Honestly asking, why spend the time?
Did a PC abuse people in any way during the 00's or maybe the 90's?
The Orange Pi is generating steps >400Khz and runs LCNC for cheap and
is not a PC.
I'm looking for pathology to support a theory for anti-x86 PC machine
control sentiment.
What am I missing here? Is this anything more than PC bad and
micro-controller good?
On 8/14/20 11:57 PM, Chris Albertson wrote:
What you want should be possible, if only some one would spend some months
working of the software. I use this little STMF104 for motion control
(but not for CNC) and the STM32F104 can generate pulses far faster them my
motors can move. What is the point of 300,000 steps per second if the
motos cn't move that fast.
Most 3D printers use Marlin software that runs on an 8-bit Arduino and it
runs all four axis up to the limit of the motors speed. The STM32F103 isi
and order of magnitude more powerful them an 8-bit Arduino. There is not
reason an STM32 can't drive a full size 5-axis milling machine. We just
need someone to decide to do it.
On Fri, Aug 14, 2020 at 1:21 PM cogoman via Emc-users <
[email protected]> wrote:
On 7/23/20 2:15 PM, Chris Albertson wrote:
What is really needed is for someone to write firmware for the common
STM32F103 "Blue Pill". These have the hardware to do things like step
gen
and quadrature decode at MHz speeds and talk to the PC over SPI I2C or
USB
and cost under $3 from 100 different vendors. I use these for motion
control when I can but not with Linux CNC. It is really "just a
matter
of software" but I'm not about to spend months of my time to save the
cost
of a 7i92.
May I suggest the shoulders of KevinOConner to stand on.
Check
out https://www.klipper3d.org/
He has set up a 3D printer program that uses python on the Raspberry PI
for the G-Code interpreter. He used OctoPrint to feed the USB serial
port, but the arduino doesn't have to interpret G-Code, so with an
arduino he can max out at 102,000 steps per second, with the STM32F104
in the Blue Pill Board he can max out at 360,000 steps per second, both
with 3 steppers stepping.
https://www.klipper3d.org/Protocol.html
https://www.klipper3d.org/Features.html
The benchmark is a little cryptic to me, but I think it says the blue
pill board can step three steppers at less than 10uS per step max (for a
$3 board). If I read this properly, you can sink up 2 blue pill boards
and control 6 to 8 steppers. Yes there are limit switches connected.
It also handles the slow speed PID control of a heated bed and extruder.
I would like to see LinuxCNC set up to control my router through the
blue pill board (or two) without needing the OctoPrint, or the Klipper3D
python interpreter. I have been running my router with GRBL, and though
it's an amazing feat of programming prowess, sometimes wrestling with
FreeCAD and either FreeCAD's PATH toolbench or JSCUT makes me wish I had
the subroutines, named variables, math, and looping constructs LinuxCNC
makes available. Many times I've spent hours or days doing something I
felt confident I could do in LinuxCNC in about an hour.
I would like to see LinuxCNC able to control a small machine through
a $14 CNC controller.
http://www.zyltech.com/arduino-cnc-kit-uno-r3-shield-4x-a4988-drivers/
The protocol and the microcontroller software is already done, though
I'd guess it doesn't use synchronous transfers, but just buffers the
steps to keep latency from stalling the steppers.
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
