I have done a lot of experimentation in the past month with various recipes and dilutions and taking a lot of notes. I keep thinking that one of these days I will sit down and put all the notes together and write a comprehensive summary of what I have learned so far. But there is always another experiment that I want to do before drawing any conclusions -- and the "tests" that I am using are still being refined. So, I feel like it is best to apply the refined versions of the tests to the solutions that I have put together.
... that being said ... I have found a recipe which so far has performed as well as (and probably better than) Cricket Hill Powder which many people think is one of the more consistent solutions for creating big bubbles. I haven't tried to do massive bubbles as I have only made a 48-inch top string wand and haven't felt ready to move up to a bigger wand.
My current favorite recipe for big bubbles is 8 parts (by volume) tap water (I live on the SF Peninsula which has pretty good tap water although it has a pH somewhere in the 8.0 to 8.4 range), 1 part Ultra Dawn detergent and 17.5 grams (per liter of mixture) of a mixture of J-Lube, surgilube and tap water. (The lube recipe is 16 ounces water, 16 grams jlube, 2 ounces (by weight) of surgilube). The latter recipe was suggested on the Soap Bubble Fanciers Yahoo group by Brian Lawrence whose opinion is highly respected on the list and whose postings are recommended for anyone looking for good bubble recipes.
For making bubbles with small wands, a mixture of equal parts (by volume) of tap water, dawn ultra, glycerine (I have used both Walmart brand and whatever they had at Whole Foods) and generic ky jelly-type lube (I have used CVS' and Safeway's brands with similar results) trounces the competition. Gazillion and Amazing Bubbles are the commercial solutions for small bubbles that have given the best results of commercial solutions. The homemade solution does even better -- a huge number of bubbles per wand dip and they are very colorful. However this solution is very very very viscous and pretty expensive to make due to the large amounts of glycerine and lube. But it does work great -- I will post some videos of the small wand tests. This recipe works far better than recipes that use only glycerine or only lube and it is much thicker. So, there seems to be serious synergy between the glycerine and the lube.
The 1:1:1:1 solution can also be treated as a concentrate. It works well with anywhere from 1 to 3 parts water added to 1 part of concentrate. I haven't tried it with big bubbles but it does nicely with everything from a small plastic wand to a string wand with 24-inch top string.
I plan on dialing in the glycerine/lube ratio as I suspect that a smaller amount of lube is probably sufficient.
ADDENDA
June 4, 2010. Last night, I did a trial run of the size potential test and (surprisingly) the 1:1:1:1 juice diluted 3 to 1 with tap water did better than everything else. I just tried it outside with my tri-string wand and it did pretty well. It is still pretty viscous at the this dilution. I need to test it head-to-head with a known quantity before drawing any conclusion -- and need to try it under better conditions. It made bubbles at least 4 feet in diameter but tubes seemed to break earlier than I like -- but that could have been because I was using a single string wand (rather than my usual double-strand homemade rope) and because there was a lot of direct sunlight. The colors were awesome.