Both are on USB power (I stopped ordering AAA batteries every week)
Firmware is december 23rd version (don’t have Arduino software at hand now)
Switched them, now it drops below 7
Will see what happened tomorrow … (wife says cold Nodes should not be my priority now 
Interesting to hear the comments about the cold. Not tried the RN2483 myself but LoRa devices such as the RFM9xs seem to work quite well in high altitude balloons where it can get real cold.
I have temperature cycled my own trackers quite a bit in the freezer down to -20C or so, never had a problem. I use 62Khz bandwidth and so the capture range is circa 15Khz, never see them drift that much.
One issue to be aware of is that the crystals used have a fairly low drive, for low current operation, and as such could be sensitive to small amounts of condensation on the modules causing significant frequency shifts, I presume you are all conformal coating the modules to reduce the risk of this ?
And to comment my own post, once a receiver has picked up one packet, its easy enough to use AFC to keep it locked. I have had my trackers running at 7.8Khz bandwidth from room temperature down to -20C (the freezer again) and it works just fine. The AFC keeps the receiver within circa 1khz or less of the transmitter frequency.
The RN2483 ‘cold problem’ existed in the first generation… haven’t hear anyone with an rn2483A based node that experienced problems below -5 … so I guess that problem is solved.
Teaser for @BoRRoZ STM32 Blue Pil based Things Network Citizen Sensors. 
The antenna connectors may need some Liquid Electrical Tape for weather proofing.
Temperature/Humidity/Barometric Pressure Sensor inside the enclosure ? … hmm interesting
I notice at least one hole in the bottom, below the BME280.
I have a Seeeduino LoRaWAN based node with a 3.7V LiPo battery and a BME280 module, currently reporting -6.2°C, so based on my experience, a LiPo battery should be OK there. But indeed, it’s protected from rain 
Here is my ‘first generation’ result:
Stopped transmitting temperature once it went below 0
Hi DannyE
I suspect that’s a bug in your code somewhere, not handeling the negative values right because I have never seen a module stopping at exactly zero degrees.
Tx = RN2483 until now no problem
I am not doing anything (myself) with code in Nodes.
Loaded the default LPP sketch on them in december, and that’s it
I am first going to finish my DB & Dashboard, then will see what to do with the remaining 6 virgin Nodes on the shelf (which, I suspect, are not A Nodes)
today I expect the Pololu A-Star 328PB… an arduino pro mini look a like, but with the new MICROCHIP ‘extended’ 328PB processor.
I have some use case ideas and also expect some problems, because its not yet fully suported in the arduino IDE
Pololu A-Star 328PB
Pololu A-Star 328PB vs Arduino Pro Mini
seems that there’s a new rn2483A firmware update coming
support and updating process for RF explorer products is excellent !
still need to test further the RAK magic stick 8 dbm 868 antenna.
first things first… installing and updating the spectrum analyzer/ rf generator and rf explorer.
updates under win 10 … pfffffffff ,seems that my second screen don’t work anymore on my laptop
worked fine under ubuntu…
still having a problem with lowpower on this board.
can you show some code how you get it that low ?
Sure BoRRoZ,
I originally used some code from here - Adafruit Feather32u4 with Lora
With a couple of additions:
#include "LowPower.h"
...
void loop()
{
...
// Sleep radio
rf95.sleep();
//USBDevice.detach(); doesn't work on 32u4
// Disable USB clock
USBCON |= _BV(FRZCLK);
// Disable USB PLL
PLLCSR &= ~_BV(PLLE);
// Disable USB
USBCON &= ~_BV(USBE);
// ATmega32U4 sleep
LowPower.powerDown(SLEEP_8S, ADC_OFF, BOD_OFF);
// Reattach USB to have Serial working again
USBDevice.attach();
}
Found in RocketScream forum here. The most important bit being disabling the USB, which saved a lot.
I’ve also had some attempts at removing the voltage regulator and bridging the 5V/3.3V pins. (Note: that you really need to program via ISP after this)
Here are the results with the Adafruit code above:
I have been able to replicate the same power usage with LMIC lib. The only additions I’ve made to the TTN ABP example code is changing pins, adding LowPower lib sleep and disabling the USB with same code as above.
Hope this helps.
BoRRoZ,
You wouldn’t be able to explain a little more on how you used the dip switch interrupt timer? I haven’t been able to get very far with them. Thanks.
sure (and tnx for your info) , but the external timer prototype is … uh gone, this is what I have left
these are the two I’ve tested, the one without the dipswitches is cheap, but not very accurate and I had to buy a seperate switch
the whole idea was to test if a customer/installer could set a node interval from the outside, but there are better , more accurate, ways over the network
where is that coming from… which lib ? or can you post the contents of that routine.
I am using the HPD13A btw from BSFRANCE LoRa32U4II and using the code below
what am I missing… lowest consumption I get with this code is
''c++
// --------------------------------------------------------------
// basic sleep test 2
// BSFRANCE LoRa32U4II - HPD13A
// --------------------------------------------------------------
#include “LowPower.h”
void setup()
{
pinMode(LED_BUILTIN, OUTPUT);
}
void loop()
{
USBCON |= _BV(FRZCLK);
PLLCSR &= ~_BV(PLLE);
USBCON &= ~_BV(USBE);
LowPower.powerDown(SLEEP_8S, ADC_OFF, BOD_OFF);
USBDevice.attach();
digitalWrite(LED_BUILTIN, HIGH); // i’m awake
delay(3000);
digitalWrite(LED_BUILTIN, LOW); // back to sleep
}
