Jerry, the ac voltage scale should be at the 2 scale, the 20 scale is not low enough for precise readings.
You measure at the speaker binding posts on your amp.
Stingreen, it would be interesting to measure your crosstalk with a DMM. I contend the DMM method is by far the most precise method because of production variations in tonearm mounting pads, cartridge mounting surface, cantilever, flatness of record, and in your case the Myrtle wood blocks. In the case of VPI's method you also have the possible variations in straightness of rod, ruler precsion and reading it.
My initial reading after having set azimuth according to the VPI method was around 20% deviation in crosstalk between channels. I also found that inperceptable movement (with the naked eye) of the counterweight resulted in quite different crosstalk measurements, ie., what may measure equal using any physical means may still not be within the 5 to 10% crosstalk tolerance level. Also, there may be added value in getting it down to about 1% vs. 5 to 10% tolerance level(I didn't listen to musical passages at different crosstalk valuations).
Your suggestion of tightning up the counterweight set screw close to its stop could be very helpful, I had the screw set loose which may have resulted in counterweight moving about too easily. Also, I wasn't setting VTF and azimuth in discrete steps, I already had the VTF set correctly. It was in trying to move counterweight about it's axis that I was inperceptably moving counterweight forward and backward which changed VTF, your suggestion of pre-tightening some may help in keeping this from happening.
Slowhand, I took it for granted the VPI allignment jig was correct. In the Fremer DVD he does allignment with the Wally tracker and it jibes perfectly with the VPI jig. I will have to go back to my DB protractor and double check allignment.