I am powering a Boron with alkaline batteries via VSUB and the LiPo charging light is flashing continuously (even during sleep). To minimise current draw I would like to disable the LiPo charging and LED.
The documentation recommends using the Power Manager API rather than directly changing the PMIC settings. But disabling charging only seems to be available in the Power Manager from version 3.0.0 (I am using 2.x LTS).
The older technique will probably work most of the time. However, the reason the feature was added to the power manager is that sometimes the setting will change at unexpected times.
If low power is your goal you would need to power the device via the Li+ pin, otherwise you’ll see excess current draw from the PMIC - this also solves your charging LED problem.
VUSB will add a couple of mA draw vs. the quoted uA draw on the Li+ pin. The BQ24195L (PMIC) has a recommended limit of 4.4V on the Li+ pin, but you could get away with 4.5V as the maximum is 6V, thus 3x AA batteries should work.
Thanks @Rftop and @no1089. I hadn’t appreciated the additional current draw from the BQ24195L when powered by VUSB. I’ve taken a look at the datasheet and have a clearer understanding of this now. Interestingly, the Boron datasheet states an absolute maximum input on the Li+ pin of 6.5V vs the BQ24195L datasheet maximum of 6V (unless I’m missing something). Either way, there’s a bit of headroom for an initial cell voltage above 1.5V when using 3 cells. I wonder whether I might get away with using 4 cells…?
I’ve been using alkaline D cells (Duracell Procell Intense, nominal capacity 15k mAh), partly for their high capacity and partly because some systems will be deployed in Africa where access to replacement lithium cells may not be possible. The plan was to use a 4S2P configuration to maximise capacity. One reason I’ve been using VUSB is the greater flexibility it gives with input voltages, and the ability to use the full discharge curve of the batteries. If I supply power with 3 cells on Li+ it will only allow the cells to drain to 1.1V (as the minimum voltage on Li+ is 3.3V). That’s not an issue with the lithium Energiser L91 cells, but is with the discharge curve of alkaline cells. Again, I wonder if I could get away with using 4 cells?
@Rftop I note that you had planned to test a 4S and 3S2P cell setup (Low-Power Boron LTE, using the EN Pin - #2 by Rftop). I can’t find any reference to the results of these tests (sorry if I’ve missed it). Are you able to give any feedback on these?
Shortly after the 10,000 publish Trial, I realized that wasn't a great idea (constantly connecting to a Cell Tower) - not very professional
So I never repeated the torture test with the 4S or 3S2P L91 configuration.
Brief testing of 4S didn't show any problems, but no exhaustive testing was performed of 4S L91's.
After you posted more details about your project, my thoughts would be to nail down your power requirements for your power budget. Decide on your Duty Cycle for the SPS30/Boron, your efficiencies (regulator, supply, etc), and your required Battery Life Cycle.
Then select the best/appropriate battery chemistry, capacity, and arrangement.
Once you have a proposed battery pack, be sure to check it's discharge curve with your particular operating conditions (max Current of any Boron/Sensor operations), to determine your usable capacity.
Be sure to re-calc the current at the lower battery voltage too. That's easily overlooked by many (myself included) when estimating battery life for the total system. The pack must be able to meet the higher current at the lowest voltage to define it's usable capacity for any particular system. That's when Brown-outs start.
I cant help myself......so I must ask: is a small solar panel possible ?
That's good to know, and gives me some confidence that I won't toast my development devices by giving this a try.
Quite possibly. The problem is that some devices will be installed in locations where the temperature is sometimes below 0C, so LiPo charging will be a problem. One of the use cases is on a glacier surface, and others in the UK during the winter. It may be possible to reduce sampling intervals at low temperatures to save battery, and only publish when the temperature is >0C. The long term solution is probably to utilise Lithium Titanate cells and a dedicated charging circuit. But I don't have time to work on that at the moment.
But you don't necessarily have to rule-out Lead Acid (SLA) or Li-Po for freezing conditions.
You can add a Temp Sensor and disable charging for a 5,000 or 10,000 mAh Li-Po when frozen.
Or a 7-12Ah SLA at 12V can operate a Sleeping Boron for a LONG time during freezing conditions.
I've used this $30 12V Panel/Controller & SLA w/ a Boron many times when Space wasn't an issue.
Agreed, there are several folks on this forum with a lot of experience w/ those cells.
I seem to remember a post about monitoring glaciers, but maybe that was you
Which country in Africa? We definitely have Lithium cells down here in South Africa.
The PMIC does have an OTG function that will output 5V on VUSB - although I don't think we have guidance on doing that. A couple of folks have used this afaik. But otherwise boosting from the LiPo would be better.
The D cells are closer to 1.5V than the Lithium cells if I remember correctly, so four should be fine. I'm not sure why the 6V/6.5V discrepancy is there to be honest. @rickkas7 Do you know?
You'll get much longer run times on Li+ than VUSB, even without draining the cells that far as the PMIC won't go to sleep and use a couple mA vs. the uA draws from Li+.
The 2G/3G devices can pull 1.8A of current when connecting, so be sure to add some capacitance - it's another reason why the Li+ is a good idea. The lower your cells drain the harder it will be to get that amount of current out of them.
On charging in low temperatures - the TrackerONE simply disables charging when temperature drops below 1°C (Not sure on exact temperature). It works reasonably well. The bigger problem is that depending on the battery, there is a level where you need to stop pulling current from the cell in order to prevent damage. (-10°C for LiPos generally).
Those are all very valid points. My device has an internal temperature sensor, so adding some code to manage the charging won’t be a problem. I hadn’t come across single cell LiPos with such large capacities, but there do seem to be a few of them around now and they are likely to carry the system through cold spells in the UK. Those Eco-worthy solar systems look like quite a good bet too.
It could well be my post about glacier monitoring that you’re referring to. I’ve been working on this, on and off, for a while
Just a reminder to anyone : Check the polarity on Li-Po cells purchased outside of Particle. I've had to swap the pins on many larger Li-Po batteries before using.
There's a 100% chance of a Disaster, 50% of the time
The plan is Kenya. I'm sure lithium cells are available, but I'm not sure whether they'll be as easy to obtain as they are here. I know that alkaline cells are readily available.
I hadn't appreciated that. I found this forum post: How to enable OTG on the PMID to provide 5v from battery. It seems to suggest it's as simple as pmic.enableOTG(); This sounds like a great feature, and I'm surprised it doesn't seem to be documented (other than a statement that that command exists). Are there pitfalls I should be looking out for? Is it likely to work above the recommended operating voltage on Li+ (the BQ24195L datasheet only gives a minimum input voltage, as far as I can see)? I'm not opposed to using a separate regulator, but it adds cost and complexity, so it would be great if I could avoid it. The current draw of the SPS30 is 30mA max, if that makes a difference (again, I can't see anything in the datasheet that suggests this is a problem). I'll experiment with this next week.
That's very reassuring
Thanks again all of you who have helped with this. Every day's a school day
I’ve also tried using pmic.enableOTG(); in loop, in case the setting is being overwritten somehow, but with no success.
Unless anyone can spot anything I’m missing, it looks like I’m going to have to use a separate voltage regulator to power the SPS30. I’m thinking that stepping up from the 3.3V output is likely to be better than using a more complex buck/boost regulator directly from the cells (I only need 30mA max).
Boost mode will not work on the Boron and Electron because, while the 5V boost will be present on PMID if enabled in software, PMID isn’t connected to anything and is not exposed on a pin, so it’s essentially unusable.
In theory you could use it on the E Series because it is exposed on a pad, but you have to be careful because PMID will exceed 5V when VIN is greater than 5V, which isn’t a great quality in a boost converter output. It’s often easier just to add an external boost converter.
