On 12/6/24 05:24, Gregg Eshelman via Emc-users wrote:
Leadshine has been doing advanced stepper drivers for several years. I couldn't
find one of their early demo videos now but I remember it had a smooth metal
knob mounted to a stepper motor in a box. It was demonstrated with a
conventional open loop driver and one of theirs.
The conventional driver was noisy and when the person grabbed the knob it would
stop and stay stopped. When holding position, the demonstrator turned the knob
and it stayed in the new position.
With the leadshine driver, the demonstrator could grab the knob and stop it,
then it'd start again when released. When holding position, he could force the
knob to turn but when he let go it snapped right back to where it was
originally. At all times, the motor was much quieter with the leadshine driver.
Did the leadshine driver need an encoder on the motor? These do, but
the 42C motors aren't optical, they are hall effect. And the A/D time
they need to do that is a killer for motion. Its better than optical at
getting to a certain position, but the A/D lag makes it jerk in steps to
get there, hence the complaints of shingled appearance. People put them
on, and take them back off the same day. The optical disk encoders don't
do that to near that obvious an extent.
printhead movements at angular directions are much smoother, or in
corexy cases no "shingling" can be seen.
I have 2 corexy printers but the biggest one, with a 400mm sq bed,
hasn't made a usable print, like the E5P, it will have a 20mm sq cf
tuube for X transport, over a pound lighter than the OEM setup so that
smoothing mass is gone, as are the OEM stell plates that carried the
ends with pulley arrangements that wound up with out of square motions,
all replaced by lighter printed parts that should do square stuff
square. That printer will be harder to adequately triangulate though as
the z motors are on the base frame, in the way of just adding angle like
I did on the E5P. I don't like having psu's hanging out so the clear bed
supports, so I'm building it a floor for that stuff out of added 600mm
long 20x20's but thats been a round tuit project while the E5P is being
fine tuned.
One of the things that has made the Sheldon conversion run so smoothly
is that its motor are earlier 3 phase, turning 1.2 degrees per full
step. Those drivers are limited to 50 volts, running on 42. At /32, dead
silent and still 2x faster than the normal steppers I used at first. The
blower in the vfd and the swish of fenner spindle belts are the only
noise, like Casper the ghost is turning the cranks.
On Thursday, December 5, 2024 at 09:32:06 PM MST, Chris Albertson
<albertson.ch...@gmail.com> wrote:
On Dec 5, 2024, at 2:48 AM, gene heskett <ghesk...@shentel.net> wrote:
They stand up to investigation. No ticklish servo tuning, they simply do as
they are told. The motor has its own encoder that's wired only to the driver.
The error determines the motor currant and that allows them to run lots cooler.
At the same time, if the error goes up, the driver will hit them with every amp
the psu has, skipped steps are history in that they don't happen. Cost over
regular steppers ranges between $40 and $80 per axis. They Just Work.
This is how the motion control world is evolving — lower level control is
migrating closer to the motors. I think one reason is to make the control loop
faster. I don’t know about your motors but 10KHz PID loops are not uncommon.
They might be running close to that.
The link below shows some tiny 2-phase steppers being controlled as if they
were analog BLDC motors. I think this is how you closed loop motors work and
then they wrap a step/dir interface around it.
I don’t know but I think the closed loop “steppers” are not even being stepped.
Today we can drive the 4-wire stepper as if it were a BLCD motor and use FOC
(field-oriented control) By using analog voltages (PWM) on the leads we have
very fine control of the magnetic field orientation and se can smoothly rotate
the field.
I think this is what those closed-loop steppers are doing, they are not
stepping but rather rotating to the next commanded step.
You could find out if you could put an scope on the power leads and record a
half second of waveform.
Here is a demo video of an FOC interfaced NEMA stepper. These are the common
$11 motors you see in 3D printers but it is not being stepped it shows what can
be done for cheap today. The plastic wheels are pressure fit direct the motor
shaft. The entire little machine cost about $50. I’ve linked the best part
of the video but you can rewind and even see the C++ code on screen
I know you might not care about little demo robots, but this shows that these
cheap ’steppers” are capable of much better performance than we think.
https://youtu.be/f9GJqqUpL2w?t=817
Arduino BLDC balancer robot - Tutorial
youtu.be
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
Cheers, Gene Heskett, CET.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
