Sunday, February 04, 2007

Building the new Tommelise controller board

I'm having to stay up fitting neural nets this evening. That leaves me a bit of free time while these things grind on my dual Xeon workstation. I've been slowly building up the new microcontroller board for Tommelise. I've taken a lot of pics for documentation but I've not the time to write that up this evening, so I'll just show you where I am at the moment.

You can see the PIC 18F4610 in the middle of the board and sockets for two 754410 dual Darlington motor controllers at the bottom. The power conditioning voltage regulators and capacitors are at the upper right of the board with a two pole connector for 12v power.

To the left of the PIC chip is a 20 MHz timing crystal that sets its speed.

The little 1K Ohms resistor connects 5v power to pin 1 of the PIC chip and the black and orange lines just to the right of the PIC chip supply grounding and 5v power respectively.

The other two, longer black and orange lines carry serial comms to the two, two pole connectors at the top of the board that are connected to that multicoloured flat cable at. That cable is connected to a comms card that turns PIC serial comms into something my workstation's serial port can understand.

This PIC chip allows for direct USB communications with a PC. I haven't got around to getting that working just yet. Right now I'm more interested in getting Tommelise to actually make something useful than I am on guilding lillies, so to speak, so USB is not near the top of my priorities just now.

Add one power transistor and five more of those blue two pole connectors to attach all the sensor input lines and carry power to the Mk 2.1 extruder and the board will be complete. It's not too complicated a line, even for newbies. It does take a bit of patience and care in order to get it going properly, though. I'll be taking you through all that step by step in the documentation in a few months.

BTW, I've tested the board and it responds to serial comms properly. Now to get the dual Darlington chips hooked up and running the gearmotors. I got tired of wasting dual Darlington chips so I bought some cheap sockets for this board. Tommelise doesn't draw a lot of amperage to run its motors, so I can get away with this, I think. Ordinarily, putting dual Darlington chips in sockets is not a very bright thing to do.

Time for a new controller board for Tommelise

I was working at trying to find the bug in PIC-based controller board for Tommelise and finally stepped back and looked at the thing.

This is ridiculous. The whole thing is patches from various things that I tried that didn't work just right for one reason or another. Not only that, it's enough to scare the wits out of anybody contemplating making their own Tommelise. The Tommelise motherboard only has three integrated circuits (IC's), One Microchip PIC 18F4610 and two Texas Instruments 754410 dual Darlington chips. How simple can you get?

I don't make elegant boards, as you can see. The new one I'm building, however, ought to be a LOT less intimidating that this one.

Saturday, February 03, 2007

Mounting the Mk 2.1

I stopped and did a little documentation photography of the final mounting of the Mk 2.1 to the Tommelise positioning system.

I created a mounting bracket out of a foot-long piece 1-1/2 x 1-1/2x1/16 inch aluminum "L" section. I sawed a section out in the middle to accomodate the Mk 2.1. Notice that the spring loaded studding bolts that maintain pressure on the filament in the polymer pump stick out behind the polymer pump so that you can't mount it flatly on the xz positioning table, thus the need for the aluminum "L" section bracket. The Tommelise's Mk 3 extruder will have an integral mounting bracket on the back side of the the polymer pump which will eliminate the need for the "L" section.

Here is a closer view of the mounting. You can see here that rather than saw the piece of flange off I simply bent it down. That was mere convenience on my part. You can also see the #10 machine bolts with wing nuts that I used to secure the brass mounting flange to the aluminum bracket. I have some longer #10's which I intend to spring load so that if there is a z-axis positioning mishap the Mk 2.1 extruder barrel will simply push upward against the springs rather than be thrust through the tempered glass xy working surface.

Thursday, January 25, 2007

Documenting the basic structure of Tommelise

Recently, I built up the basic structural elements of Tommelise in the open source Art of Illusion (AoI) 3D modelling package. I took the opportunity to correct a few dimensions that I would have built somewhat differently, like making the xz vertical positioning stage a bit narrower, than I actually did in practice.

Other than a few things like that, it is shown in AoI pretty much as built. Mind, I am not particularly happy with the 1/4 inch steel guide rods that I used with the xy positioning stage, either. I may yet change them over for 1/2 inch ones. The 1/4 rods are entirely too flexible for my taste, though they don't seem to have caused trouble yet.

In any case, if you are interested in using the Tommelise design as a starting point for your own homebrew bootstrap 3D replicator, this AoI file will give you the dimensions of the basic structure. Keep in mind that AoI is not a CAD package so you have to set up a grid to get dimensions. The measurements are done in inches.

I will be building a more complete structural description of Tommelise in the Documentation section of the Clanking Replicator Project website in a few days.