lockup solenoid diagnostics

thanks to this group and its collective wisdom, i think i finally have
my lockup solenoid problem licked. a while back, in the cold weather,
the lockup clutch wasn't coming on and i was worried my 89 civic auto
transmission was giving up the ghost at only 155k miles. at the same
time as hunting that down, i'd replaced the ecu for other reasons and
got a solenoid code. once the solenoid was replaced, the code
disappeared, and everything appeared to operate correctly. however,
yesterday morning was a little chilly, and the lockup clutch wouldn't
come on again! even after the motor appeared to be at full working
temperature. i was wondering if previous efforts to fix were flawed and
that this was a hydraulic pressure problem - a symptom of a weak pump,
especially as this car has been abused in the past and flushed with
non-honda atf. but to be absolutely certain, i decided to do one more test.

thanks to this group, the archive showed people testing their solenoid
operation with appropriately wired led's to show when the ecu gives
signal. following this suggestion, i built a led tool to do the same
thing, plugged it in and went for a test drive. the solenoid worked
flawlessly each time it was getting signal, so the solenoid and the ecu
were fine. but, driving with my hand on the shifter, something i don't
normally do, i noticed that the /slightest/ pressure toward the d3
position, and the ecu would switch the solenoid off!

conclusion:
cold weather = very slight thermal shrinkage in the cable, just enough
to throw an imperfectly adjusted shift position out! and that kind of
component doesn't warm up as quickly being as it has no direct source of
heat. anyway, that will now be re-adjusted and hopefully this problem
will not recur. in the mean time, i'm driving about with a very cool
led signal on the top of my dash!

parts list:
* 1x red led.
* 1x 2.2k-ohm resistor.
* wire.
* appropriate ecu plug terminals cannibalized from a junkyard.

method:
popped the plug from the ecu, and having identified the correct wires
from the workshop manual circuit diagrams, released the leads for the
solenoid output and ecu earth. [a8, yellow, if you're interested.]
plugged "take-off" leads back into the plug and the former ecu
connectors. the led [and series resistor] is basically wired parallel
to the solenoid from these two leads. appropriate shielding to prevent
accidents and taping wires into position, and you're in business!

again, thanks to the group for this little nugget - saved me a lot of
money on mis-diagnosis and transmission replacement.

 

followup for the archive:

it wasn't the selector switch, it was simply the thermostat. didn't
expect that - i'd put in a new honda thermostat last year, but this one
gave up the ghost early. new thermostat, instant solution.