Replace change gears with a stepper?

T

ttabbal

H-M Supporter - Gold Member
H-M Supporter Gold Member
Joined
Jun 12, 2017
Messages
2,139
To be clear, I do NOT want to go full CNC. I wonder about losing the gears and using a stepper to drive the main gearbox, which looks like it has 1:1 as an option. Maybe even using a pair of gears for reducing the required torque. There's already an RPM pick up so I could use that for timing or add another hall effect sensor. For bonus points, add a timing system for metric threading and auto stops at the ends of travel via limit switches or input from a DRO scale.

All the controls would still be manual, but I could easily switch turning speeds and thread pitches. Sort of like an electronic QCGB.

The idea seems simple enough, but I don't see it done. It would also make a nice indexer if I had it drive the spindle instead.

Random musings from what passes for a brain in my head... :)
 
ttabbal said:
To be clear, I do NOT want to go full CNC. I wonder about losing the gears and using a stepper to drive the main gearbox, which looks like it has 1:1 as an option. Maybe even using a pair of gears for reducing the required torque. There's already an RPM pick up so I could use that for timing or add another hall effect sensor. For bonus points, add a timing system for metric threading and auto stops at the ends of travel via limit switches or input from a DRO scale.

All the controls would still be manual, but I could easily switch turning speeds and thread pitches. Sort of like an electronic QCGB.

The idea seems simple enough, but I don't see it done. It would also make a nice indexer if I had it drive the spindle instead.

Random musings from what passes for a brain in my head... :)
Click to expand...

It sounds like you will do threading. It will work to go stepper, but you absolutely have to keep the lead screw in sync with the spindle if you do any threading, or the threading will fail miserably.
 
No reason it wouldn't work. As long as you have a way to electronically gear the stepper to the spindle at the proper ratio it would work fine. This really implies some kind of computer or at the very least some kind of adjustable counter system to set the gear ratio. A QCGB is just a mechanical computer.
 
Thanks for the info. I was worried I'd missed something about it. The plan would be to use a microcontroller to control the stepper and watch the spindle speed to maintain sync. I am far more familiar with electronics than machining. :) I'll probably have more trouble adapting the stepper shaft to the lathe.

Any ideas on how much torque one would need for something like this on a PM1127?
 
ttabbal said:
Any ideas on how much torque one would need for something like this on a PM1127?
Click to expand...

That is a very good question. And the answer is: I have no idea. It would seem that threading would put the greatest load on the system. So the question is how do you read the torque on the lead screw with a threading setup? I'm thinking a couple of spring scales, one pulling on a string wrapped around the gear on the lead screw to get the torque, but I'm not sure what to do with the other one. :)

Maybe just one spring scale would work as above, but set up for a heavy cut and just feed the lead screw with the spring scale and the string. Anything to get a rough idea of the max torque required.
 
You would be looking at a system similar to the Babin Hardinge HLV-H toolroom lathe with servo electronic digital threading, so would need more precise spindle position and probably some form of feedback with an encoder. The encoder could also give positional control if the half-nut stayed engaged, otherwise you would need some form of electronic scale or carriage drive for positional control. Also see Electronic Lead Screw (ELS) below, they also have a Yahoo forum. At the end of the day, you may have more cost tied up in the threading system then the cost of the lathe. It is doable, but may be a bit more complicated then anticipated. I have been toying with a similar idea to cut tapers with an electron feed on the cross slide coupled to some form of carriage positional control, use a PLC or small computer to set the angle and the stepping motor increment.
http://www.babinmachine.com/index.php?HLVELECLEAD
http://www.autoartisans.com/ELS/
 
A retired electrical engineer myself. With (sorta) machine shop capability for modeling. There is no way I would trust a stepper motor to handle threading. A stepper has a fixed rotation, so many jerks per revolution. At best you would need an analogue servo. But that would still need to be geared into the system. And I personally still wouldn't trust it.

Looked at end on, a thread is a smoothly tapering spiral around the shaft. It has a small but measurable step for each "Nth" degree of rotation. If you could figure out some way to make a stepper with 720 steps per revolution, it might conceivably work. Maybe...... Then imagine the number of steps to do a really coarse thread over say a 3/4-10.

Now, I work with smaller threads, more like 0-80. 1/2-20 is huge to me. In any case, I would want the threading train directly coupled to the spindle so a bobble in speed would show up as the same bobble in the thread.

Just an opinion, of coarse, but quite staunch.
 
You might take a look at a used Parker Compumotor SX6 stepper controller (very cheap on ebay). They are an all in one stepper driver, micro controller and powersupply. A SX6 will give you a lot more capabilities than a simple stepper indexer (at least the simple indexers I have worked with)... like calculating load from the current the stepper is using and allowing you dynamically adjust based on load. From the work I have done with them I think they would have enough intelligence to do what you need. They have several programmable I/O's and their own movement control programming language that wasn't too hard to master. I specifically remember a trace function that you can use to set the movement of the motor being controlled based off the movement of another motor... helping with the bobble issue... if the control frequency slows then slow the indexing proportionally.

Toothed belts with a good amount of reduction should give you the torque and resolution you need with almost no backlash.

It has been a while snce I have programmed one so I can't guarantee they will do what you need but I believe they will.
 
Last edited:
Oh it would work just fine Bill
until lightning struck nearby while you are threading that 5000$ NASA part
Mark
 
It's true that steppers have downsides. One I had overlooked a little is missed steps. Not a huge issue if you have sufficient torque, or position feedback, but a possible issue anyway.

I also might be underestimating the timing requirements. For feeding, I think it would be fine. Threading could well be another issue. I need to consider the timing and accuracy requirements more as it's obvious from some of the responses that there may be more precision required than I'm considering.

Resolution is an issue. 200 steps / rev is common. 400 is available reasonably. And you can gear them, but that introduces speed limits and potential errors. Large servos are a possibility, but those get expensive fast. I don't see myself doing 80 tpi anytime soon, but it's important to know what the limitations would be.
 
You must log in or register to reply here.
Back
Top