I thought that you could just start using small cell foundation whenever - never heard or thought that you'd have to order small cell bees to go with.
It's the 'won't fit' problem which is the rationale behind performing regression in a step-wise manner.
By the sound of things, you had enough cells on your "mixed bag of combs" for their acceptance. With foundationless comb, there will always be a small spread of sizes anyway - as natural combs aren't as regular in their cell-sizes as many people assume.
It may be worth commenting that this so-called 'small-cell' incompatability issue is completely man-made. The size which is commonly called 'small-cell' (suggesting that it's something unusual and/or irregular) is actually
natural cell size. It's the 'standard' stuff which really ought (imo) to be described as 'large-cell size' - which many people have grown up with, and so assume that it's what the bees are supposed to work with.
I first became acutely aware of cell sizes when trying to figure out why a Nicot Cupkit laying cage wasn't producing the goods for many beekeepers (including myself), and yet many were having great success with them. I set about trying to solve this mystery, and eventually discovered that the cell size that Nicot has chosen is 5.6mm, which is neither worker, nor drone size - hence my queens (which are used to 4.9mm worker cells) appear to have been confused by this.
Queens which normally lay in 5.4(ish)mm worker cells, presumably don't notice the difference - so a provisional explanation (all other aspects being equal) is that those using conventional foundation will have success with the laying cage, whereas those using 4.9mm cells will have very mixed results.
To confirm these suspicions, I've modified a Nicot Cupkit laying cage to 5.0mm (+/- 0.1mm) cell size, and will be reporting on what results later on this year. Here's hoping ...
LJ