Running off of a LiPo battery - what is the cutoff voltage to sense?

I’m running my Photons on a LiPo battery. The battery is connected to the 3v3 pin.
My questions are?

  1. If I’m sensing the battery voltage in order to sense an alert when the battery voltage is getting low, what it the minimum threshold I should be looking for?
  2. Is there a LDO or a ULDO connected to the 3V3 pin? If so, what is it and what is the drop out voltage for this device?

Does the LiPo not damage the photon when you power through the 3V3 pin?

The lower end of the LiPo should be 3.0V but i don’t recommend going that low.

There’s not LDO/ULDO at 3V3. That goes directly to the STM32 microcontroller

I’ve been running the Photon all day today on a LiPo 2500mAh cell starting with it fully changed. I read somewhere on here that one could do that so I tried it. FYI the Adafruit Feather boards also work directly off of a fully charged LiPo.

Thanks for info on the lower end. I was thinking of using low threshold of 3.2v.

Watch out for bricking core in battery application due to brown out condition which can happen if you let the battery run low.

It happened to me when I thought it would be simple to just run a Spark off a Lipo and solar charger, without thinking about it too much -> Ended up with a Spark flashing green (best I could tell) and completely unresponsive.

My most recent setup is configured Lipo -> Buck/Boost 5V -> MCP120 4.5V Voltage Supervisor [pull Reset to low if <4.3V] -> VIN to Photon. It’s been working so far.

Yes, thanks for the heads up on LiPo running too low.

My most recent setup is configured Lipo -> Buck/Boost 5V -> MCP120 4.5V Voltage Monitor [pull Reset to low if <4.3V] -> VIN to Photon. It's been working so far

Interesting. What's the "mileage" you get and the current drain in your setup? Pulling reset to low is interesting. How are you switching to take the Vin to Photo (I'm guessing the 3V3 pin) directly?

I actually just returned to this project only a few days ago, so it hasn’t been running long enough to be sure what kind of mileage I’m going to get or what glitches I’ll encounter. The tests all worked.

The supply voltage is 5V to Pin 1 on the Photon, which is then converted on the board to 3V3. (So yeah, it’s going battery 3.7V -> buck/boost 5V -> photon 3v3, but from other discussions I’ve seen here, given high efficiency conversion, that’s fine, and it gives me more control (i.e. to monitor the 5V supply, which should stay at 5V)).

I actually did measure the current drain, so I can give you those numbers: Photon + display + probe (my configuration): 70mA (~115mA momentary pk when it comes up and does Wifi). That’s not going to give me much in terms of lifetime from a 2500mAh Lipo if always-on. But when I put the Photon into System.sleep(SLEEP_MODE_DEEP, long seconds) with the peripherals powered, I measure 2.5 mA, so that’s good enough for now since I have a solar panel to keep charging the Lipo (using this). Without the peripherals the draw is <1mA, less than I can measure with my multimeter.

Long term I plan to put together a board that shuts off power to the display and the probe when the Photon goes to sleep. For me, for now, though, it’s working fine - allowing me to take measurements for 5 minutes every hour, upload the values to Ubidots, then go back to sleep. The sleep and wake are working exactly as expected. I’ll know more in a week or two.

A fully charged LiPo will deliver up to 4.2V and under some external circumstances even above that when fully charged.
So directly connecting it to the 3V3 pin that directly feeds into the STM32F doesn't seem to be the safest IMHO (even if it works for a while)
STM32Fxx datasheet

And the Photon datasheet does limit that even further, so driving that high will definetly be using it out of specs at your own risk.

https://docs.particle.io/datasheets/photon-datasheet/#pin-description

https://docs.particle.io/datasheets/photon-datasheet/#recommended-operating-conditions

And the Photon datasheet does limit that even further, so driving that high will definetly be using it out of specs at your own risk.

Got it. Thank you for digging through the datasheet (I should have thought of that). Someone on these forms said one could connect the LiPo directly to the 3V3 pin on the Photon (there was a long discussion about using LiPo batteries with the Photon)

1 Like

Could you link us to the topic to put a warning up there too?

Not voltage but rather Joules counter chips. out-in = unused.
Charging replaces.
Temperature monitoring.

Chips for battery management do this.

@dfreymann, thank you for all of the information you’ve provided. It’s good to know that sleep modes work as expected.

I’m at a similar stage. I have everything working but not power consumption optimized yet. That will require a lot ancillary components that are controlled via the MCU. The big difference for me is that I can’t actually turn of the Wifi for any length of time, since the device also receives commands (via MQTT) besides having sensors that needs to publish events in real-time.

So I’ll probably have to bump up the mAh of my batteries. I currently have 2, 2500mAh batteries in parallel (of course with my new found knowledge I’ll have to do the boost-buck thing or find a ULDO device that’s affordable).

I’m currently at 100mA drain, perpetually.

Turns out there’s a failure mode for my setup.

Found my Photon this morning blinking 3 orange, pause, 1 cyan(?). It was unresponsive to pressing RESET. (???)

Best I can tell, this is a failure to connect to the wifi/cloud, but I’m not sure yet.

But because it was unresponsive to RESET, my cunning MCP120 voltage supervisor circuit - which pulled !RST low - failed to do anything.

The Photon apparently kept blinking all night. The battery was drained (since the device had been running and blinking for 12 hours+ when I found it).

I was only able to recover the Photon by disconnecting power (and reconnecting to USB power for a bit).

So, there you go. More stuff to figure out.

@dfreymann,

Thanks for the update! Do you happen to know at what voltage this happened? Or what is the trip point you’re using?

Nope, not yet. Thing ran for ~four days before failing. There could have been another problem with my voltage supply - I’m still dissecting things.