Hi Matt,
I wanted to follow up on with some more information and a users perspective of the generator described in my previous post.
This unit is skid mounted, approximately 51"LX31"W and weighs about 1000 lbs fueled. It's also rated at 26A for the 240V single phase, I don't remember it's 3-P amp rating. It is also an 1800 rpm unit, electric start, and starts on a 24V battery system. I'm not sure of the fuel tank capacity, but looks to be between 5 and 10 gallons.
One day, I wanted to run it long enough to burn up the fuel and refill with fresh diesel. Of course, with no load it isn't doing any work and just sips fuel. After four hours and no noticeable change on the fuel level gage, I started loading it. My shop is wired so that I can disconnect from the outside world and plug the entire shop into the generator.
First off I started my 5 hp RPC (it had the wrap cord around shaft, pull cord, and throw switch, starting system). Then I turned on my 3hp mill, 3/4hp lathe, 1/3hp drill press, 1hp T&C grinder, all are three phase. Then I turned on all the lights in the shop (1kw) and turned on the 1/2hp and 1hp pedastal grinders. Granted, all of these machines are spinning free without doing any work. At this point, I had the generator running at half it's rated load. An hour later, the fuel gage had dropped by 1/8 a tank and I gave up, deciding the fuel in the tank was just fine.
The 3hp mill didn't start as snappy as it does on line power, but it wasn't bad. I think if you slowed a vari-speed system down before turning it off, you could restart it on a smaller generator, since it's "geared down".
One other time, I was testing a 6X18" surface grinder, don't remember the motor size, but probably 1hp. The generator was hooked up 3-P directly to the grinder. The grinder started right up and ran fine.
However, imagine a grinding cycle. The table reciprocates beneath the grinding wheel. During each pass, the motor momentarily requires additional amps for the grind, then it runs free until the next pass. In practice, at every pass, the additional amp draw made the generator surge to meet the demand. By the time the mechanical system, "governor" responded to the electrical demand, the grinding pass had already been completed and the demand removed.
The result of this situation is a perceivable slowing of the grinding wheel during each grind pass and a loping generator trying to keep up with the intermittent demand, but not reacting fast enough. It worked ok for our testing purposes, but I don't think this is a good situation. I don't know how powerful the generator has to be before this "surging" would not be noticable.
Hopefully, my experiences will give you some insight in choosing the best generator for your own shop.
John