Looks nice in red ;)

Looks like a neat module indeed.
Never tried these, but yes I guess if the external antenna and the wifi antenna are too close it may leat to some interferences.

Would also need to check that the 3.3V is powerful enough to properly feed the esp and this chip at the same time, I guess it may depends on the antenna plugged to it too.

@fabien Never heard of the "Ebyte NRF24 PA+LNA module", do you have a link for it somewhere?

So far I've been using a few of these with the "regular no name" PA+LNA from Aliexpress and did not have any problems. What are the symptoms you are observing?

Forgot about this guy... Updated the project a little, the PCB now have footprints for both normal and PA+LNA modules.
The old boards are still available on the github repository if needed.

Even with the extra length added to the PCB with the dual footprint, it is overall still smaller than a regular non SMD NRF24L01 module.

Actually was not too bad to update.

I pushed the updated example sketch to github and openhardware, compatible with IRremoteESP8266 v2.0 and up.
I did a quick test here, no errors, and the vacuum cleaner already escaped to the bedroom, so good to go!

As for the board, feel free to share any sketch here or on github.

Short and not perfect answer, they changed everything when they updated the library to >2.0.

For a quick way to test your board etc, you can compile the example sketch by reverting to version 1.2.0

Ideally you may want to upgrade your sketch to be compatible with version 2.0 according to this https://github.com/markszabo/IRremoteESP8266/wiki/Upgrading-to-v2.0

I wanted to do it, but completely forgot about it... Will try to have a look at it tonight, but hopefully you can start testing everything using the old library version.

May also ask what minor modifications you did? Always open to suggestions to improve the board.

@Mynasru unfortunately the only other suitable pin would be D1, SCL.
Wemos do not have enough pins to do gateway + I2C, you will need to switch to a nodeMCU or similar.

Also keep in mind that D2 is a MySensor default for ESP8266, I tried to have it moved when support was added https://github.com/mysensors/MySensors/issues/408#issuecomment-191403862 but without success, meaning you will need to reassign CE in your code.

@bilbolodz Actually a PA+LNA version is on github (untested)
https://github.com/emc2cube/SMDnRF24_adapter/tree/master/PA_LNA

But yes, I did not update the openhardware part yet.

The choice of MOSFET can be tricky. Seems that irlz44n was discontinued so you may need to find one adapted to the voltage and current you want to handle.
In all the cases you need to get a logic MOSFET too, meaning that they are fully open usually around 1 to 3V.

Also I'm sort of confused as it seems to me that L7812CV is a Voltage regulator, not a MOSFET. So if you are actually using that in the MOSFET spot it will definitively not work.

Especially if you are handling high current or voltage I would recommend using one from a reputable source for your MOSFET, I recently switched from mouser / digikey / aliexpress to mostly use http://www.arrow.com as you get free regular shipping, event if you order a couple of components (no affiliation to them whatsoever).

@reinhold I bet it's an IRQ issue:

If you are using version 1.1 of the board, with IRQ support, you need to compile the software on the pi with the --my-rf24-irq-pin=18 option (see Install section on openhardware / github)

By default IRQ is on pin 15 so my bet is that if it's working by following https://www.mysensors.org/build/raspberry you just need to recompile the software with this option and it should work

@scalz said in 💬 HALO : ESP32 multi transport GW/Bridge for Mysensors:

So far (but not digged a lot) i've only seen PA smd modules with ipex (but would be better if i could fit sma) or chip antenna (lol). Are you thinking to an other reference, or have a favorite module??

This work quite well with the IPX PA+LNA SMD modules

Congratulations to everyone, and a big thanks to PCBWay!

@Nca78 Not sure either to be honest, but technically I only did the calculation for a specific type of LED, the ones in the BOM.

I do use TSAL LEDs, and used http://ledcalc.com/#calc to do the resistor calculation for 7 LEDs with numbers extracted from the datasheet, leading to the ~5ohm resistor calculation.
I tried without any, or with a 4.7ohm resistor and I had significant better range results without resistor. Not a surprise, looking at the IR signal with a smartphone camera you can definitively see a brightness difference too, that's really a blaster :)

No burned LED after 48h of testing (sending one of my IR code every 5s). I also tried, for 10min, with regular red LED (http://www.nteinc.com/specs/3000to3099/pdf/nte3019.pdf) which have lower Vf and If requirements, no problems either.

I don't say that this is right, I truly don't have the background to do more than empirical experimentations, but it seemed good enough for me at the time. I did not had any problems since then, both with real TSAL from mouser and probably fake ones from aliexpress.

I hope you will like this board when you receive it!

True, no current limiting resistor in that configuration for IR LEDs. One of my previous version had one and IR signal was quite dim. By having 7 LEDs the current in each one is actually quite low, and they are only ON for a few micro seconds. I have been using this specific configuration for 3 months now (breadboard to various prototypes) and no burned LEDs so far.

It may differ for other IR LEDs, but TSAL series is pretty resilient and by using 7 LEDs it can take up a few amps without any problem. (If you do the calculations you will find that you could add a 5ohm resistor to limit to 100mA, or 2.5ohm for 200mA, so as you can see it's safe to skip)

Even using ONE regular (visible light) LED for a quick test to send a signal is actually ok, but leaving it on for more than 5s is sufficient to burn it. Having all seven LEDs, they were still working after 1 minute being ON, which is already far from any real case use.

Thanks @hek , yes I did saw it. I did try to port it to work on ESP8266 with IRremoteESP8266.h but without success so far.

That's totally why my first point is "be able to record, save and reuse an IR code without editing the code. Plenty of EEPROM space to store data", which is what this sketch was able to do. I have the feeling that someone that know what he is doing can probably do it easily, but for me it's a nightmare.

Arduino sketch is really basic and even if it works perfectly for me, there is probably a way to improve it greatly.

On top of my head it would be good to

• be able to record, save and reuse an IR code without editing the code. Plenty of EEPROM space to store data.
• have a webpage showing the serial console to see received IR codes, or other info.
• have a dynamic webpage automatically displaying all codes stored in such EEPROM.
• have more than 10 codes that can be used with domoticz.

But I do not know how to do this (even if it's possible).
If you are interested in improving this code shoot me a private message, or answer here. I have a few extra PCBs from a previous version that I will gladly send for free to anyone willing to improve this device software.

Was modified in 2.1.1, you need to change

metric = getConfig().isMetric;

to

metric = getControllerConfig().isMetric;  // was getConfig().isMetric; before MySensors v2.1.1

Glad to see you have it working!

Let me know if something was unclear in the instructions or if it can be improved to prevent people to solder them flipped.

Ah yes right.
For non-smd they both have the same pinout, so you should be able to design it using the regular footprint, and then just plug a PA+LNA module in the header.

Not sure. I do not have any RFM sensors, I tested the gateway for a little more than a week with a mock sensor and it was working. Maybe some logs could help?

You can also hook up the RFM shield to a regular arduino nano to do a serial gateway, this way you can check if it's radio dependent or not.