RAK 4631 Solar Repeater

Hi, I have and want to setup a RAK 4631 as a solar powered repeater in a high location. I noticed, while testing, that when the battery drained, and later recharged, that the node did not power back on. I had to turn it on by a short press of the reset button. This isn’t ideal for a weatherproofed unattended deployment. How have others ensured that their RAK 4631 powers back on when the battery level goes back up? There must be a way to make this happen?


Same thing here. Just took a repeater of a mountain. However the location should have provided enough solar ooomph as the other repeater is doing fine. I’m hoping the charger unit on the board is okay and its just a battery or solar problem. I’ve just ordered some 250mA panels as the current 150mA are only just cutting it (I’ve yet to turn off bluetooth).
Good if they could restart on battery levels coming up.


There is not an addressable power chip on the rak boards, this is not something software can fix. With the low power consumption of a rak if your battery is going totally dead you need a bigger panel or battery.


this is really disappointing. It makes the RAK boards unsuitable for deployment in a critical or unattended location (such as high up on a tower). Like solar garden lights, that do come back on when charged, one would reasonably expect the RAK hardware to do the same when the battery level passes a certain threshold. Seems really silly that a device such as the RAK 4631 does not have hardware within to sense that battery voltage had been restored and to automatically power on. Yes that is not a software function and is in the domain of hardware (electronics). Maybe a ticket needs to be opened with RAK Wireless?

In my experience with this problem, the only way to get the RAK 4631 to power back on after recharge, is to remove the battery, then replace the battery and then hit the reset button. Hitting the reset button without removing the battery after charging does not allow the board to turn on. Really stupid.


Sent a support request to RAK Wireless - I guess we’ll see what they say?


Assuming you properly size your battery and solar panel, this shouldn’t be an issue. There are folks that have run their rak solar device continuously for months without getting close to the battery dying.

But for those running undersized panels, and small batteries it makes sense how that can be an issue. I know there was some folks talking about this in the discord the other night and possible solutions.

1 Like

yup, but for a mission critical install, the criteria requires the deployed device to self-recover. So that makes this (in it’s present form) a non-starter. For my need, a device must be able to self-recover when charged, and require no human intervention. Device also needs to be discreet, therefore a smaller panel (5watt) and 3000mAh Li-Po battery. These are the physical criteria for our deployment.

1 Like

If it’s mission critical, wouldn’t an appropriately sized battery/panel be a key criteria of that to avoid it dying in the first place?

There’s still discreet options available. For example this setup: Printables. Relatively discreet based on the dimensions, and depending on batteries use it could be ran on battery alone for 15-30 days completely covering potential inclement weather periods that may be experienced.

But hey, use what you have, I get that. Hopefully Rak sends you a response. I’m sure that’s something many of the others that use undersized batteries/panels would be interested in hearing about. Good luck!


panel and battery can be oversized, but still some more/less expected things can happen that put the panel in the dark for some time and the battery will be empty …
Snow, cover by leaves, dust on the panel /dust in air…

I can’t remember how the quote goes, but something to the effect of “We can do anything, but we can’t do everything”. If RAK built their devices for every possible scenario then the hardware becomes more expensive, complex and power hungry. Someone correct me if i’m wrong, but once a device has shut down from a low power event, it would still need reliable power to monitor its state until it has decided it has enough reliable power to power back on fully? Again, this is why it makes more sense to just build the solar and battery setup to statistically account for and budget for extended low power events. For some things, RAK has add on modules, for other things there are either workarounds or people coming up with their own solutions (see LTO channel on Discord to deal with charging in -40 degree weather). If i’m not mistaken, the RAK units will cycle back on after detecting power to the USB charging port, so adding a second solar unit with it’s own charging timer might be a place to start for finding a work around?


It is not something that RAK has any control over, details are in the NRF data sheet. This is not a feature of the NRF52.

1 Like

Isn’t this mitigated by using a solar panel connected to the USB port, rather than the solar JST input? I recognize there is still a problem with the solar connector, but as others have said it’s not something we can change unfortunately.


I have an open ticket with RAK on this. The issue is well documented in the use case of solar powered remote deployments. They have been responding on the ticket and discussion with them is ongoing. WIll update once a conclusion is reached.

For the record, the issue is seen when:

  1. battery is allowed to go flat - I did this in my testing to test recoverability, as once these are placed, I will have no access.
  2. battery slowly charges via solar panel, unit does not recover, even when battery is fully charged.
  3. battery is charged via USB-C from flat to full, unit does not recover when battery is fully charged.
  4. in both cases, where the battery has been recharged from a flat state, to full, the 4631 does not recover until you disconnect the battery and reconnect, in which there is no way to remotely do this, hence makes this hardware unrecoverable should the battery go flat and recharge.

In the discussion so far with support, they agree that the expectation of self recovery on recharge is reasonable and that the issue seems to be related to something known as the Power On Rise time. When a battery recharges, the rise time on the voltage is obviously long, and because of this, does not allow the unit to self-recover.

1 Like

This is a problem with the ESP devices as well. On the ESP-8266s I was testing if the voltage dropped below around 2.3V the ESP entered a “locked” state and it could not recover until power was removed. I believe it can be overcome by making a small change to your power supply circuit. I was using a LDO with an chip enable pin, I was going to use that with a power supervisor chip to turn off the LDO when the voltage dropped below 3V and when the sun came back up it would come back on. Designing the circuit and selecting the parts is more than I have time for right now. Maybe I can pass it on to someone else.

Sounds exactly like what @garth linked to, which is a nordic issue.

It is a little better, same with a green power board, still not controlable without managing it explicitly.

I also still have the same problem, also discussed in discord… if i deploy a node in a hardly acesibel spot…
i am not good in electronics, but maybe putting a small “Varistor” or combination of small CAP and Varistor coul help to “switch on” the solar panel that is connected to USB.

so the voltage on USB will not rise slowly, but is beeing switched on at 5V… so the device will start…
i have yet not found any varistor below 5V…

VC 1206-05 SMD-Varistor 1206, 5,6 Volt

it can also be used to protect the Rak device from solar overvoltage :slight_smile:

1 Like

You can fix that by setting Super Deep Sleep timeout to a more sensible value than default (e.g. few hours).

P.S.: ESP32 devices could be revived by pressing the middle button (on a T-Beam), no need to disconnect and reconnect power

I’ve got a RAK 4631, WisBlock Core (RAK4631) and WisBlock Base (RAK19007) with a 900mAh LiPo and the RAKBox-B2 Enclosure with solar panel. The solar panel and battery are connected via JST and the RAK board is the older version with Micro USB.

Having stored the unit indoors for 2 weeks, the battery was completely flat. I placed in the full sun and after several minutes was automatically running and visible on the mesh.

Firmware 2.1.13, Bluetooth OFF, Power saving mode OFF, configured as a router.