With the oven now ready to run, it’s possible to do some test runs. Yay! Taking the cautious approach – mainly because the stencils haven’t yet arrived – this is a first stage of the testing, i.e., simply a number of repeated reflow runs where the temperature is monitored per the thermocouple data from the ControLeo2 board. Easy enough to get – just plug in a USB cable and log the data as it’s received.

Convection fan enabled

Figures 1 and 2 show the absolute temperature (trace 1 – blue) and the delta temperature (trace 2 – orange) for two reflows done in succession. The oven was let to cool down to about 30 °C between reflows.

I’m not sure whether this is a great result or not, but it’s repeatable which is a good sign. While the graphs for only two reflows are shown, these are representative of many runs performed.

Figure 1 – First reflow example with the convection fan enabled

Figure 2 – Second reflow example with the convection fan enabled

Convection fan disabled

Maybe the the noise in the delta temperature traces is caused by interference to the thermocouple signal from the convection fan? So figures 3 and 4 illustrate the same tests – same conditions – this time with the convection fan disabled via the menu in the ControLeo2 controller.

Figure 3 – First reflow example with the convection fan disabled

Figure 4 – Second reflow example with the convection fan disabled

Thoughts so far…

The positive temperature gradient generally seems a bit low at about 0.5-1.0 °C, but there’s no insulation around the oven so maybe installation of this will improve things(?). Otherwise the thermocouple signal is certainly noisy, and it wasn’t particularly improved with the convection fan disabled. A couple of thoughts here:

  • The traces could be cleaned up by low pass filtering them, but of course this would only be for a prettier looking graph. Still it might be interesting to do this.
  • More meaningful would be to understand whether there is any filtering of the thermocouple itself. Even something simple. It surely doesn’t make the controller’s job any easier although it might indeed be tolerant to it. Something to be investigated!

And what are we looking for?

Figure 5 shows the recommended starting profile (quoting the datasheet “designed as a starting point for process optimization”) for the ChipQuik SMD291AX50T3 no-clean SN63/Pb37 183 °C melting point solder paste. Lead-free would be nice but let’s not make things too complicated at this stage. Besides this is what I have available.

Thus how are things looking? Basically, not sure. The profiles are clearly not the same, but will it make a difference? Again not sure. I guess it’s time to run some test boards through the oven to see what happens!

Figure 5 – Starting profile for the ChipQuik SMD291AX50T3 no-clean Sn63/Pb37 183 °C solder paste