Hi guys,
I'm writing this not because I have a problem but because I was finally able to solve some. I think it's worthwhile posting my experience here to maybe help some of you.

For a long time I was experiencing random unreliable radio communications with NRF24L01+ nodes powered by power adaptors. Some of my nodes worked perfectly while some others were receiving data really inreliably and generally had problems communicating. After a lot of try and error and reading some tips I think I have now found a reliable solution: Power supply input filtering.

I knew for a long time that the transceivers we use are highly sensitive to the cleanliness of their power supply and that typical cheap AC / DC switching power supplies (like e.g. a 5V phone charger) produces pretty noisy DC. But due to lack of an oscilloscope I didn't really know what to do about that.

After I've read this awesome post I tried putting such a recommended LC filter between the output of my LDO and the transceiver to clean out the power supplied to the transceiver. Unfortunately it didn't really work and I still had the same problems. After a lot of trying and playing around I have developed the theory that some linear power regulators are very pedantic about their output capacitors (specifically their respective ESR values) and you have to be really careful about what to put on the output of those things. I also learned that LC filter design is not as trivial as I thought: In bad case scenarios such a filter can start to oscillate and make things worse. So if you don't really know what you're doing and don't have an oscilloscope it is difficult to accurately design a working filter.

So my proposed solution is the following: Put such an LC filter directly between the output of your AC/DC power brick and the input of the power regulator. Linear LDO regulators typically already output pretty clean DC voltage, but only if their input is also kinda clean. The parameter that describes this correlation is called the "PSRR" or "power supply ripple rejection". In other words: shit in --> little less shit out.

By putting an LC filter (or Pi filter, whaterver) in front of the power regulator you are pre-filtering most of the bad output ripple of the cheap AliExpress 5V phone chargers and other power bricks. This "kinda clean" voltage is then supplied to the LDO which can do the rest of the work and filter out the remaining ripple or oscillation.

I don't know if my description is absolutely accurate and if it is the best solution that also a professional electronics design engineer would choose but at least for me it was the first thing that really worked.

In the end my proposed "optimal" (in respect to complexity vs. benefit) power chain can look like this:

I don't know if the capacitor and inductor sizes are optimal, but something approximately to this is running flawlessly in my home on several nodes. As an LDO I have been successful with the MIC5205 (same as Arduino Pro Mini), AMS1117, MCP1700, LE33 and LP2950 (this was the one I initially had the most problems with).

Also if you want to use such a design, keep the components of the individual sections as close as possible together. Between the blocks you can (within reason) do what you want, but you will get the highest quality out of short distances within the sections.

edit: I played around with different filter values in circuitjs and updated some things in the proposed circuit diagram. I decreased the inductor size from 333uF to 33uF (Low-pass cutoff frequency is higher but should be sufficient for AC adapters with a switching frequency higher than ~60kHz) and roughly doubled the size of the LC filter capacitor. Also I implemented a damping resistor (1 ohm resistor in series with 470uF capacitor) to optimize the behavior. Here you can see the result:

Guys, I found a solution for my problem! I tried everything from soldering the two GND pins of the Arduino together to an additional small condenser but nothing worked.

So just for the heck of it I connected my buck converter to the RAW pin of the Arduino and adjusted it to something between 3,5 and 4V. Now my stepper driver and the transceiver both are supplied with power from the Arduino VCC pin and it works 100% reliable. So my conclusion is that the Step down converter of the Arduino conditions the 3.3V better than the buck Converter and this solved all my receiving problems.

I don't like that I now have an additional conversion step in my power supply but honestly at this moment I don't bother anymore because I'm just happy to have it finally working.

I thank you all very much for your Input!

@titvs said in Introductions and Range Issues:

I was wondering if its ok to set the MY_RF24_PA_LEVEL to HIGH on all powered nodes (sensors, repeater and gateway), or the performance will be degraded?

Talking about the frequency ranges, i have a mesh of wifi Ubiquiti Unifi AP's in my house, which are able to do a RF scan and show me the channels that are more utilized. It seems the more crowded channels here are (20 MHz): 5 (2421-2443), 6 (2431-2453 MHz), and (40 MHz): 3 (2401-2443), 5 (2411-2453) and 9 (2431-2473).
With this in mind, im setting up my nodes on channel 2 (MY_RF24_CHANNEL 1 if im correct) where there is 0% utilization.

Hi titvs,

• I can remember that initially the MySensors library set the NRF24L01s power to "HIGH" by default. I just glanced at the API page to verify if this is still the case but it seems that now the default value has been set to "MAX" which is the highest possible power setting. In my experience if you know that you have enough supply power for the transceiver (in my case it even works with 100 mA 3.3V regulators) you can set the power "MAX" without a problem. In my case it increased the range a little bit and did not decrease the network quality. But keep in mind that the PA version of the NRF24L01 (the version with the power amplifier and large external antenna) is using significantly more power. So I assume you need a power source that can handle approximately >200 mA or even more.

• I personally would just leave the MySensors network on its default channel. In the USA this channel is already out of the range of wifi. In Europe you theoretically could have an overlap but I personally have not experienced problems with this. IMO the problem with using lower channels is that wifi routers set to automatically search for the best channel could choose to just hop on your precious MySensors channel. Then your effors was just a waste of time and you have to manually change the firmware on every MySensors device again. If you use the MYSBootloader to flash FOTA updates, you are tied to use the default channel anyways.

--> Just to follow up on my previous post with my quality and reliability problems of the NRF24L01 transceivers: I think I was able to solve the problem as now my communication has been running flawlessly for over two days. The trick that seemed to work was to decouple the whole device from its power source input as much as possible by using an LC filter before going into the 3.3V LDO that supplies my transceivers. Previously I tried out everything to clean up the power supply AFTER the step down converter (also tried RC, LC filters, various capacitors...) but it didn't work reliably. I assume the problem with this is that some power regulators are absolutely pedantic about the ESR of their output capacitors. Playing around with various filter configurations on the output of the regulator maybe screwed up those values and caused the regulator to be unstable. Also I have no idea what I'm doing when I'm "designing" those filters, so I just randomly tried out various combinations of inductors and capacitors. So maybe this also caused more harm than good.

But now by putting the LC filter in between the output of the switching power adapter and the input of the LDO and keeping the output capacitors of the regulator exactly as specified in the data sheet, it basically solves all those problems. Basically the LDO doesn't care at all about the ESR or capacitor values at its input and also the chance that a badly designed LC filter is causing problems is much lower as the power regulator can dampen those influences afterwards.

So in my experience this configuration can solve most problems of bad communication due to dirty power supplies:

120V / 230V AC Mains --> Switching power adapter (5 - 24V DC with loads of ripple) --> LC / PI Filter (less ripple) --> linear LDO power regulator with output capacitors as specified in data sheet (clean 3.3V DC) --> NRF24L01+

That's a good question. I assume that you have used an electrolytic capacitor, right? If it would have been a ceramic type: no worries, they don't have a polarity.
The answers I could find in the internet for electrolytic capacitors varied quite a bit, but the consensus seems to be that they like to blow up in such cases, provided a high enough voltage. So I have two theories why yours didn't: either you have used a capacitor rated for a higher voltage and the 3.3V wasn't enough to destruct it. Or the current through the capacitor was limited because the 3.3V regulator could not provide more.

Anyways I have read that you should replace such a capacitor that has been subjected to a reverse polarity to be sure because it could have taken long-term damage.

Hello Kai,

this is such a coincidence! I am developing the same device with the same sensors at home. I am currently writing the software so it is not functional yet. But maybe I can look into what you did.

What I can offer you: I designed a custom PCB with all the components directly SMD-soldered onto it. It is very compact but due to the form factor of the VL53L0X sensors you definitely need a reflow oven or heatgun to solder it. I could try to find a way to share the PCB layout. Then you can define this task as completed.

Greetings,
Martin

@yveaux I try to hack something together to get serial debug data If I find some time this week. I'm personally really curious what I will see there.

@yveaux Thanks for your answer. I am using my own designed PCBs with a bare Atmega328 soldered onto it. The interrupt traces are short with no sharp corners. I have this problem across multiple PCB layouts so I would rule out a design fault. Some of my boards run the Atmega on its internal oscillator at 8MHz and some have an external 16MHz crystal. Both with the same behaviour.

My gateway is also based on the Atmega328, coupled with a Wiznet W5500 ethernet module. As the NRF24L01+ has to be driven by soft-spi in this configuration, the IRQ pin is not used on my gateway.

Unfortunately I have not mapped out the RX / TX pins on my PCBs, therefore I have no possibility of logging the serial data

My send rate is not high. One use case e.g. is an LED dimmer with a controllable ramp-time. Therefore, each time before my home automation server sends a new dim value it also sends the desired ramp-time. So it's only two messages directly one after another. Admittedly, I do not know how fast my HA server is sending those messages. But in my understanding the MYS gateway should care about the send timing, right?

Could it be plausible that the second interrupt fires while the µC is still reading the first message from the transceiver, thus getting a messed up message stack? Or maybe there is a problem with clearing the stack after a message has been received. Consequently, it fills up and after that everything is rejected... Unfortunately my C skills are very limited, so I cannot look for a probable cause in the code myself.

Hi,

My NRF24L01+ based MySensors network is struggling with nodes that randomly stop receiving messages. Until this week I was not able to find some pattern in this behaviour.

Now I think I found a link: The nodes that use the MY_RX_MESSAGE_BUFFER_FEATURE with the connected IRQ pin of the transceiver all fail after some time. They are still able to send messages but don't receive anything anymore. Only a reboot fixes the problem. I found that I can replicate this behaviour when I'm sending messages to more than one Child ID of the node in fast succession. But even without doing this the node fails after some time.

I have played around with different buffer size values (between 10 and 35). It seems like the node is working longer with a higher buffer size.

As soon as I disable the buffer feature the node works OK and does not stop receiving.

I can observe this behaviour with version 2.2.0, 2.3.0 and the test version of the 2.3.0 with the modified CE handling. Also on my gateway I am using the PA version of the transceiver, on all of my nodes I have the normal version with PCB antenna.

Does anybody also have the same problem? Is there maybe a bug in the buffer code? I really would like to use this feature as I have a much quicker and more reliable acknowledgement handling as long as the nodes are running.

That's a good question. I assume that you have used an electrolytic capacitor, right? If it would have been a ceramic type: no worries, they don't have a polarity.
The answers I could find in the internet for electrolytic capacitors varied quite a bit, but the consensus seems to be that they like to blow up in such cases, provided a high enough voltage. So I have two theories why yours didn't: either you have used a capacitor rated for a higher voltage and the 3.3V wasn't enough to destruct it. Or the current through the capacitor was limited because the 3.3V regulator could not provide more.

Anyways I have read that you should replace such a capacitor that has been subjected to a reverse polarity to be sure because it could have taken long-term damage.

@gohan that's a pity. Do you know what the root cause could be? Temperature? Charging cycles? Bad part quality?

@alowhum looks interesting. In case of the power supply for the NRF24L01+ they are the same as the small breakout boards: 1117 regulator. But it looks like there is another regulator for the 5V supply of the rest of the board. I assume that the 3.3V regulator is hooked up to the 5V output of the other one. So you have two regulators in series, which gives you a very good ripple rejection of the input supply. I assume they will work quite nicely.

@รอเร-อ Yeah, maybe it would be a nice project to design such a breakout board. Currently I'm all tied up in designing completely integrated custom PCBs. So the filter is sharing the same board as the Atmega328, the NRF24L01, as well as whatever specific components are needed.

Most of the time I see those NRF24L01 power breakout boards getting recommended. I have never used them but in the end it's just an AMS1117 power regulator with input and output capacitors. I can imagine them working fine when using a high quality 5V power supply, but it will have the same problems if you have ripple and general noise on your power supply.

@pihome said in Build a reliable power supply chain:

what i have and it is working so far but not for W5100 gateway.

5v--10uF----1uF----(1117 3.3v)----10uf----1uf---->3.3v for nRF

have you tried your schematic for W5100 gateway? i m having some issues with W5100 waiting for some parts to get delivered. another 3 weeks to go ....

Hi pihome,
I am successfully running a W5100 gateway with an AMS1117 for the Transceiver and rarely have problems with it. But on my gateway I have an NRF24L01+/PA Transceiver, so the larger version with amplifier and external antenna.
What I have learned about the large module:

• You have to shield it. When I was using it without shielding I had a lot of communication problems. In my experience, shielding doesn't make such a difference on the smaller versions of the transceivers, but on the large ones it does.
• It is less sensitive to power supply ripple than the smaller modules, but need more power. I assume this comes from the inherently higher receiving sensitivity due to the signal amplifier.

Therefore it would be interesting to know which transceiver version you are using. If you are also using the large transceiver with amplifier and external antenna, I would first advice to shield it. If you are using the standard module without external antenna I could imagine that our power adapter has a high output ripple that the AMS1117 is partly outputting to the NRF24L01+. In this case my proposed filter design could help with your problems.

After my findings with the LC filter connected to the input of the LDO, I also upgraded my described gateway configuration with this circuit. I can't say if it improved its performance because it was already working quite well, but it also didn't degrade it. Therefore I would suggest that you try it out and report back if it could solve your problem.

edit: What also came to mind is your actual power adapter. How much current can it provide? I know that the W5100 Ethernet module is consuming a relatively high amount of power as it gets warm when it's powered. It also has an on board AMS1117 regulator which can supply 800mA. If you are using a small 500mA power supply or something like this maybe your power adapter is too small.

Hi guys,
I'm writing this not because I have a problem but because I was finally able to solve some. I think it's worthwhile posting my experience here to maybe help some of you.

For a long time I was experiencing random unreliable radio communications with NRF24L01+ nodes powered by power adaptors. Some of my nodes worked perfectly while some others were receiving data really inreliably and generally had problems communicating. After a lot of try and error and reading some tips I think I have now found a reliable solution: Power supply input filtering.

I knew for a long time that the transceivers we use are highly sensitive to the cleanliness of their power supply and that typical cheap AC / DC switching power supplies (like e.g. a 5V phone charger) produces pretty noisy DC. But due to lack of an oscilloscope I didn't really know what to do about that.

After I've read this awesome post I tried putting such a recommended LC filter between the output of my LDO and the transceiver to clean out the power supplied to the transceiver. Unfortunately it didn't really work and I still had the same problems. After a lot of trying and playing around I have developed the theory that some linear power regulators are very pedantic about their output capacitors (specifically their respective ESR values) and you have to be really careful about what to put on the output of those things. I also learned that LC filter design is not as trivial as I thought: In bad case scenarios such a filter can start to oscillate and make things worse. So if you don't really know what you're doing and don't have an oscilloscope it is difficult to accurately design a working filter.

So my proposed solution is the following: Put such an LC filter directly between the output of your AC/DC power brick and the input of the power regulator. Linear LDO regulators typically already output pretty clean DC voltage, but only if their input is also kinda clean. The parameter that describes this correlation is called the "PSRR" or "power supply ripple rejection". In other words: shit in --> little less shit out.

By putting an LC filter (or Pi filter, whaterver) in front of the power regulator you are pre-filtering most of the bad output ripple of the cheap AliExpress 5V phone chargers and other power bricks. This "kinda clean" voltage is then supplied to the LDO which can do the rest of the work and filter out the remaining ripple or oscillation.

I don't know if my description is absolutely accurate and if it is the best solution that also a professional electronics design engineer would choose but at least for me it was the first thing that really worked.

In the end my proposed "optimal" (in respect to complexity vs. benefit) power chain can look like this:

I don't know if the capacitor and inductor sizes are optimal, but something approximately to this is running flawlessly in my home on several nodes. As an LDO I have been successful with the MIC5205 (same as Arduino Pro Mini), AMS1117, MCP1700, LE33 and LP2950 (this was the one I initially had the most problems with).

Also if you want to use such a design, keep the components of the individual sections as close as possible together. Between the blocks you can (within reason) do what you want, but you will get the highest quality out of short distances within the sections.

edit: I played around with different filter values in circuitjs and updated some things in the proposed circuit diagram. I decreased the inductor size from 333uF to 33uF (Low-pass cutoff frequency is higher but should be sufficient for AC adapters with a switching frequency higher than ~60kHz) and roughly doubled the size of the LC filter capacitor. Also I implemented a damping resistor (1 ohm resistor in series with 470uF capacitor) to optimize the behavior. Here you can see the result: