WalterT I've been using the line that runs from East to West at zenith.
You just want to be far from the poles, Walter. In the horizon coordinate system, the East-West line through the zenith is actually quite close to the pole for Chicago (I am some 3 degrees even north of you -- Capella transits almost through Zenith :-).
Strictly speaking, I think declination of 0 degrees is even better since it gives you a maximum movement on the guide plate for each degree of the sky. I.e., most movement on the guide frames for each RA guide step. I often do that (i.e., go to declination = 0 to calibrate)
Conversely, the worse is when you try to calibrate near the pole. When you move the mount's RA with a guide pulse by, say 1 arc second, the star may not move by even one pixel when you are close to the pole. I think this is the reason some people don't get enough movement. They listen to old wives' tales about calibrating where your target is located at, and their target happens to be M81.
Basically, stay away from the pole, but don't go past 0º declination by much, since the scale starts to shrink from there again.
Just remember that calibration moves RA and declination, and you are trying to measure the movement in camera x-y pixels. If you are pointed at the pole, you can move RA by 12 hours, and there is no movement of the center of the plate.
By the same token, for a constant amount of periodic (or guiding) error in RA, the stars will hardly bloat in the RA direction when you image near the pole.
I don't know why some people are so hung up on guiding. It is simply not difficult to get close enough. You don't hear such angst among people using PHD2.
If you get within 1 or 2 arc seconds RMS, most people who use ASIAIR will probably also not have an OTA where 2 arc seconds RMS error is a problem. Most people don't even have "seeing" that is better than 2 or 3 arc seconds. If you can guide better than 2 arc second RMS, turn off the guide graph. Just turn guiding on, and stop looking at the graph.
One of the problems is that some people are insisting on using short focal length guide scopes, where an arc second already represents less than a tenth of a pixel on the guide sensor -- and yet expect to guide better than 1 arc second. Especially if the quality of the optics is terrible, and out of focus. Imagine trying to find the centroid of an irregular blob 6 pixels in "diameter," and still expect to come out with a centroid accuracy of better than 0.1 pixel.
Guys like Olli and myself were experimenting with how much you can push a mount with a huge periodic error (the RST-135 has a peak-to-peak error of a whopping 70 arc seconds) with ASIAIR guiding (we all know that PHD2 on a laptop never had a problem). And even we could pretty much consistently get below 0.4" RMS with ASIAIR (starting at 25" RMS -- the equivalent of a 70 arc second peak-to-peak sinusoid). After those experiments, I don't even look at the guide graph anymore.