Jump to content
43oh

300m Range RF PCB discussion


Recommended Posts

A member recently asked me a bunch of questions through private messaging.  I'm posting the dialog here in case anyone else is interested.

 

 

I'm trying to make a circuit for a low power and long range rf link and I would like to ask you for your help.  Please could you recommend me for a module at 433mhz or 868mhz? Do you think I could reach a good range with NRF24L01? 

 

 

there's nothing at 2.4 GHz that is going to give you more than 100m range.  The NRF24L01 is good for about 20m.  What are your requirements for power & long range?

 

 

 I want to cover a distance of about 300 meters and with a low power usage. I'm looking for a cheap solution.  I'm located in Europe so I should use 868Mhz or better 433Mhz  The MRF49XA would be suitable?

 

 

To communicate further than 300m, you need to worry about multipath.  Multipath is interference coming from the signal bouncing-off things.  The receiver sees copies of the same signal, and it can be difficult to decode.

 

There are some ways to reduce multipath interference.

1. Use lower frequency.

2. Use a modulation that has redundancy, like wideband FSK.  Alternatively, use DSSS.

3. Use lower data rate.

 

I've never heard of anyone using MRF49XA.  You may want to try Semtech SX127x, which is indeed the device used by SigFox.  There are many, new, low power RF IC's coming out.  The ones most interesting are: Semtech SX127x, SiLabs Si4463, TI CC1200, and ST SPIRIT1.  All have slightly different advantages, although I use SPIRIT1 primarily. 

 

 

Thanks for your indications. I'm learning a lot more.  I have been looking the specifications of the Semtech SX1276 which have FSK and GFSK and a bit rate up to 300kbps.  What would you reconmend me? SX1276 vs CC1200 vs SPIRIT1?  Why do you use SPIRIT1 primarily? Which are their advantages over the others?

 

I have seen that CC1200 have a excellent support.  Regarding to the antenna, I have no space limiations. It would be better to use external dipole antenna right?

 

 

 

You can probably forget about communicating faster than 50kbps, maybe 100 at most.  Higher data rate reduces range.  You can do some tests to figure out what works best for you -- multipath is your biggest problem, and it is nonlinear, so you might find that there is 10% packet loss at (for example) 50kbps and 90% packet loss at 60kbps.  You should just test and find out -- make sure to adjust digital RX filter bandwidth together with datarate!

 

Yes, use an external dipole.  Monopole is OK too, but only if you know how to tune the design on your board.  So just use the dipole.

 

Of those chips, use whichever chip is easiest for you.  Make sure to use wideband FSK (set Fdev larger than baud rate) with FEC enabled on CC1200 or SPIRIT1.  On SX127x, make sure to use a combination of spreading (DSSS) and FEC.  Without going into great detail, I will say that SPIRIT1 is the best chip to use if you are an expert.  It is like a race car that is extremely fast, but difficult to drive.  SX127x is like a fast car that is easy to drive, so I recommend it for you.  If you cannot achieve 300m with SX127x, then you are doing something very wrong :).

 

 

Thank you very much. Now I understand a bit more about rf links and multipath. If I get 50 or 100 kbps I will be happy.  I dont undertand what you mean when you say adjust digital RX filter bandwidth together with datarate. You mean to send and receive at the same bandwidth?

 

Ok I will use external dipole to keep the design simple then later I can try other design with a monopole antenna.

 

I will use the SX127x for my design. Comparing the SX1272 and SX1276 which one have better range? Because I read the SX1276 have 100 dB blocking inmunity and both have +14 dBm power amplifier.

 

 

 

I think the only difference is that 1272 is for 862, 866 MHz bands only, and 1276 can use 169, 433, 862, 866.  1272 is probably cheaper.  For small volumes, 1276 is probably better, because you might want to try different bands.

 

RX filter is a setting on the transceiver.  If your RX filter is narrower than the emission spectrum (known as power spectral density, PSD), you will not receive all the power of the signal.

 

 

It's neccesary to use 4-layer board for this design or with 2-layer board would be ok?

 

 

In systems with monopole antennas, it is important to have an uninterrupted ground layer.  It dipole systems, this is less important.  However, you also want a ground layer underneath the transceiver and front-end analog circuits.  For 866 MHz you will want the spacing to be no more than 0.8mm.  For 433 you can get away with 1.6mm.  You also want to sink a lot of vias between ground layers.  This prevents ground loops.

 

Therefore, 2-layer is OK, but it is difficult to design a PCB that is good for RF with only 2-layers.

 

 

You say for 866Mhz the spacing no more than 0.8mm. This means a 0.8mm PCB? and with 433Mhz a normal 1.6mm PCB?

 

 

Yes, 0.8mm PCB if you use only 2 layers, and you use 866 MHz.

 

You should probably use 433 MHz, anyway, since range is important.  169 MHz would also be an option, although it is a new band so it can be difficult to find ready-made parts for it.  169 MHz on a dipole will go 300m easily!

Link to post
Share on other sites

If you could use 2.4GHz in Europe then I would recommend using the Synapse RF200 module. It will get you well beyond 300m distance at 2Mbps.

 

http://www.digikey.com/catalog/en/partgroup/sm200-and-rf200/20491

Probably not -- the specs of that module and the science of flat fading path loss doesn't add up to 300m in anything but dry (no rain/snow), open-field conditions.  2.4 GHz band is like a mosh pit.  Stay out unless you have a damn good reason to use it.  There are better alternatives for low-data-rate wireless.  2.4 GHz is best for indoor, short range, high data rate usage.

 

This module is an 802.15.4 module, and it lists 2Mbps, but that is a bit facetious.  The "chip rate" of 802.15.4 PHY is 2MHz (2Mbps).  However, there is 32:4 DSSS, yielding an data rate of 250kbps.  If you want to have any hope of traversing 300m, at 2.4GHz, with any degree of reliability, you will need that DSSS.  A typical 1mW 2.4 GHz 802.15.4 link can cover about 50m.  This module can do 100mW, but due to the way path loss works, plus the way flat fading makes path loss worse, you will be lucky to get 10x.  Still, that might meet 300m if there are not WiFi and Bluetooth interferers making a mess of everything -- the reality is that in developed areas, you have a -75dBm noise floor with 2.4GHz.  That absolutely sucks.  In USA and EU, the noise floor in the 433 band is more like -130 dBm.  In EU, 866 is about -100dBm in most places (it used to be a bit better). Why bother with the 2.4 GHz mosh pit when you can much more easily reach 300m with sub 1 GHz devices?

Link to post
Share on other sites

Excellent @@jpnorair !

Your personal experience is very valuable because sometimes the datasheets are very vague in some areas.

For example I would like to see a sticky post with the RF devices/modules we have used in the past with the range, speed, power usage, ease of usage. This could be of help for others in case of interest. It could be possible @@bluehash? ^_^

Who would be interested?

Link to post
Share on other sites

@@zeke I used to do work with WiFi and Bluetooth.  I'm glad you are able to get what you need from the hardware you have, but with RF you can't just say "I get 1600m, so will you."  Conditions are very critical.  Data rate, modulation, error correction methods, and receiver design are all important parameters that must be specified.  Antennas are important, too.  There are just too many externalities with 2.4 GHz band -- interferers, water, snow, walls, etc.  There are reasons why, for the most part, WSN has moved to over bands.  For a while, some people tried hard to do WSN mesh networking at 2.4 GHz, to work around the limitiations, but it proved not to be cost effective.

Link to post
Share on other sites

I think you are over reacting.

 

You said your OP needs a long range low power radio link between two points at 300m distance. That suggests several points that you are not addressing.

 

1. No mention of physical installation. Indoors or outdoors? At 300m he is going to need antennas outside.

2. No mention of antenna requirements. Yagi/Uda, Monopole, Dipole, Rubber duckies? Two Yagi's would work well

3. No mention of power consumption by radio. How much power can be put into the band?

4. No hard spec on modulation technique. CW/AM/FM/QAM/DSSS/FHSS/Custom? Synapse is SS

5. No mention of datarate.

 

Your responses all suggest to me that you have decided:

1. The antenna must be omnidirectional,

2. The modulation technique must be FM,

3. The noise floor must be pristine at -106dB or better

 

Just ask yourself how NASA gets data back from The Voyager sattelites. What power are they transmitting at? What modulation scheme are they using? What datarate are they at?

 

I contend that 2.4GHz is not a problem for this guy's measly 300m application. I know that I could make it work well. I submit that with careful antenna selection and placement he could too. The Synapse module just makes the task all the more easier and cheaper.

 

Start focussing on all the must do's to get the job done. Grounding, ground plane, power supply filtration, feed line impedance, antenna selection and placement, RF power budget and margin. Lastly, the money budget. If all it takes is two $30 modules plus two good antennas to achieve success then that proves that it is a viable solution to the original problem. It may not be the only solution but it IS a solution nonetheless.

 

Hating on a specific RF band is unwise.

 

BTW, the Synapse modules mesh quite nicely. You should check that out.

Link to post
Share on other sites

@@zeke You are free to write up a similar response.  As you acknowledge, just saying "use this module" is not enough.  You seem intent on proving that I am wrong, which is pretty common in discussions between engineers so I'm not terribly bothered.

 

I am making three assumptions:

- the user wants to achieve 300m with minimal other concerns.

- I know more about RF communication than the user does.

- I know more about RF communication than you do

 

Maybe my assumptions are wrong.  Don't take it the wrong way.  43oh.com isn't exactly a huge repository for knowledgeable RF communications engineers.  Even on the websites that are, there is a lot more empirical-backed recommendations than than theory-backed recommendations.

 

Anyway, I will expand on some points I can tell you are interested in.

 

Why not 2.4 GHz:

My recommendation for 2.4 GHz usage is: your mileage may vary.  From a theoretical perspective, it has very unattractive path loss metrics.  From an empirical perspective, it can be useless in one place vs. another, depending on the co-presence of public WiFi.  From an industry perspective, I can attest that WSN is moving away from it.  So it should not be the default choice.

 

Why FM:

Wideband FSK is a good modulation choice for a flat-fading channel.  For the most part, any channel you can use is flat fading.  PSK is sometimes even better, but I can think of no low-power, low-cost transceivers that do coherent PSK.  In a flat fading environment with a noncoherent receiver, any constant-envelope angle modulation (like FSK or PSK) is a canonically endorsed choice.  Wideband FSK is essentially an analog-spread signal, so it's really easy to do with almost any IC, and why it is almost universally supported.  SX127x doesn't use wideband FSK by default, but it combines MSK with DSSS.  Low-cost receivers implement MSK as very-narrowand FSK, so the distinction is pretty-much naught.  For what it's worth, every low cost 802.15.4 transceiver I've ever noticed also approximates the offset-QPSK from the spec with MSK, which itself is actually just implemented as FSK.

 

You also brought-up Voyager:

Voyager uses coherent PSK with a 1/2 rate, k=9 convolutional code and a 4/5 rate interior reed solomon code, both for error correction.  Aggregate error correction coding rate is 0.4.  If you apply this to an embedded 1mW device, it still works very well, but not in crowded bands.  The CC+RS coding scheme is great, and it can work for bursty, variable length packets.  More modern schemes like turbocodes and LDPC perform better at the cost of greater computational overhead and, in most cases, fixed-size packet requirements.  My product HayTag uses a code very similar to the voyager code.  Maybe in the future I will move to LDPC, but the hardware isn't quite there yet.

 

And finally DSSS:

DSSS is a funny subject, because most hardware using it quotes the sensitivity of the channel in analog terms (symbol)  -- i.e. "high sensitivity down to -146.5 dBm" in the SX1276 literature -- rather than digital terms (bit).  The boltzmann noise at 433 MHz is around -142 dBm, so 1-bit decoding more than 1.6dBm below that (shannon limit) is not theoretically possible.  But Semtech is spec-ing the receiver sensitivity according to the symbol rate rather than the bit rate -- everyone does this, despite the fact that it is horribly facetious.

 

DSSS is good for combating multipath interference due to the way the bit is correlated from the symbols.  This also allows some degree of error correction, and furthermore some additional ability to use different spreading codes to multiplex communications in overlapping channels (again, YMMV, but in general it works).

 

Nonetheless, the easiest, best way to get long range is simply to drop the frequency as low as you can, to use no DSSS, and to use some type of 0.5 or 0.4 rate error correction coding.

Link to post
Share on other sites

 My product HayTag uses a code very similar to the voyager code.  

Your AD say:

 

Unlike WiFi or Bluetooth, DASH7 operates at a radio frequency which provides for both long range (1km in the U.S., further in other markets) and excellent indoor signal propagation.

Source: http://haytagstore.com/portfolio/our-technology/

So this guy asking you just need to use 433MHz and coherent PSK with a 1/2 rate, k=9 convolutional code and a 4/5 rate interior reed solomon code.

I wonder if this can be done with those cheap dumb $1.25 SAW 433MHz FR modules? 

Link to post
Share on other sites

So this guy asking you just need to use 433MHz and coherent PSK with a 1/2 rate, k=9 convolutional code and a 4/5 rate interior reed solomon code.

I wonder if this can be done with those cheap dumb $1.25 SAW 433MHz FR modules? 

Implementing this is a lot more difficult than simply using the SX127x, which has some other built-in features for achieving long range.

 

HopeRF sources Semtech components, or possibly pirates the HW -- I have no way to verify.  Either way, SX127x might be used in a HopeRF module fairly soon.

 

The SAW-based transceivers certainly do not supply coherent receivers, PSK, or integrated convolutional code, so you need to use a different device.  Implementing RS can be done in firmware, but it is not easy.  It is even less easy to do it fast in firmware.

Link to post
Share on other sites

Nope. I am not that Engineer

 

You said:

 

there's nothing at 2.4 GHz that is going to give you more than 100m range.

 

 

You are incorrect. I suggested an RF module that can do many times that range. Namely, the Synapse RF200 module. Yet, you dismissed it because it is 2.4GHz.

 

 

You said:

 

Probably not -- the specs of that module and the science of flat fading path loss doesn't add up to 300m in anything but dry (no rain/snow), open-field conditions.

 

My response is Synapse Wireless' family product brief. The RF200PD1 module is spec'd at a max distance range of 3 miles (~4800 meters)!

 

I have repeatable RF testing and validation reports to corroborate these findings. How can you blithely dismiss this?

 

 

You said:

 

2.4 GHz is best for indoor, short range, high data rate usage.

 

Clearly, you have never heard of the company called Wi-LAN. There are so many 2.4GHz DSSS and FHSS radios installed outside in industrial applications that it would make your head spin.

 

During development of the dual diversity antenna system, the engineers at Wi-LAN suffered with multipath fading mostly while indoors. They called it Canyoning. It was all the metal wall studs in the building.  Outside, the system worked freaking awesome. I worked two feet away from the VHDL engineer on the project. I worked 15 feet away from the RF engineer on the project. This was back in 1997.

 

Your reaction to my RF module suggestion just seem out of character - like you are simply not familiar with the technology available today or you have been burned by a past experience.

 

 

The User asks "Which solution is obtainable in the least amount of time for the least amount of money?"

 

The Developer asks "How can I extract the most amount of money from my client for the least amount of my input?"

 

The Experimenter asks "How can I make this whizzy cool thing happen with the least (or greatest) amount of learning for the least amount of money?"

 

 

In the end, my opinion does not matter. You can always argue the technical superiority of one solution over the other. And since I have nothing to prove and it seems that you have already made up your mind about 2.4GHz then that's okay. We do not have to have the same understanding about a topic and we can still get along.

 
 

But our boss will always make the decision based on money and time. You know it will always be about money.

 

Our client is the boss.

 

We have to focus on making him successful with whatever skills, talents, gifts, abilities, knowledge, wisdom that we have available to us.

 

If we concern ourselves with their interests above our own then we will be rewarded.

 

Our reward is in the satisfaction that he will become one of our Raving Fans and return as a customer - repeatedly.

 

The money we earn is the evidence of our good intentions towards our customer's interests.

 

So go forth and Be Awesome!

Link to post
Share on other sites

there's nothing at 2.4 GHz that is going to give you more than 100m range.

One bad assumption I made is that the power should be low.  With 1mW, it requires a lot of finesse to get it.  Sure, it's possible to find corner-cases, or to use more power.  That was my mistake.

 

 

 

Your reaction to my RF module suggestion just seem out of character - like you are simply not familiar with the technology available today or you have been burned by a past experience.

Are you referring to the Synapse module?  That's an 802.15.4 module.  I know that technology well -- for the record, it uses MSK with a 4:32 spreading sequence.  In most places it is limited to 1mW, hence the bad assumption from above.

 

 

 

 

 

My response is Synapse Wireless

family product brief. The RF200PD1 module is spec'd at a max distance range of 3 miles (~4800 meters)!

Repeat test with two nodes, at ground level, in an NLOS environment.  In the dry, then in the rain, then with the target node concealed by snowfall.

 

High frequencies are good for high data rate because antennas can conduction larger bandwidths at higher carrier frequencies.  Using 2.4GHz for low-data-rate seems a bit foolish, unless there is no other unlicensed band to use.

Link to post
Share on other sites
  • 1 month later...

This is an interesting discussion. I am caught in a struggle between using 2.4GHz or sub-GHz frequency. Both range (ideally above 300 m) and size is important to me and communication speed is not important. I am just sending a few bytes of data occasionally.

 

The transmitter and receiver will be used in an indoor environment, with lots of machinery and several walls. I can afford to use the transmitter with a power amplifier and a long antenna. But on the receiver device, it needs to be small, almost like the size of a watch. With 2.4 GHz, I could possibly have a PCB antenna or chip antenna to make the board really small to fit into my design, however, I might run into problems with range and multi path interference? On the other hand, using sub-GHz frequency seems to give me more range, but I can't afford to have a huge antenna protruding from my design.

 

Any suggestions or comments will be most appreciated! Thanks in advance!

Link to post
Share on other sites

What kind of data rate do you need?

 

If you are in North America, 915 MHz is probably the best choice for this, because the transmitter power can be extremely high (even up to 1W), and it is allowed to use DSSS.  In south america, some countries have this band also, but others do not.  In the rest of the world, there is nothing like it.  Your best bet in the rest of the world is to use 433 MHz at 10mW, which is quite alright for sending 300m at relatively low data rates.  Building a 433 MHz compact antenna can be difficult, but certainly not impossible.  Plus, in a couple months you can just buy HayTag boards, and all that work is done.  :)

 

2.4 GHz is an option, of course, but it doesn't have anywhere as good indoor range as you will get with the sub 1 GHz frequencies.

Link to post
Share on other sites

Thanks for the reply jpnorair.

 

Most likely I will got for 433MHz. I just need to send a few bytes of data, like a string message with up to 20 characters. 

 

The Haytag boards looks great. Again, I wonder how they manage to fit the antenna into such a small form-factor. Any tips on the antenna design part?

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...