RF Communication...stranded

Many transmitters, one receiver...

All transmitters transmitting at the same frequency...

I can almost hear Oliver Hardy..."Well, Stanley, this is another fine mess you've gotten us into!"

The only way that I can see you getting around this problem without too much modification is to stagger your transmitters' transmitting time. What I mean is that you program each transmitter to transmit one at a time (transmitter #1 sends, then transmitter #2 sends, followed by transmitter #3, etc. etc.).

I don't think there is a way for your system to encode their data so that the receiver can differentiate between stations. In any case, your receiver probably cannot process more than one signal anyway so the sequential transmission method is the only way to go.

Thanks Vulcan... That is what we were looking at as our best option....

like sending messages in two or three bursts... one instantly and the other two at an interval specified by their system number(ID)..

We cannot put a real time clock in just yet... so I think that should do...

Well in case for a later version of the design... coz this wouldn't work if the network components are large say... 100+ in number... what would you suggest... feel free... no bounds...yet... Later the customer would like us to cut down the frills and costs too....

Well thanks for the input... It is reassuring to find that more people have come to this timing conclusion... Thanks again...

I read this post earlier, and was thinking about it, when a thought came to me, I came back and find that Vulcan had the same thought! Just to elaborate on what he said, you could use a "sequencer" that would only allow a transmission when it's turn comes up. The complexity of this is going to depend on how your system currently operates. This will require all of the transmitters to be connected to the "sequencer", but it still allows "one-way" communication. The other option is to have a "receiver busy" signal that would go to each transmitter, and this would hold off the transmission until the receiver is not busy. I guess you could also link all the transmitters together, and do the same thing, if one is transmitting, it prevents the others from trasmitting.

As far as a software only fix, the only way I can see this working is if you have perfectly synchronized clocks in each transmitter, then each transmitter could be programmed to have it's own time slot. Good luck!

Tom

Hello tdesmit,

well you are right.. it is the most logical solution....

As far as a sequencer or a real time clock is concerned it is out of the picture...

Well about the receiver busy signal... would that require a receiver on every transmitter unit.... if so how costly a modification... (a complete receiver section is my guess..) and if does not need a receiver...do tell me more about it... please....

Also we are looking at incorporating a receiver... but my Hardware engineer friend is worried about the antenna length... (the customer does not want any protrusions..well...aesthetics and some safety norms....) The transmitting antenna is a simple loop back... but can the receiver also have a loop back antenna instead of a long Ariel or wire....

And well a hardware change is necessary and up in the near future.. so do let me know.. if you have any inputs.... they would be most welcome...

But right now... there are issues like demo due date.. first prototype investment and client commitment.....

take care....

Well about the receiver busy signal... would that require a receiver on every transmitter unit.... if so how costly a modification... (a complete receiver section is my guess..) and if does not need a receiver...do tell me more about it... please....

I guess it was too early in the morning for me to be giving an answer, I guess I was thinking about a signal wired from the receiver to transmitter, not a good solution for a wireless setup! Sorry about that. I could be done by placing a scaled-down receiver on each transmitter unit. I say scaled down because it would only need to be a "it's there/it's not there" detection. The busy signal could come from the receiver, or you could detect when another transmitter is talking. With the proper filtering and detection, this could be done without a complete receiver, it would be a simple RF detector. Good luck with the project, I'll be watching this thread.

Tom

could be done by placing a scaled-down receiver on each transmitter unit...would only need to be a "it's there/it's not there" detection

That's a good improvement idea! Actually it has a name, "collision detection", and it's used in fieldbuses and LAN networks. The way it works is that the transmitter waits for a while (well, a few microseconds I guess) and listens for traffic. If it's all clear, it sends its data. As soon as it sends, the other transmitters hear it and hold off on their own transmissions until it's all quiet again.

There will be times when two or more transmitters will transmit at the same time. The collision detection works here as well by stopping all transmissions dead in their tracks. The transmitters have a wait period (each is different) and the first transmitter to send, shuts out everyone else.

This doesn't remove the requirement for internal clocks. For now, the clocks can serve to program the transmitters to send sequentially. In the future, you can use them as the "wait and listen" timers I was talking about.

Damn it... too slow, you guys beat me!

Basically it's down to synchronisation, either manual or automatic. Other alternative is add more recievers and transmitter to make it all bidirectional.

It sounds like it is never going to be robust.

Even with 2 microcontrollers talking to eachother on the smae board it makes sense to have some method of acknowledgement. Maybe give the operators flags to wave or mobile phones, or set transmission times.

Thats ok Del... it is not who got first but the journey itself.. well seems a lot of us agree...

The mobile phones, flags wont do... nor will the transmission time... as no realtime clocks or...well MACs... just plain transmitters for now...

But the transmission time slots or sync. of some kind is the priority in the immediate upgrade.. Well the next version alternative is to dump RF... but really do you have any other option that is also cheap and would not lead to much trouble....

Once again...thanks a lot Del....

Tell us more about the distances involved etc...maybe we can come up with a cunning plan...

Signals down the power lines at different phase angles ?

.... what is the volume of data you are sending...?

Are you using big blue shiny electrons or just cheap dull grey ones ?

(sorry that's just silly... they are all blue and shiny everyone knows that)

There are such things as DTMF tones that can be incorperated into the system.

Crazy idea. RF, if FM then transmit at different tones on the carrier and recored the music(?) at the receiver, write a program to unentangle your ad hoc multiplexing.

or not

Two bits from Brad

Hi:

There's been a lot of discussion of time multiplexing, but not much of frequency multiplexing. If you have the bandwidth, use of individual subcarriers, keyed "audio" tones at the transmitter end and filters at the receiver end, might be useful. Other techniques, such as Frequency Shift Keying, would work but may be unnecessarily complex and expensive. Good luck.

DickL

I'm curious about why the frequency multiplexing suggestion is being ignored. All it requires is an audio oscillator at each transmitter modulating the transmitted signals with different frequencies and a row of sharp band-pass filters at the receiver to sort out the signals. A low transmission duty cycle and a varying keying pattern would prevent a lot of interference. Cheap and simple. What criterion did I miss in this problem?

DickL

The requirement was to give suggestions that will need the least amount of modification. Read it as "fastest/easiest to implement and cheapest".

I like your suggestion, however. It doesn't involve installing receivers in the transmitters. The transmitters all transmit at a center frequency of, say, 10MHz. Each transmitter modulates their carriers at different frequencies. The receiver gets these transmissions and, using bandpass filters, separates the data from each transmitter in separate buffers where the microprocessor can access them.

I'm not sure if there are complications but this can be easily built and experimented on.

Along that line, my suggestion in post 12 would require no hardware modifications at all which is what I thought he desired since he already had hardware built.

If you adapt your software to accommodate redundant transmissions and then apply a different delay between the first and second transmission for each transmitter, spacing the transmissions far enough apart will allow you to ensure that if every transmitter were to transmit coincidentally that all second transmissions would be in the clear then you can avert your problem.

The more devices you have the greater the spacing required to ensure all devices have a guaranteed clear time slot without any form of synchronization. You may need to imbed a sequential message code along with each transmitter ID so you can be sure to only cull redundant messages when both transmissions are received. You must also use a mandatory inhibit period following any transmission event that is greater than the maximum redundant transmission period to prevent a second transmission from any one transmitter from asynchronously stepping on another transmissions redundant emission.

This method has its limitations and may not be applicable to your situation depending on your requirements. I was able to use it successfully in a diffuse optical system where there were numerous completely asynchronous transmitters to a single receiver and it was very robust but you have to plan it carefully so you are aware of all the possible event sequences and allow for them.

I doubt that this system can be made to operate smoothly without backing up and punting. Incorporate receivers at each transmitter site, using the ACK function. The same antenna can be used by using an antenna switch (could be electronic using PIN diodes). Good luck!

One thought I haven't seen here is for the main station to poll each data site via RF. Just a thought. Good luck.