Oueen size is somewhat independent of queen cell size, but larger doesn't hurt. Small queen size can, but does not necessarily affect queen quality. Queens can grow and shrink in a matter of minutes.
The main concerns are: 1) Begin with good genetic stock; 2) commence queen rearing during a nectar flow; 3) be sure the queen cell builder colony is strong and populous - bees boiling out of the top when the cover is removed (combine colonies and add capped brood 10 days before you start); 4) Make absolutely certain that there are NO OTHER QUEEN CELLS in the queenless portion of the queen cell builder colony; 5) make sure they are well-fed with 1:1 syrup and a pollen patty as well as 1 or 2 frames of super-fresh pollen; 6) have your nucleus or increaser colonies already built before the queens hatch; 7) separate the queen cells before the first one hatches.
The idea is that massive numbers of 5- to 15-day-old bees that have nothing to do but eat and feed babies are made queenless and are given nothing from which to make a queen until you put in your grafts.
Do the above listed 7 things, and your queen cells will probably be fairly large. I usually let smaller queens prove themselves to be poor egg layers before I cull them. Sometimes they get much bigger later.
Swarm cells are always welcome, BUT NOT IN A CELL BUILDER COLONY. Those I cut out during routine Springtime work get planted into splits in short order and marked "SC" and the number of the colony they came from. Then go break up the brood nest of the parent colony that is trying to swarm. I add 2 empty frames to the brood nest of the mother colony to delay swarming, giving the brood frames to the splits.