Having used the T962A reflow oven and determining it was pure junk, even with mods the search for a reflow oven was on going. Seeing a 1/4 size convection oven on craigslist locally for sale got the idea to get a controller and see how it worked.

The oven was nasty and after some cleaning and installing the Controleo3: https://www.whizoo.com/controleo3

I was impressed at how easy it was to setup the controleo3 and then it had a learn button. I figured this would be easier than I thought... Boy was I wrong.

Since these metrics made little since to me I figured well maybe the controller is smarter than I am, and lets just run a lead-free profile and see what happens. So I ran the lead-free profile and I noticed right off the controller was not turning heating elements on fully, rather it was using around 70% PWM. Of course with 70% PWM there was no way this oven was going to meet profile curve, and sure enough the controller stopped when it was 25C away from curve. However why was it only 70% duty cycle? A brief look at the source code made me realized that the developer had no idea what PID really was and had really messed up the machine.

At this point I realized that my oven was not good enough. That is I could not hit the profile curves because I could not heat fast enough. So I added more heating elements. So around 1200W was added to oven, and bang bang controller was implemented in firmware. I could now almost meet profile but still could not get 1C/second ramp rate.

Education is what you get when you did not get what you expected!

Taking a break from oven for a night I thought about the system and mentally modeled it as a capacitor for the thermal mass, a resistor across capacitor as the insulation and then Wattage as a current source trying to charge up the cap to certain voltage (temperature).

Using this model helped me understand why I was confused about the "Learn" results, what the heck is "power", "inertia" and "insulation". So after much reading on the whizoo website and source code I gain more knowledge. First off the "power" is the measure of the PWM percentage the heater needs to maintain 120C. If you had the total wattage of the machine then this would be a measure of the insulation of the machine. For example assume ambient is 20C and it requires 13% of the total wattage to maintain 100C. So assuming your oven is 1200W, 13% would be 156Watts. So the temperature difference is 120C-20C =100C, hence the thermal resistance (insulation) would be 156Watts/100C=1.56W/C. What this would mean is to have oven at 250C you would need a minimal of (250*1.56)=390W just to overcome the thermal insulation.

The "inertia" appears to be how quickly the unit will heat up 120C to 150C (30C delta). Hence for my oven 30C/62S is ~0.5C/second which is really bad. You need around 1C/second to be good enough for a reflowing. Note this is labeled inertia is that it is really more about the thermal mass than anything else which provides thermal inertia, yes resistance does play a role in the way Controleo3 measures "inertia"

The insulation metric is just bad, as it is a measure of how long it takes for oven to cool from 150C to 130C with heater off. However if you oven has huge thermal mass you are not really measuring insulation. Rather the "power" measurement is closer to the insulation measurement.

So based on this data and my mental model I realized that my problem was that the oven inside had way too much thermal mass to get the ramp rate I needed. Note I did increase heater elements to 2200W and it still was not enough to overcome the thermal mass.

Note using the electrical circuit model pictured above you could make a simple learn/test procedure that would quantify the insulation and thermal mass in SI units. Maybe one day I will modify the Controleo3 firmware for this.

The Redesign

So I ordered some 1/4 ceramic high temperature insulation blankets and started the redesign.

I first took out the inner oven hot box, and built a new one with wire mesh.

Then I put ceramic insulation on the inside and outside of the wire mesh. The wire was used to reduce the internal thermal mass, insulating the inside of the wire mess further reduces thermal mass.

Note I should have made the wire box around 1" larger inside than the original metal oven box to allow room for the ceramic insulation.

I then coated the inside with metal tape and/or gold insulative material.

The oven now has around 1200W watts of heat, and a temperature ramp rate of almost 2C/second.

The next steps are to replace the weak servo door opening system with a linear actuator, add an external fume vent to the outside of shop and reprogram the Controleo3 with a good PID controller that has integral overflow. However with just a simple bang/bang control, instead of PID, the oven matches curve within 2C of desired temperature.

A bang/bang controller is basically turn heaters on fully if we are under temperature, else turn heater off. This type of controller works very very well when you have large thermal mass and/or fast control loop update rates.

Conclusion

The Controleo3 is a good reflow oven controller, it is far better than the T962A, however the firmware on the controleo3 could be so much better. The controleo3 firmware caused me more confusion and delay and I would have been better off buying a reflow oven that worked. However finding a reflow oven you can buy which works is a huge problem.

If you do want to build you a reflow oven, the 1/4 size convection oven was a good form factor (bigger than toaster oven) to start with. However you will have to redesign everything to make it work, also external venting is really needed which will have to be designed into system. The net result again is that it would be better to buy a working commercial unit with external venting.

I did learn a lot from the experience, knowing what I know now I think with some modification to heating elements, convection fan, and some insulation even the T962A could be made to work, but then again it would be easier to buy a working oven.