RN2483 problem temperature


#21

I've had a Microchip board with the RN2483 working fine at -28C. Was sat in a freezer for 2 weeks without a hitch outputting it's temperature. It was connected to a laptop via USB for power.


#22

I tested at -20°C indeed RN2483 works flawlessly


#23

Some Lipo don't like low temp, see Felix's post

https://lowpowerlab.com/2015/02/03/chinese-lithium-cells-freezing/


#24

Those are the batteries i'am testing with right now.
From L->R : saft LS 14500 Li-SOCl2 3.6v, GP NiMH 3 x 1.35v, Standard Alkaline 3x 1.5v

I will be testing those batteries with same unit at several temperatures (+50 / -20) for several hours.


#25


This PCB and enclosure will be used.

Live update on Thingspeak -> https://thingspeak.com/channels/113020
(first 3 graphs) Currently testing with standard alkaline, just outside for 1 hour least.


#26

Unit seems to behave quite well. Now moving to the fridge.

Stopped working @ -17 C ??


#27

@veen10
Nice enclosure, where do you get these nice waterproof connector ?


#28

You can get them here:
https://www.conrad.nl/nl/brandlist/weipu.html?category=%1FElektromechanica%1FConnectoren

For prototyping, they are pretty good. For the real thing i would use better connectors with full metal housing like Binder or Amphenol.


#29

I was able to reset the unit, via the spare connector i use for those purposes.
It is working again. I'll now put only the sensor in the fridge, to exclude any problems on the sensor side.

I'am using Sensirion SHT21 for humidity & temperature. This sensor has also an operating range down to -40 C. The unit is working over I2C communications, so i want to test also communication problems due to low temps with this sensor.


#30

When you look at the graph, battery voltage drops from 4735mv to 4432mv at the moment of signal loss. Maybe this gives an indication for the faillure. Still glad that i have te opportunity to dive into this issue, cause when units are out in the field, cost for replace and reputation damage, will be much higher.

Temperature is now -20 C + , so we passed the critical point. Lets wait for at least an hour, to be sure there are no problems with this sensor.

Battery is slowly recovering, but voltage level is still much lower than normal. Temperature outside the fridge will be around +10 C.


#31

when you look @

It's obvious that the battery never recovered to the starting value (around 4.7v), after the -17 C failure. It's also clear that the sensor is working correct at low temps, at least until -23 C. Next i will change the battery (now cheap alkaline type) to the saft LS14500. This battery is rated for -60C, so when there are problems @ -17C, it's definitely not the battery.


#32

wich battery are you talking about now ? AH.. that were the alkaline battery's I see now

and.. could it be the sensor (you have a spare?)


#33

Morning, yes that's correct. i'll now test with the saft (before it was cheap alkaline).
I do have a spare sensor, but as said before, the sensor is not the problem. I putted only the sensor in the fridge and lora device outside. Sensor worked down to -23C. (unfortunately fridge is not going any lower ).


#34

morning , so:

  • the processor (ESP?) + rn2483 are outside the fridge and the sensor is inside

can you power your setup with a powersupply near the fridge instead of the batteries.
you know that esp uses a lot of power to connect to your router (hence dropping voltage hence disconnected)


#35

Yes , that was the second measurement i did yesterday. The first measurement failed and it was with the sensor and the box all in the fridge. (failed @ -17C)

It's not an ESP , processor is the Feather M0 from Adafruit. Lora PCB is lorawan-featherwing from Dan.

The interface board is designed by me. It has power management on it (unit sleep will be < 26uA) and all I/O is ESD protected / filtered.


#36

and the code ?

how do you get your temp to thingspeak ?


#37

The main code, i wrote myself. Of course i use common libraries of others too. The main goal for this unit was to get the power down as low as posible. This greatly increases the usability of the unit. Design consideration was also, that it needed to be modular. So all pcb's can be used independently. When you don't use GPS, you can simply leave the unit out. You can also put in another Feather, fi the wifi or bluetooth variant. Look for more info on the Adafruit site. https://www.adafruit.com/product/2772


#38

Via Node RED & MQTT.


#39

Ok, unit is equipped with new battery Saft LS 14500. I will put it in the fridge now.


#40

but you are testing different batteries in the fridge, because you suspect - the batteries fail @ certain negative temps (true for lipo) or the rn2483 fails @ certain temp.

other possibilies are maybe

  • bug in (connection) code - again,how do you connect to thingspeak ? did you change singlegatewaycode ?
  • bug in hardware
  • the 2483 uart doesn't connect after sleep .. timings
  • voltage is not stable enough when coming out of sleep and transmitting (test with external supply)

I am not working with rn2483 for the last couple of months (switched to LAIRD rm186) but had a quick look and I'll have some parts.. it seems a recurring problem, hope to find time to at least test the rn2483 myself in the fridge :slight_smile: