The throttle valve cable is adjusted at the throttle body. First, loosen the screws holding the cable bracket to the intake plenum, and get the throttle cable just short of tight (warmed up engine, please, so all the metal expansion is done) by sliding the bracket on the bolts. Think the slack spec is 1mm or less. Then loosen the two lock bolts on the threaded sleeve on the throttle valve (TV) cable. Slip off the oragne cable boot so you can see the cable exit the sleeve. Note the crimped on brass cable stopper just outside the sleeve. Operate the throttle body lever by hand to wide open throttle, and measure the distance from the end of the cable sleeve to the crimped-on stopper. FSM spec for the distance is 1 5/16"-1 3/8". Mine behaved as yours until I got it dead right. Seems to be very sensitive to small outages. The longer the measurement, the harder the pull on the throttle valve, and the later the upshift, the earlier the downshift, and perhaps (more later) the 3-4 shift doesn't happen at all. Mine is at the upper limit, and I'm going shorter to 1 5/16". It tends to not upshift after a hill downshift while in cruise, unless you tap the decel/set button and let the throttle flop toward closed.
The 3-4 upshift is both hydraulic and electric controlled. My 87 a/t kills 4th gear on overheat, controlled by a thermoswitch on the water neck. Later models kill the upshift while the motor is cold for emissions, again, thermoswitch controlled. It will be a single wire switch, I think, and I think in both cases controls a relay ground path for the relay that controls the OD solenoid valve on the tranny. I bet FrankR has the ODSV electrical diagnostic on the tip of his tongue...
Check the electrical function of the shifter switch, too.
You can hear the hubs hook up, and release, too. A little "chunk, chunk" sound. You can get them to lock up by hand by hand-turning the front driveshaft in either direction. Reverse rotation to check release. In my 85 truck, the directions said opposite direction travel after 4wd disengagement should be 12' to unlock the hubs. Later model says 5'. The former is just under 2 wheel rotations, or about 9 driveshaft revolutions. The later is about 2/3 of a wheel rotation, or about 3 dshaft revs (if the diff is a 4.625). Static testing like this will probably only lock one hub unless the front wheels are airborne so the locked hub can turn while the tardy one locks up. They never seem to do it at the same time.
