Most of you might recall I was working on converting my Logan lathe over to CNC last year, and got it pretty darn close to... well, not 'done', but at least functional.
My main problem, however, is my near-total lack of familiarity with Mach 3, G-code in general, and the whole process of trying to model a part, convert that to a toolpath, transfer the toolpath to the machine, etc.
It's all stuff I need to learn, of course, but with all the other things drawing my attention these days, it's hard to sit down and concentrate much on it.
And in any case, I always knew Mach 3 was kind of a "placeholder" program- one of M3's biggest drawbacks is that it can't support more than a single-pulse encoder. That is, just a single signal "bit" every revolution of the spindle.
That's sufficient to tell M3 the spindle is turning, and how fast, but it's barely enough resolution for basic threading (assuming you don't have a slippery belt or weak motor) and nowhere near enough data for rigid tapping or multi-lead threading.
The original plan was to switch to something like LinuxCNC, which can do all of those things, but of course is just another layer of complexity for me- I've never dealt with Linux, and since the whole thing is open-source, there's a lot of manual work you have to do to add 'modules' and other capabilities to the basic OS.
However, earlier this past summer, I heard about a new option by a company by name of Centroid, who have been producing well-respected industrial CNC controllers for some 35 years now. They've ported their industrial software over to a new DIY-type CNC controller called an Acorn, which has a lot of very nice features.
The first and foremost of which is that it can accept a proper quadrature encoder, so that the controller can not only know how fast the spindle is turning, but also what direction. Centroid recommends a 2000-line encoder, which produces 8,000 pulses to the controller, so that it can tell- and adjust accordingly- if the spindle is, for example, slowing down during a heavy cut, as well as being able to determine start locations for multi-lead threads.
I waited patiently until last week when the Lathe software was finally released, and ordered one, along with several other goodies in order to almost completely rebuild the entire control system for the lathe.
It's not all in yet (tracking says my new speed controllers won't be here 'til the middle of the month) but late last week I got in the spindle encoder and some generic pulleys in order to build an offset 1:1 drive for it.
I spent... okay, maybe a bit too much time on it this weekend but I've been champing at the bit to finally get this thing properly running. It's been sitting there in the corner, not entirely dead, but might as well be. (I've managed a few very simple operations on it, but it usually takes so long to make it work right, I might as well just make the parts manually.)
I knew the encoder drive would take a bit of work, so I wanted that completed and tested before the actual controller got here, so I can hopefully jump right in once all the parts are in.