Some things to consider, especially when back-firing is involved, is camshaft timing, meaning timing belt. A belt that's off as little as one tooth will cause running problems including erratic idle speed.
You need a scanner to view the AIS step number. If it is holding steady, (step 32 is typical for a well-running engine), but idle speed is erratic, look for vacuum leaks, fouled spark plugs, bad plug wires, things like that. There was a fairly common problem with the gasket under the throttle body, but if memory serves me right, I think that applied to the engines with carburetors. Regardless, dribble some water around the base of the throttle body while the engine is running to see if it gets sucked in.
With erratic idle speed the AIS step number should be changing rapidly as the computer tries to correct it. What you need to do then is try to see if the idle speed goes up when the step number goes up, or if idle speed goes down when the step number goes up. If they both go up at the same time, the computer is calling for an increase in idle speed in response to something else. If they go in opposite directions, the computer is trying to correct a wrong idle speed but is not having success.
With no scanner to view the idle steps, you can do almost the same thing by listening to that sucking sound. When idle speed drops below the target idle speed, the computer will run the AIS motor to open the valve and let more air in. You'll hear a substantial increase in the volume of that sucking sound. You can mimic the same thing by unplugging one spark plug wire momentarily. That will cause a huge drop in idle speed, with the increased number of steps the response, with its increased air flow past the AIS valve. Viewing the step number, or listening to the air flow, is an effective way of comparing cylinder contribution to see if one cylinder is not producing as much power as the others. For example, it is typical for step 50 to be called for when one cylinder is misfiring. If you disable a spark plug, (or disconnect a fuel injector on those engines that have one for each cylinder), you should see step 50 called for when any cylinder is disabled. If the steps increase very little or not at all for one cylinder, that cylinder is not contributing any power, and the cause must be diagnosed. Step "50" is just an example. The point of my sad story is whatever the step jumps up to for one cylinder should be close to the same for every cylinder. We used to do a similar cylinder contribution test on carburetor engines by watching how much idle speed dropped when we disabled a cylinder. The exact rpm drop was not important. What was important is that it was the same for every cylinder.
Tuesday, September 27th, 2016 AT 4:03 PM