Hesitation

The Engine Computer adjusts fuel metering plus or minus up to 10 percent based on what it sees from the front O2 sensors. Adding more fuel shouldn't cause an increase in engine speed unless there is also extra air to go with it. That could be coming from a vacuum leak. The clue that a defective O2 sensor is the CAUSE of the problem vs. It's just reporting the RESULTS of the problem is whether it occurs right after startup, although even that can be hard to tell on newer cars. O2 sensors do not work until they reach 600 degrees. Regardless of what is wrong with the sensor or its circuitry, the computer won't modify fuel metering, engine speed, or anything else based on that sensor until it reaches 600 degrees. That may never happen during prolonged periods of idle so there is a heater inside the sensor to get it to that temperature sooner.

If the engine speed surges right after start-up, the O2 sensor is most likely not the problem. If the surging starts a few minutes after starting the engine, the sensor might be suspect. By using a scanner to view live sensor data, you can see when the O2 sensors start working, (the system will go into "closed loop"), and you can read whether the sensors are switching properly between "rich" and "lean" a couple of times per second, or if one is staying at too high a voltage.

The next step would be to connect a scanner to read live sensor data and hopefully find a clue. There will be some diagnostic fault codes from unplugging the O2 sensor. You'll have to just disregard any that are related to that.

One thing to look at is "idle steps". The Engine Computer will command the idle speed motor to one of 256 positions to control idle speed. You will typically find it at around step 32 when everything is working properly. If you find it going up and down a lot, say from 15 to 45, and engine speed isn't changing, suspect a defective automatic idle speed motor, or a plugged air passage, (very uncommon). If the step is staying low, typically from 0 to 20, the computer sees that idle speed is too high and is trying to bring it down, without success. Suspect a vacuum leak. Try pinching various vacuum hoses or unplug them and cap the ports. If you find one that steadies the idle speed, follow that hose until it branches into multiple lines and try that on each one.

Sounds like the water pump is pulling very nicely against the closed thermostat. Lower hoses actually have a wire spring inside to prevent them from doing that.

Very possibly. Everything so far sounds like a vacuum leak. I don't think there's coolant running through the intake, but even if there is, that doesn't mean it has to be leaking out where you can see it or into a cylinder. The leak could be a gasket that is cracked allowing air to enter from below the intake.

If your mechanic has a smoke machine that will help in identifying a leak. White smoke is injected into a vacuum hose at 2 psi, then you watch to see where it comes out. If the bottom of an intake gasket is leaking, the smoke will come out through the oil fill cap or the dipstick tube.

Not very common on this engine but a leaking head gasket will usually cause white smoke from the exhaust when coolant is burned. The engine will often overheat from the combustion gases forming air pockets in the cooling system that cause the thermostat to close.

I'm sensing some confusion on these gaskets because there are two totally different problems with the same terminology. The focus has shifted to leaking head gaskets. The most common symptom is white smoke from the tail pipe. That is from antifreeze being burned. Usually the combustion gases push into the cooling system, then the overflow reservoir and gives the appearance of overheating due to the bubbles. While an air pocket CAN cause the thermostat to close which leads to overheating, the clue most people miss is there is no steam associated with those bubbles in the reservoir. The coolant might even be cold, so the bubbles are from the leaking head gasket, not actual overheating.

There is a test that identifies combustion gases in the cooling system. Air is drawn from the radiator through a glass cylinder with two chambers partially filled with a special dark blue liquid. If combustion gases are present, the liquid will turn bright yellow.

Originally we were discussing a potential vacuum leak, and one of the possibilities is a leaking intake manifold gasket. Very few engines have coolant flowing through the intake manifold but some cylinder heads have passages that are blocked by the manifold. Those ports are only there as a result of the casting process, but they have to be blocked and the intake manifold and gaskets do that. That's why it is possible for coolant to leak from those gaskets. GM wrote the book on that problem with their trucks.

A vacuum leak at an intake manifold gasket is much more common than a coolant leak. The 3.3L / 3.8L engines have very little trouble with leaking head gaskets OR intake manifold gaskets, but when we're looking for something elusive that's where we look. One way to find a vacuum leak is by injecting a white, non-toxic smoke from a special machine that produces that smoke. These came about with the advent of fuel supply leak detection systems that are built into every vehicle since 1996. Tiny little pinhole leaks can be detected by the system, but they would be impossible to find without the smoke machine. The machine also works very nicely for finding vacuum leaks. A rare but possible place for that leak to occur is the bottom of the intake manifold gasket where it can't be seen. If it was on top, it could be found by spraying water on it while the engine is still cold. That wouldn't work if the leak was on the bottom where it can't be seen. It still wouldn't be able to be seen by injecting smoke, but the list of possible places to look would be narrowed down a whole bunch if it came out the oil cap or dipstick tube. When the smoke is pushed into a vacuum hose, it goes into the intake manifold, and there is no place for it to go accept through a leak or out the tail pipe if any cylinder has both valves partially open. If that happens, the tail pipe can be blocked to help a vacuum leak show up.

Yes, there's an obstruction. It's called the thermostat and it's doing what it's supposed to do. I'm not convinced there is a problem unless there is some other objectionable symptom. You might want to check to see if the hose is mushy and should be replaced, or try a different radiator cap. There is a valve in it that lets coolant flow back in from the reservoir to prevent a vacuum when the coolant cools down. That should prevent the hose from collapsing.

Nope. All radiator caps must have that vacuum valve to let coolant back in when it cools down.

I agree it's not normal but unless you saw it while you were under the hood, I don't think you would even be aware of it. My first reaction is there is still an air pocket in there that allows the pump to push coolant against it and compress it. That would explain how the pump can move enough coolant to collapse the hose without moving other coolant from the engine back into the hose. I assume this doesn't happen if the cap is removed when the engine is still cool.

Eventually any air should work its way out if that was related to the problem.

I wonder if the thermostat is not fully opening. Also, any chance there is a kink in the hose going to the overflow reservoir? It's odd that the upper hose would collapse before coolant would be drawn back in from the reservoir.