💬 MySensors Nrf24/RFM Serial/Ethernet GW Shield (for Uno)

[sundberg84]

@Nick-Willis @Nick-Willis - Im also trying out the RFM version, but not finished yet. I need to run the signals through a signal analyser, Either i messed some soldering up or something else i up with this... Im getting that MySensors does not recognize the radio, but I have to use softspi and not sure yet how this works with MySensors and Rfm.
R1-R3 (led current limiting resistor = 330 - 1k
R4 - R13 = 10k pullups
C1 = 0,1uF, C2 = 10uF C3= Nrf24l01+ resistor, 4.7µ - 47µF

[Nick Willis]

@sundberg84 Thanks for the update and information.
The soft spi with RFM issue is a concern. Have you tried using the new RFM driver instead of the default?
https://www.mysensors.org/apidocs/group__SoftSpiSettingGrpPub.html

[sundberg84]

@Nick-Willis - yes I did, and when I used the new driver everything compiled and the node started, but I got a strange serial log message saying radio = 0 (not radio failed) or something.

[hlehoux]

@sundberg84 Thank you
My knowledge being very limited, i understand this might be difficult for my to go on the RFM69 version ?

Hopefully, my nrf24 network based on your designs works well :-)

[hlehoux]

Thanks to @sundberg84

My Nrf24 gateaway is know quasi professionnal !

Can't wait to test the RFM69 version

[Nick Willis]

@sundberg84 I've now built both the NRF gateway (working perfectly) and the RFM gateway but not working for the RFM one.
I get this in my serial output:

0 MCO:BGN:INIT GW,CP=RPNGA---,FQ=16,REL=255,VER=2.3.2
55 MCO:BGN:BFR
76 TSM:INIT
89 TSF:WUR:MS=0
105 RFM69:INIT
121 RFM69:INIT:PIN,CS=6,IQP=2,IQN=1
3230 RFM69:PTX:LEVEL=5 dBm
3259 !RFM69:INIT:SANCHK FAIL
3289 !TSM:INIT:TSP FAIL
3314 TSM:FAIL:CNT=1
3336 TSM:FAIL:DIS
3354 TSF:TDI:TSL
3372 RFM69:RSL

And this is the important bit of my sketch. I'm using it as an MQTT Ethernet gateway:

#define MY_RADIO_RFM69
#define MY_RFM69_NEW_DRIVER
#define MY_IS_RFM69HW
#define MY_RFM69_FREQUENCY RFM69_868MHZ

#define MY_W5100_SPI_EN 4 

#define MY_SOFTSPI
#define MY_SOFT_SPI_SCK_PIN 14
#define MY_SOFT_SPI_MISO_PIN 16
#define MY_SOFT_SPI_MOSI_PIN 15

#define MY_RFM69_CS_PIN 6

#define MY_RFM69_IRQ_PIN 2
#define MY_RFM69_IRQ_NUM 1

#define MY_GATEWAY_W5100

Is this the same problem you get too?

[sundberg84]

@Nick-Willis - I have not had the time yet to investigate this. It is not the same serial output as you (and im not using an ethernet shield for my RFM gateway). It seems in your serial output that the radio cant be initialised at all (no contact).

Thi is my code:

include <SPI.h>

// Enable debug prints to serial monitor
#define MY_DEBUG

// Enable and select radio type attached
#define MY_RFM69_NEW_DRIVER
#define MY_RFM69_FREQUENCY RFM69_433MHZ // Set your frequency here

#define MY_SOFTSPI
#define MY_SOFT_SPI_SCK_PIN 14
#define MY_SOFT_SPI_MOSI_PIN 15
#define MY_SOFT_SPI_MISO_PIN 16

#ifndef MY_RF24_CE_PIN
#define MY_RF24_CE_PIN 5
#endif
#ifndef MY_RF24_CS_PIN
#define MY_RF24_CS_PIN 6
#endif

// Enable serial gateway
#define MY_GATEWAY_SERIAL

#include <MySensors.h>

void setup()
{
    // Setup locally attached sensors
}

void presentation()
{
    // Present locally attached sensors
}

void loop()
{
    // Send locally attached sensor data here
}

[sundberg84]

So, today I have been making some test for RFM radio + this shield.
With the new driver, it seems ok - but I suspect TSP=NA isnt a good sign.
This happens both with SoftSPI with and without the shield attached.

0;255;3;0;9;0 MCO:BGN:INIT GW,CP=R-NGAA--,FQ=16,REL=255,VER=2.3.2
0;255;3;0;14;Gateway startup complete.
0;255;0;0;17;2.3.2
0;255;3;0;9;359 MCO:BGN:STP
0;255;3;0;9;365 MCO:REG:NOT NEEDED
0;255;3;0;9;368 MCO:BGN:INIT OK,TSP=NA

With a logic analyser, I cant seen any activity at all from the UNO on the defined softSPI pins. All I see here is some activity on RX and some on A3 (for signing - atsha chip)

By removing softSPI from the code, using default pins (and bridge the uno to my shield with dupont cables) i get exactly the same as above, TSP=NA

Could it be that SoftSPI uses analog pins, and im using logic level converters on the shield...

By just replacing the new driver with //#define MY_RADIO_RFM69 i get a more correct looking log:

0;255;3;0;9;0 MCO:BGN:INIT GW,CP=RRNGAA--,FQ=16,REL=255,VER=2.3.2
0;255;3;0;9;108 TSM:INIT
0;255;3;0;9;111 TSF:WUR:MS=0
0;255;3;0;9;114 TSM:INIT:TSP OK
0;255;3;0;9;117 TSM:INIT:GW MODE
0;255;3;0;9;120 TSM:READY:ID=0,PAR=0,DIS=0
0;255;3;0;9;123 MCO:REG:NOT NEEDED
0;255;3;0;14;Gateway startup complete.
0;255;0;0;18;2.3.2
0;255;3;0;9;129 MCO:BGN:STP
0;255;3;0;9;135 MCO:BGN:INIT OK,TSP=1
0;255;3;0;9;139 TSM:READY:NWD REQ
0;255;3;0;9;145 ?TSF:MSG:SEND,0-0-255-255,s=255,c=3,t=20,pt=0,l=0,sg=0,ft=0,st=OK:

I can also see things happening all over all SPI logic channels.

So whats up with the new driver? Did i not define it right?

// Enable and select radio type attached
#define MY_RFM69_NEW_DRIVER
#define MY_RFM69_FREQUENCY RFM69_433MHZ // Set your frequency here

Or is it using other than default pins?

[Nick Willis]

@sundberg84 I'm also still having problems with this.
I got the same result if I define only MY_RFM69_NEW_DRIVER then TSP=NA so no transport. I think the MY_RFM69_NEW_DRIVER option is not defining the radio type but is just an option (like the frequency setting) so you need for the old driver:

#define MY_RADIO_RFM69

And for the new driver - both options set:

#define MY_RADIO_RFM69
#define MY_RFM69_NEW_DRIVER

However it still doesn't work with SoftSPI. I've also used dupont cables and wired directly to the hardware SPI pins on my Mega then removed all the SoftSPI defines. I'm also trying a simple serial gateway as follows:

// Enable debug prints to serial monitor
#define MY_DEBUG
#define MY_DEBUG_VERBOSE_RFM69

// Enable and select radio type attached
#define MY_RADIO_RFM69
#define MY_RFM69_NEW_DRIVER
#define MY_RFM69_FREQUENCY RFM69_868MHZ // Set your frequency here

// Enable serial gateway
#define MY_GATEWAY_SERIAL

/*
#define MY_SOFTSPI
#define MY_SOFT_SPI_SCK_PIN 14
#define MY_SOFT_SPI_MOSI_PIN 15
#define MY_SOFT_SPI_MISO_PIN 16

#ifndef MY_RF24_CE_PIN
#define MY_RF24_CE_PIN 5
#endif
#ifndef MY_RF24_CS_PIN
#define MY_RF24_CS_PIN 6
#endif
*/

#include <MySensors.h>

void setup()
{
    // Setup locally attached sensors
}

void presentation()
{
    // Present locally attached sensors
}

void loop()
{
    // Send locally attached sensor data here
}

My log:

0;255;3;0;9;0 MCO:BGN:INIT GW,CP=RPNGA---,FQ=16,REL=255,VER=2.3.2
0;255;3;0;9;5 TSM:INIT
0;255;3;0;9;7 TSF:WUR:MS=0
0;255;3;0;9;10 RFM69:INIT
0;255;3;0;9;12 RFM69:INIT:PIN,CS=53,IQP=2,IQN=0
0;255;3;0;9;1286 RFM69:PTX:LEVEL=5 dBm
0;255;3;0;9;1289 !RFM69:INIT:SANCHK FAIL
0;255;3;0;9;1293 !TSM:INIT:TSP FAIL
0;255;3;0;9;1296 TSM:FAIL:CNT=1
0;255;3;0;9;1299 TSM:FAIL:DIS
0;255;3;0;9;1302 TSF:TDI:TSL
0;255;3;0;9;1304 RFM69:RSL

So something is wrong. I've checked my soldering and connections and it all looks ok (well as far as I can tell). There was one odd thing though - if I check resistance (on an unsoldered board) between the 3.3v pin on the RFM and the 3.3v on the regulator there is no connection. If I test against other 3.3v locations they are connected.
Looking in KiCAD I see a white line between the two 3.3v rails that are not connected. Is some connection missing here?

I tried soldering a wire to make the connection but still it won't work (same log as above). Really not sure what to try next.

[sundberg84]

You are correct, it should look like this:
I will update this to the final version. Can you comfirm you have 3.3v on the RFM module?

I will try with both #defines - that might work! Thanks for that.

One thing to do (quickfix) if you want to use it as a serial gateway is to cut the pinheaders for all SPI channels. Then hardswire the SPI channels to the normal pins 10-13 instead of softSPI which is currently using. You ofcourse needs to solder them from the pinheader and not RFM module itself since you need to logic level conversion. I will give this a new go as soon as I have time. I really really want a RFM gateway over ethernet.

[Nick Willis]

@sundberg84 Yes I can confirm I see 3.3v on the RFM.
I'd thought about cutting the pin headers+ hard wire for hardware SPI but wanted it to work with the dupont cables first. I've tried a Mega and now an Uno but still it doesn't work. I guess either something is soldered wrong, the radio is broken or something else. Checked it all many many times so perhaps time to replace the radio.

[Nick Willis]

Oops fixed it by remelting all the solder joints on the mini regulators. So will cut the pin headers and hard wire for a serial gateway (which I'm starting to think is a good idea anyway to keep down system latency of going via ethernet).

[sundberg84]

@Nick-Willis - great to hear.
By using dupont cables and "normal pinout" for RFM radio i now also get the new driver to work thanks to your input.

#include <SPI.h>

// Enable debug prints to serial monitor
#define MY_DEBUG

// Enable and select radio type attached

#define MY_RADIO_RFM69
#define MY_RFM69_NEW_DRIVER
#define MY_RFM69_FREQUENCY RFM69_433MHZ // Set your frequency here
#define MY_IS_RFM69HW

//#define MY_SOFTSPI
//#define MY_SOFT_SPI_SCK_PIN A0
//#define MY_SOFT_SPI_MOSI_PIN A1
//#define MY_SOFT_SPI_MISO_PIN A2

//#ifndef MY_RF24_CE_PIN
//#define MY_RF24_CE_PIN 5
//#endif
//#ifndef MY_RF24_CS_PIN
//#define MY_RF24_CS_PIN 6
//#endif

#define MY_SIGNING_ATSHA204
#define MY_SIGNING_ATSHA204_PIN A3
#define MY_SIGNING_REQUEST_SIGNATURES
#define MY_SIGNING_WEAK_SECURITY 

// Enable serial gateway
#define MY_GATEWAY_SERIAL

#include <MySensors.h>

void setup()
{
    // Setup locally attached sensors
}

void presentation()
{
    // Present locally attached sensors
}

void loop()
{
    // Send locally attached sensor data here
}

⸮S[0;255;3;0;9;0 MCO:BGN:INIT GW,CP=RPNGAA--,FQ=16,REL=255,VER=2.3.2
0;255;3;0;9;108 TSM:INIT
0;255;3;0;9;111 TSF:WUR:MS=0
0;255;3;0;9;114 TSM:INIT:TSP OK
0;255;3;0;9;116 TSM:INIT:GW MODE
0;255;3;0;9;119 TSM:READY:ID=0,PAR=0,DIS=0
0;255;3;0;9;123 MCO:REG:NOT NEEDED
0;255;3;0;14;Gateway startup complete.
0;255;0;0;18;2.3.2
0;255;3;0;9;128 MCO:BGN:STP
0;255;3;0;9;135 MCO:BGN:INIT OK,TSP=1
0;255;3;0;9;138 TSM:READY:NWD REQ
0;255;3;0;9;145 ?TSF:MSG:SEND,0-0-255-255,s=255,c=3,t=20,pt=0,l=0,sg=0,ft=0,st=OK:

But with SofSPI engaged - not working-

0;255;3;0;9;0 MCO:BGN:INIT GW,CP=RPNGAA--,FQ=16,REL=255,VER=2.3.2
0;255;3;0;9;0 MCO:BGN:INIT GW,CP=RPNGAA--,FQ=16,REL=255,VER=2.3.2
0;255;3;0;9;277 TSM:INIT
0;255;3;0;9;279 TSF:WUR:MS=0
0;255;3;0;9;282 !TSM:INIT:TSP FAIL
0;255;3;0;9;285 TSM:FAIL:CNT=1
0;255;3;0;9;288 TSM:FAIL:DIS
0;255;3;0;9;290 TSF:TDI:TSL

• Continuity on all logic level converters are OK
• Not much traffic at all from the UNO in SPI. I guess the intial contact with the radio is faulty so thats why.

It must be either connection Uno > Logic Level Shifter > Radio, code / defines or softwareSPI not working.

[sundberg84]

I have now tested all SPI lines and its logic level converter individually.
To be absolutely sure the hardware worked, I desoldered the radio and wrote a arduino program sending 5v in 5 sek over the same sofwareSPI lines and 0V over 5 sek - all behaved correctly, 3.3v and 0V.

I also swapped radio, but same same...

Not really sure how to proceed. I need to get the softwareSPI working or else I cant use it as a ethernet gateway.

Is there someone here that got RFM gateway working with softwareSPI?
I dont think its the hardware anymore...
Code:

#include <SPI.h>

// Enable debug prints to serial monitor
#define MY_DEBUG

// Enable and select radio type attached

#define MY_RADIO_RFM69
#define MY_RFM69_NEW_DRIVER
#define MY_RFM69_FREQUENCY RFM69_433MHZ // Set your frequency here
//#define MY_IS_RFM69HW

#define MY_SOFTSPI
#define MY_SOFT_SPI_SCK_PIN 14
#define MY_SOFT_SPI_MISO_PIN 16
#define MY_SOFT_SPI_MOSI_PIN 15

#define MY_RFM69_CS_PIN 6
#define MY_RFM69_IRQ_PIN 2

#define MY_SIGNING_ATSHA204
#define MY_SIGNING_ATSHA204_PIN A3
#define MY_SIGNING_REQUEST_SIGNATURES
#define MY_SIGNING_WEAK_SECURITY 

// Enable serial gateway
#define MY_GATEWAY_SERIAL

#include <MySensors.h>

void setup()
{
    // Setup locally attached sensors
}

void presentation()
{
    // Present locally attached sensors
}

void loop()
{
    // Send locally attached sensor data here
}

[sundberg84]

What im noticing is that when I connect the radio, the IRQ line on D2 is 1.0v when high and not 3.3v. Anyone who knows this is normal behaviour? Same with two different (and comfirmed working) radios. SPI logic level converter is comfirmed working and changed with same behavoir.

Edit > This is the same using normal pins and does not have anything to do with SPI. It works when not using softSPI and IRQ has this low voltage...

[sundberg84]

This post is deleted!

[sundberg84]

Anyone good with SPI channel analysis?

Hardware SPI (Looks great):

Software SPI (Same setup, just changed code and pins).

In the first, I can clearly see MOSI and MISO communicating on 5/6. Enable on 7.
Software SPI does not make any sense.

[hlehoux]

@sundberg84 Hello, coming back to this RFM GW, so i will try it

I have soldered between the two 3.3v rails , i confirm now i have 3.3v on the RFM module

I will test with the softwareSPIs and report.

I will give a try to your "quickfix" for hardwire : could you please explain or post a photo of what needs to be done for the hardwiring ?

Thank you!

[sundberg84]

@hlehoux My quickfix means you cant use the ethernet shield.
All I did was solder with wires to the normal pins and made it a serial gateway.

[hlehoux]

@sundberg84 thank you.

i have ordered the components such as BSS138 so i will test when they arrive.I don't need ethernet