kekiefer
commented
8 years ago Thanks for the updates!
This looks pretty good to me from a code review perspective.
My only gripe is that the default SPI CS for this application has changed from 0 to 1, so that will break people who swap out their L2 for an L3 on the same system. However there is now better plumbing in place to address it properly. What do others think about this?
Also, any suggestions on if/how we should maintain backwards compatibility with L2?
Rewrote the code in LeptonThread.cpp to accommodate a slower SPI speed (20MHz). Switched from simply using an array to store the read to using segmented structs. An image is created from four segments, which hold 60 packets of data each. Each packet is a 164 byte line from SPI. This version reads 20 packets into a frame buffer at a time, which is as close to the 4096 byte limit as we could go. The Lepton runs just as well as before, with very few dropped frames.
Andrew Myers
Novacoast, Inc. amyers@novacoast.com