ia4421 antenna
I wish to design a telemetry data collection system, that will pull data off vehicles to a base station connected to a PC. The units in the vehicle have to be very low cost, and I am thinking of implementing with some type of 433MHz FSK radio (there seem to be many good chips available). Now, what I am not sure of and would like to understand, is how to go about designing this, both from a RF point of view and also software.
The requirement is that once a vehicle enters a parking lot, the system will sense it, and download the data from it. This sounds like master/slave communications where the vehicle only answers when the hub speaks. But, the vehicle should also be able to send unsolicited data to the hub (which could in theory be implemented by polling also from the hub).
What I would like to understand is, once you pick up one of these ISM chips, what type of addressing scheme and/or transmission protocol has to be developed? Should FSHH be implemented?
Sorry about the so general nature of the question but my company wants me to do this all on my own (meaning not to contract out) and I have to learn...
Thanks!
Blapcb
Maybe i'm not the right person to give you all answers (I'm an RF IC designer) but I am interested into this subject as I am starting new product development process in sub 1 GHz low data rate field.
For RF issues, we have to separate air transmission (for thet you have to do a link budget) and pcb (antenna to IC connection).
Useful informations about protocol & pcb usually can be found in transceiver's company web pages as application notes. Read carefully this infos is my suggestion.
I have a question for you: once you will build your system from the beginning, you can choose to use 2 ICs (MCU & transceiver) or only one (some company sells MCU & transceiver in a single, very small solution). Can you comment and explain why you will prefer first or second solution?
Mazz
Hello,
I think, apart from the two solutions mazz suggested, there is also the option of using a RF transceiver module, e. g. EasyRadio http://www.lprs.co.uk. This would be generally interesting for low volume projects (below 1K items) where the RF design effort probably doesn't pay or for evaluation and fast prototyping.
Regarding the application, you already pointed out that communication would be bidirectional, utilizing transceivers at both sides. For the <1GHz range there are no standard protocols, I think (execpt for RFID protocols, which could be basically also used here). So you are free to design communication on your own. The RF channel implies a frame with a preamble, also a CRC could be regarded as part of the transport layer. At user layer you have a transparent data transport without addressing, each transmission is a multicast to all receivers.
You can assign different RF channels or transport coding to separate both directions, you could use an address field that allows individual access to devices. Generally you have to avoid telegrams that could cause simultanous answers.
Additional questions arise, if the application could need mutiple hubs or repeaters.
Reviewing your post, I think, from the said parameters, the application could also be unidirectional, using only transmitters at the mobile side, but there can be additional requirements.
Regards,
Frank
Maaz - to answer your question, I would prefer single SoC or Mu+Radio based on cost and performance parameters. There are very interesting SoC out there these days, but they do not all fit the bill. It "depends", so I don't think there is one answer. Cost is very important though.
But my question is another one, and I would appreciate if you or someone else could help me out (it is more basic).
Firstly, once again the objective: to design a very low cost unit (BOM < S10) that can communicate with a base station and send some data to it. Communication is basically two way.
What I want to understand is how to design/implement an appropriate communication scheme between the base and node units (a one-to-many scheme, or point-to-multipoint).
I would like to understand the design parameters and considerations that one has to look at for such an implementation. For example, how do you govern the communication between the base and the nodes? What if two units transmit at the same time?
I realize that this amounts to designing a wireless communication protocol and I guess that is what I am trying to understand (not the application protocol, which is on top of that, but the protocol that governs the wireless communication).
I hope I managed to clarify what I am looking for, sorry if it sounds confused...
Thanks!
Added after 31 minutes:
I would further like to add to my question.... Basically, it would be best/good to be able to chose different components as they become available and/or less expensive. After all, if the frequency and communication scheme are the same, why not? But, I think that many of the radios or SoC's, do not use "plain vanilla" schemes and often have something proprietary. So to what extent do I have to restrict myself to a single solution? Or can I use any two ISM band radios? Just thinking out loud....
By the way, here is a compendium of ICs I found that might be of interest...
- AMIS-52000 (AMIS)
- MLX71121, TH7121, TH7122 (Melexis)
- IA4420-DS, IA4421-DS (Integration)
- NA1TR8 (Nanotron)
- nRF905 (Nordic)
- XE1201A, XE1202A (Semtech)
I would consider TI/Chipcon CC1100 et al, too.
TI CC1111 may fit to your budget. You can build an application with few more parts.
On communication protocol side, I think I cannot help you very much: I think you have to make your software that manages also collisions, for example.
It depends, I think, on how many nodes you have to manage respect to available channels.
In CC1100 you have certain informations about LQI (link quality indicator), and, in case of fail, you can ask a re-transmission of data from your vehicle.
Make also calculation on probability of this event: having a certain number of available channels and using a very small duty cycle, this probability could also be very small.
Maybe you can get some ideas from sympliciTI network protocol (made for 256 nodes) or studying the shockburst protocol from Nordic (very funny names, isn't it?).
To give an answer to your last question, let's say...no. I think you need to go with one solution, there is no standard here, so compatibility of different radios is really difficult.
I hope it can help.
Mazz
p.s.: on Dec 2007 IEEE communication magazine there is an interesting article on wireless sensor network.
When I think of Chipcon ICs, where nearly everything is configurable, except for transmission datarate which is restricted to an integer part of crystal frequency, it should be able to copy other protocols, don't know how about the reverse case. Another question is, if there's any in using different chips in the design, at least when creating a product with a proprietary protocol.
The purpose of standardized protocols is mainly interoperatebility. It may seem if adopting a standard could ease your work, but that's often only half the truth, it can be more work to adapt your design to the standard and overcome it's hidden flaws.