I came into the Hackspace full of determination to get some movement out of the CNC machine. The new motor is fitted along with the new four pin barrel connectors. The first job is to test the new motor by connecting it up to a known working driver chip on the CNC shield. Laszlo was keen to help out.
To also improve things I have commandeered one of the 5 Dell desktops that were donated from London Hackspace (Thanks Jon!) to be a dedicated machine for CNC work. I installed Cambam, Arduino IDE, Universal Gcode Sender and a load of other essentials (Java, 7zip, Sublime etc)
We plugged in the new motor and as soon as the Arduino was powered up the motor started clicking away like the old broken one did. Crap, I hope I didn’t pull and refitted the same broken motor. So we plugged in what I thought was the broken one and it also clicked. It could be that the replacement motor was broken too or that the driver chip was duff. To test the driver chip hooked up the X motor to driver chip A and all worked so the driver is ok. After much head scratching and some motor disassembly, Laszlo noticed that two of the wires going into the white motor connector plug were in a different order to all the rest. This would have meant the coil pairs would have been crossed over and would explain the clicking. Flipping these wires fixed the click and the motor ran fine. The faulty plug was being shared between the motors that seemed to test as failed which explains the symptoms.
After reconnecting all the motor and getting the directions correct we sent some basic move commands from the Universal Gcode sender. The Y axis was nice and smooth but the X & Z were very lumpy and crunchy. They were also moving far too fast which suggested the micro-stepping wasn’t working. The micro-stepping is controlled by jumpers under the driver chip and sure enough X & Y had no jumpers set. Putting these back in smoothed out the motion and slowed down the movement.
First run on GRLB
To give it a decent test, I plotted out a 30mm circle in CamBam and set a pocket operation with a 1mm tool. I exported the gcode and sent it to the Shapeoko via Universal Gcode sender. All axis moved smoothly and quietly. We were not cutting at this time as we have no end stops connected so cannot run the homing procedure.