It finally got cool enough tonight to venture outside with my welding gear on.

First thing I thought I would tackle as a cube, so I cut 12 pieces of steel to the correct length and started to join them together. The first lesson I learnt is that these magnetic guides from Harbor Freight:

are *the* best thing to use when trying to get stuff perpendicular. I did try a right angle picture frame clamp, but I ended up with a distorted square, so rapidly abandoned it.

The steel was 16 gauge, so I set the welder for 4.5 feed, and E power setting. Worked really well.

When you are MIG welding you have to sure of two things:

1. You have the gas on before you start…
2. Clean all joins with a flap disc in an angle grinder – it really can’t be too clean.

Also *only* tack weld until everything is in place. I made a mistake early on, and I’d added some solid welds, and so I had to cut through them with a hacksaw to fix it.

If I can make about a dozen or so of these cubes, then I think I should be ready to tackle the Lotus Seven chassis.

Finally got the build table up and levelled – which is harder than it sounds because the floor has a composite angle sloping towards the center of the room.

It’s built from 2x4s and a 4×8′ MDF sheet – again thanks to Pete K for trucking the sheet up the street from Home Depot.

The advantage of getting it (spirit) level is so you can use plumb lines and levels to get things truly parallel and perpendicular.

The first step is to build a cockpit mockup from lightweight wooden timber just to ensure that the Miata seats can be made to fit (and so can I!)

Then the next step is to obtain the tubular steel to build the space frame chassis:

The above thumbnail was taken from the McSorley 442 plans available from here : http://www.sevenesque.com/plans/

See those two bolt heads on top of the long metal frame?

They don’t look like much, but they WILL NOT turn.

At all.

I’ve tried an impact wrench, breaker bar (4′ long!), PB Blaster, “smoker” wrench, and the Dalek – and nothing, NOTHING, will make them turn.

So…. I’m going to have to cut ‘em off.

With explosives.

Pete just picked up the rear end and dumped it on the trailer.

I got a whopping $65 for the scrap metal.

This represents about 2 weeks work (in almost 100F heat, and the intense humidity that Indiana likes to lay on as a special treat).

I still have a third of the car left, as the trailer wasn’t really big enough to hold the whole car – but there’s a lot more room to move in the garage now…

Next step is to make the build table for the chassis, then we get some serious welding in…

So I was driving along 96th street in Indianapolis this morning, and when I pulled up to the Meridian Street stop light, I saw this:

Wasn’t expecting to see that – it had the big ’7′ logo on the front, so it could be a Caterham – anyone out there know who owns it?

Just bought a 95 Miata…

IT’S ON!

The goal is to turn this:

into this:

(c) Caterham Cars - www.uscaterham.com

I “test fitted” the sensor circuits to the gantry tonight and I’m pleased to say that they work with a surprising degree of accuracy.

The sensors are wired to input pins 11 (X), 12 (Y) and 15 (Z). I used the Mach 3 “auto set” feature to configure the pins as home for M1, M2 and M3.

A neat feature of the homing functions (Ref) in Mach 3 is that it homes past the switch limit, then backs up to the on position, so the home is always referenced from the same position.

UPDATE : After giving this some thought it occurred to me that the home position doesn’t really have to be at the origin at all. It just needs to be a fixed reference point, so that the machine can “pick up where it left off” from a known location. Changing the tool on the router will typically nudge the motors and cause the software to get out of sync with current position.

I believe it is possible to script Mach3 to move to a fixed offset away from the “home” position, and use this to set the start point for the cut.

I’ll have to put together a movie that shows how Mach3 is configured too.

One thing missing from our CNC routers is an automatic axis home circuit. This is something that signals the control software that the gantry is in a known position (usually at the origin). A lot of CNC hobbyists like to use the end mill as part of the homing circuit – this is OK, but does not take into account different sizes of bit diameter. A better solution is to have the X/Y home switches attached to the linear axis. Then the home position will be the same regardless of bit diameter.

Initially I wanted to use Hall Effect sensors, but the detectors themselves proved difficult to get. I’ve had some QRD1114 reflection sensors for some time, so I thought it was about time I put them to good use.

Initially I used a PNP transistor to amplify the switching signal from the sensor, but this proved to be way too analog. The resulting signals were not very clean. So instead I used an LM324 opamp configured as a comparator, with a simple potentiometer providing the reference voltage. A 1M resistor is used for feedback, providing some hysteresis around the switching point. Eliminating any potential chatter that can occur.

Once we have the system fitted and working – I’ll be publishing a howto guide.

This switch will be used to home the X and Y axes, the Z axis will be using a touch plate.

Pete’s house has a great view of the downtown area of Indianapolis. Humidity was high, but we I managed to get a couple of decent shots of the festivities…