I also have a 1024VSL and, like most of us, just a single phase domestic supply. In no particular order, these were my thoughts/actions (and possibly specific to my machine – S&B do not seem to have been entirely consistent about the electrics in these machines):
Forget the star-to-delta motor conversion – the motor does not appear to be suitable for this. So I needed a 415V 3-phase supply.
There is an electro-magnetic brake seemingly hard-wired to the motor which only releases when the motor is under power.
A VFD hard-wired to the motor means that there is no supply to the control electrics without the motor “live”, so start/stop buttons will not function. My machine uses 110V “control” power for contactors, etc. Normally, this is provided by a transformer in the control box fed by the incoming 415V.
What I did was to use a 240V to 415V VFD. Actually, this was a bit of an electrical fudge built by a company that no longer exists and consists of a good quality ABB 415V single-phase in/three-phase out with a voltage doubler circuit on the input. The best thing I can say about this is that it mostly works – see below. As an aside, I am grateful to Clive who has posted above; this “VFD” was one of the things that his much better solution replaced and he kindly passed on to me. There are now better-looking solutions that do the same thing relatively cheaply from companies like Ecogoo. The VFD connects directly to the motor.
I wanted to keep the existing variable-pulley system and run the VFD at 50Hz output. I disconnected the internal transformer and fed the control box from an external 110V site transformer. Note that you need to remove the internal centre-tap earth in the site transformer. Note also that the speed-change motor, along with the oil pump and suds pump, are also three-phase.
Fortunately, I found that the main fwd/rev contactors had a spare contact (N/O) available. I used these to connect to the fwd/rev inputs on the VFD. I now had a system that would start/stop the main motor controlled by the existing buttons. I left the motor-to-control box connections in place. Note, when one of the motor contactors operates, the VFD is also feeding (in a sense) in the opposite direction to the internal three-phase bus, which is also used to feed the speed-change motor, etc. The clever bit about this is that although I reverse the main motor via the VFD, when the phase-swap happens (which is how a three-phase motor is reversed) the corresponding original contactor also switches that feed back into the control box with a similar phase-swap. The overall effect is that the speed change motor is always fed correctly and up/down speed change works consistently whichever way the motor is running. This takes a bit of thinking about, but it does work!
I could not see much use for the run/inch switch so I repurposed it. It now switches another VFD input which sets the output frequency to 25Hz for half-speed operation. I use this sometimes for thread-cutting. It can be switched on the fly while the motor is running.
I have set a ramp-up time for the VFD output which brings up both frequency and voltage during motor spin-up. I was concerned that the motor brake might not release early enough during the ramp-up but in practice this does not seem to be a problem.
And it all works! However, there is just one gotcha. The voltage doubler on the VFD input side feeds the VFD with what is, in effect, 25Hz. This is fine most of the time, but the machine load above about 2K rpm is such that the voltage doubler can’t cope and the VFD trips out with an input phase error. It’s OK when you test it with the spindle in neutral but the supply can’t cope with the bearing drag, etc, at the top end of the speed range. I keep thinking about replacing the VFD setup with one of the Chinese 240-in/415 out boxes (which are not a silly price and have had a good write-up elsewhere) but have not got round to it.
By the way, you really, really, need to know what you are doing when you take the covers off the control box. There is a lot of high-voltage “stuff” going on in there!
Having said all that, Clive’s solution is technically the better one!