-
Welcome back Guest! Did you know you can mentor other members here at H-M? If not, please check out our Relaunch of Hobby Machinist Mentoring Program!
electronic lead screw
- Thread starter greenail
- Start date
That chart doesn't make sense for a 9Nm stepper. that is showing .7Nm at peak?
anyway, I'm using a 2.2Nm hybrid stepper direct drive on a 2mm pitch leadscrew and it happily stalls out my spindle motor before it faults. I've threaded hardened shafts and the insert is more of a problem than the stepper. I'd guess that Clough's firmware would also stop driving the lead screw when the spindle stopped. I guess if you are taking huge cuts and or cutting huge threads you might need such a beast but for < 2mm pitch I can't see much need to go bigger. Are you planning on cutting acme threads or something?
Finally consider that motors don't just start stop instantly, there should be an acceleration curve (or trapezoid). I doubt the controller can compensate for crazy spindle speeds and the sync issues decelerating at the end of your thread. The larger the motor and the more inertia it has which makes that acceleration curve flatter. at some point it can't keep up no matter how much torque you have unlike a gear train which just breaks when you run it too fast. everything is a trade off
math
decimal go too far.... chart good! still odd it only goes to 7Nm...likely micro stepping steps, most steppers come in either 200 or 400 unless they happen to be super special and 3 phase, or come with a gear box.
total steps = micro steps * motor steps (200)
Thanks for the input!
Yeah, I was confusing the 1000 lines of the built in encoder with the motor steps.
I've not read the docs for that device but one additional config confusion thing is also the quadrature mode for the "decoder", assuming the highest resolution algorithm for quadrature you get 4 signals per encoder line. It could be 2x or 1x though. HomoFaciens has a good video on how it works here. PPR (or LPR) is the line count, counts per revolution CPR is the quadrature counts per revolution (4xPPR in 4x decoding mode).
You can actually make a very decent optical encoder pretty easily with about 0.50 worth of parts but it will not likely be high enough resolution for this application. The "dexter" robot has a pretty interesting 3d printed encoder setup that uses a combination of quadrature encoding and sin/cos encoding of the analog photo diode signal.
I've played with a similar approach using hall effect sensor arrays and refrigerator magnet strips for linear position encoding which I should take back up again (it has been sitting on the shelf for a year now
) since it would make a fairly useful and inexpensive DRO.
- Joined
- Feb 1, 2015
- Messages
- 9,623
It's 700 N-cm or 7 N-m. A stepper torque is highest at zero rpm. Stepper torque is also dependent on driver voltage. I believe the literature says that driver is capable of running a 100 volt supply.That chart doesn't make sense for a 9Nm stepper. that is showing .7Nm at peak?
anyway, I'm using a 2.2Nm hybrid stepper direct drive on a 2mm pitch leadscrew and it happily stalls out my spindle motor before it faults. I've threaded hardened shafts and the insert is more of a problem than the stepper. I'd guess that Clough's firmware would also stop driving the lead screw when the spindle stopped. I guess if you are taking huge cuts and or cutting huge threads you might need such a beast but for < 2mm pitch I can't see much need to go bigger. Are you planning on cutting acme threads or something?
Finally consider that motors don't just start stop instantly, there should be an acceleration curve (or trapezoid). I doubt the controller can compensate for crazy spindle speeds and the sync issues decelerating at the end of your thread. The larger the motor and the more inertia it has which makes that acceleration curve flatter. at some point it can't keep up no matter how much torque you have unlike a gear train which just breaks when you run it too fast. everything is a trade off
The question at hand was how much torque does the OP need. In my experience, I would have no problem cutting any threads. The problem may come when using power feed for roughing cuts. However, I wouldn't be making a roughing cut at 2400 rpm. For one, my motor would stall out would stall out.
As to the acceleration and deceleration, the stepper is synced to the spindle and as long as the spindle is turning, the stepper is turning. When cutting metric threads on an imperial lathe or vice versa and you are forced to stop the motor and reverse for the next pass, the lathe motor inertia should overide any acce4leration/deceleration issues with the stepper. This may not be true if you have clutch and mechanical brake on the lathe though.
- Joined
- Feb 1, 2015
- Messages
- 9,623
I believe that Clough used a hybrid servo on his lathe which is why he set the resolution at 1000. The encoder on the hybrid steppers performs a different function. It monitors lost steps and attempts to inject additional steps to bring the stepper back in sync. Additionally, there is a user selectable error limuit which, if exceeded, shuts the driver down requiring a power on reboot. The default setting on my driver was 1000 encoder steps or one revolution of the stepper motor. When threading, an error like 999 counts would be excessive but I don't believe that I would encounter that situation. Using power feed, loss of sync isn't an issue. Loss of steps would be due to an insufficient force meeting a insurmountable object and shutting down the drive could be viewed as a good to have safety feature.
I set my ELS up with separate power switches for the ELS control and the stepper drive. This allows me to power down the drive while still having the ELS operating. Thus, my lead screw isn't turning unless I want it to turn. It's also more convenient for rebooting the stepper drive.
I set my ELS up with separate power switches for the ELS control and the stepper drive. This allows me to power down the drive while still having the ELS operating. Thus, my lead screw isn't turning unless I want it to turn. It's also more convenient for rebooting the stepper drive.
I've not read the docs for that device but one additional config confusion thing is also the quadrature mode for the "decoder", assuming the highest resolution algorithm for quadrature you get 4 signals per encoder line. It could be 2x or 1x though. HomoFaciens has a good video on how it works here. PPR (or LPR) is the line count, counts per revolution CPR is the quadrature counts per revolution (4xPPR in 4x decoding mode).
You can actually make a very decent optical encoder pretty easily with about 0.50 worth of parts but it will not likely be high enough resolution for this application. The "dexter" robot has a pretty interesting 3d printed encoder setup that uses a combination of quadrature encoding and sin/cos encoding of the analog photo diode signal.
I've played with a similar approach using hall effect sensor arrays and refrigerator magnet strips for linear position encoding which I should take back up again (it has been sitting on the shelf for a year now
Thanks for the reference, I’ll check it out.
That’s a very clear explanation.
BTW, I ordered the motor, driver, and power supply last night. I switched gears again and got the Automation Technologies stuff. The Stepperonline shipping charge was going to nearly double the cost and they have a week delay now for a holiday. For about 20% more, I’ll have it much sooner and, and I suspect, better support.
Also got motor cable, a variety of connectors, IEC power cord, case fan, etc. I’ll wait until I lay it all out before getting an enclosure.
Today, I shop pulleys and belts.
One odd thing about the lathe. I finally got around to measuring the leadscrew and it is imperial! The feed rack and pinion, and cross slide screw, are metric. (They have inch “conversion” dials, approximately 4” per turn, with a small accumulated error).
This explains the odd collection of bushings and couplings that attach the leadscrew to the QCGB. I guess a shop converted it in house. If so, they did a nice job with the half-nuts, doesn’t look bodged up at all. Or maybe it was a factory option?
Anyway, the leadscrew is 4tpi and 1-1/8” dia. Half-nuts are bronze.
The feed shaft has a single spline, and drives much more freely.
Fiiiinally got all the parts in.
There’s a new development that brings up a problem. Of course.
My initial plan was to mount the hybrid servo motor within the lathe bed, similar to Gus' idea (as posted in a Github issue), and drive the gearbox input with the servo via a M5 belt. This would let me put an idler on the change gear banjo and I'd have an easy way to adjust the belt tension by moving the idler up or down. But two things interfere. Most importantly, the QCGB is so badly damaged that I only have a couple of working ratios out of over a hundred possibilities. The simplest one is 32:15, gearbox input:leadscrew. I was worried that adding those factors to the numerator and denominator calcs might be too much. (James had advised Gus against using his 120 and 127-tooth change gears between servo and leadscrew, because the math might be too much heavy lifting for the microcontroller). Then I found that my initial measurements were wrong, and placing the motor inside the lathe bed would be extremely challenging, so it's all moot anyway.
So I decided to add an extension to the tailstock end of the leadscrew and drive it from there. I gutted the gearbox, leaving only the gears which connect the leadscrew and feed shaft. They run 1:1.
Here's the problem - the feed shaft is geared within the carriage apron to a ratio of (approximately) 19 turns per inch of carriage travel. So the leadscrew and feed shaft produce different rates of carriage movement when turned at any given stepper value. I don't want to abandon the feed shaft and do feeds with the 4TPI Acme screw, so I'm trying to figure out the best way to use both. I have a few ideas.
Solution 1: Alter the Feed formula by 19/4ths. Actually, I'd just round it off to a factor of 5 in the numerator, that's only a 5% error in feed rate.
Solution 2: Define a new variable (say, "feedshaftrate") to replace the leadscrew TPI variable in the Feed formula. Set the new variable to 19 TPI and I'm done. However, I have Zero coding experience, so I'm not sure I'm up to it.
Solution 3: Change the Feed tables, so the the value displayed on the LED panel is low by a factor of 5. So for example, selecting .010"/rev (as displayed) would actually be telling the controller to feed at .050"/rev, but the slower feed shaft rate would actually result in a feed of .0105. Close enough. I'd have to add some rows to the table at the top end to get the faster speeds, and delete the slowest speeds. Again, Absolute Zero coding experience.
Okay, what would you do? Am I missing an easier answer? Any input will be appreciated, so thanks in advance.
There’s a new development that brings up a problem. Of course.
My initial plan was to mount the hybrid servo motor within the lathe bed, similar to Gus' idea (as posted in a Github issue), and drive the gearbox input with the servo via a M5 belt. This would let me put an idler on the change gear banjo and I'd have an easy way to adjust the belt tension by moving the idler up or down. But two things interfere. Most importantly, the QCGB is so badly damaged that I only have a couple of working ratios out of over a hundred possibilities. The simplest one is 32:15, gearbox input:leadscrew. I was worried that adding those factors to the numerator and denominator calcs might be too much. (James had advised Gus against using his 120 and 127-tooth change gears between servo and leadscrew, because the math might be too much heavy lifting for the microcontroller). Then I found that my initial measurements were wrong, and placing the motor inside the lathe bed would be extremely challenging, so it's all moot anyway.
So I decided to add an extension to the tailstock end of the leadscrew and drive it from there. I gutted the gearbox, leaving only the gears which connect the leadscrew and feed shaft. They run 1:1.
Here's the problem - the feed shaft is geared within the carriage apron to a ratio of (approximately) 19 turns per inch of carriage travel. So the leadscrew and feed shaft produce different rates of carriage movement when turned at any given stepper value. I don't want to abandon the feed shaft and do feeds with the 4TPI Acme screw, so I'm trying to figure out the best way to use both. I have a few ideas.
Solution 1: Alter the Feed formula by 19/4ths. Actually, I'd just round it off to a factor of 5 in the numerator, that's only a 5% error in feed rate.
Solution 2: Define a new variable (say, "feedshaftrate") to replace the leadscrew TPI variable in the Feed formula. Set the new variable to 19 TPI and I'm done. However, I have Zero coding experience, so I'm not sure I'm up to it.
Solution 3: Change the Feed tables, so the the value displayed on the LED panel is low by a factor of 5. So for example, selecting .010"/rev (as displayed) would actually be telling the controller to feed at .050"/rev, but the slower feed shaft rate would actually result in a feed of .0105. Close enough. I'd have to add some rows to the table at the top end to get the faster speeds, and delete the slowest speeds. Again, Absolute Zero coding experience.
Okay, what would you do? Am I missing an easier answer? Any input will be appreciated, so thanks in advance.