Issues when powering thingy:91 from solar panel

[Jonathan E]

Hi!

I am trying to power the thingy:91 from solar panels and have some issues with the thingy:91 restarting when the solar panel receives a medium amount of light. The solar panel charges the device properly in full sun and does not interfere with the device when there is no sun. However, it starts rebooting when there is some light. I assume that this is because the USB solar panel outputs a voltage lower than 5V, but high enough that the regulators turn on so that the nrf9160 is supplied with too low voltage. I have tried with multiple USB-solar panels which most often are used to charge surveillance cameras.

Do you have any tips on how to run the thingy:91 from a USB solar panel properly?

Best regards,

Jonathan

[Kenneth]

Hello,

I would likely have measured the VDD on startup, and made sure it's stable over a few seconds and (sufficient high) before starting to do something useful, also periodically measure VDD and enter power down immediately if VDD start dropping below operating conditions:
https://docs.nordicsemi.com/bundle/ps_nrf9160/page/recommended_op_conditions.html 

Kenneth

[RichM]

I would also like to know this outcome. It seems like the PMIC is designed to handle this situation. I also tested with PPK2 and everything continues to work. Having an inexpensive solar option (like the ones available) would be a huge boon to our customers (beekeepers).

Thanks!

[Jonathan E]

Hi again,

I have actually tested this one:
https://voltaicsystems.com/solar-usb-regulator/?srsltid=AfmBOopYRxC9XLNVjKzA9O73vR363_SWk2QQ4Y-b0l6lik7y8aQQfee1

That also had the same issue as the solar panel.

I contacted Voltaic and they suggested something like this:
https://www.pololu.com/product/4085

Currently, I am testing an external voltage regulator/solar panel charger:

https://www.digikey.no/no/products/detail/dfrobot/DFR0559/9356334

This didn't work either and also has the same issue.

Now I am testing using DFR0559, an external battery and direct contact between the LIPO connected to the DFR0559 and the LIPO input on the thingy 91. This seems to work pretty well but is not optimal as it requires more installation steps and probably an EMC test.

[RichM]

Rats,I was hoping for a cheap/easy solution. Currently we use Voltaic V25 plus solar panel. That works well, but is a bit pricey and requires a larger weatherproof enclousure. We use the Thingy91 inside a solar shield and it makes a great weather station / BLE receiver to read our sensors.

Maybe I will see about a custom solution from Voltaic. They have been a good partner for us.

It is hard when you see the very cheap panels that _almost_ solve the problem.

[Jonathan E]

We have the same setup as you, where we use the Thingy:91 as a BLE to LTE-M/NB-IoT gateway. I have also considered using the V25 USB Battery Pack, but the whole setup becomes quite expensive.

I have also bought one of the new Thingy:91 X and will test if this has the same issue. The Thingy:91 X is also cheaper, so if this can be powered directly from a cheap 5V solar panel it is probably a nice solution.

[RichM]

I have the 91X and unfortunately it means new firmware... plus we use the hci_lpuart which has not been ported to the 91X yet. Something more to do... just a very busy time of the year for our small company with beekeeping season starting. Our group is BroodMInder.com, what is your's?

BTW, it took us a long time to get the T91 stable. We are (finally) quite happy with it and it is reporting all the connection parameters back to us as of about 4 months ago. Since we are installing globally, this is a huge help in problem areas.

[Jonathan E]

I have also been worried about the difficulties in upgrading the firmware, but hope it will be fine. However, I am worried that it will introduce more complexity as the nrf5340 has a network core and an application core, adding one more layer of things that can fail.

We also had a lot of issues getting the thingy 91 stable especially when using hci_lpuart. Since we use Bluetooth very frequently, we had to discard hci_lpuart and ended up using lpuart between the two devices, programming the nrf91 and nrf52840 separately, and only sending simple command messages between the chips.

Your products look really cool btw. We have started a company called botanist.ag, creating low-cost, accurate soil sensors for industrial use.