To learn to walk before I run, I am making a simple XY plotter as planned here. I have spent weeks on this. But it is a good exercise, testing design alternatives and so on. The last attempt is in real metal.

Works! (6th July)

It works, yes. In all honesty, “barely”. A nudge in the wrong place and things are out of alignment. This is partly the nature of the constructor kit, and partly the mechanical design was just made to be sufficient to exercise the electronics, firmware and host-software.

Videolinks: 1 and 2. The frame and axis stuff (cartesian X-Y) is “temporary” made  of the building set. It is to wobbly, but with building kits it is easy to rebuild. Needs “tightening” a bit. Nothing here (apart from experience) is reusable for the 3Dprinter.

The motors are the DC motors from the building kit with a feedback switch. The mechanical switch is not really fast enough and causes vibrations, I’ll need to get some photointerrupters or maybe even an encoder disc. The motors are a bit slow. It is unlikely, though, I’ll try stepper motors (I’d need to find some free junk ones somewhere) to avoid damaging the building set parts when trying to mount them. A bit more power may help, though.

The electronics is now properly put on a prototype board (2nd attempt), soldered, for two motors and a solenoid/servo.  Each attempt took two days to layout and solder. Diagram here. The first version had errors, and to top it, the L298N connections were mirrored 😦   This board is only for the XYplotter.

Firmware runs the motors, handles serial input and output in a non-blocking way, includes a simple command interpreter, which uses the G-Code syntax model. Code fills 11K of 30 possible at moment.(Published here ) The XY firmware ought to be be 80% reusable when going 3D. I noted looking at the RepRapGode specs though, that I have implemented blocking exactly the other way round. Need to think about that.

Host-pc software. I have serial connectivity from a Delphi program with polled input/output. I can edit a small command list or spool a file (of han-edited commands). It really needs some more bells&whitles, in particular some buttons for manual (re-)positioning. Thinking of doing something graphic-like, to allow doing a simple drawing (lines and circles).

Lessons learned

The point about the XYplotter was not to make a well functioning plotter in Fishertechnic, but to see what the challenges are, what wont work and so on.

Accuracy/strength in the mechanics are – surprise, surprise – essential. I have come close to the constructor sets limits. A total rebuild might get a little more accuracy. When doing the real thing, therefore everything has to be as superaccurate as I can make it.

The plotter could be improved by replacing switches with photo interrupters, smoother speed movement, increasing resolution (more vanes detecting the motor movement), but the slack in the gearing makes this pointless)

The commandset in the firmware could be rewritten, changing the flow control, using an active Ack-response to every command (inspired by the RepRap specs)

Making the chip(s) socketed as well as all connections,was good; This eased debugging the soldered board and it’s connection a lot.

Doing the electronics took twice the time and multiple times the effort to what I expected, even when subtracting the “stupid” errors. The firmware and software were as easy as expected.

I still need to test/experiment with speed, ie. specify a slow draw and I am not sure the “diagonal” code is correct/optimal. Scaling should be in the microcontroller.

A second attempt (16th July, 31st July)

Rebuild it partly, tried to redo the mechanics, but it is not going to get better (accuracy/resolution) until I use “real metal”. I will still use it to verify/improve the firmware. The switches can’t take much more pounding, so urgent I use photointruptors soon.

With Real Iron

I am now making the XY plotter in Real Iron(tm) 🙂 at my new evening hangout place, a hacker–space. Got the base plate frame workdone now, and have borrowed two stepper motors to just try. Working in real metal is slow work (drilling sawing and all that with some precision). Fortunatly they have some tools, in particular a bench drill.

8th Sep – a picture update:


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: