I've always wondered why it's a top down design instead of a bottom up? It seems to me like if you were to freeze a container bottom up, the impurities would be pushed out towards the top, and eventually out of the water into the air above, instead of making bulging shapes as ice expands and wasting half your ice with trapped impurities.
I should test the theory later at home.
Top down you can put the the water in a small ice chest and the insulation will slow the freeze in the bulk. otherwise it freezes outside in
That ice floats may play some role, I am not sure if for or against.
I made a couple of attempts, just for fun, and found the ice chest necessary but never made a systematic study of the process.
Probably better if you don't expose the water to air after heating it until it is frozen.
Start with good water.
Freeze in a gradient(top down is easiest) half of your ice will still be garbage.
Explicitly: by gradient I mean put it in an insulated container so that it freezes starting at the top and ending at the bottom.