Often a chip will be available in several different packages. Sometimes QFN which has a thermal pad, and TQFP which has no thermal pad. The justification for the thermal pad is that it helps conduct heat away from the IC. If this were the case, then why doesn't the TQFP need the thermal pad?
The reason I'm moaning is that the thermal pad gets right in the way of the layout. Tracks and vias can't be placed under the device (except in some cases), making fanout tricky in space confined PCBs.
Is the thermal pad just traditional, or is there a good reason I'm not aware of?
Answer
The answer is that "it depends". Also, I would like to point out that there are TQFP's with thermal pads, so this isn't a QFN vs TQFP issue. It's a Pad or No Pad issue.
Pro's:
- It's easier to cool a chip with a thermal pad. Some parts absolutely need this, while other parts can have an extended ambient temperature range with a pad.
- Some parts don't need cooling, but use the "thermal" pad as a very low impedance connection to GND. This allows for lower noise and EMI.
- QFN's have a lower lead inductance and are generally smaller than the equivalent pinned TQFP. This improves signal integrity and lowers EMI.
Con's:
- Makes hand soldering much harder. Not as bad as a BGA, but you'll certainly need a hot-air solder station and possibly a preheat device for the larger parts.
- Makes PCB routing more difficult, especially on 2 layer PCB's.
I personally would always choose a thermal pad over no pad, and always a QFN over TQFP. Of course most of my PCB's are 6 or 8 layer boards with 4 mil traces, so routing and fanout usually aren't a huge problem. Cooling is usually a problem for me. I do benefit from the smaller QFN's. And the better signal integrity and lower EMI is a huge plus for me.
The other side benefit is that nobody asks me to rework a QFN! :)
No comments:
Post a Comment