7
\$\begingroup\$

I am prototyping a potential future product, and I am struggling to get my head around the complexities of using Wi-Fi components in it.

I understand that Wi-Fi (or 802.11x) is primarily just a 2.4 GHz signal modulated in a specific way, as is Bluetooth and ZigBee. Is it possible therefore to just use a basic 2.4 GHz transceiver in my project and configure it for the 802.11 protocol rather than having to use some expensive branded one that comes pre determined? Or is it not as simple as that?

I presume in industry, when doing design for large production runs, something similar takes place? Or do they indeed have to use a transceiver that is preconfigured to be Wi-Fi? All the pre-configured components I have found so far seem quite expensive (even when purchased in volume) to make a commercial design viable.

I have looked at devices such as the Lantronix WiPort and the Roving Network WiFly GSX and comparing them to devices such as the Microchip MRF24J40 - is it possible to use a Microchip MRF24J40 and then configure the rest of the hardware and software to enable the device to work on a Wi-Fi network?

For further detail, here is what I want to do at the most basic level:

Part 1: I want to make a patch that contains a small buzzer, an accelerometer, a PIC and a Wi-Fi transceiver that can be stuck to an object. This patch would be able to communicate with "The Cloud" though the users' home Wi-Fi network. When the accelerometer detects movement, the PIC will send a message, via the Wi-Fi link, to a server in the cloud to register that movement along with a time stamp.

Part 2: Through a web interface I want to be able to send a message to the patch setting the buzzer to make a sound next time the accelerometer detects a movement.

I have already made a rudimentary version of the system which works over a 433 MHz RF link and a laptop serial port with a local piece of software running on my laptop. I know how to do the web coding (PHP and MySQL), but it's the replacing the RF link with the Internet that is causing the issue.

\$\endgroup\$
  • 1
    \$\begingroup\$ This buying question is off-topic and will be closed! \$\endgroup\$ – Leon Heller May 3 '11 at 16:35
  • 5
    \$\begingroup\$ I don't think this is really a buying question though it was phrased that way. I would suggest you reword it w/o words like, buy, cost, price, etc... and the the word/question police will stop nagging. \$\endgroup\$ – kenny May 3 '11 at 16:59
  • 1
    \$\begingroup\$ Thanks for the feedback chaps, I have read a couple of the meta questions related to shopping type questions and hopefully I have reworded it well enough to focus on the technical aspects of the problem, cheers \$\endgroup\$ – SimonBarker May 3 '11 at 21:16
  • 1
    \$\begingroup\$ @Simon - Nicely done! The initiative you showed to read up on Meta and fix the question as well as you did is much appreciated. I just had to touch up the title a little, but I think it's ready to be reopened now. \$\endgroup\$ – Kevin Vermeer May 5 '11 at 13:32
  • 1
    \$\begingroup\$ However, I'm curious what transceivers you've found that are so expensive. Are you comparing, say, the Lantronics WiPort and Digi Connect to stuff like the Microchip MRF24J40? That's not apples-to-apples at all! \$\endgroup\$ – Kevin Vermeer May 5 '11 at 13:36
4
\$\begingroup\$

802.11x is significantly more complicated than Zigbee, and the TCP/IP stack you need to make it work on top of all that is similarly complex. If you purchase something like the Lantronix WiPort or Digi Connect WiMe to make 802.11x networking as easy as communicating over a serial port, you're paying for a lot of stuff (there's a whole ARM based server in those modules!) that you don't need if you're willing to do a lot of RF hardware design and software integration.

If you're interested in taking that path, you should get some background understanding of the IEEE 802.11x network architecture. This is one of just a couple IEEE standards available for free through the IEEE-Get program.

Once you've got an overview of the networking system, look up the Maxim MAX283X line of transcievers. From the datasheet,

The fully integrated transceivers include a receive path, transmit path, voltage-controlled oscillator (VCO), sigma-delta fractional-N synthesizer, crystal oscillator, RSSI, PA power detector (MAX2831), temperature sensor, Rx and Tx I/Q error-detection circuitry, basebandcontrol interface and linear power amplifier (MAX2831). The only additional components required to implement a complete radio front-end solution are a crystal, a pair of baluns, a BPF, a switch, and a small number of passive components (RCs, no inductors required).

That's about as close as you'll get to a generic 2.4GHz transceiver that can be sanely integrated into an 802.11x network.

They cost about $5 in small quatities at the moment. As you requested, these chips only implement the PHY layer of the protocol. You still need to handle the data link layer (MAC and LLC), network layer, and transport layer on top of that before you can begin communicating at the application layer level.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Wow thank you for your answer reemrevnivek - it's what I am after. So there is a lot more to this issue than I first thought then!?!? At least I can now see why the Lantronix etc stuff is so pricey. I feel like yours and Davids answer has helped me lift the lid on a very large can of worms. I will have a good read through those documents you have pointed me to and then see about getting stuck into the maxim hardware. \$\endgroup\$ – SimonBarker May 5 '11 at 15:31
2
\$\begingroup\$

You don't say exactly what you want to do with Wi-Fi. You use the word "transceiver", but I've learned that people use that term somewhat generically. Thus, forgive me if the rest of the answer isn't quite what you were looking for.

To directly answer your question (paraphrased), "Can the Microchip MRF24J40 be made to do IEEE 802.11a/b/n?", the answer is no. It's made to do IEEE 802.15.4, or ZigBee, and cannot be forced through software or hardware to do Wi-Fi.

But to address the bigger issue: Unless you're an expert in 802.11 there is little or no chance that you'll be able to use generic chips and make it do 802.11. The RF modulation and software protocols alone are quite challenging-- enough for people to make just that into a career.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ thanks very much. I have added my intended use in the question. I didn't initially as I thought it may be to detailled. Your answer has helped start to clear up my confusion - hopefully my edit gets my intention across better \$\endgroup\$ – SimonBarker May 5 '11 at 15:27
  • \$\begingroup\$ In general, at this point, 802.11xx is available to high volume manufacturers as a small RF sub-block with a corresponding digital controller. TI/Broadcom//CSR/etc all have highly integrated parts that aren't available to the small player. You'll need to go through a 3rd party company like Roving Networks to get access to the parts, and you'll end up paying for it. If you get to the point that you'll be ordering a couple of million parts a year, then you'll get access to the lower pricing/higher integration parts. \$\endgroup\$ – rfdave Aug 10 '12 at 1:30
1
\$\begingroup\$

Is it possible therefor to just use a basic 2.4GHz transceiver in my project and configure it for the 802.11 protocol rather than having to use some expensive branded one that comes pre determined?

You seem to be basing this on a backwards assumption. There certainly are universal (within some bandwidth) RF devices - that's basically what a software radio is, and they are available in reconfigurable forms.

However, they aren't cheap.

What is cheap are the highly specialized, huge volume production devices intended for consumer products. These generally have their flexibility limited both by optimization for a given target (frequency, compute power for digital modulations, etc), and manufacturer's desire to not release more programming data than absolutely needed for the intended application. An additional problem to small quantity users is that it can be quite hard to buy chips unless you are purchasing in massive quantities.

Probably your little tags won't have USB host capability, so leveraging the lowest prices of generic USB wifi adapters won't be an option, so you'd be in the next class of embedded modules that talk spi or asynchronous serial or similar.

| improve this answer | |
\$\endgroup\$
1
\$\begingroup\$

found some interesting articles on

warpproject.org/trac/wiki/802.11/PHY

and

www.eirp.org/webtut.pdf

based on the '2.4 ghz transceiver' notion we are probably looking for a 802.11 PHY and we try to implement the 802.11 MAC as well as the IP stack in software. i.e. the PHY give us the bits & we deal with the data frames for 802.11 MAC & IP stack - this is pretty complicated in itself

the 802.11 PHY is apparently complex to implement (see the 2nd link above) 1'd need to cater for FHSS (4GFSK, 2GFSK), DSSS (DBPSK, DQPSK, DQPSK-CCK, DQPSK-PBCC encoding/modulation etc)

however, if 1 is keen on on the heroic efforts to do a 'software PHY' as well i.e. decode all that FHSS, DSSS in software, e.g. using FFT, DSP algorithms etc there are some interesting chips deemed as 'RF frontends' e.g.

www.maximintegrated.com/en/products/comms/wireless-rf/MAX2830.html ww1.microchip.com/downloads/en/DeviceDoc/75028A.pdf

in theory, if we have 'RF frontends' i.e. all analog signals, we mix them say produce IF may be add some PLL stages etc we can use some superfast ADC and ADC convert them all to digital & we do software DSP, FFT & decoding to convert that into bits (i.e.the job of a PHY), then we take those bits assemble them into frames (the job of a MAC) and we then take the frames and handle them as IP datagrams.

i'd guess if that's possible, there might be a possibility to do 2.4ghz anygram it'd seem the first link

warpproject.org

is trying to do just that - a software radio using FPGA :)

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

I was also looking at something similar. If you want to run 802.11 and 802.15.4 on a single transceiver / chip

It would in someways impossible if the chip cannot support DSSS and QPSK. Even if it does you would be looking at rewriting 802.11 stacks to make it work.

look at the following product for a better and a easy implementation of two protocols on a single chip

GainSpan G2000 SoC

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.