Wednesday, 2 September 2009

Balls to the rescue

Since I disassembled my MyVu Solo glasses, I've been looking for some optics to enlarge the image to a useful size. The stripped displays are only a few mm square.
From MyVu

I thought about trying a jeweller's magnifying loupe (the thing they use to look at gemstones), but they looked a bit heavy.

What I really need is some kind of curved mirror. After looking up in wikipedia, I reckon I needed a short section of concave curved mirror with a focal length of only a few cm. OK, where the hell do you get one of those?
Well, if you extend a curved mirror, you end up with a sphere. So, a mirrored sphere - a few cm across? Well, your local christmas shop!
From HMD

To cut a small square section, take a diamond-coated file and gently score the lines you want to cut....
From HMD

Or just explode it in you hands, like I did.
Taking a couple of the larger sections, the internal surface is pretty much mirrored. Too far away, and there is no image:
From HMD

But, at just over half the radius away:
From HMD

An enlarged, reflected image is formed. Since the glasses originally had plane reflectors in, I think it probably won't even need reversing - now to mount them in a test pair of glasses.

NOTE : safety warnings apply - don't play about with razor-sharp wafer-thin slivers of glass near your eyes. Like I am. If I stick glass in my eye it's my own stupid fault. Don't try this at home.

Laser Scanner

Not strictly RepRap related, but I've been playing with the David Laserscanner software from

It does seem pretty useful for generating source data. If it's good enough, you can scan an item, tidy up the data a bit, then print it put on your RepRap. The system then becomes a 3-d copier.

So, is it up to the task?

Initial setup:
You need a webcam, a laser line (either a laser level or a laser pointer), and the software.

I used an old webcam (640x480), and a red laser pointer (5mW) from ThinkGeek - but they're pretty readily available.

The software comes with a few different templates - basically a dot pattern you print out. You then attach the pattern to two sides of a 90 degree corner - like the inside of a box, or two card sheets at right-angles.
From laserscanner

You then point the webcam at the box, and 'paint' it with the laser line. The webcam can see the shadow of the line on the object and the background, and can work out the 3-d profile of the object.

To turn a laser pointer into a line, you need to bounce it off a curved surface (I made a line with a shiny metal knife handle), or shine through a glass cylinder. Wandering around Bath, I care across a shop selling christmas decorations - and they had a nice glass raindrop - shape:
From laserscanner

Shining the laser through the glass cylinder produces a nice line:
From laserscanner

I snapped off a short section of the glass and polymorphed it to the front of the laser. I've now got a simple laser line.
Initial playing with the software gives an idea of the result - I think it needs a bit of careful configuration to get the best scans out of it, but the potential is definitely there.
From laserscanner

Quick-release extruder

I pulled out my reprap at the weekend and fired it up. After levelling the bed, checking calibration, and fixing the z-flag that was loose, I started the extruder....

Nothing came out.

After some fiddling, I disassembled and reassembled the heater/nozzle - it was fill of brown/grey ABS, absolutely rock-solid, and had to be drilled out.

Since this is about the 50th time I have disassembled/reassembled my heater/nozzle assembly, I tried a modification to make it easier to change. My PEEK insulator is bolted to a penny washer with an m8 thread and two half-nuts.
From quickrelease

I cut a slot in the penny washer and I can now loosen a bolt and slide the extruder barrel/heater out to change it.
From quickrelease

This came in really handy, because as I restarted, warmed up to 225C, and fed the ABS through - it stuck fast. At least I could disassemble to find the problem.

The ABS appears to form a solid plug of the brown overheat mess as soon as it hits the heater, going solid and blocking the tube.

I think I've got a broken thermistor, or maybe the heat transfer is bad. It seems to warm up as expected, and reports the right temp (ish) staring about 16C and rising to 225C - fluctuating about 223-226C, but the ABS is definitely overheating.

I'll swop it out with another one and see if that helps.