Adapting Morgan firmware for other SCARA Configurations

How difficult would it be to adapt Morgans firmware to run a SCARA arm robot like this? Black boxes are the stepper motors.

Different math, so writing some code will be neccessary.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Hello,

I hope it is Ok for asking a similar question in this Thread !?

My name is Mike, I live in Germany, 29, and still do not (yet ;-)) have a Reprap machine.
I was also thinking about using a SCARA geometry, just like "Evil Monkey" (EM) writes about in here: [forums.reprap.org]
The principle scara geometry is very interesting and there are only few parts that move, increasing reliability (at least I think so).

So, It is nearly exactly the same question as EM:
How is the inverse kinematics implemented into the firmware?

As far as I understand, for a cartesian machine, the x (y)-value (desired position) is divided by the step-count per unit lenght and the µC will send the correct step count subsequently to the driver.
By using an inverse kinematics, the arm angles are calculated, which is an (analytic) function in dependence of x and y.
So, the angles are divided by the step-count per degree...et voila, same principle.

Nevertheless, in which file do I have to implement this formulas? I mean, when looking at other Scara-projects (unfortunately I only know 2 projects), such as the
Morgan:
[github.com]
and the
Armstrong:
[github.com]
I do not understand what has been done there.

As far as I understand, both developers (Quentin Harley, ttsalo) have taken the state-of-the-art "Marlin"-Software and modified it. E.g. when looking at ttsalo´s code, within "configuration.h" there is a small paragraph about scara, but it does not look as "ugly" as the inverse kinematic should look like.
There are so many files within the github repositories, from whose I think they are mainly needed for things like pid-control, LCD- and stepper driver-,...-stuff. In addition, is there an explanation of the used variables and all these things?
Wouldn´t it be useful to implement a "standard" scara geometry into the state to the art software where one can select? OK, that´s a little bit to far by now, but...
Perhaps somebody knows and can tell us the principle how to implement "fancy" kinematics into the Software??

OK, one problem using the scara geometry is the curced travel, and therefore the curve is intersected into small, linear sections. But is this the reason that the inverse kinematics can not be found?

Thank you in advance,
Mike

Oha,

I just found this one here:
[forums.reprap.org]

BR,
Mike