Newbie- Help me retrofit a lathe with EMC please!

[RP Designs]

Ok, I am advancing with my retrofit of the lathe portion of a Smithy 1220Xl. The mill portion will be scrapped entirely as it is pretty worthless (and I have a cnc mill already). I really want to use EMC2 but I have to admit that it is probably my single largest source of anxiety for the whole project. I am not a programmer, I have no programming skills and I am most comfortable in the windows environment. I have no Linux experience and everything Linux just looks so foreign to me. I think that this is where I will need the most hand holding.

In terms of experience, this is my second build, the first being a retrofit on a full size CNC mill. I did take the easy way out and went the "kit" route with an AjaxCNC control.

For this build I want to do the whole build from scratch but will also need to try and keep the costs fairly contained, especially considering what the base machine is. My original plan was to try and use a servo for the spindle and to try and incorporate indexing into the spindle for use at a later date. I have seen high speed brushless DC servos that seem like they would be pretty good for the application and cost wise would be pretty competitive with an AC motor and VFD. My biggest concern with this set up is torque and I am just not sure that the servo would have enough torque even with gear reduction.

Thinking about the spindle more though, I will really be limited by software and most of the hobby software doesn't really support indexing on a lathe, so I really won't be able to take advantage of it, so is there a benefit? So right now I am leaning back towards an AC induction motor and VFD. Realistically I think mostly I will be doing basic lathe work with threading, so I really need to build that functionality in. Looking into some older threads I have seen some issues with single index pulse sensors versus encoders with EMC for threading. Has there been a consensus on what is the best hardware set up is? Obviously the single index pulse has an advantage in cost and encoders in accuracy, but what is easiest to integrate into EMC? I am in the planning stages right now so planning for it now is much easier than trying to later.

Has anyone done a similar build and want to share any advice as to hardware and software etc?

[chester88]

I have not done what you want - this is my opinion:
It is possible to have EMC control the spindle as a C axis (indexing, contouring).
It's just not commonly done so there are few example to look at.
The biggest pain (that I can see) is the mechanism for switching EMC from turning mode to C axis mode (spindle as axis) .
There seems to be no standard way yet, but a custom G code can do it.
Others have made two configure files one for using the spindle as a spindle, one for using it as a C axis.

Threading using an encoder will always be better.
Threading using a single sensor is possible (many big irons did it) - the bigger (heavier) the machine the better because the 'flywheel effect' keeps the spindle speed steadier.
Interfacing either to EMC is not particularly difficult - the encoder has more examples.
You didn't mention what interface you will use between computer and lathe nor what kind of motors you will use for the axis.
If your using a parallel port with software stepping and encoder reading, then that limits you a quite bit.

[RP Designs]

Thanks for your reply Chester. Nothing is set in stone at this stage since I am just in the planning stages. I will most likely do servos on the axis as for the drives and interface I have not decided yet. Obviously I need to keep the limitations of the base machine in mind here and not try and build something that will twist itself into a pretzel when I push it, reasonable accuracy and speed is all I am looking for. High productivity I don't expect.

[chester88]

Will you use step and direction servos? (eg gecko drives)
or analog servos?
the three main interfaces are:
direct parport - cheap but a little limiting
Mesa products - more expensive very flexible
Pico products - more expensive well proven

[RP Designs]

I am leaning towards the Pico system stuff to minimize integration. I just spent some time reading over on the Linux forum and Pico's website. I haven't ruled out the Mesa stuff yet though. I plan on using analog servos, I am looking at using BLDC.