Dan
Forum Replies Created
-
Posts
-
Joram,
– connect robot
– open settings
– change size
– close settings
– DO NOT DISCONNECT
– open settings againare the numbers wrong?
Hm. I was just about to release an update. I guess I’ll have to investigate. Is there any error in the log.html file?
It seems more quiet with the TMC2100. I’d like to know how much. Do you have a sound meter app on your phone? I’d love to see the hard numbers.
A giant paste is not concise. What is the URL to the github version of this code? Can I see a photo of your machine? I have several stewart platforms, so I’m not sure which one you are using.
Probably the first thing you need is R71 (set EEPROM defaults) and R70 (save EEPROM) once. Then you can G28 (find home). Once the robot is home it will start to listen to commands.
Thanks!
a 200 step-per-turn motor would be 360/200= 1.8 degrees per step. a 400 step motor (like we use) would be 0.9 degree per step.
http://reprap.org/wiki/TMC2100 might also provide more information about how to set the hardware to 256 microstepping to match the software setting.
> So should I just define machine dimensions as where the outer pulley/pins are located?
yes. Then you can run a “border” test to see that it draws a nice rectangle with straight edges where you expect on the machine. that validates your settings and confirms where your drawing will appear.
I don’t get it.
both ends of the belt are going to the pen holder? that won’t work.
the loose end of the belt (with counterweight) is going to hand down and interfere with the end going to the pen holder? that will work, poorly.
As for the imaginary point… sure, that looks pretty close. I’d just calculate from the center. if you want to be really picky you could work out the point on the curve of the guide for every x,y position on the board, and then fine tune accordingly. it’s a difference of less than (circumference/4).
This I a great post. Thank you!
So… I have always treated the math as being from center of each motor shaft. That said, it is *not* the most accurate solution. It’s been merely “good enough”.
We know the spacers on the motor shafts are 6mm diameter and that the hole centers are 31mm apart (http://reprap.org/wiki/NEMA_17_Stepper_motor), so it should be possible to adjust for that. This gets extra fun when the pen holder is in the bottom corners and the belt stops touching the corner post. Whee….So… I fudged it close enough. If you have a better way to *very quickly* calculate a more accurate positioning then I am very interested in putting it the software.
The TMC2100 is 1/256 microstepping driver. it should be moving the pen 1/16 the expected distance if you use the default makelangelo firmware. this is because the firmware is set to 1/16 stepping, so the physical behaviour and the robot’s mental perception of its actions are diverging. The fix is to teach the robot it is now at 1/256 microstepping.
https://github.com/MarginallyClever/Makelangelo-firmware/blob/master/configure.h#L79
change
#define MICROSTEPS (16.0) // change this. microstepping on this microcontroller
to
#define MICROSTEPS (256.0) // change this. microstepping on this microcontroller
and upload the code again. As long as you are using 1/256 drivers you will have to make this change every time a new makelangelo-firmware is released.
Uh… we tested homing before it left the shop so I don’t know how that’s possible.
Anyhow, the fix:
– confirm arduino serial window is closed and makelangelo-software is disconnected.
– disconnect power from robot
– locate motor turning wrong way
– follow wire from motor to control board
– disconnect said wire at board
– turn it 180 degrees
– plug it back in
– reconnect power and home againSince your machine is backward it may have shipped without QC (Quality Control). You may find that on homing the homing ends and the pen stays at the bottom of the drawing area. If that is the case, stop and email us, there are more steps that have been missed in QC.
> it prompted me to make the Makelangelo-firmware folder and move the Makelangelo-firmware.ino file into it.
All the other files in the project must ALSO go into that folder, then restart Arduino and open Makelangelo-firmware.ino in that folder. The tabs should then appear.
updated original instructions.
* download the firmware (https://github.com/MarginallyClever/Makelangelo-firmware/releases)
* extract the files
* make sure the folder that contains Makelangelo-firmware.ino is also called Makelangelo-firmware.
* open Makelangelo-firmware.ino in Arduino.
* confirm that there are many tabs at the top of the arduino program. If your window is narrow there should be a down arrow on the right that exposes many more tabs. Each tab is a file in the project.
* confirm that in configure.h the MACHINE_STYLE is POLARGRAPH.
* confirm that in robot_polargraph.h the MACHINE_HARDWARE_VERSION is 3 or 5, depending on your machine type. If your machine has automatic homing it is a 5.
* confirm tools > board = mega 2560
* confirm tools > port is set to your usb connection. it will probably say something like “COM4 (mega2560)”
* click upload (the second circle from the left, with the arrow pointing right)
* if upload bad, report back here.
* open serial window (last circle on the right, magnifying glass)
* set baud = 57600
* set newline = newline
* text should appear on the screen.
* type “M114” and hit enter. robot should report it’s position.
* close serial window
* start Makelangelo-software.
* proceed as normal.Are you certain you have the latest software and firmware?
You’re not the first to suggest it. I would post that publicly to the projects subforum and get other users talking. I’m only doing support and basic maintenance on the project at this point, my focus is on other robots at the moment. If you do post it publicly I will retweet it to help promote the idea.
In makelangelo firmware configure.h change MICROSTEPPING from 16 to 256 and upload it to the board again. You’ll have to do that on every update.
I hear 1/256 stepper drivers are really quiet, they’re used in Prusa i3 MK3.
here’s one such driver. We’re considering getting them for stewart platforms and the robot arm prototype.
with power off, unplug the right motor, turn the plug 180 degrees, and put it back in. done!
one of your motors is now backwards. this is because of a change in firmware.
Drive will only work after set home.Does that help?
1. Try other settings and report back. I like this setting. good quality drawings, and 1/32 seems excessive.
2. You’d have to change the dip switches under each a4988 to match the software setting.
4. distance from the origin to the top, bottom, left, and right sides of the machine. For most people L is -(width/2), R is (width/2), T is (height/2) and B is -(height/2). This lets me flip the drawing by swapping left and right, and it lets me move the origin around (tho in practice i don’t at the moment).
5. the equation should be in the polargraph.h file. circumference of spool / (normal steps per turn * microsteps). gt2 pulleys with 20 teeth have a pitch circumference of 40mm. if your motor is 200 steps and you have 1/4 microstepping then it’s 40/(200*4).
Note that string on a spool was abandoned because the string winding changes the diameter in an inconsistent manner, which really messes up the drawing quality.
is (A) servo failing sometimes, or (B) is servo too slow or (C) both?
When you press pause… is it from software? is it from lcd panel? software is supposed to return pen to previous state. lcd panel… maybe it forgets.
Well my first one was merely a proof of concept.
Were I to do it today I’d use a 4 axis motor controller like the RUMBA or RAMPS.
The Makelangelo firmware already supports crude skycam kinematics.
We don’t currently have a kit before skycams because demand is too low.But feel free to ask questions and post your progress, we’re very interested and happy to advise.
