I just did a little testing of my own in cheat mode and with everything else set to all abundant in a huge galaxy I got 754 with scattered systems, 736 with loose clusters and less than 500 with tight clusters.
I repeated the test and got 483 with tight clusters, 763 loose clusters, 746 scattered systems.
So it does look like there's no hard limit and that tight clusters clearly yield fewer planets. But between loose clusters and scattered systems it's close to a wash although based on very limited data. An average of 749.5 for loose and 750 for scattered seems to give a slight edge to scattered which also seems to yield the more consistent numbers, assuming you're willing to base the results on only two samples.
Since I tend to prefer the more even distribution of scattered, scattered it is.
The debug.err file gives the following in all cases:
Debug Message: Star Frequency: 4 Max Num Stars: 320
Debug Message: Planet Frequency: 4 Max Num Planets: 1280
and for loose and scattered:
Debug Message: Number of Unplaced Stars: 0.
and for tight:
Debug Message: Number of Unplaced Stars: 140.