I’ll run a power measurement of 1.2.59 in router mode in my lab over the weekend. Will see if the long term draw meets our expectations.
1 Like
It drew 55ma average. It should draw less, that means we have a power consumption bug somewhere. Hmm…
Sound much better than in my case. Could you provide me with some more information? Was the measurement done with power supply via battery or USB? Which hardware was tested, the t-beam and which settings did you use, the is_router true setting?
Do you have any hint for me, what I might do to be able to get also the 55mA as result? Any suggestion is helpful, since the 55mA would already give me one more day of battery life.
It was a tbeam set as a router powered by the battery terminals at 3.6v. The result was the average consumption over about 30 hours of run time.
I just startet a similar test with my tbeam also in router mode and powered by the battery terminals. Actually, in may case, 3.6 Volts didn’t work properly. The devices only reacted when turned on manually and switched off automatically after few seconds, but with 3.8 volts they turned on automatically and stayed on.
I will continue the test until tomorrow morning (approx. 20 hours) might be sufficient to get reliable results.
They’ll have to be manually turned on. That’s the behavior of the AXP chip. Sometimes they’ll just turn on, but consider that a fluke.
Well, with the 3.8 Volts my device turned on when connected to the power source.
After nearly 2 hours of testing, I get an average of 93 mA, but I also recognized, that the display is always turned on, and the LED is flashing every second, which isn’t necessary for a router. I will continue, but I guess, the consumption will stay this high.
I wonder if that’s because the axp assumes that going above that threshold must mean the battery is being charged. Mmmmmm…
So, I finished my measurements, with 3.8 Volts I got on average 92,4 mA, the display is always on, which is surprising, since I thought the router mode would turn this off after some minutes.
I did another run with 3.6 Volts (when reducing the voltage from 3.8 to 3.6 V while device is turned on, it works and devices stays on), the average current is 94,9 mA.
Unfortunately, I don’t see the 55 mA, but I would guess that something with the router mode is not working properly (display always on, what I observed with battery powered devices as well).
Note: with supply voltage of 3.6 when turning the device off and on again, I see the same behavior as before, device turns on for few seconds and goes off again.
If you don’t have is_always_powered turned on then the screen is turning on when there is lora activity, your device is never sleeping at all because of mesh activity.
Sounds reasonable, just asking to be sure, shall I use the command:
meshtastic --set is_always_powered false
Or should the argument be true here? I will run the test tomorrow, would be great, if I would see lower power consumption.
I actually thought, this setting was part of the is_router setting.
Thank you @garth .
On this page I don’t see how to disable screen. There is the “screen_on_secs” setting, but unfortunately setting it to 0 will set it to 60 (default, one minute). Do you know how to disable screen, at least after bootup, please ? Thank you
Did you try to hold the button on the device for a while? That works for me.
0 is default in protobufs, did 1-5 not do what you want?
Hi Garth,
thank you. No, I did not try 1-5. Is this a special range that really disable screen ? Or is this a workaround to limit screen ON time ?
Cheers
That is the right way to limit the time the screen is on.
I ordered a multimeter that includes data logging. Original plan was for collecting data on my solar node setups, I think it may help with this area as well.
For power limited routers, is it possible to shut off, and keep off everything (screens, leds, GPS, WiFi, Bluetooth…) not needed to repeat packets and set the esp32 CPU to the slowest clock speed that can keep up with the lora radio?
I can’ remember for sure what the clock speed is currently set to but I found this table
There are functions to manage wifi power consumption, anyone know if they are being used? Middle of this page talks about wifi https://www.mischianti.org/2021/03/06/esp32-practical-power-saving-manage-wifi-and-cpu-1/
Since the is_router mode doesn’t affect the power consumption on my side, I followed the recommendation from garth and ordered two WisBlock RAK4630 (WisBlock Meshtastic Starter Kit | The Basic Kit for Meshtastic – RAKwireless Store)
I followed the clear instructions and uploaded the newest Meshtastic firmware for this system, which works very easy, since the device shows up in the finder/explorer window.
Thereafter, I was able to pair the RAK4630 device with the Meshtastic app and communication with another LORA T-Beam was possible.
This morning I started a measurement of the power consumption, after 2.5 hours I got 4.9 mAh/hour, which is 22 times less than what I measured for the T-Beam. This would allow 28 days on-state with the 16850 battery.
@grath thanks for the recommendation, now I will start check, if the RAK4630 board is also capable of charging a battery through the solar panel connection.