electronic lead screw

bumping this back to the bottom, still looking for feedback on how to leverage haptic feedback/input controls

I wanted to post an update and ask for some feedback from folks. I've been playing with haptic feedback and I think you could make a pretty intuitive ELS controller. My requirement is to limit the buttons/knobs and be something you don't have to look at to operate. The idea is to have a single "smart lever" which can act as a toggle, or 3 position switch, or variable detent rotary encoder, even have variable spring load per mode (e.g. rapid mode would have a stronger "virtual spring") etc etc.

I'm wondering what folks think about the control setup I describe below?


P.S. This morning I have fixed the oscillations in the PID loop in a very robust way.
 
I like the whole concept of what you folks are working on, but it’s not for me. My lathe is easily switched to whatever feed and pitch needed just using levers, and after spending my entire career designing computers, processors, printers, disk drives, internet edge devices, control systems and software for all this, I am in no mood to mess up my hobby by pulling in a bunch of work issues. :)

But I applaud your progress and enthusiasm!
Some of us just like building stuff :), and in my case, my Takisawa TSL-800 came without any change gears, so it was a choice of not cutting threads, having gears made, or implementing an ELS. Like you, my electrical engineering and software background made me wary of complicating an otherwise simple tool. But I like mechanical things, and money-wise, it made sense to go this way, and it's working great.
 
I would agree that simplifying the controls is highly desirable when running the lathe or any machinery for that matter. The switches Clough I/O panel are too small and too close together to avoid diverting your attention from the machining. However, the only controls that I might be using while actually machining would be the FWD/REV and +/-. Separating the FWD/REV into two separate switch would be nice in that it would eliminate a double press and/or the need to verify the state visually.

I would like to see a OFF position as well so a three position switch would work for me. A rapid function would be useful for power feed. It couldn't happen with loss of sync for threading though.

Ideally, those controls that I would use while machining would be conveniently located with large buttons sufficiently separated from each other to permit "blind" operation.

As to your smart lever concept, I would think that it could work once you developed the muscle memory. I like to keep controls simple and having a one control does all is a bit troubling for me. As an example, my Logitech mouse has the standard left click/right click/ center click/scroll functions but also has a left and right center click and two left side buttons. Way too much for me. I have disabled all but the basic functions.

However, you are designing a control for your own use so what I think is immaterial. I do like the idea of a software defined switch. Presumably, it wouldn't be that difficult to program in various modes so the control could be fit to the desired task at hand. One thing that has to be considered when adding complexity to a machine is that a machine gone berserk can do a considerable amount of damage and /or be a safety hazard. The Boeing 737 MAX is a good example of this.
 
My weird situation is “I wish I could build an electronic lead screw but I don’t want to”.
 
I would agree that simplifying the controls is highly desirable when running the lathe or any machinery for that matter. The switches Clough I/O panel are too small and too close together to avoid diverting your attention from the machining. However, the only controls that I might be using while actually machining would be the FWD/REV and +/-. Separating the FWD/REV into two separate switch would be nice in that it would eliminate a double press and/or the need to verify the state visually.

I would like to see a OFF position as well so a three position switch would work for me. A rapid function would be useful for power feed. It couldn't happen with loss of sync for threading though.

Ideally, those controls that I would use while machining would be conveniently located with large buttons sufficiently separated from each other to permit "blind" operation.

As to your smart lever concept, I would think that it could work once you developed the muscle memory. I like to keep controls simple and having a one control does all is a bit troubling for me. As an example, my Logitech mouse has the standard left click/right click/ center click/scroll functions but also has a left and right center click and two left side buttons. Way too much for me. I have disabled all but the basic functions.

However, you are designing a control for your own use so what I think is immaterial. I do like the idea of a software defined switch. Presumably, it wouldn't be that difficult to program in various modes so the control could be fit to the desired task at hand. One thing that has to be considered when adding complexity to a machine is that a machine gone berserk can do a considerable amount of damage and /or be a safety hazard. The Boeing 737 MAX is a good example of this.


Thanks for the feedback. Looking to brainstorm with folks so any ideas you have are welcome. I also hope my lathe isn't quite like a boeing 737 but I do want it to be safe!

Reading your post i wonder about a 5 position switch mode for jogging:



Rapid Left--------Jog Left--------Do nothing--------Jog Right--------Rapid Right
|​
|​
|​
|​
torque 10​
torque 3​
torque 10

It is hard to describe in the video but the little gimbal motor puts out decent torque and it is good feedback.

For threading you could just omit the 2 left detents and keep the jog right/rapid right...
 
Last edited:
Thanks for the feedback. Looking to brainstorm with folks so any ideas you have are welcome. I also hope my lathe isn't quite like a boeing 737 but I do want it to be safe!

Reading your post i wonder about a 5 position switch mode for jogging:

Rapid Left -------- Jog Left ------- Do nothing -------- Jog Right ------ Rapid Right
| | | |
torque 10 torque 3 torque 10

It is hard to describe in the video but the little gimbal motor puts out decent torque and it is good feedback.

For threading you could just omit the 2 left detents and keep the jog right/rapid right...
Jogging isn't a concern for me as I can disengage the half nuts and crank the carriage manually. If it were a full CNC machine, then it would be.
While threading, none of the above can be done without losing sync. It would be nice to have a rapid return feature while threading and I had given some thought to it but couldn't come up with anything that would work.

Not a problem when Imperial threading on an Imperial lathe or metric threading on a metric lathe as the time proven method of releasing the half nuts and cranking the carriage manually works but this doesn't work when mixing inch and metric.

I had thought about using the enable function on the stepper driver along with a sensor to detect the end of a pass but gave it up after realizing it effectively took the electronic gear train out of gear and sync was lost. I could still use it for power feed though.

For the purpose of accurately threading to a shoulder, I decided to go with a mechanical solution. Actually, electromechanical as a sensor would activate a solenoid which would disengage the half nuts. The design is done and I have to start cutting metal.
 
Jogging isn't a concern for me as I can disengage the half nuts and crank the carriage manually. If it were a full CNC machine, then it would be.
While threading, none of the above can be done without losing sync. It would be nice to have a rapid return feature while threading and I had given some thought to it but couldn't come up with anything that would work.

Not a problem when Imperial threading on an Imperial lathe or metric threading on a metric lathe as the time proven method of releasing the half nuts and cranking the carriage manually works but this doesn't work when mixing inch and metric.

I had thought about using the enable function on the stepper driver along with a sensor to detect the end of a pass but gave it up after realizing it effectively took the electronic gear train out of gear and sync was lost. I could still use it for power feed though.

For the purpose of accurately threading to a shoulder, I decided to go with a mechanical solution. Actually, electromechanical as a sensor would activate a solenoid which would disengage the half nuts. The design is done and I have to start cutting metal.
Hiya Bob, lost contact w you! Glad u r ok:)Watched Greenail's vid. Clever, like the haptic concept very much, was always a nag in my mind re imminent machine crash, esp after having a parting off tool explode.
But I dont get how it would work with threading..
If the pass is say, too deep into the work piece, & the ave stepper current is too hi, what would the haptic sense be used to do? Slow the spindle, & therefore the stepper? Not necessarily a good thing to do, is it?
...
BTW, re rapid feed, my lathe has a power feed drive train, that could be used as a rapid feed.
But for those that dont have that feature, running the leadscrew at a fast pace for rapid, wouldnt that simply add to premature wear of the half nuts?
 
Last edited:
Hiya Bob, lost contact w you! Glad u r ok:)Watched Greenail's vid. Clever, like the haptic concept very much, was always a nag in my mind re imminent machine crash, esp after having a parting off tool explode.
But I dont get how it would work with threading..
If the pass is say, too deep into the work piece, & the ave stepper current is too hi, what would the haptic sense be used to do? Slow the spindle, & therefore the stepper? Not necessarily a good thing to do, is it?
...
BTW, re rapid feed, my lathe has a power feed drive train, that could be used as a rapid feed.
But for those that dont have that feature, running the leadscrew at a fast pace for rapid, wouldnt that simply add to premature wear of the half nuts?
As I said, I don't believe that it will work with threading. You can't change the relationship of the spindle with the lead screw without losing sync. That includes changing lead screw speed as well as reversing the lead screw. This would be akin to changing the gearing in the gear train. The half nuts are unique in that regard because we have a mechanism that keeps track of screw rotation and we are disengaging after the lead screw so it doesn't lose sync.

For threading, any remedy for problems caused by too aggressive a cut would have to be at the spindle, before the gear train in order to maintain sync. Basically, slow or stop the spindle. For power feed, we can decrease the feed rate or reverse it at will without penalty.

I am really not concerned about having a rapid power feed. It is a simple matter to disengage the half nuts and manually crank the carriage. I actually don't use power feed all that much because historically, it was a PITA to change gears in the gear train so for fifty years, I just moved the carriage manually. To a certain extent, I like the control that fully manual operation provides. The ELS has made it much more convenient to change gears and I am slowly adjusting to using power feed but for many operations, I don't even activate the ELS. (My system is set up so the power to the driver is switched separately so I can use the spindle speed display without turning the lead screw.)

If more advanced controls were desired, the Rocketronics system would be a better choice. It is a quasi CNC, detecting spindle orientation and adding power cross feed, enabling a fully automatic operation for both turning and threading. But then, you are close to full CNC. The disadvantage of the Rocketronics system is that you can't easily set up a new routine as you can by writing some G-code.

For a brief moment, I had considered going CNC. It would increase the versatility of the lathe. I resisted beaus I like manual operation. I am fortunate that I have a CNC mill and if I really needed to have the capability that a CNC lathe offers, I can run the mill as a vertical CNC lathe.
 
Hiya Bob, lost contact w you! Glad u r ok:)Watched Greenail's vid. Clever, like the haptic concept very much, was always a nag in my mind re imminent machine crash, esp after having a parting off tool explode.
But I dont get how it would work with threading..
If the pass is say, too deep into the work piece, & the ave stepper current is too hi, what would the haptic sense be used to do? Slow the spindle, & therefore the stepper? Not necessarily a good thing to do, is it?
...
BTW, re rapid feed, my lathe has a power feed drive train, that could be used as a rapid feed.
But for those that don't have that feature, running the leadscrew at a fast pace for rapid, wouldn't that simply add to premature wear of the half nuts?

Just to clarify my thinking: The brainstorming around a haptic input controller was not intended to solve all problems. The main problem it should try to solve is the complexity of the interface.

1607267571054.png


7 tiny buttons vs 2 or 3 buttons and 1 lever is the idea with most of the interaction being with the lever.

1607268540253.png

This one has 30 buttons!

The complexity traded off is perhaps the fact that the fewer controls become modal, meaning they behave different depending on the mode. Many modes are possible but too many modes creates it's own complexity. You "could" imagine a threading mode which would not fix issues with your depth of cut but would:

1. let you keep the half nuts engaged.
2. physically move to the neutral position when the thread pass is done, prompting you to retract the cutter.
3. allow you to jog (one detent right) or rapid (two detents right more torque on the 2nd detent) back to your thread start without having to stop/reverse/stop/reverse the spindle
4. physically move the control lever back to the neutral position when the thread start is reached if you have set a virtual stop.
5. the ELS syncs the thread start, no thread dial needed.
6. Offsets the cutter each pass to obviate the need to fiddle with a compound.

Ideally you would not have to ever look at the controller, your hands would tell you everything you need to know and you could focus on the thread you are cutting. Not having to look at it is the goal here.

This may not work well in practice, it is just an idea exploring what you can do when you can position the control lever and reconfigure it's detents and the detent strength arbitrarily. If you didn't want to rapid you could use at thread dial and disengage the halfnuts. That choice really has no impact on the discussion i'm trying to have. Saying "that is stupid, i've been doing it this way for 50 years" isn't really helpful when trying to develop something new. It may in fact turn out this idea doesn't go anywhere. I'd like to explore it a bit before I give up on it.
 
Just to clarify my thinking: The brainstorming around a haptic input controller was not intended to solve all problems. The main problem it should try to solve is the complexity of the interface.

View attachment 346275

7 tiny buttons vs 2 or 3 buttons and 1 lever is the idea with most of the interaction being with the lever.

View attachment 346276
This one has 30 buttons!

The complexity traded off is perhaps the fact that the fewer controls become modal, meaning they behave different depending on the mode. Many modes are possible but too many modes creates it's own complexity. You "could" imagine a threading mode which would not fix issues with your depth of cut but would:

1. let you keep the half nuts engaged.
2. physically move to the neutral position when the thread pass is done, prompting you to retract the cutter.
3. allow you to jog (one detent right) or rapid (two detents right more torque on the 2nd detent) back to your thread start without having to stop/reverse/stop/reverse the spindle
4. physically move the control lever back to the neutral position when the thread start is reached if you have set a virtual stop.
5. the ELS syncs the thread start, no thread dial needed.
6. Offsets the cutter each pass to obviate the need to fiddle with a compound.

Ideally you would not have to ever look at the controller, your hands would tell you everything you need to know and you could focus on the thread you are cutting. Not having to look at it is the goal here.

This may not work well in practice, it is just an idea exploring what you can do when you can position the control lever and reconfigure it's detents and the detent strength arbitrarily. If you didn't want to rapid you could use at thread dial and disengage the halfnuts. That choice really has no impact on the discussion i'm trying to have. Saying "that is stupid, i've been doing it this way for 50 years" isn't really helpful when trying to develop something new. It may in fact turn out this idea doesn't go anywhere. I'd like to explore it a bit before I give up on it.
You cannot interrupt the drive train while threading with the Clough ELS; either mechanically or electronically. Doing so will result in the stepper driver losing its reference. My understanding of the Clough logic is that counts are totaled from the start of threading, with packets subtracted from time to time to keep the count within range and some error correction is used. It does not track actual orientation od the spindle.

If you think about it what is required to successfully cut a thread, when the spindle is ina particular orientation, there is a define set positions of the carriage allowed. That set is uniformly spaced by a distance equal to the thread pitch. If you disconnect the lead screw driver from the chain, you will destroy that link. The same if you reverse the direction or change the feed rate/pitch. You can verify the latter two easily with the existing Clough ELS. When you reach the end of a pass, reverse the feed and back the cutter out and back the carriage up for another pass. The consequences of changing the feed rate or pitch is obvious enough not to require an actual run.

This is not to say that there isn't merit in your haptic control system. Just to say that for threading, there is no need to mess with any controls once set. As to adding extra utility,

I would like to see an electronic auto stop feature. My attempt to do this in a straight forward manned using the driver enable failed. For my driver at least, the order of operation is enable followed by direction followed by pulse. If pulses are being sent when the driver is enable, it goes haywire. It may be possible to gain access within the Clough firmware so that the pulse and direction signal are disabled when the enable output is disabled. In fact, it is probably being done for the forward/reverse function already. I'm not literate enough in the software language to ferret this out.
 
Back
Top