1989 Accord LX warm up problems

My wife's car (89 accord LX carburated motor) has had this odd problem for the last year and i'm just investigating it now.
When she initially starts the engine (wintertime) it sits at about 2000-2500rpm to warm up. It takes it's time (about 5-15mins, depending how cold the morning is) and when you finally kick down the motor, it drops to about 750rpm and drives away just fine.
BUT, while the motor is warming up, if you try to kick the motor down before it's ready, it'll stay at it's normal warm up idle rpm, and out of the blue, it will shoot up to about 4000rpm!! And stay for like 5 mins! If you try to turn the car off, it just sputters forever and wont shut off and even backfires sometimes. It's as if the car knows it's not warm enough yet and punishes you for it. Has anyone ever experienced this or have heard of it before?
She knows how to warm it up just fine but when I drive it, it does that to me every time and it bothers the heck out of me.

Easy solution: don't drive it. ;)

It is a fairly common Honda carb issue...but it's been so long since I've had to work on carbed models, I can't remember if there is a typical problem that causes it.

Auto chokes are a delicate ballet of several things happening...

But in a nutshell, cold start idle setting works something like this:

You pump the gas a couple of times. The auto choke has a spring and a weighted linkage that act on the throttle stop cam. When cold, the spring tension overcomes the weight, allowing the cam to rotate into position to hold the throttle open more, and closing the choke plate...this action happens when you first pump the gas pedal.

Now the throttle plate shaft is resting on the highest step of the throttle stop cam, and the choke plate is fully closed, so when you start it, it idles up as it should with a rich mixture. As the engine starts to run, a diaphragm called a choke pull-off moves inward and pulls the choke plate open slightly, so there is a very rich initial mix, then as the engine begins to run the mixture leans out slightly...

As the engine warms up, the auto choke spring tension relaxes, allowing the throttle stop cam to rotate to progressively lower steps when you blip the throttle pedal, until it is finally overcome by the weight, at which point the throttle stop cam is no longer in play, and the throttle shaft is resting on the base idle stop, not the fast idle cam, and the weighted choke plate linkage is free to allow the choke plate to fall into the full open position....

If you get a runaway idle speed when you attempt to kick it down, something is out of sync. You'd have to identify all the parts I mentioned and watch to see what's not working correctly. Honda carbs are kinda nightmarish to look at, but if you can identify the critical parts, it's not so bad.

First order of business would be to make sure all vacuum lines are intact and plugged in....especially ones on the underside of the air cleaner canister, which often get pulled loose during servicing and left unconnected...and make sure there is no debris, wiring, or anything else that might be interfering with any linkage movement.

With the air cleaner completely removed, spray all linkages and moving parts with carb cleaner, particularly the throttle shaft, fast idle cam and choke linkage, which are sometimes not real visible...they are down low one the carb base.

Also grab the carb body and try to twist it...loose carb to manifold bolts or loose throttle plate to bowl section screws often cause strange problems...you'll notice if any thing is loose when you do this.

I don't know if Honda uses a heat riser tube (small pipe that brings exhaust heat up to the choke mechanism)...I just can't remember...but a clogged riser will make it act strangely too.

I'll try to think of what else...been too long, and I usually tried to avoid Honda carb issues unless I had no choice...they aren't my favorite by any means...always seemed terribly over-complex...