Jan Rychter

K22 family observations/feedback

Discussion created by Jan Rychter on Jan 7, 2015
Latest reply on Jun 18, 2015 by Zhe Tian

I recently took a first look at the K22 family and I have some observations. I don't know where else to post this, so I'm posting here, hoping it will reach the right people.


I've been looking at migrating from K20 to K22 in a low-cost USB device project. I looked at the K22 family and… well, I'm not happy. Here are my main observations:


  • The K22 is supposed to be "optimized for cost-sensitive applications". Sounds great! I was so happy to see crystal-less USB operation, the crystal is always a significant cost component, both in terms of BOM cost and board space! But then I discovered that the small K22 devices with FP that I wanted to use (MK22FN128) do not have the USB VREG! Why, oh, why? This means that I can save on the crystal, but have to add the cost of an external LDO! Not good! What's even worse is that it's very difficult to find out that the VREG isn't there. There is even a bug in the datasheet: the front page of K22P100M100SF9 says "USB LS/FS OTG 2.0 with embedded 3.3 V, 120 mA LDO voltage regulator", which is not true. Only the KINETISKMCUSELGD shows the lack of integrated LDO if you look closely (the lack of a (2) footnote next to four devices). Why was the VREG removed from those 4 devices? It makes no sense.
  • Packaging. The entire K22 family doesn't have a single QFN device. It's a disaster for people designing small cost-sensitive devices: you can't use BGA because of the PCB requirements increasing costs significantly, so you're stuck with the huge LQFP packages. The smallest I can use is 64LQFP, which is 10x10mm, while right now I'm at 32QFN (5x5mm) with the K20.


The K22 devices would be perfect for my applications if it wasn't for the above problems.