Increasingly newer chips are only available in packages that are designed to be solder reflowed. There are a lot of tutorials and web pages out there where people describe ways that they have managed to reflow at home. The best I have seen is probably sparkfun’s tutorials which talk about their experiences with a number of techniques including trying to use a toaster oven for reflowing.
After looking at a lot of tutorials and blog postings about people’s experiences I opted to try reflow soldering using a toaster oven. Probably my favorite tutorial on reflowing using a toaster oven was published online by the Seattle Robotics Society.
Even after reading all of the tutorials I could lay my hands on I still had a lot of questions. Since I prototype out of a home lab my biggest question was if I could use lead free products in a toaster oven since they have a higher melting point by about 60 degrees Fahrenheit.
We purchased a convection toaster oven. They are slightly more expensive but the convection’s circulation of the air in the oven should allow for a more uniform heating of the parts. For the solder past I used SnAgCu lead free past from CHIPQUICK.
The first test just used some solder past on a piece of PCB to test the reflow profile. Just for kicks we tossed on a chip resistor to see how the past would flow around it.
Reflow test on a bare PCB
Initially the results looked very nice – but with the solder reflowing at a much lower temperature than initially expected.
A friend who was over at the house during testing pointed out that we were shooting an IR temperature gun through the glass front of the toaster oven to record temperature – and that if the glass was slightly IR opaque it could impact the readings. We tested at cold temperatures and the IR gun seemed to work fine. However the error caused by the glass seems to increase with temperature. Unfortunately it took a while to realize just what was going on.
The results from the initial test looked promising enough to try reflowing an actual board. The test boards were the $33 each boards from Advanced circuits. I highly recommend their service. We tried reflowing a MSP430 in an IDW package.
Reflow test of an MSP430 board from PCB express
The maximum thermal heating rate for the MSP we were testing with was 2 degrees Celsius per second and the toaster oven seemed well below that. So we initially set the temperature dial to 320 degrees to heat and soak the board – then turned the temperature up to ramp up temperature and reflow.
This reflow is when we learned just how much hotter things were getting inside the oven. The boards blackened and suddenly started bubbling an up a black goop and out gassing a horrible smell. I think a friend put it best when he said “ if that board was a pizza – that stuff would be the cheese”. We opened the doors in the shop to help with ventilation but man it stinks. Initially I thought the toaster oven reflow might be a wash until we hit the inside of the oven with the temperature gun and found the temperature significantly over our target.
Burnt board
Most of the discoloration seems to have occurred in the part of the board in the center of the oven – so it may well be that even with convection there are significant hot and cold spots on the board. That or all the VIAs on the left side of the board prevented the bubbling.
Cooling time measured both through glass and with door open
Once we figured out that the temperature gun did not work so well through the glass we decided to take multiple readings as the board cooled to try and gauge the delta of temperature between readings through the glass and those taken with the door open. As you can see the difference at higher temperatures was extreme.
For the third test we repeated our ramping up the temperature – testing outside the glass and then turning off the heating of the board 60 seconds after the solder paste visibly glassed and turned liquid. We did not open the door during heating as the regular inrush of cold air would likely have introduced errors of its own. Measurements taken during the cooling cycle for the third test seem to indicate that at its peak there is over a 200 degree difference in internal and external temperatures.
Reflow profile - test number 2
The solution to the heating problems seems to be using a temperature sensor inside the oven like the SRS tutorial suggests. Cooling is still an unknown problem as we need to bring the board temperatures down at a controlled rate.
Paste application test
The next thing we wanted to look at was hot to apply the solder past. Most tutorials seem to use a bead of past over the pads – but they also have shorting problems. So we tried using a bead of solder paste vrs smearing the paste over the pads. Both sides came out amazingly well – but with a few solder bridges to clean up – so it is still to early to tell which approach works better.