pins as

> 1 shield
> 2 data out
> 3
> 4 bridged to 5 in some cables
> 5 bridged to 4 in some cables
> 6 GND
> 7
> 8
> 9 data in

Does the Cadmus use the 9-pin end or the 25-pin end?  If the 9-pin end,
this could be part of your problem; connecting that cable to that
pinout leaves me wondering how it echos anything at all, since pin 3 is
not a data pin.  My notes do show an "Apple modem" pinout for DE-9 that
puts TXD on pin 9, but that one puts DSR, not RXD, on pin 2.  I have no
pinouts in my notes that put data on pins 2 and 9.

> Their drawing about pin locations is the following:
> 1 to 5 are the long upper row from left to right
> 6 to 9 are the short lower row low from left to right
> when viewed from cable side.

That's the usual numbering for one of the two sexes, though I'd have to
check which one (I don't have an example with pin numbers handy).

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse at rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B