@gohan these values are just the ones i used and should filter lowpass at about 5.9 kHz.
The ripple on the LM2596s is usually arround 25-50kHz. (Depends on which clone you get. ;))
I'm no electrical engineer. Maybe @AWI can comment if there may be a more suitable component selection.
@ronnyandre The main reason for using a filter would be a crappy/cheap switching power supply.
I needed to use one of these filters on a cheap LM2596s (clone) module to filter the output to a nrf24l01+ pa/lna module.
Unlike @AWI 's solution i just added a second stage after the first stage (that is on the module).
http://blog.blackoise.de/wp-content/uploads/2016/03/lc-filter-schem1.jpg
L1: 3.3µH
C1: 220µF
Worked fine enough to get the nrf24l01+ pa/lna module working.
Short: probably. TrY soft spi for the nrf24
@karl261 with the original nrf24l01+ modules this is correct, yes.
To save battery use the lowest setting that still works without transmitting errors.
@karl261 the MAX Setting is the HIGH setting + a extra bit. The extra bit is ignored by the original nrf24l01+ chips. Some nrf24l01+ clones (also known as the SI24R1) are using this bit to switch into a higher mode.
I wonder if i ever solved this. 
I remember i began drawing a schematic in KiCAD. Maybe i should look if i will find this and which mosfet i used in the schematic.
I like the magic blue smoke hole directly above the cap on your RPi case. XD
@AWI nice finnish 
I wonder if the nrf24 shielding is still necessary in a metal enclosure.
@alexsh1 is this now officially called "the ugly fix"? 
I should reconsider my naming choices.
@Mark-Swift does "appear" mean you actually measured it all the way to ground or that you believe that it should be well grounded?
@Mark-Swift there are basicly effects that could occur while you holding the module:
1.) You are forming a very small capacitor between you and the module.
Much unlikely to be the reason.
2.) You are functioning as an antenna for the module.
Also much unlikely, you would probably bring more noise then signal in the system.
3.) You function as a shield to the module.
More likely. I got the same behavior with insufficient shielding.
The shielding works ungrounded in low noise environment but needs to be grounded in a high noise environment.
This happens, as example, if a insufficient shielded switching power supply is nearby.
Please make absolutely sure that your shield is properly grounded!
@AWI there also 2 versions of the PA/LNA Modules out there. Which one do you tested?
hackaday caught wind on the tinfoil method.
http://hackaday.com/2016/05/31/fixing-the-terrible-range-of-your-cheap-nrf24l01-palna-module/
Maybe there also some points hidden in the comments that would help?
@Mark-Swift well.. maybe you should give the module a little bit more current. 
I don't know which regulator the Uno uses but i would assume that it don't deliver enough current.
About the grounding.. better be sure, take your multimeter and check if the shielding has continuity to GND.
@AWI ah okay... well.. i'm still unsure if and which modules need the lower 3.3V level on CE.
It just happens that i only use 3.3V arduinos.. and the raspberry pi, which also has 3.3V logic.
@Mark-Swift and @AWI i never worked with the adapter board, but doesn't shift the adapter board also the signal levels down?
@Mark-Swift need to hold the module is often a sign for non solid shielding or the shield is not grounded.
Out of courosity what power supply are you using in front of the adapter board?
@Xyee-Cheah please post it on github or something similiar. This is unreadable. :S
@parachutesj as Jokgi said, you can't measure correctly the current of the modules with just a multimeter, you are losing peaks in the process.
Did you tried yet how far the shielded versions of the module reach? Would be great to have some sort of comparision.
@parachutesj i'm reading pretty much the same values on my multimeter, so nothing to worry about. 
@Jokgi it may be true that the ground filling seems to be quite good, but it seems like it needs also be shielded.
Without the shielding the module is influenced by its own signals and external signals.
There are also modules out there that have actually a shield added: http://www.aliexpress.com/item/Shielding-Case-10pcs-lot-nRF24L01-PA-LNA-wireless-communication-modules-with-antenna-2-4GHz-2Mbps-1000m/851058052.html
But i have not tested them yet.
By quantitative numbers you mean range?
On max transmitting power i don't got even 5meters because the PA TX seems to overload the RX and ack will never be received.
With the shielding i got arround 1000 meters free line of sight.
Both also delivered in their own post-car with security guards? 
@Jokgi hard to say to which module excatly it belongs.. would need to track the traces on the pcb to be 100% sure.
Isn't CE pulled down by the MCU while not transmitting or receiving? I do not have really any problems with the module going power down.
@parachutesj Well i got most problems with this sort of module, if you have the one pictured.
They tended to not run at all without adding a grounded shield, add the LC-Filer and maked sure to run the signals with 3.3V.
@Jokgi: https://www.squirrel-labs.net/wp/wp-content/uploads/2015/02/nRFa.jpg
I think that is what most of the boards are doing. But i also seen schematics in wich the 1kOhm resistor isn't there.
So there could maybe be problems if you try to control the nrf24l01+ PA/LNA with a 5V Arduino because the PA can't handle the 5V signal.
@parachutesj you probably don't benefit from a LC-Filter that much, because you are already using a linear regulator (LD1117 ).
Make sure that you used the shielding right and grounded it.
@parachutesj: expect the RF24_PA_MAX no. I need to say that i still use the 1.5 library and set the PA_MAX setting in the MyConfig.h.
Hi @parachutesj,
you should test it with a loop sending data. It has the highest consumption while sending data.
Seeing your PA/LNA you should also shield it, to extend range.
@hek thanks!
I'm really happy with this article and this is great but i also hoped for a little more constructive criticism on the project.
@drock1985 would be great!
tweet tweet
And now finally part 5 online:
http://blog.blackoise.de/2016/05/building-a-connected-nesting-box-network-part-5/
Which finally includes links to the actual pictures the network takes.
Hello and welcome @xydix,
the back of the board is to mark the actual model you have, it is often not marked by the factory.
The description clearly says that it is the 3.3V 8Mhz Version.
The VCC pin is to directly feed the microcontroller and should be regulated to 3.3V.
The RAW pin runs into the oboard voltage regulator and can be feed 3.3V - 12V (Be aware that most china clones don't withstand the full 12V)
@signal15 probably talking about http://blog.blackoise.de/2016/02/fixing-your-cheap-nrf24l01-palna-module/?
I've not written it, but the foil touches the ground on the antenna socket. So yes, it needs to be grounded.
In most cases it even works ungrounded though. Even a ungrounded shield has an effect on the circuit.
Part 4 posted: http://blog.blackoise.de/2016/04/building-a-connected-nesting-box-network-part-4/
Now with 90% less text and 60% more pictures!
After adding a capacitor AND changing the Arduino to a Nano it got to life.
Maybe it didn't liked the signals from the UNO. You would maybe see something on a oszi but it's probably not worth the hassle.
Other guess: 5V Uno 3.3V nano?
There are also many topics on this forum how to get the PA/LNA module in a working state, can be quite tricky and frustrating.
Also contra: no signing
@mfalkvidd said:
Doesn't the OTA stuff use som sort of stream concept? Maybe that is a good solution for your data as well?
I don't know about the MYSBootloader 1.3 but the old one was more like a request <-> answer thing of blocks, this is fine but not very fast and not very "streamy" ^^
@mfalkvidd said:
Nice. Not fantastic, but a pretty ok speedup.
Well it was enough to have a reasonable transfer time. There are still more things someone could do to speed things up.
By example: Disable the auto-ack and send more then 24bytes at a time. (Keeping radio on and send multiple packets)
But thats all a little bit more complicated with the mysensors lib.
My memory says: 5kb/s without modification and arround 20kb/s with modification.
In my nesting-box project i have the case that i need to send a lot of data over the nrf24-link.
The normal mysensors way of sending the data would be to define a message object outside the sending loop and then set the payload in the loop followed by a send.
The actual "problem" with this is, that a call to send will each time set the last, sender, destination, version, command and requestack.
To speed up i declared to methods "prepareSend" which sets all the stuff to the message object and a "send" method that just takes the message object and straight call radio.send.
Pro: Properties of Message can be set one time -> faster loop
Contra: Routing changes etc. will have no affect in the loop.
What do other developers think of this, and is it maybe worthy to also implement it for real?
@tekka first off: great work!
great.
Any number, test cases on that? In which situations does that apply?
Also great.
But as far as i see some features needed also to go away?
@Mark-Swift let us know how they perform. The supplier looks very professional.
Would be interesting if they really reach the promised 2.1km
(I wonder with which antenna)
even better to use 3.3v (hence on my next one to play it safe I'll use a spare 5v -> 3.3v level shift converter too).
Or use a 3.3V pro mini.
@ted you are probably searching for this: https://github.com/Oitzu/mysensors-Arduino-attiny/tree/master/libraries/MySensors
But i need to seriously warn you, it is horribly outdated. The repository do not got the love it needed. ^^
Oh i see its now on hackaday also. http://hackaday.com/2016/04/17/are-you-in-bed/ 
Also check ce and csn pin configuration correspondence to your wiring.
Maybe post a picture of the build in troublesome state. To get a better understanding of your setup.
@Mark-Swift as everything in your setup seems shielded. Maybe an idea: Could it be the power-source that causing the interference?
A switching regulator sends out a good portion of interferences itself.
@flopp said:
Left is GOOD right is BAD
Just to note: That is a little bit to simple. The left is a linear regulator the right one a switching regulator.
That are different technologies that both have there up and downsides. But for sensors a linear regulator often fits better, yes.
it started to work, even without shield
Never the less i would recommend the shield for even better performance, especially if you want to crank up the pa_level for higher range.
@flopp said:
I have no knowledge about that but I don't think so, because I can turn on a screw to change from ~1.5-4 V
Oh, sounds like a switching regulator to me, but can't be sure. Does it have any name? A switching regulator could produce some significant ripple in your supply, that maybe needs to be filtered out.
Would try the shield first, and then see.
@flopp said:
through a step-down to 3,3.
I suspect linear regulator? Then it should, normally, no problem.
@flopp said:
Yes, different computers both GW and REP is powered by USB.
But same source the pa/lna module or also different?
Add shield try again. Even electronics near to the pa/lna module can induce enough noise that it stops working.
Different power supplys on the ethernet gateway and the uno?
Totally agree with that. I had some contact with him in before about his Esparducam board.
Left him a message, about MySensors.
Maybe something for the guys that use RFM69 as radio and are looking for a very small temp sensor:
http://johan.kanflo.com/the-aaduino/
@sundberg84 in most of my cases it failed due to spi interferences.
@Denke said:
I get alot of strange behaviours and are not able to upload.
You not yet have added the nrf24, or? With the nrf24 module added the upload via isp will not work.
Can you describe any more behaviour? Log output, led flashing...
@moskovskiy82 said:
I am using external chargers 5V for powering my sensors (some are old branded nokia ones, others are a cheap chinese replacement).
Power goes to arduino VCC (arduino pro mini 5V) and also the NRF24 adapter board (containing the voltage regulator and some capacitors - got it off aliexpress).I wonder if adding a decoupling capacitor just before power is split to arduino and NRF24 board will help in any way. If yes what is the best value for the capacitor?
Well, the arduino should have some decoupling on the board itself, also the decoupling on the NRF24 adapter board should be sufficent.
You could add a low-pass filter after the power supply (are your supplys switching regulators?) to reduce the ripple produced, but the frequency rejection of the linear regulators should also handle this fine.
Thanks. I'm also still a beginner in electronics and try to fit as many infos as possible into the posts.
If some question arise i'm happy to answer them, or expand the posts about them.
I need probably 3-4 more posts to cover the whole project.
PS: Some not so obvious credits, as i seem to forgot them:
The connected nesting box project has the target to observe the endangered boreal owl in the free wild and is initiated by the NaBu Siegen-Wittgenstein Germany (NaBu = Naturschutz Bund =~ nature conservancy organisation). It is also partially fund supported by the Umweltministerium NRW (Ministry of environment NRW, Germany).
I probably need to compile a bigger lists of credits ^^
As the title said, i'm currently building/build a set of solar powered nesting box nodes, to observe the inner of said boxes. (By Picture)
The Gateway (a RPi) is also solar powered and connected with a UMTS stick to the internet.
I'm currently writing a series of blog-post of this MySensors powered project.
@TRS-80 you may want to look at RTL-SDR http://www.rtl-sdr.com/about-rtl-sdr/
If you just want to scan on the nrf24l01+ channel range, this is maybe enough for you: https://maniacbug.github.io/RF24/scanner_8pde-example.html
@Anticimex said:
What's wrong with a few herbs?
Maybe he just wants to change the lights, music etc. in his living room to get a more fitting scene for his current activities?
This can probably marked as solved in http://forum.mysensors.org/topic/797/pimatic-mysensors-controller-plugin/85 ?
@fahrettine said:
What is my USB port? USB0 or USB1 or other else?
dmesg | grep tty
Will give you an output like:
ch341-uart converter now attached to ttyUSBXX
@fahrettine said:
{
"plugin": "mysensors",
"driver": "serialport",
"protocols": "1.5.1",
"startingNodeId": 1,
"driverOptions": {
"//": "'/dev/ttyUSB0' if using serial Gateway",
"serialDevice": "/dev/ttyMySensorsGateway",
"baudrate": 115200
}
}
You may want to double check that config.
It should be more like:
{
"plugin": "mysensors",
"driver": "serialport",
"protocols": "1.5.1",
"startingNodeId": 1,
"driverOptions": {
"serialDevice": "/dev/ttyUSB0",
"baudrate": 115200
}
}
if, ttyUSB0 is your connected arduino nano.
If someone has one of these left in germany or nearby i'm willing to buy one for daugther-board dev purposes.
@TRS-80 said:
And then he would be saving on power supplies for each and every node (he would only need one central power supply / step down transformer in the basement)
Energy efficient wise it would be even interesting to run 12/24V through the house and step that down on the sensors. He could run led-lights etc. directly from this power source.
Its much more efficient to use one big transformer to do the 230V AC -> DC conversion instead 10-20 small transformers all over the place. (Think: Sensors, USB-Charger, LED-Lights, different other small gadgets)
If a solar panel is planned for his house he could even source the DC directly from it instead of doing the costly 230V AC transform.
@Anticimex forget it, i wrapped my head arround it again and i'm totally wrong.
Sorry for the inconvenience caused. Just wanted to get it right.
@Anticimex: yes i read that about the inwards facing. But if i look at the MySensorsNode-Board and the Easy-PCB-Board.. the MYSX seems to be different orientated so that it would face on one of the boards outward...... or i'm wrong?
The possibility is high that i'm wrong. XD
@sundberg84 i think this would also fit. Thanks.
Didn't find it on the openhardware.io site.. maybe due to lack of keywords.
@Anticimex and @sundberg84 do i see that right that the orientation of the pin header on both boards is different? So that a daugtherboard, depending on the design, would stand out on one board and would sit above on the other board?
@Anticimex hm.. the 5x5 Board looks like it could fit. Would it be possible to lower the pro mini and the nrf24l01 to fit a daugtherboard above it? Can really see on the renders what is beneath it.
@sundberg84 yep. Already read that. But if i would design a daugtherboard i would also need a "Motherboard" to plug it in with MYSX 1.1 or above.
Maybe @Anticimex knows what is available.
Hello,
i was looking into creating a MYSXconnector daughterboard, and asked myself what are the option for the "Mainboard"?
I looked arround a bit on openhardware.io but couldn't find much about it. Maybe the community knows what are the options?
Can i "mate" that with my fitness tracker? (pun intended) 
However, reading carefully the descriptions state that some of the "5m strips may be soldered together" LOL,
Normally not a problem, but the adhesive on the backside of the soldered parts can be weaken.
also, they show pictures of 3M backing, but do not explicitly state that the backing is 3M.
Well they are shipped with the 3M backing... but usually they can just support their own weight and not a gram more. I suggest to obtain double sided adhesive for parts that are more stressed.
@Reza what controller are you using? Its all a matter of, if the host-system can handle the ch340 ic.
The FT232 and the CH340 are both serial to usb ICs.
The CH340 is often used on Chinese Arduino clones.
There is no difference on the user side, expect that you need to download and install drivers on windows for the CH340. (On Linux included)
Personally i'm happy that they are switching to the ch430 instead of using ftdi ft232 clones.
@ahmedadelhosni
http://www.ebay.com/itm/KEMO-M028-Power-control-110-240-V-AC-2600-VA-Made-in-Germany-/251096815198
as example..they are relativ expensive.. see it more as an example.
@kr0815 yes, tmrh20 confirmed this.
https://github.com/TMRh20/RF24/issues/233
Did you already tried the other recommendations?
@jemish maybe have a look at following blog-post: http://www.homemade-circuits.com/2012/04/how-to-use-triacs-for-inductive-loads.html
With this information you should be able to "improve" above schematic to be also able to handle inductive loads.
If you have no clue about all that... maybe have a look at this module:
https://www.kemo-electronic.de/en/Light-Sound/Effects/Modules/M028-Power-control-110-240-V-AC-2600-VA.php
It can control AC Motors,
in combination with this module:
https://www.kemo-electronic.de/en/Transformer-Dimmer/Converter/M150-DC-pulse-converter.php
You can connect it safely to the arduino.
@bigjimjeedom said:
@Oitzu
do you use thoses settings (max max)?
Yes. I use the max/max settings. I also use as datarate RF24_2MBPS but RF24_250KBPS should be better for good range.
because some people says that it was not the good way with this module but maybe it is because of bad power supply
Yes this is due to low power supply -> The module doesn't get enough current for the MAX setting.
OR
Bad power supply -> The module needs more current -> bad supplys tend to be noisier on higher current.
OR
Bad shielding -> The output from the module is that high that it feedbacks (without the shielding) back into the IC.
i use this for supplying (with a 5v from USB):
Never used them, but on the datasheet they look okay.
@bigjimjeedom you should receive the best results with:
#define RF24_PA_LEVEL RF24_PA_MAX
#define RF24_PA_LEVEL_GW RF24_PA_MAX
#define RF24_DATARATE RF24_250KBPS
As long as your power supply is clean and your module is shielded.
@jemish which module do you mean exactly?
The schematic in the first post is not suitable for inductive loads.
Pinging @marceloaqno i remember he had something in the pipline.
@flopp in a free line of sight positioning the alignment 3 should be the best.
But you comment that you have about 3 walls between the gw and node.
In such a envoirment it is not simply foreseeable how the rf signal will propagate. Many factors of reflection and absorbation are in play in such a envoirment.
@noelgeorgi: My experience with this: Power supplies differ. Even modules that seem to be the same can have different effective output. So maybe trying another supply is the way to go, yes.
Just want to help.
@noelgeorgi have a look here: http://blog.blackoise.de/2016/02/fixing-your-cheap-nrf24l01-palna-module/
Tackles most of the nrf24l01 pa/lna module problems.
In short: Power/Noise/Shielding.
Addidtional caps don't help in any case.
@noelgeorgi if it works with the normal nrf24l01+ modules, it is more likely that it is a power/noise/shielding problem.
Just to get an idea how they perform: I got, with the right supply etc. about 1000m free line of sight out of this modules max.
Currently i have a stable link of about 270m with 200 packtes/s in a non line of sight environment (forest).
@moskovskiy82 there are several topics about this:
http://forum.mysensors.org/topic/3353/nrf24l01-with-antenna-not-working/2
http://forum.mysensors.org/topic/1877/solved-2-x-nrf24l01-pa-lna-with-rf24_pa_max/2
as example. There are several problems with the nrf24l01+pa+lna.
@LastSamurai depends on the used onboard regulator and the used nrf + random sensor.
But yes, this works also, most of the time on non pa/lna modules.
@AWI hm... okay. Will probably order a few and test them.
@LastSamurai https://www.adafruit.com/datasheets/MP2307_r1.9.pdf
seems to be more complete and well.. a filter or a second stage would be recommend.
@AWI i'm actually have a project running from 12v lead batteries, best way would be to use 2 stages?
ATM i'm using an additional lc-filter to get the ripple out.
@Igor-Katkov eh... now you just sound like my gf.
No, you're right. But it may come handy in some other situations.
@Igor-Katkov said:
DSO138
Was thinking of something like the Hantek 6022BE. Most EE guys would say "eh... crap", but oh well.. it's just a hobby and i'm not willing to pay hundreds of dollars for it.
It would be so useful to be able to measure the supply voltages with an oscilloscope. So far all my transmission problems have been related to the power supplies. Only after using a scope I fully realized that. No more freezing or poor function of the nodes.
You are totally right! Unfortunately i also don't own a oscilloscope. But a cheaper usb oscilloscope is on my need to have list because of a all the trouble and testing i had with this...
@LastSamurai can't really say for this modules.
The datasheet for the LM2596 says you should use a post ripple filter for low noise, because it generates sawtooth ripple at its switching frequency (150khz). The china-clones of the lm2596 seem to be even worse generating even greater ripple with 50khz.
Whats the ic on the mini dc-dc converter called? The ebay offer says it has a output ripple of 30mV bute more then often these values are the best case. Maybe a datasheets is available.
@LastSamurai also keep in mind that if you are using the LM2596 buck converter that you need to use a post ripple filter.
I explained that here: http://blog.blackoise.de/2016/03/building-a-lc-filter-for-your-nrf24l01-palna-module/.
@LastSamurai Which nrf24l01+ modules are you using on which side?
Whats the typical error pattern you get? (Failed transmission from gw to node or from node to gw?)
Need to know your exact construction, before guessing what could be wrong.
@gmccarthy don't give up. You seem to be on the right track and already got great results.
Maybe go bare bones and try the rf24 scanner,
https://maniacbug.github.io/RF24/scanner_8pde-example.html
to find the best nrf24 channel for your.
@Ironbar ahh... now i remember you.
Interesting findings. Your problems on higher PA levels could also be problems with the used power source.
Part of the design of this enclosure was the idea that you could wrap the interior of it with aluminum foil to shield the radio. Overall, I think the results between the two would be similar.
Maybe, maybe not, depends on how much is actually inside the box next to the nrf24l01+.