sorry, I know it must be posted somewhere but I can't seem to find it

what is the likely cause of one button triggering an entire row of
buttons?

I saw rows and columns being affected by the ribbon cable having a bad 
connection.  the two ribbon cables send row/column info to the LED 
board, row 1, column 1 being on means the LED at 1,1 being on.

hmmm.

maybe the easiest would be to just replace the cables entirely.

that wouldn't be a huge surprise, if one got tweaked.  now I have a 
makeshift temp case but for a long time things were just dangling all 
over

one line of pins on each cable is for LED's and the other for buttons, 
correct?

it must be a problem with the button line, because the LED's will always 
match the ghost presses

check your soldering near the connector to the ribbons, both on the 
logic board and keypad kit.

try switching the cables. if it changes, one of the cables has a 
problem. we can ship you a new one, or you can unhook the head and 
recrimp it further down the ribbon.

I switched the cables this morning.  The behavior is the same. i.e. not 
transposed onto a column  (seems slightly less frequent, but that is too 
un-scientific to count on for diagnosis)

So it must be the soldering of the cable headers then, eh?

could be that handling the 40h without a case one of the solder joints 
cracked on the cable headers.  touch the soldering iron to connections 
to remelt the solder and let if flow again, add more solder per your 
judgment.  other than that it would be a pin on the 7221 LED driver but 
that's not likely since I'm sure you've been careful not to touch or put 
pressure on the logic board chips.

a REALLY close look at the solder joints might show a hairline fracture 
in the solder, a black line thinner than a human hair.  maybe turning on 
all the LEDs and jiggling the headers could excite the fault, make it 
more obvious.

I re-heated all the solder joints on the cable headers.

Still had the problem

Then I found a small ball of solder sitting on the button ground trace 
for the offending row. It was right near the header soldering so I'm 
thinking it may have allowed shorting to the adjacent pin

Cleaning that seems to have helped, although I have got one bad press 
since, but over several hundred presses.

Does that seem like it could have been the root of this problem?

I think without a faceplate the buttons can effect each other.  when I 
first started using my 40h I thought there were bad presses, now I'm 
more purposeful when I push them and all seems OK.

also keep doing visual inspections, if there's a problem it'll be 
something obvious that you just haven't seen yet.