Hardware

@thecricketer some ideas here https://www.youtube.com/watch?v=FqtNsIKpZAw

@thecricketer There are so many things to consider when selecting components but for us hobby-tinkerers there are a few that are most important e.g. Max input voltage: Vinmax Max output current: Iout Size or package Price You will find all these answers in the data sheets. Normally available from your components supplier or else on the internet. Below is a simple table: Type Vinmax Iout Package Price (SEK, single units) LD1117 15 V 800 mA TO220 5:50 LE33 20 V 100 mA TO92 9:- LP2950 30 V 100 mA TO92 9:- LM317 40V delta 1500 mA TO220 9:- (There is also an 100 mA version for 6:25) If you are using the Sundberg newbiePCB the design is for an TO92 package so either LE33 or LP2950 is the choice. The LD1177 is TO220 package and you do not need 800 mA for the radio. It only consumes about 30 mA when transmitting. The LM317 is an adjustable regulator so it can regulate to 5 V or 3.3 V or whatever you like but it takes two more resistors to set the voltage. In this case it is overkill and does not fit the Sundberg board. If you plan to run your design on batteries I think the LP2950 is better than the LE33 because the LP2950 consumes less power internally, only 75 microamps. The LE33 can use as much as 500 microamps internally. This is just scratching the surface of component choice but keep it simple in the beginning. Happy tinkering!

@korttoma Adafruit library. It communicates over UART.

https://forum.mysensors.org/topic/6961/nrf5-action/1659 That fixed it.

@tsjoender thanks I will try your advice Dick

@vladimir there is an example for using rotary encoder on this page: https://www.mysensors.org/build/dimmer

@otto001 Hello! Did you manage to implement this node?

![Thanks for your replies. Actually, the problem is to trigger the pulse of the relay with the function "delay". It is working but sometimes during the "delay" function, the gateway has not received the information of the first pulse yet. So, the second pulse "digital write..." will not work. I will have to press several times. It is a coding optimization i think. For the status i have used a ACS712 sensor to know if the light is on or off. i attached for you the electrical diagram of the "standard" impulse relay on a french house installation which is located in the main board in the garage. Push button are connected. This relay is trigger with 2 quick pulses when you press the push button, then the light will turn on or off. To insert Mysensors project to this original electricall installation i directly connected the mysensors Node with an NO NC relay directly on the same wires than the push button. From the gateway, i trigger the relay with 2 pulses. [image: 1535352299463-telerupteur.jpg]

@petere the INA219 readings are erratic because the current is erratic. When the PWM pulse is off, there will be roughly 0 current (but negative right after the cutoff). When the pulse is on, the current will be the same as if you didn't use PWM. This will change about 500 times per second (on a normal Arduino PWM, not sure if the Ruideng uses a different frequency), so INA219 will read different current at different times. See if setting a longer sample time helps. Not sure which library you are using but the Adafruit library defaults to 532us sample time. The longest sample time for INA219 is 69ms which should give you readings that are stable enough. EDIT: I see now that you already got a similar (but slightly less detailed) answer 5 days ago at http://forum.arduino.cc/index.php?topic=564517.msg3846445#msg3846445

Looking at it more closely, I noticed that TI's design is meant to is to attach directly to the optical chamber of the Sharp Microelectronics optical dust sensor (part number: GP2Y1010AU0F).

@marceltrapman You could desolder the LEDs and solder some wires or even a header to the pads. Then regular LEDs can be connected to the wires/header. Will require a steady hand though...

@alowhum Now I remembered that I once shot a video about this sensor. In the description below the video there are links to sketches that work with such sensors. But for this video I did not do subtitles in English, in sketches comments also seem to be in Russian. But I think it won't be a problem. https://www.youtube.com/watch?v=Q0sxeT0VsVg https://www.youtube.com/watch?v=NdrU0byfbvI

@hakha4 I take it back, the sketches work. The problem was solved as follows: connected the radio to the output of nodemcu 3.3 V. before that, the radio was on a separate power line. I think the esp32 is very power sensitive, I think it was a level mismatch on the SPI pins. Try it, I think you have the same problem.

@alexsh1 current draw details in sketch header. 10-20 uA I had some problems with recovery after gateway power cycling. See discussion here. This issue has been fixed with a workaround.

@alowhum use --spi-spidev-device=<DEVICE> to use the other spi. Documentation: https://www.mysensors.org/apidocs/group__RaspberryPiGateway.html

@rozpruwacz I guess it shouldn't. Maybe if one incoming byte is treated as two, but the kernel should still be able to process that I think. It would decode the byte incorrectly, but it shouldn't cause an overrun.

@raptorjr If you can figure out what the filter differential pressure should be when requiring serviced (perhaps from the filter manufacturer or a manual?), it should be possible to find a pressure switch which triggers alarm on the supply side of the filter. If I recall correctly, some water supply pump pressure switch controllers can be turned down to trigger between 5 and 30 psi, perhaps an old one can be found and re-purposed if it fits the range you are aiming for, however fugly they are...

I noticed 2 suppliers that sell shielded versions of the small modules for about $2,60 (including shipping) https://www.aliexpress.com/item/Wireless-transceiver-digital-RF-module-Full-original-device-nRF24L01-2-4G-shield-wholesale/32842485649.html