My wio-e5 mini is officially lost in transit, but i did some experiments with RAK3172 breakout boards.
Interestingly when you use correct pin config but with wrong LP/HP define it still barely transmits.
testing across bench yesterday(1.2m) I got:
RSSI -109 at +20dBi and

//rak3172-e(no TXCO)
static const uint32_t rfswitch_pins[] = {PB8,  PC13, RADIOLIB_NC, RADIOLIB_NC, RADIOLIB_NC};
static const Module::RfSwitchMode_t rfswitch_table[] = {
   {STM32WLx::MODE_IDLE,  {LOW,  LOW}},
   {STM32WLx::MODE_RX,    {HIGH, LOW}},
   {STM32WLx::MODE_TX_LP, {LOW,  HIGH}},//CORRECT
   END_OF_MODE_TABLE,
};
bool lp_supported = true;
bool hp_supported = false;

at glance above code should work, this pdf at page 7 clears things up a bit.

RSSI -101 at +20dBi and

//rak3172-e(no TXCO)
static const uint32_t rfswitch_pins[] = {PB8,  PC13, RADIOLIB_NC, RADIOLIB_NC, RADIOLIB_NC};
static const Module::RfSwitchMode_t rfswitch_table[] = {
   {STM32WLx::MODE_IDLE,  {LOW,  LOW}},
   {STM32WLx::MODE_RX,    {HIGH, LOW}},
   {STM32WLx::MODE_TX_HP, {HIGH,  HIGH}}, //INCORRECT
   END_OF_MODE_TABLE,
bool lp_supported = false;
bool hp_supported = true;

RSSI -23 at -3dBi with :

//rak3172-e(no TXCO)
static const uint32_t rfswitch_pins[] = {PB8,  PC13, RADIOLIB_NC, RADIOLIB_NC, RADIOLIB_NC};
static const Module::RfSwitchMode_t rfswitch_table[] = {
   {STM32WLx::MODE_IDLE,  {LOW,  LOW}},
   {STM32WLx::MODE_RX,    {HIGH, LOW}},
   {STM32WLx::MODE_TX_HP, {LOW,  HIGH}}, //CORECT
   END_OF_MODE_TABLE,
bool lp_supported = false;
bool hp_supported = true;

RSSI -60 at -3dBi and antennas removed, just SMA sockets on PCBs

looking at stm32wl datasheets(stm32 in general) it should be possible to write MySensors compatible bootloader capable of rewriting entire remaining flash(similar to nrf24(l)01 bootloder), or do dual bank firmware upgrades.

P.S it seems that RLSX126x driver from your repo uses wrong mux pins for wio-e5-le, that would explain abysmal range you're getting:

#if defined(WIOE5LE)
// Set RF switch configuration for Seeed WIO-E5-LE module
// E5-LE is always Low Power Transmit. Antenna setting PA4 1; PA5 0 = receive. PA4 0; PA5 1 = transmit.
static const uint32_t rfswitch_pins[] = {PA4,  PA5, RADIOLIB_NC, RADIOLIB_NC, RADIOLIB_NC};
static const Module::RfSwitchMode_t rfswitch_table[] = {
	{STM32WLx::MODE_IDLE,  {LOW,  LOW}},
	{STM32WLx::MODE_RX,    {HIGH, LOW}},
	{STM32WLx::MODE_TX_LP, {LOW,  HIGH}},
	END_OF_MODE_TABLE,

everywhere i check for wio-e5-le it shoudl be(edited as i remebered it wrong at time of writing, sorry)

{STM32WLx::MODE_TX_LP, {HIGH,  HIGH}},

my modules didn't arrive yet so i cannot test it myself.

I got about 400m using rfm69w at 433MHz and +14dBm tx using pece of wire as antenna. And that was through at least 3 houses. I would expect LoRa 868/915 MHz to at least match 433MHz FSK range...

ordering wio-e5 for testing - the small small CPU+RF package makes for perfect small nodes retrofits. inside light fittings/sockets/plugs etc... especially if OTA could be done without external memory...

If you're using insulated wire your antennae length will drop, sometimes dramatically, for PVC insulted single core wire a 73mm is a good starting point, then trim down to ~65mm in 0.5-1mm steps... miulticore wire is another box of worms. I Use wire pulled from cat 5/6 and velocity factor seems to be 0.83% in this calculator: https://m0ukd.com/calculators/quarter-wave-ground-plane-antenna-calculator/

@OldSurferDude
yeah, 10 virgin nodes powered up at the same time is a bad idea I seem to remember reading somwewhere here warning aginst that.
If yu have node ID trust issues with Mysensors you should move your address to different locaion as 0x00 is used by Mysensors... just sayin..;)

@OldSurferDude
Yeah I see the special use case here.
but.
A) MySensors stores node ID in epprom without any of your code
B) Sketch without

#define MY_NODE_ID 

will use eeprom stored ID, If eeprom is empty, addres 0x00=0xFF node will request ID to be assigned by gateway and save assigned ID at eeprom address 0x00
C) to retrieve eeprom stored ID you should use MySensors function :

getNodeId();

To put it bluntly your code to read/write eeprom is redudndant as everything it does is natively handled by MySensors.
Don't get me wrong, your code looks tidier than most of my "efforts" but you're reinventing the wheel here, a bit.. ;)
Neat trick with skipping compile before upload by the way.

To replace dead node all you have to do is upload simplest node sketch with definded desired node ID, let it boot to update eeprom, then load your node program using

getNodeId(); 

to retrieve node id if you need it in your code.

Or am I still missing something? :thinking_face:

easiest way would be to add thermostat "sensor2 to youd node and adjust that to desired temperature. But since your node is sleeping you will have to request the set value when it wakes up.
Here is an example of how my power meter was getting last KWH from controller after power cut, I removed irrevelant code for clarity:

loop(){
	if (!daySet) {
		request(dayID, V_KWH);
		wait(3333, C_SET, V_KWH);
	}
	
	if(!nightSet){
		request(nightID, V_KWH);
		wait(3333, C_SET, V_KWH);
	}
}

void receive(const MyMessage &message) {
	
		if(message.getCommand() == C_SET){
			if (message.type == V_KWH) {
				if(message.sensor == dayID){
					dayCount = MeterRatio * message.getFloat();
					dayCountOld = dayCount;
					dayCountWatt =dayCount;
					daySet = true;
				}
				else if (message.sensor == nightID){
					nightCount = MeterRatio * message.getFloat();
					nightCountOld = nightCount;
					nightCountWatt = nightCount;
					nightSet = true;
				}
					
			}
		}
}

It was taking approx 2 seconds wits domoticz(2019 version) to reply to my requests.

You may run into errors with this method as your sleeping node won't send back ACK when you set new offset though.
Some controllers can disable requirement for ACK, i think.
More advanced way would be creating dummy thermostat to set new value(so you would have live and dummy thermostats) and using controller scripts to update the live thermostat on receipt of a temperature, you would have to wait after sending temperature or humidity.

@OldSurferDude
no offence but rtfm ;)

#define MY_NODE_ID 3

is all you need to set predefined node id's, change 3 to any number from 1 to 254, it's worth having a spreadsheet with assigned ID's to avoid doubles, or at least write ID on the node itself.
If you ommit above define Mysensors store ID in eeprom after it gets autoassigned by gateway/controller on first start.

@GLAB Uno-Nrf comms are ok, your node woulnd't get to sending radio packest if radio chip comms weren't ok.
Instead of reflashing working node and gateway try backdating MySensors library in arduino. You should see what version is used in working gateway somewhere in your contrloller, or by connecing yor serial gateway directly to your PC. in case of ethernet/wifi gateway you can alco use MySController from here:
https://www.dropbox.com/scl/fi/prn9p2cnheyteopokszor/MYSController.zip?rlkey=up54kg4lt2blmff66t2jlgeo6&e=1

if your working node/gateway in on 2.00 and your new node works with that, then you can update them all with minimal risk of borking it all.

it seems your node never received any reply, are you setting you radio correctly? I'm not using nrf24l01s but you are probably using pa+lna version witn non pa+lna code...

@OldSurferDude jus request counter states from controller on each reset, worked fine with Domoticz...
or save in eeprom but increase location every write, on power up/reset search eeprom for largest value stored and carry on from that location, one long integer will fit 256 times in atmega328(p) eeprom, is that long enough for you ;) ?

Over 4 years on, and my node ran continuously sending 4 messages every 5 minutes.
It's 700mAh LiFePo4 battery voltage dropped from initial 3.45V to 3.24V, according to LiFePo4 voltge/SoC tables It's just under 40% left.
700*60% = 420mAh. 100 mAH per year. using library 2.40 and blob message I could cut usage to ~25mAh/year, not bad

@NeverDie Linux and antivirus... I say no since only one I can recommend is only available for windows and mac.

@mfalkvidd said in Best password manager?:

@NeverDie for travel, I would say the largest risk is a border search. US does it, so I would suspect Russia does as well. Good guide: https://www.eff.org/document/eff-border-search-pocket-guide

Say whaat? Border officials in US can confiscate my laptop willy nilly? I'm glad I have no plans to travel there, and if I do I'll encrypt the hell out of everything I carry, even my wrist watch will need password to show time ;)

@NeverDie said in Best password manager?:

Here's a case study that examined how 1Password, Dashlane, KeePass and LastPass could leak data: https://www.ise.io/casestudies/password-manager-hacking.

As I said before no password manager is safe on machine crawling with viruses and/or malware
That's where antivirus software comes to play if you skip that there is no escape from being eventually "hacked".

Main advantage of password manager is not reusing passwords for different services. When one gets compromised it's that. One password leaked, one account compromised, rest is safe. Alternative would be pen and paper notepad, reusing one password, or using memorable password combinations all bearing much greater risks than any decent password manager.

IMHO you are getting a bit paranoid, I do agree with BearWithBeard, use as few extensions as practical.
Firefox+adblockplus+kepassxc+cookiecleaner is my "daily driver"
For weird side of internet I use TOR browser.
For testing downloads use VM.
Plus good paid antivirus, so far so good.

@NeverDie

The keylogers I mentioned are either usb plug extensions(hard-ish to spot) or Bluetooth dongles that listen to wireless keyboards, some of them use very weak or no keyboard<->dongle authentication.

Intercepting passwords between browser and website/server is possible but requires:
a: MITM attack => access to local network easy peasy on café WIFI
or
b: DNS poisoning => admin access to ISP infrastructure or local network router.

On top of that stupid/not paying attention user who will ignore lack of SSL/https connection or add exception to accept website certificate signed by ROOT CA not trusted by os/browser.
Or physical access to victim's computer to add own ROOT CA to trusted CA's database, malware can do it too.
Only way to rule user error, malware or physical access out it to fake ROOT CA and sign certificate to dodgy server, but it's not possible without access to some serious brute forcing compute power - we are talking exascale supercomputer for couple of years here.
It is very personal attack, and unless you are VIP you can forget about it anyway. I'm not so sure about forgetting...

No password manger is safe on a machine crawling with malware, Antivirus/antimalware are a must!!!
Although there is a plugin for original keepass that auto-types password out of order to fool potential malware keyloggers. It's called floating panel.

AFAIK all browser password managers use windows user password to encrypt them, and changing password automatically changes encryption key too SIC!!
Considering that one can overwrite windows password or disable it temporarily for covert hack!! in 30 seconds with simple bootable usb stick(windows password unlocker) such form of encryption is more than useless in stolen laptop scenario. And Mac's have passwords reset feature baked into o/s so resetting it is even easier.

From time to time there is malware/loophole in browser that leaks passwords, last one was opera on the news.
NEVER USE BROWSER SAVE PASSWORD FEATURE!

+1 to keepass, store your database on Google drive/Dropbox/nextcloud and secure it with password+ yubikey and you have bulletproof solution. Just remember to have clone youbikey in a safe.
Keepassxc on windows/Linux, keepas2android and keepasium on Android and iPhone respectively.

@mfalkvidd ever heard of hardware keyloggers? You can buy ones that log every keystroke on any wireless keyboard(wired too).

That's why I'm using yibikey and keepass. Even if my master pass leaks out it's useless without youbikey. And stolen/lost yubikey without pass is just a piece of plastic.

I see it this way:
NodeA sends C_REQ of Var1 to NodeB
NodeB processes C_REQ message in receive();
NodeB replies with C_SET of Var1 to NodeA just as controller would
NodeA processes C_SET message in receive();

You can use echo/ACK functionality or not, it doesn't matter.

If you want both nodes to synchronise their variables, then it seems best to:
a)request values on boot
b)On variable change send values without prompt

That way you cut traffic by half compared to request-reply approach.

@alowhum said in House renovation, how to a good electrical system oriented to MySensors?:

What do you mean with running it in a loop? Where could I learn more about that?

wires in the loop:
PSU-----GW----Node1----node2
|____________________________|

Use switching regulators to keep current down.

I made pcb for it it's on the roof of my flat, will dig trough my archives to see if i have kicad project for it.

cough cough , so much dust here in the archives... ;)

@NeverDie SPV1050 is what you are looking for, only Maximum Power Point Tracking IC from all IC's you mentioned/tested in both your topics. And trust me, it makes a whole world of a difference. ST even have online component calculator for matching it to solar panel and battery used. My usage scenario was outdoor rechargeable coin cell battery and calculator solar cell. Only drawback: you have to determine maximum power point of your solar cell.

just a thought, since resets are happening under heavy trafffic:
What is your serial speed? 9600 could lead to buffer overflows under heavy traffic