retrofit on EMCO PC turn 55

chuck1024 wrote:

Wow, that looks like a perfect fit. I removed the sensors but I could put them back in. How much do you think it would improve the accuracy of the homing sequence?.

In the case of an encoder index, rather a lot, as those are very sharp, typically in the 1/1000 rev range. I think a Hall sensor might be a little more variable, but even if it is as variable as the home switch it improves matters proportionally to the leadscrew pitch.

What I am not at all clear on is where in HAL you would connect that input. www.linuxcnc.org/docview/html/man/man9/motion.9.html suggests that it relies on the index-enable functionality of the encoder components. In a stepper system there are no encoders to which to net that input.

It might be that the way to get the function is to AND the home switch with the Hall sensor in HAL. There will be problems if the Hall sensor on range is less than the axis stopping distance. But it probably won't be.

Perhaps the easiest solution is to wire the home switch and the proximity (index) sensor in series. Both home switches on the Emco are roller switches and the actuator is a ramp so they can over travel quite a bit. As you eluded to, if you over run the index pulse, things will go haywire. I think the home search speed is slow enough this would not happen though.

cs

I got the lathe! Unfortunately, it was dropped along the way, but it seems like $100 will cure what at first seemed like terrible damage. The carraige had become detached. However, getting this re-attached amounts to just replacing $30 worth of retaining strips.

It turns out my motors are different from Chuck's. They are 3 phase NEMA 23 motors, VRDM 366/60, which are probably driven sinusiodally. I should be able to swap them right out, except for needing to drill a hole in the shaft for a retaining pin.

I spent about an hour trying to work through the stepper board. I didn't make much headway--it seems like every trace has at least one through hole. There's a 16-pin connector that somehow translates into up to 3 axes but is configured with 2 motor power boards, each with 6 transistors. This would probably yield its secrets in a few hours if I were disciplined about tracing it through and writing eveything down.

jrkeat wrote:

They are 3 phase NEMA 23 motors, VRDM 366/60, which are probably driven sinusiodally.

You can quite possibly drive those with a generic 3-phase motor drive.
Part of the alignment sequence for brushless servos can be a phase of open-loop drive, basically running the motor as a very low pole-count stepper.
It would all be a bit unusual, as you would have infinitely variable current control in software, for example.
A single Mesa 7i39 could probably drive both axes. Or perhaps not…

I really do not understand the obsession with reusing the original motors. Most likely it will lead to a lot of wasted time and expense buying and experimenting with hardware that may never be used in the end. Best case ending is OEM performance of 30 IPM which isn't all that great.

Replacing the steppers and using a G540 is a very well defined solution that is very straight forward and easy to implement. The cost is known and the results are predictable and the performance will be double the OEM speeds.

chuck1024 wrote:

I really do not understand the obsession with reusing the original motors.

A good point, well argued. The only reason I am suggesting it is that you said that the flange and shaft sizes were not actually NEMA23, and that modifications would be needed.

The three-phase steppers sound like they might be rather nice though.

If I was converting one of these machines (and one was rather tempting me on eBay last week) I think I might look at a servo conversion, rather than swap the steppers. Mainly because I have the servos, though.