Multichannel RF receiver question

[Jimbz]

Hello all, I have a small question regarding RF receivers...

Say we have a 4 channel transmitter and a 4 channel receiver all operating at the frequency of, for example 900 MHz. And we transmit our data on all four channels, my question is, how does the receiver differentiate between the channels ? How does it know that T1=R1, T2=R2 etc ? Are the carrier phases different or is there a slight difference of frequency on each channel ? As in T1=899 MhZ, T2=900 MHz ? Could anyone please explain this witchcraft for me, I am not the sharpest knife in the drawer when it comes to RF....
Also if you are transmitting data on all four channels at the same time, how will one single antenna know to seperate them ?
Would be nice if someone could give me a short answer o nthese questions, and maybe perhaps even some readables.

Good day gentlemen

[ve7xen]

Most devices labelled "900MHz" (or 2.4GHz or whatever) do not operate at precisely this frequency. They will have several channels nearby the quoted frequency. Maybe the 4 channels are 903MHz, 904MHz, 905MHz and 906MHz or whatever. The channel spacing is going to vary widely depending on the bandwidth required for the transmissions etc. 802.11 WiFi channels for example are spaced by 5MHz, with the band centred at 2.442GHz (so channel centres at 2.442, 2.447, 2.452 etc.), but the transmission is something like 20MHz wide, so neighbouring channels overlap and interfere with each other. This is frequency-division multiplexing, and is probably the simplest and cheapest scheme to implement, and is continuous in the time domain, so most consumer ISM gadgets do it this way. It gets more complicated when you think about something full duplex (transmit and receive simultaneously) like a cordless phone - it can't transmit on the same channel it's receiving on, so it might use a separate uplink and downlink channel. Usually spaced by some fixed spacing much larger than the channel spacing (for filtering). For something like a cordless phone, for example, you might have 4 uplink channels at 902MHz, 903MHz, 904MHz and 905MHz while the corresponding downlink channels are +10MHz, so 912, 913 etc. A cordless phone probably wouldn't have a 1MHz channel spacing, but you get the idea.

There are other techniques too, such as code division multiplexing (aka CDMA) you have probably heard of, which relies on each transmitter modulating their transmission with a special code that allows fancy math and signal processing to extract multiple overlapping signals. This is used in modern cellular networks and GPS, among others. Time division multiplexing is also possible, and is more akin to frequency division multiplexing in that each transmitter is assigned a short time slot (by some arbitration system), where it is the only transmitter on the channel, in this way multiple transmitters can share a channel and not overlap. I believe this is used in Bluetooth, DOCSIS and some other systems. It was also used in many early wireless digital networks like GSM, but these have mostly moved on to CDMA now.

Anyway there are lots of ways to skin this cat. Chances are your consumer gadgets are just using a different frequency for each channel. Some fancier things, like the better cordless phones, are using spread spectrum techniques and probably CDMA. In these schemes, the transmitter makes a wideband but very weak transmission (intentionally near the noise floor), and the code known to both transmitter and receiver is used to extract the signal from the noise.

I don't have any good articles at hand, but some of the wikipedia stuff isn't bad. Here is probably a decent place to start: http://en.wikipedia.org/wiki/Channel_access_method

[Jimbz]

Most devices labelled "900MHz" (or 2.4GHz or whatever) do not operate at precisely this frequency. They will have several channels nearby the quoted frequency. Maybe the 4 channels are 903MHz, 904MHz, 905MHz and 906MHz or whatever. The channel spacing is going to vary widely depending on the bandwidth required for the transmissions etc. 802.11 WiFi channels for example are spaced by 5MHz, with the band centred at 2.442GHz (so channel centres at 2.442, 2.447, 2.452 etc.), but the transmission is something like 20MHz wide, so neighbouring channels overlap and interfere with each other. This is frequency-division multiplexing, and is probably the simplest and cheapest scheme to implement, and is continuous in the time domain, so most consumer ISM gadgets do it this way. It gets more complicated when you think about something full duplex (transmit and receive simultaneously) like a cordless phone - it can't transmit on the same channel it's receiving on, so it might use a separate uplink and downlink channel. Usually spaced by some fixed spacing much larger than the channel spacing (for filtering). For something like a cordless phone, for example, you might have 4 uplink channels at 902MHz, 903MHz, 904MHz and 905MHz while the corresponding downlink channels are +10MHz, so 912, 913 etc. A cordless phone probably wouldn't have a 1MHz channel spacing, but you get the idea.

There are other techniques too, such as code division multiplexing (aka CDMA) you have probably heard of, which relies on each transmitter modulating their transmission with a special code that allows fancy math and signal processing to extract multiple overlapping signals. This is used in modern cellular networks and GPS, among others. Time division multiplexing is also possible, and is more akin to frequency division multiplexing in that each transmitter is assigned a short time slot (by some arbitration system), where it is the only transmitter on the channel, in this way multiple transmitters can share a channel and not overlap. I believe this is used in Bluetooth, DOCSIS and some other systems. It was also used in many early wireless digital networks like GSM, but these have mostly moved on to CDMA now.

Anyway there are lots of ways to skin this cat. Chances are your consumer gadgets are just using a different frequency for each channel. Some fancier things, like the better cordless phones, are using spread spectrum techniques and probably CDMA. In these schemes, the transmitter makes a wideband but very weak transmission (intentionally near the noise floor), and the code known to both transmitter and receiver is used to extract the signal from the noise.

I don't have any good articles at hand, but some of the wikipedia stuff isn't bad. Here is probably a decent place to start: http://en.wikipedia.org/wiki/Channel_access_method

Thanks for the reply, managed to clear some things out, but a question remains, say we have 2 or more channels transmitting data/singal at the same time and only one antenna. How will the frequency look like even ? I can see this being done in CDMA/TDMA as you said, but in FDMA, how will the carrier react to multiple frequencies transmitting at the same time ?

[ve7xen]

Thanks for the reply, managed to clear some things out, but a question remains, say we have 2 or more channels transmitting data/singal at the same time and only one antenna. How will the frequency look like even ? I can see this being done in CDMA/TDMA as you said, but in FDMA, how will the carrier react to multiple frequencies transmitting at the same time ?

In the case where you have two transmitters connected to the same antenna, you're more into RF engineering concerns. You need to be careful that the transmit power of one transmitter isn't getting dumped directly into the other transmitter and wasted (possibly causing damage). This is generally accomplished with filters to isolate the two transmitters. Another technique is to combine the two carriers (if possible) while they're still at small signal levels and that can be easily done, and then amplify the combined signal for final output to the antenna. Once it reaches the antenna, the signal is now a superposition of the two transmissions. When both signals are pure, unmodulated sine waves, the output in the time domain will have a main frequency component that's the average of the two frequencies, and an envelope that's the difference. This is illustrated by this animated gif I borrowed from a site that describes wave superposition. In the frequency domain, and an ideal case (with unmodulated sine waves) you'd simply see two peaks at the transmission frequencies.



On the receive side, you would then use filters to separate the two transmissions again.

[w2aew]

Hello all, I have a small question regarding RF receivers...

Say we have a 4 channel transmitter and a 4 channel receiver all operating at the frequency of, for example 900 MHz. And we transmit our data on all four channels, my question is, how does the receiver differentiate between the channels ? How does it know that T1=R1, T2=R2 etc ? Are the carrier phases different or is there a slight difference of frequency on each channel ? As in T1=899 MhZ, T2=900 MHz ? Could anyone please explain this witchcraft for me, I am not the sharpest knife in the drawer when it comes to RF....
Also if you are transmitting data on all four channels at the same time, how will one single antenna know to seperate them ?
Would be nice if someone could give me a short answer o nthese questions, and maybe perhaps even some readables.

Good day gentlemen

In most cases, the channels are on separate carrier frequencies.  The channel spacing will depend on the modulation and information content.  The antenna doesn't "separate" them, just like the antenna on your car doesn't separate the signals from the various radio stations that you can tune to.  The receiver will separate them, either through filtering or other processing.  For example, it could be a narrow band receiver that is tuned to one channel at a time.  Or, it could be a receiver with a wide enough bandwidth to receive all the channels, and the separation is done deeper in the receiver path.  Lots of ways to skin this cat...