Part 2 - Machine Design

1. Frithjof

   Install of the Gestalt framework proved to be surprisingly easy. The laptop used to drive the virtual machines had a full python installation so all libraries where available and 'python' generated a command prompt from the DOS shell. The interface board was build and connections made.
   Both XY and single line test programs worked with the caveat that the temporary file (*.vmp) must be removed whenever changes are made to node ordering. Potentiameter adjustment on each board was necessary to reduce current draw as the driver chips would heat quickly and initiate thermal shutdown. Once correctly adjusted, no overheating was observed and motors could be left poered on for extended periods of time.
   
   The gestalt system requires XY coordinate data to action the stepper motors. The coordinates are generated using the source code supplied to drive the axidraw printers. The software provides an interface from which the drawing coordinates generated using inkscape are captured without having to manage the intricacies of the SVG format. Axidraw creates the drawing path as an intermediary step and this was used as raw input data.
   
   The axidraw source code is available on Github (https://github.com/evil-mad/axidraw). After understanding the general flow, a file was opened and data in the form of XY coordinates where written to a file by placing write statements within appropriate code sections. It should be mentioned that the use of python and the associated interpreted permitted a fast, iterative development cycle that required no recompliation or restart of inkscape when changes were introduced into the Axidraw code.
   
   The file needed was manually edited to remove certain formatting elements and loaded into the Gestalt framework directly by opening a file pointer and reading data into a python array.
   
   Some slight conversions where added to permit scaling where necessary. Due to difficulties in integrating the paramaters for the DVD motor, pen up/down functionality was not integrated in the first iteration of the project.
   
   
   Plotting text requires the use of special plotter fonts. Most fonts in modern computing systems are rendered as two sets of lines to provide a thickness for the stroke. This is altering depending on the font size. The thickness may not be uniform over all the paths of the character itself. Plotter fonts depend on the diameter of the writing head to manage thickness which in turn requires characters be desribed as single strokes. A well known font family is called 'Hershey Text' which are a series of fonts designed in the late 20th century when electronic printing technologies where not able to render thickness aside from changes in the pen size. Tests for text printing used these Hershey Fonts to plot readable text. The results were acceptable considering pen up/down functionality was not integrated.
   
   Next, some clip art was used to generate an image. This worked surprisingly well so long as the clip art was suitably modified to removed lines that would create many connected strokes.
   
   It was originally decided to integrate a third motion axis to control pen up/down but was not achieved due a lack of documentation on integrating new parameters into the gestalt framework. This could eventually prove to be interesting as it would permit a more aesthetically pleasing output. Additional switches to limit the travel of the motors could also be useful to limit the motors working agains the end stops. Axidraw does this partly in software but hardware failsafes could protect motors in case of incorrect input or unplanned bugs affecting motor functioning.
   
   Test:
   
   Code for reading coordinates from file:
   f = open('TdMCoords.txt','r')
   coords = []
   for line in f.readlines():
   x, y = [float(str) for str in line.split()]
   coords.append([x, y])
   Code for normalizing coordinates:
   xmin = min( x for x,y in coords)
   ymin = min( y for x,y in coords)
   print xmin,ymin
   moves = []
   for x, y in coords:
   moves.append([int((x - xmin)*10), int((y - ymin)*10)])
   This was all the code necessary to permit plotting the data, the rest of the data was standard pyGestalt.