Logan Model 200 Servo Threading

[Briney Eye]

Just throwing this out there to see if anybody is interested. I see a fairly straightforward way to set up a Model 200 for servo threading that would be easier than retrofitting a quick-change gearbox while also being much more flexible.

Conceptually it's simple. Put an encoder on the spindle and a servo motor on the lead screw, with some digital intelligence in between and a user interface to select the ratios. Feed rate would become continuously variable, with simultaneous display of both axes. The noisy change gears would go away. Metric would become easy. The function of the forward/reverse lever, half nut lever, and crossfeed plunger would remain the same.

I would avoid "modes" and go with a full-time display of feed rate and pitch in inches and millimeters, with a "highlight" when you landed on a standard pitch. I see a clear-lidded box with a power switch, a display, and a knob, mounted where a gearbox normally sits. When the box was on, the lead screw would behave as if there were still gears coupling it to the forward/reverse lever. You would still use the thread dial. And of course the lathe could still be returned to its original state.

I possess the necessary skills (from nearly 40 years doing this sort of thing), and think it would make a fun (and functional) project.

What are your thoughts?

 

[markba633csi]

It's been proposed but I have not seen a working example. The trick of course is to make sure the leadscrew never becomes de-synchronized with
the spindle. The circuitry should be immune to electromagnetic interference and you would need a backup system for power failure
Mark S.

 

[T. J.]

I have no personal experience, but I've heard this one mentioned on the Atlas 618 Yahoo group before:
http://www.autoartisans.com/ELS/

There is also a Yahoo group called "Electronic Lead Screw".

 

[Briney Eye]

I have no personal experience, but I've heard this one mentioned on the Atlas 618 Yahoo group before:
http://www.autoartisans.com/ELS/

There is also a Yahoo group called "Electronic Lead Screw".

There are quite a few projects like that floating around. I just think that it would be worthwhile to do something without all the bells and whistles. I'm not shooting for a CNC lathe. I want something very straightforward, without menus or modes, no programming, no tapering or jogging, just a "continuously variable" gearbox. If the spindle is turning and the forward/reverse lever is engaged, the lead screw turns at a proportional rate. What got me started was wishing that I could slow down the cross-feed. I even bought a DC gear motor and worked up some plans, but I eventually realized that this approach would be more worthwhile.

"All" I have to do is attach a suitable encoder to the stud gear and servo motor to the lead screw, make it all fit under the cover without modifying any of the existing parts, and write a program. Nothing to it .

 

[T. J.]

Nothing to it .

I start out pretty much all of my projects thinking like that...

I don't have the electronics knowledge to contribute meaningfully to a project along those lines. But I do agree that it would be an improvement to these smaller Logan lathes. I have the quick change gear box, but still the finest feed available is 0.004" per rev.

 

[markba633csi]

My thoughts, If I had the coding skills to do it (which I don't, yet) would be to make it a closed loop rather than an open loop one.
In other words, instead of just one optical encoder on the spindle, have two; one on the spindle and another on the screw. Then your
microcomputer would act as a comparator in addition to a ratio-changer hence it would never get out of sync. I'm sure it's been done.
Mark
ps if you didn't care about cutting threads then an open loop system might suffice, or simpler still, just a geared-down motor drive; no micro.

 

[Briney Eye]

My thoughts, If I had the coding skills to do it (which I don't, yet) would be to make it a closed loop rather than an open loop one.
In other words, instead of just one optical encoder on the spindle, have two; one on the spindle and another on the screw. Then your
microcomputer would act as a comparator in addition to a ratio-changer hence it would never get out of sync. I'm sure it's been done.
Mark
ps if you didn't care about cutting threads then an open loop system might suffice, or simpler still, just a geared-down motor drive; no micro.

Thanks, I should have been more clear. Any servo motor that I used would certainly include an encoder. I'm enamored with the Teknic Clearpath motors. This project would be an excuse to try one. Their standard models have, I think, 800 counts per revolution, so I would want a spindle encoder with at least that resolution.

I'm pretty sure that one of the Clearpath motors will do pretty much all the hard work, in fact, by connecting the spindle encoder directly to it and programming the ratio. I just have to get one of the models that I can "talk" to and then the control box really is just a front end.

 

[Briney Eye]

My thoughts, If I had the coding skills to do it (which I don't, yet) would be to make it a closed loop rather than an open loop one.
In other words, instead of just one optical encoder on the spindle, have two; one on the spindle and another on the screw. Then your
microcomputer would act as a comparator in addition to a ratio-changer hence it would never get out of sync. I'm sure it's been done.
Mark
ps if you didn't care about cutting threads then an open loop system might suffice, or simpler still, just a geared-down motor drive; no micro.

I apologize for my initial misunderstanding of your comment. A true closed loop system with the controller monitoring encoders on both the spindle and screw might be ideal, but I think I'll have to leverage the Clearpath motor's built-in brain to tell me if it gets out of sync. But that's a "feature" of this approach, anyway. I will be able to program the Clearpath so that I have an "overload clutch" on the feed, something the Model 200 sorely lacks (I know this from personal experience).

Also, after reading the Clearpath documentation more closely it's become clear that I will still have to do most of the work. While they have quadrature input and scaling capability, it's limited to multiples of 800 encoder counts.

I'm going to start by trying to implement this on an Arduino with a small touch screen, and see if I can do quadrature decoding with interrupts. At 1440rpm a 100ppr encoder will be interrupting almost 10,000 times per second, which I'm sure is at the limits of what an Arduino can do. If it can't keep up I'm looking at a programmable logic chip, a PIC32 microcontroller setup with dedicated quadrature decoding hardware, or a Cypress "Programmable System On a Chip" micro which I know supports quadrature decoding logic blocks. Years ago I did a lot of quadrature decoding with programmable logic for robotics applications. Dedicated hardware is really the way to go, so I probably won't spend too much time trying to make the Arduino work before I punt.

 

[Briney Eye]

Well, the first small step is a success. I've hooked up a 200ppr (pulse per revolution) encoder to my Arduino Uno. With quadrature encoding, that equates to 800 counts per revolution. Starting with the SpeedTest sketch from the Encoder library I kludged up something that seems to be working fine at over 1400rpm. First I tried SpeedTest straight, which pulses a pin that you can integrate with a DVM and see the average voltage decrease as the interrupts come faster, and it looks like it takes about 17% of the processor time at 1400rpm. Then I substituted a print statement for the PWM code. When I spin it up my oscilloscope says one phase is running at 4.7khz. If I run it for ten seconds, it comes very close to the 188,000 expected counts. I guess I'll have to hook up the 360ppr encoder and see how that works, and then try the 600ppr encoder (which ought to soak up about 50% of the CPU cycles).

The performance is much better than I was expecting. I'll have to study the code.

It's dead quiet, too.

This is fun!

 

[RandyM]

Excellent! I am glad it is working out for you with the added bonus of FUN. You are doing something most of us would not even consider. If you could add pictures of your work we'd greatly appreciate it. Photos really add a lot to the story, particularly for those of us that are more than out of our element. Keep up the good work.