Hot plate reflow soldering

2. January 2014 12:33 by Jens Willy Johannsen
Categories: Uncategorized

Until now I've been soldering SMD components by hand only (here's an image of my first hand-soldered SMD board). However, for my next project I will need to solder an MMA8452 accelerometer (datasheet) which is a 3 x 3 mm QFN package (0.5 mm pitch).
And while it is possible to hand-solder QFN packages (I've successfully hand-soldered an ADXL335 accelerometer in a 4 x 4 mm LFCSP package – which is a 0.65 mm pitch QFN), it is simply too difficult and too much bother (I'm sure people with better dexterity and eyesight than me can do it easily).

So: no way around it, time to start doing some reflow soldering.

For a hobbyist like me there seems to be three ways of doing it:

  1. Toaster oven
  2. Hot air rework thingie
  3. Hot plate

Originally I planned on using a toaster oven and this nice reflow controller from Beta LAYOUT. But after reading SparkFun's tutorial and the comments I decided on trying the simpler way first – which is getting a hot plate or electrical skillet.

They didn't have a skillet in the store and I couldn't be bothered to check other stores so I bought a simple and cheap hot plate.

And it works perfectly!

I got some (leaded) solder paste (like this one from Farnell) and applied the solder paste straight from the syringe needle (for the TQFP, SOIC and QFN packages I dragged a thin line across the pads), placed the assembled PCB on the hot plate and dialed it to the second-lowest setting and kept an eye on it to see when the solder paste started reflowing. After a minute or two, everything had reflowed and I took the PCB off the hot plate. This is what it looked like:

Click for larger version

Click for larger version

A lot of tutorials warn against using too much solder paste. If anything I should have used a bit more on the "big" components like the 1206 resistors and some of the 0805 capacitors. But both the SOIC and TQFP packages look nice and not a single solder bridge to clean up.
Of course, the real test is the QFN device. The pads are all nicely aligned with the PCB traces leading straight to the pads so it looks good, but we'll see when the board has been fully assembled (battery holder hand-soldered on to the bottom side of the PCB, for example) and the MCU has been programmed…


Even though it worked well enough, I will need to make two improvements:

  1. A thinner needle for the solder paste syringe. I will order these 25 gauge (0.25 mm) needles from Farnell to be able to apply the correct amount of solder paste to the pads.
  2. I will need an aluminum slab (about 15 cm square and 1 cm thick) to more evenly distribute the heat. Because the hot plate heats up pretty unevenly. According to this IR camera image, there is a difference of about 80 °C between the hottest and coldest spot on the hot plate:
Thermograph image of the hotplate

Thermograph image of the hotplate


A special thanks to the authors of the following articles – they were a great help:

Comments 1 comment »