I have carried out a performance review of the common available GPSs that you might use for a LoRa tracker.
In most cases you will run the GPS in hot fix mode, and unfortunately most of the common GPS breakout boards that you get on eBay etc, are not suited for this in a power miser tracker that needs to survive on batteries for a month or longer. The reasons are explained in the article.
There is data of the performance of 12 common GPSs, signal strength (antenna effectiveness), real world power consumption and hot fix performance over an extended period. There are notes on methods of reducing tracker power consumption.
Choose the wrong GPS and your tracker, powered by a 2800mAhr battery, might only last 14 days. Choose the GPS wisely and the same battery can last 200 days.
Thanks, that was quite enlightening. I was previously under the mistaken impression that the ublox modules were the best. Itâs not even a close comparison between them and quectel. I agree that once sleep current reaches the single digits microamp range, there are bigger issues to tackle when it comes to battery life. This is especially true with the large duty cycle and the relatively huge power consumption of GPS modules.
There is the UBLOX SAM M8Q, mentioned at the end of the report, that had a very low average hot fix over 24 hours, 3.2 seconds. But it can be expensive and is not as easy as the Quectel L80 for DIY soldering. The L80 has the solder pads wrapped up on the edge of the package making soldering in place easy.
I have a L76 and L86 in the post, they both look quite promising, and will add them when tested.
As for node power consumption, whilst it is easy enough to have a node power down to sub 0.1uA range (with a TPL5110) its also not difficult to arange the components so that with one of the ATMegas the sleep current is 2.5uA or so. With this approach the processor is just stalled with memory intact, ready to resume operation in 5mS or so, versus 1.5 seconds for a node that is powering up all the time.
I donât understand why it takes 5mS to wake up from sleep, unless it is a watch dog reset, not a watch dog interrupt. The datasheet makes it clear that waking from the sleep modes takes 16k clock cycles and no other delays. Coming out of reset is a different story, you have the 16k clocks (1mS at 16MHz), plus a potential 4.1mS or 65mS delay appended (depending on the value of SUT1 and SUT0 fuses). To me, this means that with proper fuse selection, the chip should wake up from âpower down modeâ in 1mS assuming a 16MHz crystal. Power down mode is the formal name of the deepest sleep mode that preserves RAM contents.
I did explain in the report what the âwakeupâ was measuring, so there would have been some bits of program for the processor to run before the logic pin was changed to register (on my scope) that the program was continuing.
Its was all tested on a 3.3V 8Mhz device, I dont use 5V devices much anymore, too much like hard work.
8MHz would certainly double the wakeup time to 2mS minimum. I guess the rest would be actual startup time of the oscillator. Iâll go back and read your report again. I need to do some testing myself, as I have the equipment, itâs just that it is in storage due to a flood. I just wanted to be sure that you hadnât inadvertently used a watchdog reset instead of the interrupt, but that would add the massive overhead of the bootloader, unless you were bypassing that by using SPI programming. Thanks for your input, Iâm grateful that you took the time to work all this out.
Edit:. Iâd be greatly interested in knowing your fuse bit settings, if you could provide those. 3 more mS of wake up would entail millions of instructions executing. Maybe the particular âcrystalâ is slow to get going.
Thanks again. For anyone interested, on Arduino 1.8.5, they are set to pad the reset minimum 16k ticks (2mS) with an additional 14 ticks (~2uS) and 65mS bringing the total minimum to a bit over 67mS. Of course this is compunded by the huge bootloader delay. I am creating a separate thread to discuss some additional power saving possibilities. If you are interested, Iâd appreciate your input. Thanks again.