开云体育

Long answer with some ELS theory.

Using this web page as a reference, if you are cutting clock parts or tiny robotic stuff with diameters below 0.25", for example, the SFM for brass and aluminium goes up as high as 300 SFM (for Al).


So pick the high end and turn a piece of 0.25" brass down to 0.1875" that will serve as a spacer and perhaps be threaded at one end. With 200 SFM we get 4200 RPM and that's with high speed steel tool bit. At the low SFM end it's 2133 RPM. The web page states carbide can run 2 to 3 x that speed.
So pick aluminium at 0.1875 with carbide at 2x the speed = 600 SFM. Yikes! That's 12,800 RPM.

Even this kit from Sherline wouldn't let you get there but normal max speed for Sherline is 2800 RPM for turning.


That's over the high end of 5000 RPM for a tooth belt coupled encoder with 3:1 ratio so we're back to wanting an encoder disk right on the spindle. Assuming once can get one with a bore as large as the outside of the lathe which for my South Bend I can't but then I also can't run it really fast either.

Say a Sherline at 2800 RPM. If the lead screw is 20 TPI it's got a pitch of 0.050". IF you wanted to cut a 20 TPI thread you turn at the same speed as the spindle so the lead screw moves 0.050" per spindle rev.

Smooth turning is usually done with a larger radius tool bit at about 0.003" per rev which is close to 1/17th the lead screw pitch so the lead screw moves at 1/17th the spindle speed and that's only about 170 RPM.

Therefore a 100 line encoder that produces 400 pulses per rev results in 18666.6666 pulses per second or one about every 50 uS. Theoretically the RELS unit should be easily able to handle a Sherline since for turning the software uses only one rising edge while for threading it uses two rising edges. It doesn't use quadrature. That means a 200 line encoder for the spindle would also work.

Not only that, it appears to be limited to handling only 4 micro-steps per step so the output frequency for the carriage isn't nearly as high as a system that needs to handle a Gecko Drive at 10 micro-steps per step.

As I understand the poorly documented RELS software, the end result is the number of encoder pulses from the spindle must always be larger than the step rate of either of the axis. Most stepper motors have trouble running more than about 750 RPM before their torque falls off so badly that they lock up and lose position.

Gecko gets around that in an interesting way that I'd really like to try one day. They use 10x micro-stepping until they pass the resonance point where the torque on the motor dips. After that point they switch over to full step mode which provides much more current for a longer time period and allows the motors to reach much higher speeds.

But again in the above example we require the stepper to only turn 170 RPM or 2.8 RPS which is only 5666.666667 steps/second with a 10x micro stepping driver. And that's lower than the 100 line quadrature spindle rate so the lead screw needs one step for every 3.294117645882353 spindle clock pulses. (186666.6666667/5666.666667).

So as long as you know what to change in the RELS software, it would be possible to put a toothed wheel with 100 teeth and an optical sensor on just about any lathe and probably run that high speed turning.

How about low speed. Turn the spindle at 60 RPM or 1 rev per second. That's in 2x mode then 200 interrupts per second or one every 5mS. To turn lead screw pitch of 20 TPI means the stepper motor also needs to turn 1 RPS; with a Gecko that's 2000 steps per second.

Oops! That's no longer less than the number of spindle clocks per second. So the software would have to issue a burst of step pulses at every encoder interrupt at a speed that results in 10 steps before the next encoder interrupt. Ideally the last step would be out before the next encoder edge interrupt.

And it gets even worse if you start turning the spindle by hand. But go back up to the pulley driven encoder at 1800 pulses per rev and a stepper motor running 4x micro-stepping (800 steps per rev) and the lead screw is always running with less steps per rev than the spindle and then the limitation on threading becomes 2x the encoder lines (3600 edges per rev and 800 micro-steps/step) or 4.5 times lead screw pitch. If the TPI on the lead screw is 8 TPI then the coarsest thread is 1.777778 TPI with one step per encoder edge.

What it all means is that software that relies on the high res encoder to create step pulses can't run high RPM or is stuck with 4x miro-stepping and therefore can't run a AC Servo with a 2500 line encoder on the lead screw.

Hopefully this all makes sense. And if I've made a mistake in the math please do speak up.

John Dammeyer

-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Richard
Sent: December-15-19 2:56 AM
To: [email protected]
Subject: Re: [digitalhobbyist] Step #1 Encoder

Thanks for clarifying that John.
Richard

On 15/12/2019 10:11, John Dammeyer wrote:

That number is what we decided on for the ELS about 10 years ago. For

example my little Unimat DB-200 as well as the Sherlines easily turn 6000
RPM. So my ELS had that as a stipulation since it was to be used for both
turning and threading.

If I'd been limited to only about 1200 RPM I'd have probably looked more

closely at encoders but both cost and capability ruled them out for a small 8
bit PIC without a quadrature encoder interface.

John

-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Richard
Sent: December-15-19 1:59 AM
To: [email protected]
Subject: Re: [digitalhobbyist] Step #1 Encoder

I am not really sure where this number of 6000 rpm for the lathe spindle
came from?
The RELS certainly has limits of operation and I am not sure what they are.
As far as threading is concerned the limits are suggested as a maximum
rpm against each pitch value.
For interest some of the max threading speeds are quoted below
?? Pitch ? ? ?? Max Speed
"0.25mm" " 999rpm"
"0.70mm" " 810rpm"
"1.75mm" " 320rpm"
"4.00mm" " 140rpm"
"80tpi " ? ? " 999rpm"
"48tpi " ? ? " 999rpm"
"20tpi " ? ? " 440rpm"
" 6tpi "????? " 140rpm"
I cannot guarantee these speeds as they would be based upon the

original

lead-screw pitch but they should give an idea of what is possible.
For feeds I am not sure what the maximum speeds are and I am not
competent enough to determine the values from the software or timing

the

interrupt.
I can say however that if you want to run your spindle and feed at
6000rpm then the RELS is not for you. What maximum speed you can feed

at

say .1mm per rev I am not sure.
With lower cost come limitations.
Richard (who has an RELS and is very happy with it.)

On 14/12/2019 22:04, John Dammeyer wrote:

Behalf Of grumpy via Groups.Io

is this usable