RepRapMicron Block Stage
The default OpenFlexure Block Stage was modified to have a slightly larger working volume roughly 4mm x 4mm x 2mm (about a quarter of the Delta Stage), reinforced stepper motor lugs, a gap for UV illumination to be added under the stage, and a hole added to the platform so that a fixing screw underneath it could be accessed. NEMA17 motors were fitted to mounting plates previously used with the Delta Stage. Block Stage Mounting lugs seemed to be shorter causing some interference with the NEMA adaptor plates, so stand-off collars were added to the plate. The corners were rounded off to prevent wiring etc. catching on them.
The Block Stage is not symmetrical and tends to tip over with the weight of the motors, so the whole assembly was screwed to a base board which was also used to anchor the USB microscopes. A white LED with suitable current limiting resistor was placed behind the block stage connected to VMOT/GND on a spare RAMPS board stepper driver socket. A piece of white paper was put underneath the stage to provide a plain white background for observations. A switched, unbranded 4W USB UV resin-curing lamp on a gooseneck and base clip is used to set resin prints, but this was discovered to be too large to fit under the stage.
Software
The firmware used was a GRBL derivative designed for CNC control running on an Arduino Mega512. To interface to the GRBL board, a host computer (Xubuntu Linux) running a selection of cross-platform software was used. The GUI used for controlling the GRBL was cncjs, and the configuration values were determined empirically. GCODE was generated either by custom routines written in Python, or by converting SVG path files created in Inkscape using jscut. Values used were in microns throughout, which resulted in some mathematical rounding issues as experimentation progressed.
Zeroing XYZ
Flexure joints do not move the Block Stage platform linearly. It was found necessary to zero the stage X, Y and Z axes for any degree of consistent manipulation on a scale of less than 10 microns. A modified Base module was designed for the Block Stage that used simple sprung wires as endstop switches. These were connected to XMAX, YMAX, and ZMAX of the controller board.
Probe Height
The probe holder from the first prototype proved to be functional but unwieldy, so an improved version using constrained flexures rather than two springs was made as a replacement. This assists in manually setting the initial probe Z height without damaging the probe tip. In practice a low-quality USB microscope aimed horizontally is used to observe the tip as it is lowered. The shadow cast by the probe and its reflected image give an indication as to height. If the probe flexes, it is in contact! A number of conductive probe experiments have been carried out to characterise the behaviour of the probe and Block Stage. These use a conductive substrate (a coin) and the standard "Z Touch" probe input to the GRBL controller connected across the T2 thermistor pins of a RAMPS board (A5 I/O pin).
This allowed testing of the touch probe with CNCjs. Probing out a grid causes the probe position to be reported in the serial console. By copying this data into a spreadsheet it could be separated into columns and used as input to gnuplot and produce a 3D height map of conductive objects. An NZ 10c coin was used as the test object, as it has lettering as small as 0.7mm high. This revealed that the system worked, but that the X axis is reversed. That was fixed in the spreadsheet.
No satisfactory way of establishing probe contact has been devised as yet, but experiments involving coating the slide with tin oxide are planned. Crude tin oxide has been deposited from homemade tin (II) chloride on hot glass, with a resistance of 800kΩ over a 10mm diameter blob (it should be 1.75 x 10^3 cm-1Ω-1 from databooks). The tin chloride was made by dissolving lead-free pewter ("Britannia Alloy") in hardware store 24% w/w hydrochloric acid at above 60C in a water bath. The antimony present in the alloy will not dissolve in hydrochloric acid. Experimentation suggests a very fine spray of acidified tin chloride onto a 400C+ glass surface works (though not optimised) with approximately 13g of tin dissolved per 100ml. Too little spray and the film has bare patches, too much and it becomes very uneven. Also the glass shatters under thermal stress.
In the meantime height was determined by wiggling the probe back and forth to see if it had made a mark on a smear of Sharpie marker on the slide, and if it had not the Z probe was lowered further.
