Sharing AM Frequency
Mike Collier
Posts: 25
Does anyone know of a method for having multiple receivers and transmitters on the same AM frequency - without interfering with each other.
My assumption has always been that this would not work.
I plan to use the frequency for (digital?)·remote control.
Any links to documents on this matter would be appreciated, I'm not a radio expert, just a humble programmer.
Many thanks, Mike.
Post Edited (Mike Collier) : 12/29/2005 3:27:48 PM GMT
My assumption has always been that this would not work.
I plan to use the frequency for (digital?)·remote control.
Any links to documents on this matter would be appreciated, I'm not a radio expert, just a humble programmer.
Many thanks, Mike.
Post Edited (Mike Collier) : 12/29/2005 3:27:48 PM GMT
Comments
The term "AM" has a specific meaning; do you really mean amplitude modulation?
One way for a receiver to distinguish between multiple transmitters is to do what GPS does - each transmitter spreads its signal by modulating it with a different pseudorandom code, and receivers detect specific transmitters by using each individual despreading code to restore the signal.
Another way would be to do what Ethernet does - have each transmitter send small packets containing specific receiver addresses, and detect collisions where multiple transmitters happened to send overlapping packets, and retransmit after random delays in those cases.
Or you could do like the previous suggestion, which is sort of like token-ring networking; each transmitter transmits only in its assigned time slot.
You could do a simple multiplex by using a FM-stereo transmitter with one FSK/PWM happening on the Left and another one happening on the Right (this would infer a·good degree of separation and some additional filtering.)
You could, I suppose, do this with mono-AM, if the FSK/PWM [noparse][[/noparse]control] frequencies are far enough apart to simply filter them (assuming that they would not "mix".)
David, by AM I mean in terms of AM/FM. I am interested in AM because I have a large number of AM receivers available, all on the same frequency. I don't have any transmitters yet.
PJ, I don't have the experience to implement the mutiplexing aproach but could probably cook up a network packet solution.
I work in a boarding school for boys under 13 and would like them to build their own low budget radio control robots. I have plenty of receivers and want to avoid any frequency clashes (over 100 boys).
How does this sound then, using the network aproach mentioned in prev thread:
1.
All receivers are on the same frequency.
All transmitters are on the same frequency.
Each robot contains a receiver.
Each hand held transmitter contains a receiver.
Each transmittion that is sent includes the receiver ID as part of the packet.
Only one transmittion is allowed at a time.
2.
A central control allocates unique ID's to every receiver - whether it be in a hand held transmitter or robot.
Each transmitter is informed of its robots receiver ID.
3.
Transmittion signals are synchronised in one of two ways (maybe others):
a) The central control sends a permit-to-signal message to each transmitter in turn (waits for end-of-signal for each).
b) The transmitters synchronise themselves (in some kind of loop).
Also..
Transmitters could be used as relays i.e. Transmitter3 sends message to Reciver2 (in Transmitter2) requesting it to send message to Receiver3 (in robot 3).
Any thoughts on this, (apart from the extra cost of the receivers)?
David B. mentioned token-ring as a possibility; with token-passing, there is only one "token" in the network. In order to send, you need to hold the token, and give it up once you are finished with your transmission. In this way, only one transmitter can be on the "net" at any time.
It would require a listening mode on the receivers, i.e., if they have data to send, they wait until the last transmitter advertises the token as available. A bit tricky without a "ring" topology to provide the latency such that not all units respond to the advertisemnt at once, but maybe each transmitter can have a random timeout before responding. Sounds like fun!
Jim
Mike.
You could mutiiplex them by using pl tones, thus increasing the amount of data sent on one am frequency at a time.
It's the same way modems send data..
Say you have 10 different tones on the same frequency, you then mutiplied by 10 the amount of data sent...
Between pl tones and a device ID code, 100 + students is doable..
With a "MASTER" control transciever, it would poll each "SLAVE" or "STUDENT" transciever on a particular pl tone with students ID code, allowing that
"STUDENT" transmitter to send out control impulses to the "STUDENTS" ROV that will only recieve and use code transmitted on the same pl tone,
and is prefexed by the students ID code....
Bob N9LVU
Post Edited (Robert Kubichek) : 12/30/2005 1:51:47 AM GMT
Daniel
Can TWS/RWS 434·modules handle multiplexing?
http://www.rentron.com/ruf-bot.htm·is an example of a project using these TX/RX modules and was the basis for my first thoughts.·
Yes they could, using Holtek encoder ICs.
Here is an example using IR as medium;
www.rentron.com/Fire-Stick-II.htm
You can do the same using RF.
By using the address pins of the encoder-transmitter/decoders-reciever.
You have 4 bits of data that are transmitted, and with 4 bits, you can have a lot of different commands....
You can control up to 128 students at one time using a Master/teacher, transmitter/receiver, that tells the Slave/student when to transmit the ROV commands.
Bob N9LVU
Post Edited (Robert Kubichek) : 12/30/2005 4:18:51 AM GMT
Whoaaaa just a minute.....
Do I understand that 50 AM transmitters will be transmitting at the same frequency at the SAME instant??......... NO WAY!!
The AM receivers won't be able to discern who is who as all will be interfering with each other. If they are time sequenced as in a token ring or some scheme where only one is transmitting at a time, then OK, you could make that work; but NOT all at once!
To do muliple transmissions at one frequency simultaneously, the only way I know of is via Direct Sequence Spread Spectrum (DSSS) and that ain't simple stuff; certainly not in line with the reason for choosing the AM to begin with.....lots of available simple receivers. DSSS receivers are far from simple, and even then 50 at a time is a stretch.
Much as you wish, I can't believe you can't get this concept of multiples at once to work.
Cheers,
Peter (pjv)
Maybe a slightly different design would help. What if the 50 or 100 or so students each had a controller that was hardwired to a computer that controlled the master transmitter. A piece of software would poll the controllers and generate messages to transmit. The master transmitter would send the coded data packets, and only the appropriate robot would respond. The controllers could be as simple as switches wired to a series of logic chips wired to the parallel port. A configuration of shift registers and latches could easily be as expandible as you'd ever want; you'd just have to take care to make reliable wires, connectors and controller boxes.
This way, you could still have free-roaming robots, but having only a single transmitter would greatly simplify the design.
Using a single transmitter occured to me too, especially since the robots will probobly be battling in a specific area of the school grounds so the boys will be in the same area and could therefore be connected.
I am leaning towards building the system now with the simplest design (perhaps the one you mention here), and then extending it towards the "token ring net" (not quite but close) as I become familiar with the technology. From a design perspective, the idea of a robot network does sound both interesting and fun (to me at least)..
Thanks to everyone for their suggestions and have an excellent new year!!
www.rsgb.org
http://www.webopedia.com/TERM/T/TDMA.html
TDMA works by dividing a radio frequency into time slots and then allocating slots to multiple calls. In this way, a single frequency can support multiple, simultaneous data channels.
Searching google (web & groups) for TDMA·returns plenty. Hope this helps anyone else·interested in this thread.
Mike.
[noparse][[/noparse]EDIT: erm.. seems TDMA is not as relevant as I thought, sorry]
Post Edited (Mike Collier) : 1/3/2006 8:48:10 PM GMT
TDMA is more for taking several channels of data and giving each channel a time slot in a continuous transmission. It is used, for instance, to put 24 or 30 channels of digitized voice onto a single T1 carrier by the phone company. Channel 1 gets x milliseconds, Chapter 2 gets the next x milliseconds, etc. Also, I'm pretty sure it requires either a transmitted or recovered clock (synchronous). Not sure it's a do-able thing for your project.
Sorry,
Jim
You seem to be changing your requirements, and I guess that's OK.
TDMA is NOT what you originally specified as all 100 units simultaneously. They are time sliced, and sequential. To keep it all synchronized it thakes some master unit to orchestrate things, OR some form of token ring method as previously suggested.
What more exaxtly is it you are trying to achieve? Pleas post details about range, latency, data rates, collision issues, target price, frequency restrictions, power source, etc. etc.
Your first simple inquiry could have a HUGE bunch of surrounding issues, and we can't tell which may be important and which one are not.
Cheers,
Peter (pjv)
I think what I am looking for is as follows:
1. To have up to 100 AM receivers all on the same frequency·(I own these already and would like to use them).
2. Have the "appearance" that all 100 are receiving and controlling robots "simultaniously".
I know almost nothing about radio - just programming, and was hoping that there would be a solution that either combined all my transmissions·into a single transmission - or time sliced them so that only one transmiters signal was in the air at a time.
I am aiming at the time slice method but don't have a protocol that manages that and was hoping that TDMA source code would be a basis.
I havn't done any detail study/design yet but I was thinking in terms of each transmitter taking turns to transmit its signal, the end of the signal would be an indication to the next transmitter (each transmitter has a receiver as I stated prev so that transmitters could sync with each other) that it could transmit. This design was based on my assumption that all transmitters and receivers were always in range and that there was some "logging on" process so a "loop"/"ring" could be registered amongst all transmitters. Each signal would include data about who was currently sending.
When I saw TDMA, I was hoping to find source code for it and port to a microcontroller, although I could write my own I am still hoping for a working design in wireless timeslicing.
I guess it is a timeslicing protocol that I am really looking for now and it would have to fit into a stamp, be AM based and work at 2KHz data rates over at least a 20 meter area and cost·under $20 per unit (cost is for communication related parts only).
I hope this clears things up, sorry I wasn't more specific at the start but it has only been since I had these replies that I have thought it through further..
I will give further details about latency etc shortly as I still have a few considerations to resolve.
Thanks again for everybodies help (I am looking into all the links given to me)
All the best,
Mike.
Can you detail the 100 AM receivers you already have a bit more? Are they truly analog AM or carrier on/off. What is their banwidth? I believe you stated they were 418 MHz. Also you stated that as yet you had no transmitters.
So, as you properly deduced, for a transmitter to "get the token", or in other words, for it to know when to transmit, it needs to have a receiver attached. That would be the normal way of doing things. Unfortunately with low power RF (presumably your application), interference from other transmitters foreign to your system, or null-fade spots can cause a loss-of-token situation, temporarily interrupting all operations until it is somehow re-established.
What might be fun, and this is PROVIDED THAT your receivers have enough bandwidth, AND your data payload (data rate) is relatively light, such as 100 bits or less per second continuous for each transmitter, is to build a Direct Sequence Basesband Spectrum Spreader. This would have an effect similar to Direct Sequence Spread Spectrum (DSSS) radio transmission, but it would be done at a baseband frequency instead of at RF. That way it is quite within reach of SX speeds to do the necessary DSP work for restoring the basesand data.
In simpler terms, and IF it worked well enough, you could have many transmitters (perhaps 100) ACTUALLY transmitting simultaneously on the same frequency, and yet all independent of each other, eliminating all "token" and time slicing, or frequency division issues etc.
Some similarity to ZIGBEE without the "mesh" concept............hmmmmm perhaps a mesh could/should be considered.............way cool.
All probably quite inexpensive; just an SX and not much more, but probably quite a bit of work......
Youre lucky...........I'd LOVE to tackle an application something just like this.
Please reply to my questions above, I'd like to give all this some more thought, and please keep us informed as to where you're heading.
Cheers,
Peter (pjv)
Post Edited (pjv) : 1/3/2006 2:42:44 AM GMT
Pin No Pin Name Notes
1 RF +Vcc
2, 7 RF GND
3 DATA IN (Ant)
10, 12, 15 AF +VCC Can be ignored on versions without pin 10
11 AF GND
13 TEST POINT
14 DATA OUT Hmm, the DSBSS does sound interesting, what is the next thing to do if I wanted to go down this route?
To begin with my aim was to use up some old (redundant?) receivers but if this project is interesting enough I would consider purchasing some new stock.·
Perhaps I will build a token network straight away (since I expect to finish that quickly) with a low number of existing·receivers and a couple of transmitters performing·larger simulations. Then try some of your DSBSS ideas when new stock has arrived..
Please keep me informed of your views, I am very interested...
Cheers,
Mike.
Peter seems to be contemplating Zigbee as a solution; I've been thinking about this since his post. I think it might be the way to go. Although it's early days, i.e. not a lot of material yet on using it, I'm going to do a project to try and collect data remotely using Zigbee modules to connect to a Zigbee equiped boat as it goes past the sensor array. It provides for control, via a controller module, it allows communication between end devices (your bots), and as far as range, far off bots could communicate up the line to the controller in a router-type functionality. Another good thing: the RF transceivers are in the <$range, less in quantity, and it takes care of the RF in a module the size of a quarter or so. Antenna included! Range, about 80m fairly dependably (OR SO I"M TOLD). These are the 1mw modules, they make a 60mw which is ~ 300 meters. I'm going with Maxstream units (www.maxstream.net), but others are on the market. Just a thought...
Jim
Just one quick comment. Heterodyning. If you have more than one AM transmitter on at one time, even if it is just the unmodulated carrier, you will get an unwanted mixing of signals. The receiver cannot decipher the heterodyned signals. TDMA, CDMA are possible but you have to have tight control over the transmitters. Only one transmitter can be on at a time. FM has the option of the "capture" effect. AM does not.
Sofa
Well, your kinda right, specifically for strict AM. That's why I initially reacted negatively to the requirement of "100 transmitters on all at the same time".
But I was postulating an SX based pseudo random spreading generator that would make the CW carrier of a transmitter turn into an orthogonal frequequency mix, and hence be NOT AM in nature; somewhat more like FM. In fact it becomes very much like simple minded CDMA. The AM device would simply be a source of RF to be spread by the baseband, and because of that, your hetrodyning goes mostly away. And that's why I was asking about the receiver's bandwidth. Unfortunately receivers for on/off keying systems leave much to be desired, and the ones that Mike has may not be suitable.
The approach would have some similarities to Zigbee, but without the mesh concept, or its huge stack overhead.
As I said, this would be a very interesting project, and not ASSURED of success, but with a reasonable shot.
I have used an SX many times for generating the necessary spreading codes, and I suspect there is sufficient horspower to do the receiver's DSP work to collapse the desired code back to baseband data signals. From what I know, gut reaction tells me this should be do-able.......but surely a big bunch of work.
We'll have to see where Mike wants to go.
Cheers,
Peter (pjv)
An animated explanation can be found here:
http://www.techtionary.com/ ...choose Z for Zigbee.
Jim