Soldering tiny chips - geekery
I owe you guys an entry that summarizes my life the past few months. But I've been lazy and not much has been happening anyway. Still, I plan on it soon.
But after seeing my project list balloon up, I decided to tackle one from the list this weekend.
Many months ago, I contemplated using 3-axis accelerometer chips in various electronics projects but I needed to get the chips. A while ago, I discovered that all accelerometer chips come in QFN packages. There are many specialized package names with their own acronyms, but QFN is the basic style name. Quad Flat No-lead. That means they are square, with pins on all 4 sides, and there are no leads on the pin. Oh, and they're small. Like, fucking tiny.
I searched online and found breakout boards that are useful for prototyping. You solder a surface mount chip onto the breakout board and then you can use the board in prototyping projects. Great, so I ordered a small batch of 8 3-axis accelerometers and 8 breakout boards.
When they came in, I looked at it all and thought it was a daunting task. I mean, look at these things:
This is the chip, the breakout board, and a ruler. Each small line is 1 millimeter
This is the bottom of the chip.
Go on, click on the pictures to get a better view.
Each of those little pips on the bottom of the chip is a pin. Huh, this was going to be tougher than I thought.
But I decided to tackle the project this weekend. Thursday night, I tried doing it the naive way I would solder any old SMD chip that I would do at work. But after a frustrating two hours, I realized that this was just not going to work.
Luckily, I found a website that detailed two 'easy' ways to deal with these chips. One is to solder magnet wire (30 gauge) to each pin. But the other way would work well with the breakout board I had. I needed to get flux, flux remover, and a heat gun.
There's already a heatgun in the house. I found flux and remover at Fry's. Fry's used to have so much more, but they're starting to disappoint me on a number of fronts for electronics.
Anyway, back at home, I tried going through the steps outlined in the webvideo, but there were problems. The first problem is that I couldn't do it on my desk. The heat gun made a small blister on the workbench. So I moved outside to a metal tray.
Then all the flux I used to help the solder flow also made the chip 'swim' on top of the breakout board, getting all twisted. It was quite frustrating to deal with it, but I managed.
After many hours, I completed the process on 8 chips. Then I tested them. I screwed up and burned 2 of them out because of a stupid wiring mistake, sending over 6 volts to a chip whose max input is 3.6 V. Oh well.
Of the remaining 6, 4 worked flawlessly! The other 2 had one axis not connected. Today I redid the process, and those work just fine! While the process seems similar, I always surprised myself when something that seems too simple for what looks like a delicate process actually works.
This is the soldered chip on the breakout board
Now I can do something with these chips. I have some plans. But first, I'll do some more testing with them to see if they'll do what I want them to.
But after seeing my project list balloon up, I decided to tackle one from the list this weekend.
Many months ago, I contemplated using 3-axis accelerometer chips in various electronics projects but I needed to get the chips. A while ago, I discovered that all accelerometer chips come in QFN packages. There are many specialized package names with their own acronyms, but QFN is the basic style name. Quad Flat No-lead. That means they are square, with pins on all 4 sides, and there are no leads on the pin. Oh, and they're small. Like, fucking tiny.
I searched online and found breakout boards that are useful for prototyping. You solder a surface mount chip onto the breakout board and then you can use the board in prototyping projects. Great, so I ordered a small batch of 8 3-axis accelerometers and 8 breakout boards.
When they came in, I looked at it all and thought it was a daunting task. I mean, look at these things:
This is the chip, the breakout board, and a ruler. Each small line is 1 millimeter
This is the bottom of the chip.
Go on, click on the pictures to get a better view.
Each of those little pips on the bottom of the chip is a pin. Huh, this was going to be tougher than I thought.
But I decided to tackle the project this weekend. Thursday night, I tried doing it the naive way I would solder any old SMD chip that I would do at work. But after a frustrating two hours, I realized that this was just not going to work.
Luckily, I found a website that detailed two 'easy' ways to deal with these chips. One is to solder magnet wire (30 gauge) to each pin. But the other way would work well with the breakout board I had. I needed to get flux, flux remover, and a heat gun.
There's already a heatgun in the house. I found flux and remover at Fry's. Fry's used to have so much more, but they're starting to disappoint me on a number of fronts for electronics.
Anyway, back at home, I tried going through the steps outlined in the webvideo, but there were problems. The first problem is that I couldn't do it on my desk. The heat gun made a small blister on the workbench. So I moved outside to a metal tray.
Then all the flux I used to help the solder flow also made the chip 'swim' on top of the breakout board, getting all twisted. It was quite frustrating to deal with it, but I managed.
After many hours, I completed the process on 8 chips. Then I tested them. I screwed up and burned 2 of them out because of a stupid wiring mistake, sending over 6 volts to a chip whose max input is 3.6 V. Oh well.
Of the remaining 6, 4 worked flawlessly! The other 2 had one axis not connected. Today I redid the process, and those work just fine! While the process seems similar, I always surprised myself when something that seems too simple for what looks like a delicate process actually works.
This is the soldered chip on the breakout board
Now I can do something with these chips. I have some plans. But first, I'll do some more testing with them to see if they'll do what I want them to.