Over the past couple of weeks, another bit of hardware experimentation has been wrapped up. In order to successfully solder some of the more delicate avionics hardware for Skydart-I, a reasonably decent soldering station is more or less a per-requisite to avoiding burnt hardware. Unfortunately, a decent temperature controlled soldering station is expensive. So how hard can it be to make one on the cheap?
When constrained for component cost, the answer can usually be found on the internet. Hako 907 handles are available for under £10, provided a Chinese clone isn’t a problem. Couple this with a slew of DC power supplies available for even less money than that, and there’s a project in there somewhere.
The main hurdle that needs to be overcome, however, is measuring the temperature of the nib. Most handles destined for use with their respective soldering stations have a thermistor built into the iron near the heating element. While this, in theory is quite simple, the reality is that the range of this thermistors’ resistance is not only quite large, but unknown for the low cost iron clones.
Some back of the envelope maths indicated that, in the ideal case, a temperature control of ~1C was possible over a 0-600C temperature range. While not bad, this wouldn’t be extremely flexible as it could be – as the accuracy of a potential divider system would deteriorate near the extremes of the range. Not to mention that the more simple systems would need a specifically calibrated reference resistor.
The solution then is to gain resolution through the tactical application of capacitors. The capacitor charge equation is quite well known, and easy to understand. Discharging the capacitor through the thermistor would allow precision measurement of the time constant of the system at any given time, and the accuracy of the timing available on low end hardware is generally quite good (in the 10s of microseconds). And since we’re only controlling a simple heating element in a system with lots of thermal inertia, we don’t need a very fast reading either. Setting the upper bound of each measurement to 100ms, allows for distinction between ~10,000 distinct states, or a 10-fold improvement of a potential divider system making use of the ADC only.
Using a TIP-121 to charge the capacitor (strictly, a TIP-126 PNP would be better suited to this) and a TIP-126 to discharge the capacitor through the thermistor is all that is required. By measuring the voltage drop and time elapsed, it is then possible to calculate the resistance of the thermistor in the iron itself, and therefore deduce the temperature. All of this can be wrapped up quite neatly in a rather compact hardware package, the PCB design of which can be found below with the files, including the firmware, available on the gumake git repository over here.
Fully assembled with some oversize wiring, it’s not much to look at – but for about as much as a decent non-temperature controlled iron off the shelf there’s not much to complain about. All that’s left to get a truly great soldering station from the project now is a quick spot of calibration (all capacitors will be slightly different, as will be the thermistors – so each unit really needs to be calibrated specially) and possibly tweaking the control algorithm. For now though, there’s some soldering to get done!