Hi Ryan,
I think there are too many variables and unknowns to recommend specific settings.
I'm not sure how homogeneous the various zircon RMs are for trace elements, but it was my understanding that some are better than others but none are great. You may want to start out just using 610 and evaluate how homogeneous 91500 and GJ1 are within and between individual shards/grains.
If both seem homogeneous and give results similar to the published values that is great. In that case, I would include one in the calibration and use the other for QA/QC. If only 1 is "good", I would stick with using 610 to calibrate and use the good zircon for QA/QC.
Ultimately, I think you'll need to play around with the settings. You might find that some of 3D TE's features, like downhole fractionation correction and/or affinity correction might work well. That'll depend on how hard you're ablating these things and whether you're able to include a zircon in the calibration.
I think normalizing the yield to 91500 (if you're using it) should only make a significant difference if you're not using an internal standard (presumably you are using an internal standard?)
My recommendation is to try all the settings and various combinations and convince yourself which works best so that you can defend that choice once you want to publish or present it.
All the best and good luck!