(please remove the previous thread, I cant seem to find the edit button, where is it?)
Hello all,

Im trying to come up with a simple method (and design) for detecting (mostly and foremost next to transmitting) a (continued) HIGH or LOW state through the 420-450Mhz RF band. Without the possibility of interference. Preferably analog, at least on the receiver side.

The range has to be a minimum of 1000m (but not much more) and the speed has to be at least around 1Mbits per second (although Im currently trying to avoid sending actual bits, as said above).

On the receiver side of things I thought, for a moment, about using a (possibly envelop detector like) design which would require receiving two separate frequencies at once (on two separate antennas) in combination with an AND gate. Which assumes a HIGH state if the two frequencies are present and the LOW state if they are not. However this still has a risk of interference.

Which might force me into a digital receiver (or transceiver) type of design with "bit" verification for the signal. The downside is that this adds complexity in several ways. It, for instance and among others, almost always requires a additional MCU on the receiver side, which is something I would like to avoid. Also the speed is not high enough on most of the receiver ICs in the suggested RF range. I didnt look at receiving transceivers for now as they always require MCUs. Again, Im not looking to receive information, only a continued HIGH or LOW state.

On the transmitter side it would likely (if technically feasible) require a more standard IC transmitter/transceiver type of design which would be driven from a MCU (CPU). However, again, these are also mostly made to send out bits, so this could turn out to be a challenge. The receiver would have to be able to translate a (possibly digital) signal into a analog HIGH state.

Does anyone have any suggestions on how to approach the (receiver and to a lesser degree transmitter) design in order to reach the requirements? Can I, for instance, add logic on the receiving side to detect the (analog or digital and logically modulated) signal? Modulation does require additional speed but it possibly cant be avoided to prevent interference?

Or should I direct my time and resources on a dual transceiver IC design (if actually available and feasible for this type of operation)? Also considering FCC regulations and such, although a simple analog receiver design in this range should not be a problem I think.

I sincerely appreciate any and all suggestions. Thanks.

 

As soon as you say, "Without the possibility of interference," you are asking for something that is not possible to achieve. So the first thing you need to do is take the time to quantify what kind of interference you need to be resistant to and how resistance you need to be to it. If you want to keep things very simple yet have high jam resistance, then one option is to just use OOK (On-Off Keying) and look for energy being present within a particular bandwidth. If the energy is below a certain threshold, then you call is LO and if it is above that threshold you call it HI. Your threshold has to be set high enough that your interfering sources are very unlikely to put that much energy into that band and your transmitter's energy has to be sufficiently higher than that so that, at the receiver, it will be comfortably above the threshold despite attenuation due to distance, fading, or interference.

 

Original thread deleted.

Spectrum techniques using pseudorandom noise can give you a boost in signal-to-noise ratio, and in this instance any radio signals within the receiver's bandwidth will be suppressed.

What is the end use of this application and what country will it be deployed (this is more than a small pair of details).

 

Thanks for the reply WBahn. I appreciate it as Im mostly new to RF design and the directions are very helpfull.

Im afraid the Neanderthal in me likes the possibly simple, reliable and elegant analog solution of straightforward HIGH-LOW switching. RF makes it less straightforward.

Regarding interference, if I understand you correctly that is.

One thing is that selecting and designing for sensitivity in a very specific frequency helps, which depends on the analog or digital design (possibly of the IC used). What is required in terms of interference is that the HIGH state or LOW state does not change in anyway other than actively by the transmitter, in whatever way possible. Now digitally there is interference but it is (mostly) worked out through receiving (accepting only) the bit or byte sequence (although with OOK it is only one bit, not free from interference?) . I did think therefore of using and detecting two (very specific) frequencies at the same time to lower the risk of (registered) interference but the risk remains, ofcourse.

Doesnt OOK send a specific bit (or byte string) to promote the HIGH state? Or is it always done directly through energy thresholds as you also seem to imply? Its both, correct? 0 bit and 1 bit at lower and higher energy states or the other way around, where the energy threshold is programmable? Working with energy levels seems to be prone to interference? Byte strings are more complex and slower but less prone to interference. OOK seems very delicate to me, depending on the frequency used, it really doesnt seem free from interference. I appreciate the suggestion though. Other protocols, Im afraid, will possibly downgrade speed and reliability in terms of receiving the signal on time and every time.

Maybe I need to check the ability of OOK more carefully as I checked mostly receivers and not transceivers. Microsecond range switching is required and OOK receivers (especially those not requiring a MCU) are on the slow side.

I should have mentioned it will be one specific frequency within this 420-450Mhz range. The bandwidth to select a frequency from is open to change for whatever works and considering the availability in the sense of regulatory requirements.

Also I understand my question is somehat unusual. But Im not looking to send data or audio, only basic states.

Thanks again.

P.S. Dick, Im not sure what exactly you mean with your suggestion on noise but that is my fault in lacking in-depth RF knowledge, not yours. Youre relating to a possibly new receiver design or is it an option within receiver/transceiver ICs? Or an option when designing the PCB? Or both? The application is essentially to drive a power state of some kind, any kind.

(The edit button is sometimes not present on the forum so I cant change the initial question or edit otherwise)

 

(please remove the previous thread, I cant seem to find the edit button, where is it?)
Hello all,

Im trying to come up with a simple method (and design) for detecting (mostly and foremost next to transmitting) a (continued) HIGH or LOW state through the 420-450Mhz RF band. Without the possibility of interference. Preferably analog, at least on the receiver side.

The range has to be a minimum of 1000m (but not much more) and the speed has to be at least around 1Mbits per second (although Im currently trying to avoid sending actual bits, as said above).

On the receiver side of things I thought, for a moment, about using a (possibly envelop detector like) design which would require receiving two separate frequencies at once (on two separate antennas) in combination with an AND gate. Which assumes a HIGH state if the two frequencies are present and the LOW state if they are not. However this still has a risk of interference.

Which might force me into a digital receiver (or transceiver) type of design with "bit" verification for the signal. The downside is that this adds complexity in several ways. It, for instance and among others, almost always requires a additional MCU on the receiver side, which is something I would like to avoid. Also the speed is not high enough on most of the receiver ICs in the suggested RF range. I didnt look at receiving transceivers for now as they always require MCUs. Again, Im not looking to receive information, only a continued HIGH or LOW state.

On the transmitter side it would likely (if technically feasible) require a more standard IC transmitter/transceiver type of design which would be driven from a MCU (CPU). However, again, these are also mostly made to send out bits, so this could turn out to be a challenge. The receiver would have to be able to translate a (possibly digital) signal into a analog HIGH state.

Does anyone have any suggestions on how to approach the (receiver and to a lesser degree transmitter) design in order to reach the requirements? Can I, for instance, add logic on the receiving side to detect the (analog or digital and logically modulated) signal? Modulation does require additional speed but it possibly cant be avoided to prevent interference?

Or should I direct my time and resources on a dual transceiver IC design (if actually available and feasible for this type of operation)? Also considering FCC regulations and such, although a simple analog receiver design in this range should not be a problem I think.

I sincerely appreciate any and all suggestions. Thanks.

All transmission / receiver systems are statistical,
i.e. there is always a risk of an error in the raw transmission / reception,

With a given Tx / Rx signal, you will have an associated SNR / Error rate.

The way to lower the error rate is to

a) increase the Tx Power
b) increase the Rx selective sensitivity
c) change the medium
d) encode the signal and use a "clever" receive algorithm.

Using (d) is the way to get a reliable Tx / Rx system.

Using (d) , we can do things like transmit from the distant planets with 5W of power and receive on the earth,

Indeed, algorithms such as that used in GPS+ , or even 5G mobile phones can be used to pick signals out that are below the noise floor.

https://electronics.stackexchange.c...on-if-the-received-power-is-below-the-noise-f

The amount of DSP you want to do is effectively the limiting factor

 

The next level up from OOK is to encode some number of bits in the signal. This can be done in a variety of ways. The more bits you have the less likely it is for interference to produce a coherent sequence. The DSN (Deep Space Network) keeps listening across all bands for coherent sequences.

https://www.nasa.gov/directorates/heo/scan/services/networks/deep_space_network/about

 

Yes the simplest way of creating an algorithm is to encode a byte string. To keep up the speed and reliability and decrease errors I think 4 bits would do however, for now, Im trying the find the elegant analog solution. Possibly a dual OOK system using two specific frequencies. With which I need to start balancing speed and reliability in receiving the signal. And also the availability of ICs (the Si4356 seems to be an option). Byte encoding methods are an option but this adds complexity and Im currently not sure if its available of the shelf (at the required speeds). Im afraid I need to start looking at transceivers not made for this purpose or discard the sub Ghz RF solution completely. It shouldnt be this hard.

Thanks again and please keep the suggestions coming.

 

What would be the optimal frequency in terms of speed and decreased risk of interference? Considering global availability it seems a specific frequency within the stated range of 420-450 MHz would be most suitable. This frequency would also be capable of reaching 1000 meters.

P.S. Thanks Andrew, let me read through the link. I`m not sure if its relevant though.

 

Just about every legal jurisdiction in the world has stringent requirements for what kind of signals can the transmitted and by whom (and a bunch of other details as well). Optical transmission (think laser) might be the most practical since in many places you don't need a permit to transmit optical signals.

Ever wonder how hundreds of GSM cellphones all operating in the same part of the spectrum can simultaneously send and receive messages clearly? That's because they operate via spread spectrum using different spreading codes.

Quite simply a digital code pseudo noise is generated and connected to the input of a balanced modulator made so that one state (like "1") causes the baseband input (the signal to be encoded) (the other input to the modulator) to not be inverted and the opposite state (like "0") to invert the baseband input. You just created double sideband but instead of a constant frequency carrier you have a carrier made of pseudo noise that covers a wide bandwidth. The baseband data is encoded as sidebands on the noise. That is the easy part : -)

Pseudo noise is a repetitive pattern, the phase of which can be readily detected. This is so that at the receiving end an identical pseudo noise stream can be synchronized (that is the hard part) this locked pseudo noise signal is then used to demodulate the encoded data.

Below is the circuit of a simple decoder. U2 generates the pseudo noise and adjusts the clock to get its pseudo noise in sync with the incoming signal's code.


This circuit was used with an optical detector and worked quite well at separating the pseudo noise from externally induced noise.

 

What would be the optimal frequency in terms of speed and decreased risk of interference? Considering global availability it seems a specific frequency within the stated range of 420-450 MHz would be most suitable. This frequency would also be capable of reaching 1000 meters.

You are setting up conflicting goals. You want a simple analog solution yet you want to make it immune to any and all forms of interference.

As I said previously, that second goal is absolutely unobtainable. Someone can always build a transmitter that will interfere with your transmission. You need to accept that. There is no such thing as jam-proof.

Having accepted that, you have to decide how much jam resistance you need and what kinds of interfering signals you need to be resistant to. Being able to resist interference from automobile ignition systems and garage door openers is one thing. Being able to resist active jamming from someone that knows how your system works and is intent on disrupting your comms is quite another.

About a decade ago I built a radio that operated in the range you are interested in (I think I was parked at 430 MHz) and my transmitter used a 7" antenna with a mean power output of around one microwatt. The receiver was in a UAV, using the same antenna, and was in an orbit between 1 miles and 5 miles way. The jammer was located about fifty feet from my transmitter and they had to get their output power up to 75 W before they prevented recovery of the signal by the airborne receiver. I was using a simple OOK transmitter and I was transmitting a pure sine-wave waveform (when keyed) at 4 Mbaud. But I was doing a lot of processing on the received signal using a new coding theory that we had developed specifically to improve the jam-resistance of signals when the enemy has your keys and knows everything about your waveforms.

 

Thanks Dick, yes juridiction is a thing. I did check frequency available however I need to look into the allowed modulation or even the continues sending of frequency signals. The latter might not be allowed. Optical transmission is not a solution due there being objects and multiple detectors/receivers all looking for the same signal at the same time.

Regarding Spread Spectrum, sure but the technology seems overly complex for what Im trying to achieve here. And will it work with multiple detectors/receivers looking for the same signal (I think yes if they are configured equally?). Let me look into it nonetheless. Interesting but complex. There is no reason it cant work in the 420-450Mhz range or is there? Its however a different ballgame and Im not sure I can pull it off.

For the most part I understand your explanation. It does however require a complex design, an MCU and therefore coding on the receiver side.

The last sentence I do not understand as you switch back from RF pseudo noise to optical methods? What is the location and use of the optical detector in this regard?

There should be a relatively simple method for sending a continued (but switched) HIGH or LOW state without interference (or with filtered out interference at the receiver side).

 

Thanks WBahn. No sorry, its not like Im looking to encrypt information. It has to be free from general unintended interference in the atmosphere. Not free from intended jamming, there is no need for that type of security at all. Its not dealing with information, only HIGH-LOW states driving a power state.

Interesting work it seems. Jamming and the encoding to prevent it is a different type thing altogether. 4Mbaud is too slow for what I need however what you are saying is that OOK keying might work if I find a smart way to send and decode it correctly. The required speed needs to be there however and encoding increases errors? It also requires an MCU and programming on the receiving side which I would like to avoid.

I will take some more time thinking about a robust analog receiver (detector) design as it does not need to protect information. I mention again that there are multiple detectors/receivers looking for the same input.

 

Read Wbahn's post #10. No matter how good you make your system it is possible for somebody else to crack it and corrupt the signal.

Encoding should reduce errors, not increase them

The baud rates you seem to require probably can't be achieved with microcontrollers. Consider using an application specific IC or a discreet logic implementation using high speed logic devices.

If you find an interference free analog receiver, please share the details with us.

 

Thanks WBahn. No sorry, its not like Im looking to encrypt information. It has to be free from general unintended interference in the atmosphere. Not free from intended jamming, there is no need for that type of security at all. Its not dealing with information, only HIGH-LOW states driving a power state.

Interesting work it seems. Jamming and the encoding to prevent it is a different type thing altogether. 4Mbaud is too slow for what I need however what you are saying is that OOK keying might work if I find a smart way to send and decode it correctly. The required speed needs to be there however and encoding increases errors? It also requires an MCU and programming on the receiving side which I would like to avoid.

I will take some more time thinking about a robust analog receiver (detector) design as it does not need to protect information. I mention again that there are multiple detectors/receivers looking for the same input.

You said that microsecond range switching is required, now you are saying that 0.25 μs switching is too slow. So how fast is needed?

What is it you are trying to accomplish? What is the underlying problem you are trying to solve? What are your actual requirements?

 

Yes Dick, true. However there is no information and therefore motivation to crack it. Its not a critical application in that sense.

Encoding at first does increase errors due to possibly incorrect or incomplete readings of the format but they are corrected, right? This often requires two way communication. Which adds to complexity. Now it seems that some types of modulation improve signal reception making errors less likely. Modulation however is not completely and always the same as encoding? Encoding deals with 0s and 1s, modulation can also deal with sending out specific patterns, frequencies or a combination of both. The terminology can be confusing, at least for now and for me.

Sorry WBahn. 4 Mbaud would be sufficient, my brain got stuck on 4 kBaud. Your setup sounds highly optimized and customized, too complex for my needs.

Ok, lets say I use a simple (possibly envelope like) detector to sniff out a specific frequency and filter out the rest. Which then outputs a voltage translated through a OPAMP configuration to provide a HIGH state if the signal is present and a LOW when its not. That is essentially all I need. Now I need to be able to somehow isolate this frequency signal from other atmospheric interference (mostly other signals at roughly the same frequency). For instance by encoding in it a (simple, for example, 4 bit) signal which I decode with analog or basic digital means (hopefully not requiring an MCU, possibly logic gates). Or by sending a more complex pattern of frequencies which I somehow detect for a duration of time around the 1 Mbaud rate, something like DSSS (Direct-Sequence Spread Spectrum, someone suggested it but for now the specific design details allude me).

Those seem to be the requirements and options. Essentially relatively low tech.