It will be referred to as an encoder and the ability to push the shaft will sometimes be an option.

Here is a place to start.

Try http://shop.polybelt.com/main.sc

Anything will work; there shouldn't be any issues with interference to worry about with a stepper. At .33A you could use modular plugs (ethernet or phone cables) if that would be convenient.

You're not very clear on how you're trying to mount this motor (or to what). It's a "NEMA 17" motor size, however, and that might give you some leads with Google.

I've played with their software which will quote whatever part you build, and I think what you'll find is that they're pretty expensive if you're doing just one or two pieces.

You're basically describing a wheel with an encoder. There's some code examples in the playground; your wheel and encoder would probably be much sturdier than what is shown but it gives you an idea of what such a project would entail.

With four wires it's a bipolar motor. The wires can logically be divided into two pairs which will have conductivity between them. Test the resistance (in ohms) between the wires of one of those pairs; that is the coil resistance that Mike is describing.

You can build a double H bridge circuit to control the stepper but it's awful cheap to buy a controller.

I'm going to steal Terry King's thunder and mention that a 555 works well as a gate driver for a mosfet or for general purpose sourcing/sinking of more current than the Arduino's pins are capable of.

Ambient light will certainly have an effect on the reading of an IR detector. In addition to Mike's suggestions you should write your program to blink the emitter and base your detection on the difference in readings when the emitter is off vs. on.

The Aoyue 936 is a very good entry-level station at ~$40. That's 35W compared to the FX-888's 70W.... or perhaps a 25 second difference in getting up to working temperature.

No soldering iron should ever spark. The above two models both have a replaceable, ceramic heating element and are ESD safe.

That Weller station you have linked does not appear to be temperature controlled. It has no indicator light to show when it has reached temperature and is only described as "variable control".

I noticed the Arduino can take up to 20V - would it not be better to power it directly from the (12V) battery?

It would be simpler, yes, and probably appropriate in most situations. The Arduino uses a linear voltage regulator which burns off the difference between the 12V and 5V as heat and thus is less efficient than a switching voltage regulator, however.

Do you mean http://www.pololu.com/catalog/product/1182 for the second of those links ("less expensive driver")?

Yeah I did.

An LN293 is not suitable for this motor, because it reaches its maximum current at such a low voltage. This sort of motor requires a regulating (or chopping) driver that not only controls the sequence but also the current. You can set the current to be anything you want it to be and there is no need to run the motor at it's maximum current.

Here's an example as Mike describes, or if you can handle sticking to 2A you can try this less expensive driver at Pololu. Both drivers use the same step/direction/enable interface.

If you read the documentation for these drivers you can see that the controller's voltage (5V) is independent of the voltage the motor is running at  and you will want to use a higher motor voltage (typically 12V to 40V) to gain more speed with the stepper. If the motor gets too hot to press a finger on it then you need to lower the amperage.

Beyond that, the FETs in the servo will run less efficiently when at slower speeds; does it seem to get warmer when run slowly?

Are you sure of this?
Do you know what the mechanism of this is?
I can't think why this should be.

Why wouldn't you expect the FETs to run hotter (less efficiently) under the PWM or linear mode that would be required for the slower speed?

Amperage is based on how much torque the servo is providing and that is the most likely reason, especially given the high gearing in the servo which is a major use of the available power. Beyond that, the FETs in the servo will run less efficiently when at slower speeds; does it seem to get warmer when run slowly?

How much of this welder is missing? The amperage control and wire feed? Are you trying to build a spool gun or do you have the original gun?

In your code you might want to verify what value you're analogReading. Pots can get scratchy and that might be the cause of your jerky motor. It would also be a good idea to use some smoothing; something like:

Code:

void loop()
{
  if (analogRead(analogPin) > val && val < 1024) {
    val += 1;
  } else if (analogRead(analogPin < val && val > 0) {
    val -= 1;
  }

  analogWrite(ledPin, val / 4);  // analogRead values go from 0 to 1023, analogWrite values from 0 to 255

  delay(10); // creates a smoother ramp up/down
}

Coding it this way, if your analogRead() values are jumping around (due to interference or such) then that won't be reflected in the motor.

I'm struggling to imagine any situation where I'd want a home made welder instead of a decent quality production one.

That situation would be immediately following seeing the price of a 290A welder.