I strongly believe whether you warm a MH up fully or not, underdriving it has an equally adverse effect.
I have a 600w sun pulse 3k MH running at 400w right now, and I can tell you, the spectrum is nowhere near what is should be. Way more white/green/blue, less orange. I know bulbs change their spectral output as they age, but this is more, enough to probably affect the internode spacing during flowering.
I have 6 lumateks. I don't know how the ignitor is set up on them, but they will start any wattage bulb on any setting. If I had to take a guess, I'd say the lumatek sends out the same KV pulse regardless of what the wattage setting is, and then limits the current draw as the bulb warms up to provide the correct wattage. Here's a link to explain the difference between pulse start and probe start MHs. http://www.lrc.rpi.edu/programs/nlpip/lightingAnswers/mwmhl/differenceProbePulse.asp
Just to let you know, the 750 is HPS, and 600w MHs are now commonly available. And, as I quoted, the new 1000w lumateks do not run the 750s anymore, they run 1000, 600, and 400 bulbs. They must be reading my mind on this shit. The best bulbs I've found so far are the sun pulse halides.
I also now have a sun blaze T5 ho I use for cloning. I was considering flowering with a T5 at one point, but unless I get some new info on this, I don't see a T5 ho being able to compete with my sun pulse halides at all. Especially when you consider the UV they put out, which takes the trichomes to another level.
I try to stay on top of all new technology, so yeah, I have heard of the sufur plasma lighting. Same with induction lighting, and a few other technologies that are elduding me currently. But I was already enough of a guinea pig with my lumateks when they came out. I destroyed multiple bulbs with them before I learned that MHs are the way to go with electronic ballasts. So I'm not too eager to be on the cutting edge, I just want to be legit.
I'm cutting my overall wattage to approximately 63% of what it was last year while hopefully maintaining my yield and quality. Can I do it? Guess we'll see..