Wow, that’s awesome work. The gnat is amazingly small. Does it require an external antenna for the GPS, or is there an internal patch? Personally, I’m really interested in the grasshopper model for a project I’m working on. Ive only skimmed through the docs so far, but it appears that it also has very well thought out low power modes for battery operation. One question I have is about the USB interfaces on all of these, doesn’t that add a fair amount to the power requirements, or is extremely low power as well when it’s not being used? Forgive me if I’ve asked a stupid question here.
The USB peripheral does indeed use a lot of power relatively speaking, but only when the USB cable is connected. There is a USB connect detect built into all of the board designs such that the USB peripheral is disabled when no VUSB is detected. In this way we can achieve the lowest power inherent in the Murata Type ABZ (CMWX1ZZABZ-078) module.
So for Grasshopper, the sleep current (once the power on led is disabled) is ~2.1 uA, ~1.65 uA for the L082 STOP mode (with full retention, 5 us wakeup) plus ~0.5 uA for the NCV8170 LDO.
For the LoRa SensorTile (AKA Cicada) this is ~8 uA since there is a battery charger and 16 MByte SPI flash as well as some sensors on the board. The Cricket asset tracker is a little more (~12 uA) since the 8 MByte flash uses ~2 uA less power but the GNSS RTC backup uses ~10 uA more or so.
The Gnat and longCricket were designed to minimize power usage even further (check the Tindie product pages for a fuller description) so that each uses as little as ~2.5 uA when the L082 GPIO powering GNSS RTC backup is disabled while the BMA400 accelerometer continuously monitors for fast wake-on-motion.
This is probably more than you wanted to know…;>