Ballasts are designed to push a constant current through a particular voltage on each of their outputs. Hence the nominal power printed on the box. Obviously, to maintain a constant current they must be able to manage that through a certain range of voltages and for modern electronic ballasts, even cheap ones, this is often a very large range. The voltage across a fluorescent tube is mostly dependant on its length but also a little on its thickness.

A 40W T12 ballast is designed to push 430ma while a 32W T8 ballast is designed to push 265ma (at ballast factor 1.0). The T8 ballasts you have been playing with are most probably ballast factor 0.88 and push even less than 265ma. The voltage across the T8 and T12 is close enough that you can consider it to be the same, so the T12 ballast will push very close to 430ma through the T8 tube. Hence even normally driven the T8 with your T12 ballast will be hotter and brighter, not far short of your original overdriven setup. Overdrive this and you will get something rather intense.

The 265ma current is below nominal for the T8 tube, and the 0.88 ballast factor takes them even further below their design levels, they are deliberately being underpowered to match the light output of legacy tubes. While your original 2x setup with the T8 ballasts is not far above the original designed tube power, overdriving it with the T12 ballasts takes it a long way above the design levels and you can expect more heat, shorter life, and decreased efficiency.

Thank you for the detailed and thoughtful reply! If I understand you correctly the current T-12 ballast in these fixtures driving 2 T-8 tubes may approach the output of the single SL-15 driving a single tube T-8? If this is the case, my problem is solved. I spoke with Sunpark today and learned one tidbit about temp rating of these ballasts. The rated temp to 4000 hours of service is 45C. For every 10C increase the life drops by half. I do not think my SL-15T s in OD configuration are exceeding that, but based on a very short experiment with the T-12 ballast driving a single T-8 lamp it would be shortening the life of the ballast considerably.

Since I can fit 3 of these fixtures into my shelves, using the installed T-12 ballast in normal output configuration I may achieve around a 1.4 ballast factor through the T-8 tubes. I am not hitting the mark for the OD SL-15T, but I am not off by much, and I save the cost of the additional ballast.

The person I spoke with at Sunpark seemed to think that running a T-8 tube with the T-12 ballast in normal configuration would actually exceed the T-8 in an OD configuration with the SL-15T. I do not see how that could add up, but perhaps he did not understand what I was asking.

I also learned that the both these fixtures actually have Sunpark ballasts that are simply re-badged by the manufacturer of the fixture itself. Apparently Sunpark is no longer going to sell ballast to them, though they will still be available to consumers on the Internet from Sunpark direct.

I think that for now at $9.00US a pop I may have to stock up on these in case I keep getting plants and do not finish the planned CFL fixtures.

Are you sure the T12 fitting does not have an SL-15 ballast in it, rebadged or differently labelled? The SL-15 is specc'd to operate two 40W T12 tubes although it would effectively be running them as 27W tubes.

Good question, and that is why I bought them in the first place even though the box was labeled for "40 W T-12 tubes with electronic ballast", I considered the possibility that they might have been using the SL-15T and I was willing to waste nine bucks on the gamble that they were. Also I figured the electronic ballasts offer this flexibility in set up, and I figured if it was a different ballast I might still be able to make something work.

As it happens, it is a different ballast - E240120 MR-01, which I understand is ballast code for 2 tube 40 Watt 120 V input, and Sunpark confirmed it is a T-12 ballast. I could not pin them down to the output or ballast factor though. When I asked he said, "for sure it is less than1". I kind of already knew that. I suppose I could test one with my digital Multimeter and get an actual read on the output side.