Thursday, August 6, 2015

4. Problems: Measuring temperature and soldering iron efficiency

I've been working on the firmware for the soldering station for quite some time now, but I seem to have completely underestimated the complexity of the temperature calculation mechanism. Currently, the software reads the ADC of the MCU, and knowing the 10 bit ADC takes values from 0 to 1024 and that the reference voltage is 5 volts, it calculates the voltage coming in from the op-amp. Based on this value, the temperature of the thermocouple can be calculated and shown on the screen. Simple, right? Not really, since this temperature doesn't seem to be the same at tip of the soldering iron.

This might sound a bit confusing, but it's true. I've hurried with the design of the station and I completely ignored the fact that there are many types of irons out there made by a myriad of companies, each having its own efficiency.

The design, materials and thermocouple positioning inside are so different between the brands that there would be impossible to make an universal formula to correctly deduce the tip temperature. For example, here's how a Gordak-clone ceramic heater looks like:

Gordak clone soldering iron
As you can see, the sensor is very close to the heater coil which means that it will heat up very fast. The tip on the other hand, having a much greater mass than the ceramic capsule will heat up more slowly. It is also safe to say that the tip will not actually reach the same temperature as the heater because of the various losses (the contact with the air and other surfaces, the external sleeve, etc.).

For a universal, do-it-yourself soldering station like the one here, this poses a significant problem in calculating the exact tip temperature. For a company (like Weller or Hakko), this is not an issue because as a manufacturer, they know the exact technical parameters and temperature curves of the soldering iron. So what is to be done?

As quitting is not a solution, the first thing that comes to mind is to somehow compensate for the losses from the firmware, a huge advantage of using a microcontroller. Ideally, the temperature measured by the thermocouple should also be at the tip, but this is not the case. What we actually have are two temperature curves, one for the heater and one for the tip. These two are proportional, but one is linear (the tip) and the other one is not. In order to see what's going on and how they relate to each other I made measurements of the tip in 25 degree C increments, starting at 150 and up to 450 degrees C. At the same time, I also noted the temperatures reported by the thermocouple. The experiments were done on a Solomon-type soldering iron (for use on ZD-929C stations):

Soldering iron for the ZD929C stations

The values were put in a table as seen below:

Table containing the 2 temperature value sets

As you can see, the tip temperature is much lower than the one reported by the thermocouple. To see how big these differences are, take a look at this graph:

The 2 temperature curves

So to get 250 degrees C at the tip, the heater must be at a whopping 405 degrees C. This is not very efficient...

Now the quest is to implement a calibration mode that will allow the microcontroller to calculate the correct tip temperature based on the data above. This will ensure that no matter what soldering iron will be used, after calibration the temperature will be diplayed correctly.

Another subject I'd like to touch is the power supply. Initially I wanted to use an old, soviet era transformer that was lying around:

Old transformer

It could only supply a maximum of 1.5 amps. This was well below the needed 48 watts the iron required and it got hot in just a few minutes. To counter any problems, I bought a new 3 amp transformer:

New 3000 mA transformer

This one can work for about an hour before it needs to cool down a bit, but that's enough for general hobby use. In case you didn't yet buy a transformer, make sure it is up to the task when shopping for it. If you need to only make light work with the station, you can get a 24 volt, 3 amp center tapped (12-0-12) transformer, otherwise if you need to solder for larger periods of time, a 5 amp one will be a much better choice.

If everything goes well the firmware will be published soon :).

No comments: