This is typically a sign of intake-leaks and/or misadjusted throttle-plate, idle-screw and idle-stabilizer. Follow the manual's procedure for disabling the idle-stabilizer and setting the idle-screw. With the idle-stabilizer disabled, setting correct idle-speed on a warmed-up car will have it not contribute any extra air to idle-speed. No catch-up game with overshooting and undershooting.

You may also have a sticking ICV. A little carb or TB or MAF cleaner down the ICV hose may free it.

You're right on about the settings. I had gone through all that before I disconnected the O2 sensor. I'll have to go through it again after I do a re-check of the vacuum for leaks again. While I had the intake manifold off last week I cleaned the ISV and rechecked the lines for leaks.

What are the pin assignments on the O2 sensor round plug (3 wire)? I've checked everywhere and can't find a diagram for which pin has input, shield, heater wire.

O2-sensor signal wire goes to pin# 24
O2-sensor coax-shielding goes to pin# 23 (also shared with shielding from speed-sensor).

There should be no continuity between these two pins. Have an assistant wiggle the O2-sensor wires in the engine-compartment while you watch the voltmeter. Also disconnect and connect the O2-sensor, shouldn't have any continuity.

While there, you'll want to open up the wiring-connector and solder a wire to the back of the #24 terminal-pin to extract an O2-sensor signal for monitoring. That's the best location to get it. Many people run a wire out to the engine-compartment and hook it up to the O2-sensor connector just before the it connects to the shielded harness wire. But then you've got an unshielded cable running parallel to the shielded cable and you won't be seeing the exact signal that the DME's getting.

With a fully-warmed up engine, you should see the O2-sensor voltage flip back and forth across 0.5v about 4-5x per second: 0.8->0.5v->0.2->0.5v->0.8 etc. It should be symmetric on both sides of 0.5v. If it's leaning heavier towards one side or the other, then you may have a problem with O2-sensor and/or wiring. Use pin#28 for sensor-signal ground (not power ground or body).

1. Also check the mechanical position of the throttle-butterfly. There's a picture in the manual on it's positioning, something about just barely after a tiny hole or something like that.

2. And check the throttle-butterfly axle for leaks, that can actually be significant enough to mess with idle-speed and the ISV. There's a write-up somewhere along with list of O-ring parts needed to re-seal.

3. check the O-ring under the throttle-bypass screw (idle-speed adjustment). If O-ring isn't fully sealing, that too can be a source of leaks.

Sounds like all the stuff that I've done over the last year. I replaced the O rings in the throttle body shaft when I did my initial vacuum checks. Bubbles came out of the TB under the Cam. Also checked the idle speed adjustment O ring...Good.

I'm not sure what you're referring to on the butterfly positioning.

The angle that the throttle-plate rests at when you release the throttle is adjustable. It should be behind a small hole on the edge of the throttle-body. That hole sends a vacuum signal to the EVAP system. If the throttle-plate isn't set properly, you could be sucking in extra fuel from the charcoal-canister. Although this would be more troublesome at idle than under acceleration.

The stumbling issue goes away when you disconnect the O2-sensor. Assuming that the replacement O2-sensor is working properly, then we have to assume that something is wrong with the O2-sensor wiring. If you've verified that the O2-sensor signal-wire is not shorted to the shielding braid, then we should test the O2-sensor signal at the O2-sensor plug itself. Then test it again at the DME terminals to make sure it's the same signal.

If this is a 944na, we should verify that the full-throttle switch on TPS is working and that terminal activates properly at the DME terminal. If a 951, should verify that full-throttle signal is coming from KLR.