Scara support ?

dc42

I think the problem is that X0 Y0 is not an accessible point on any SCARA printer, so the arm joint position calculation returns "nan" instead of a number, and that causes the JSON parsing error. Try extending both arms in the +X direction before you start or reset the Duet.

frafa

Same error on duet web control if change position

On Pronterface X+ RETURN/

[[language]]
[ERROR] Traceback (most recent call last):
  File "printrun\printcore.pyc", line 241, in _readline
  File "printrun\pronterface.pyc", line 1713, in recvcb
  File "printrun\pronterface.pyc", line 1669, in update_pos
ValueError: could not convert string to float: 

```G92 X0 Y0
return no error on Pronterface but after cmd duet bugged not move …

dc42

Try placing the print head somewhere it can reach easily. For example, X300 Y0. Then send G92 with those coordinates.

If that doesn't work, I'll simulate your printer later today and see if I can reproduce the problem.

In the next beta firmware release, I'll add error handling for attempts to set unreachable positions with G92.

frafa

Hi
G92 X0 Y0 ==> ko
G92 X300 Y0 ==>ok

Test…
move manualy arm on x300 y0
send G92 x300 y0

move with ptonterface button
x+ 100 ok (Approximately to the eye)
After +100 blocking no longer moves to x +

on position x100 y300
test move y-100 (result move Approximately to the eye on y250 x10)
Y does not move correctly, Appears y move diagonally

on Pronterface all mouvement return:

[[language]]
[ERROR] Traceback (most recent call last):
  File "printrun\printcore.pyc", line 241, in _readline
  File "printrun\pronterface.pyc", line 1713, in recvcb
  File "printrun\pronterface.pyc", line 1669, in update_pos
ValueError: could not convert string to float:

–-------------------------------------------------
Edit Upgrade ....

Just for info...
On Smoothieware i have modified
https://github.com/Smoothieware/Smoothieware/blob/edge/src/modules/robot/arm_solutions/MorganSCARASolution.cpp

Comment:
//actuator_mm[BETA_STEPPER ] = to_degrees(SCARA_theta + SCARA_psi); // Morgan kinematics (dual arm)
Uncomment:
actuator_mm[BETA_STEPPER ] = to_degrees(SCARA_psi); // real scara
and removed ==> * this->morgan_scaling_x on:
SCARA_pos[X_AXIS] = (cartesian_mm[X_AXIS] - this->morgan_offset_x) * this->morgan_scaling_x; //Translate cartesian to tower centric SCARA X Y AND apply scaling factor from this offset.
SCARA_pos[Y_AXIS] = (cartesian_mm[Y_AXIS] * this->morgan_scaling_y - this->morgan_offset_y); // morgan_offset not to be confused with home offset. This makes the SCARA math work.

After support real scara its ok
Impatient to test with Duet

frafa

To operate a Scara kinematics it must specify to him where is its zero in the space before making a G92, no?
Example in Smoothie I have 2 parameters to say where is zero in space with compared to the main axis of the first arm

morgan_offset_x 150 #my X offset
morgan_offset_y -68 #my Y offset
http://smoothieware.org/configuration-options

I'm not good at math, but it seems to me that it lacks these parameters.

dc42

@frafa:

To operate a Scara kinematics it must specify to him where is its zero in the space before making a G92, no?
Example in Smoothie I have 2 parameters to say where is zero in space with compared to the main axis of the first arm

morgan_offset_x 150 #my X offset
morgan_offset_y -68 #my Y offset
http://smoothieware.org/configuration-options

I'm not good at math, but it seems to me that it lacks these parameters.

You may have a point. I don't have a SCARA printer yet, nor have I ever used one, so I don't know how they are normally set up. Where is X=0 Y=0 normally placed on a Scara printer?

frafa

No rule in my opinion…
You can definitely define X0 / Y0 on the axis of the main arm and G92 will define the offsets, right?

dc42

Thanks, but that's not the way G92 works in RRF, nor in other 3D printer firmwares AFAIK. G92 is used to tell the firmware where the print head is, from which it works out the motor positions - in this case the arm positions. In essence, using G92 is an alternative to homing.

So G92 X0 Y0 is telling the firmware that the print head reference point is coincident with the proximal arm joint - which is not possible.

If there is a need to tell the firmware that we wish to define (0,0) somewhere else - for example, in the centre of the printable area - then I can see two possibilities:

1. Allow X and Y bed origin offsets to be configured for SCARA printers, as Smoothieware appears to do;

2. Use the tool offset facilities provided by G10 and M206.

frafa

ok
When I tested smothie I used M206 X … Y ... to redefine the offsets x0 / y0 with respect to the position of homing

Example in Smoothie we define the position x0 / y0 on the bed with respect to the axis to the main arm
Morgan_offset_x 150
Morgan_offset_y -68

And after homing I measure the distances xo / y0 and I correct the offsets with M206 and they are put in memory for recalculation
I'm really not sure how to express myself correctly in English ...

frafa

Without the homing currently you can add M206 to calculate the position x0 / y0 with respect to the axis of the first arm, No?
And after when the homing will be functional you will add the offsets …

dc42

I've decided to add X and Y parameters to 1.19beta9 to allow the bed to be offset from the proximal joint. I will also change the way that the crosstalk factor works, which will have the effect of making C0:0:0 correct for your machine and C-1:0:0 correct for the Helios.

Until homing is implemented, I suggest you manually move the arms (slowly, to be kind to the electronics) so they are in line and the proximal joint is at the half way position. Then power on. The displayed coordinates should show Y=0 and X= just below the sum of the arm lengths. If you then send a G92 command with the same X and Y coordinates as the displayed ones, it will then consider that the machine has been homed.

We're planning to display a Helios at the TCT show in September, which means that I will get SCARA homing working before then.

dc42

Firmware 1.19beta9 is out, with added X and Y bed origin offsets and a fix for G92 X0 Y0.

frafa

Oup's…
I had not seen page 2 of the post
A big thank you for your responsiveness !
I test this afternoon at fablab

frafa

Hello,
Again a big thank you for your responsiveness
My tests:

The M669 command used (X150.0 Y-61.0):
M669 K4 P200 D200 A-175:175 B-175:175 1:0:0 S200 X150.0 Y-61.0; set SCARA kinematics parameters

ok
Test moves First arm of 20 ° it's ok?
G1 S2 X10 ;anticlockwise
G1 S2 X-10 ;clockwise

ok
Test moves second arm of 20 ° it's ok?
G1 S2 Y10 ;anticlockwise
G1 S2 Y-10 ;clockwise

All movements in Pronterface in debug mode return:
SENDING:G92 X0 Y0
[ERROR] Traceback (most recent call last):
File "printrun\printcore.pyc", line 241, in _readline
File "printrun\pronterface.pyc", line 1713, in recvcb
File "printrun\pronterface.pyc", line 1669, in update_pos
ValueError: could not convert string to float:

Homing goes in the right direction x/y/z

I manually move the arms to position x0/y0
I power on with Pronterface I send command G92 X0 Y0

I'm testing x/y placements everything works but upside down
When i do x + i get an x-
And when i do a y + i get a y-

I reverse the engines in config.g and the displacements do anything?
I put the file config.g normal, and I reverse the connector of the engines,
Same results the movements do anything

In short I did not succeed in reversing the meaning, how to do?
It's almost perfect we're not far

dc42

In writing the firmware I assumed that the bed is to the right of the proximal joint, and the zero angle position of both arms is with them stretched out to the right so that they point towards the +X direction. Do you by any chance have your arm mounted the other way round?

frafa

hi
My Kinematics
for later End-stops (hall effect +on5v) are on X22.00 Y-35.00

dc42

Ok, so what i have assumed is +X is your +Y, and what I have assumed is +Y is your -X. Thinking about it, there are 4 possible orientations of the bed if we only consider the obvious ones, and I guess the firmware should allow any of them in a configuration option. Or maybe it should allow an arbitrary bed rotation angle.

For now, if you define the top left corner of the bed in your diagram as (0,0) and the axis directions as I have assumed them, then you should be able to print

Regarding homing, I was assuming there will be a limit switch at one end of the travel of each joint.

frafa

Of course I'm stupid, I did not think about it
I test this afternoon

frafa

Hi
define the top left corner of the bed, change on M669 x/y
and print…
ok

Video, sorry the machine is under development all is not clean ...
http://www.openhardware.eu/tmp/DuetWifi/VID.mp4

I still have to solve the same vibrations as with Smoothie …

My M669
M669 K4 P200 D200 A-175:175 B-175:175 1:0:0 S200 X118.0 Y-150;

My reduction motor and pulley
Motor 400 step/revolution 0.9°
Pulley 80/16
Microstepping 256
Calcul Step/° X/Y (Réduct 80/16) ==> 400256(80/16)/360=1422,2222222222222222222222222222

What can be the cause in your opinion not enough mechanical reduction, or other ?

Thank you very much !

dc42

I'm glad you are making progress.

Have you tried adjusting the segments/sec and minimum segment length in the M669 command, to see whether it affects the vibration?