Wireless needs a Brainstorm!

Wireless Design and Protocol Discussion

Moderators: seaton, strogg

Postby JonSenior » Sun Sep 23, 2007 12:46 pm

I'm very much in favour of an off-the-shelf solution that does as much of the work as possible. That old chestnut about re-inventing the wheel. It also makes it easier to offer the plans and for people to build the units themselves. If we start moving onto custom designed RF transmitters and receivers, the work load and end-user testing gets much harder.

Better to have a few (slightly pricier) off-the-shelf modules and some simple glue for the logic, than a complex design which is far harder to debug.

I see one micro, one transmitter, power supply control and a few headers for connecting the off-board stuff like displays, switches and connectors.
JonSenior
 
Posts: 60
Joined: Fri Sep 21, 2007 7:27 pm
Location: Paris, France

Postby Gordon » Sun Sep 23, 2007 2:15 pm

off the shelf for the physical and link layers seems initially the best idea - particularly if you want more than a handful of people to actually be able to build this up themselves.

It wouldn't be too hard to come up with a simple RF trigger from components that most could solder together, but if this is to do something more interesting and support more functionality, abstracting the lower levels away and buying FCC approval off the shelf is going to make life a lot easier to get to something that works, quickly.

Zigbee/xbee isn't the only option out there for this sort of low bandwidth, medium range wireless control, but it does seem to be a good option.

First off, we could just strip back from zigbee and do straight 802.15.4. I doubt we really need all the network/ security and application layers that zigbee provides - a straight physical and MAC protocol implementation might well be enough.

Then there's a question of which zigbee protocol - standard zigbee (2.4GHz band, 250kb/s, up to 50m range or the low power, alternative zigbee - over 100m range ~40kb/s bandwidth and in the 868Mhz or 915MHz bands (Europe vs US there) Freescale Semi sell dev kits and chipsets, as do xbee, atmel, TI/Chipcon.

Wireless USB might be an option - longer range wireless USB works up to about 50m or more at a reasonable data rate for this project. Cypress Semi sells dev kits for these as well as single chip solutions.

Another option is Z-Wave (Zensys http://www.zen-sys.com/ ) 9.6kbit/s should be more than enough for this I think. Range is over 100m (UHF band) Intel & Cisco are backing this protocol so it has some legs.

Or we could look at lower level protocols and get something in the ISM band, high range, reasonable data rates, but physical layer only so we could implement our own protocols, error correction etc. Less issues with delay etc, more work to implement (though it isn't that difficult)

Cirronet, Lemos, Linx, Maxstream, radiometrix all supply chipsets, kits etc for ISM. Micrel and Maxim, along with Analog Devices are suppliers/ manufacturers of chips that work in this band. (315, 492, 915 MHz range typically)
http://www.maxstream.net/
http://www.micrel.com/products.do
http://www.micrel.com/_PDF/other/RF%20Solutions.pdf
Gordon
 
Posts: 4
Joined: Sun Sep 23, 2007 1:59 pm

Postby matt_e » Sun Sep 23, 2007 6:45 pm

Hi, I'm new here, but funnily enough this idea's been kicking around in my head for ages now.

Has anybody looked at the Nordic chips - sparkfun has a great little transceiver board - http://www.sparkfun.com/commerce/product_info.php?products_id=705 - 80m range @ 250kbps, physical & mac-type protocol on-board, with an SPI interface for $19.95
matt_e
 
Posts: 8
Joined: Sun Sep 23, 2007 6:39 pm
Location: Seoul, South Korea

Postby TwoLeftFeet » Sun Sep 23, 2007 10:37 pm

Well, I'm not a big RF guy like the rest of you, just whatever I learned at school 6 years back... I'm wondering if we can NOT do ISM band :D Too much interference there. I say we should aim for a solution that will allow at least 100m of interference-free range in both outdoor and indoor situations.

Basically something that can support 9.6kbps datarate may be enough, but I don't really know, I guess it's something that'll have to be decided at a higher level requirement... But let's keep kicking ideas around...
TwoLeftFeet
 
Posts: 12
Joined: Sun Sep 23, 2007 5:52 am
Location: Toronto, ON

What prevents successful RF transmission

Postby Rudeofus » Mon Sep 24, 2007 12:23 am

Hi folks, I know I'm late to the show, but I would like to point out some issues we need to address before we settle on any RF technology, regardless of whether we use a ready system like those Xbees or whether we hack up something from scratch.
The key question is: why do those cheap triggers works so poorly? What makes them misfire so frequently (at least according to the threads, I've also read reviews which were quite fond of them ...) ?
Let me quickly rehearse the challenges we face:
  • We are not the only ones using the frequency, whichever we pick. And we do not know when and with which kind of signal anyone else will interfere
  • The 1/r^2 law only works in outer space or in the flat desert. As soon as we have buildings around us, or even worse, if we happen to be in one, reflections of the radio waves will cause erratic wave propagation. The laws of physics do still apply, but the exact location of all reflectors can't be predicted/calculated
  • Apart from specific interferers (intentional transmitters which happen to operate at the same frequency and time as we are), we also face considerable noise levels, especially in urban environments. This noise does not only limit the range or successful transmissions, it may also trigger false triggers.
  • We have a maximum trigger delay in the range of 1 ms. This is substantially shorter than the time frame in which we can hope for unfavourable radio conditions to improve in case one of the above listed issues hits us.
If we want to have better performance than ebay triggers, we need to make a clear case how we are going to build something substantially better. Let's assume for a while that the designers of these ebay triggers know their trade and did their best given the bill of materials and the manufacturing costs they were allowed to run up.
As one initial experiment I suggest we look at two existing designs: The ebay triggers and the Xbees.
  • It would be great to reverse engineer the schematics of those ebay triggers, so we can learn what can be improved (using more expensive components).
  • As far as the Xbees are concerned, we need to know whether their protocol guarantees sufficiently short delays under adverse conditions. It is hellaciously difficult to implement a smart retransmit protocol at these data rates which doesn't delay the trigger too much, and I'd be positively surprized to hear that they can do it reliably.
Rudeofus
 
Posts: 27
Joined: Sun Sep 23, 2007 11:28 pm

Postby Breamer » Mon Sep 24, 2007 7:03 am

Just a few notes:

On the frequency band, my personal feeling (not backed up by calcs) is that 2400 MHz is too high for a 100m distance requirement - the extra path loss and limited allowable transmit power are big trade offs in the link budget. As I said I haven't calculated this though but can do.

The less than 1GHz ISM bands can suffer from interference from harmonics from other licensed bands e.g TV but this can be minimised by selecting a channel on the edge of these harmonic bands. As for interference from other devices - well it's pretty hard to mitigate against that unless one goes back to using a spread spectrum technique. The good news that is in the 400MHz and 300 Mhz ISM bands, transmit time is limited to less than 5 seconds.

Back to the kit issue, previous posters were correct and I thank them for pointing this out. If this is marketed and sold as a kit with a transmitter then FCC approval is required.

So for me the potential outcome of this is a set of specifications, designs, bill of materials, potential sources, assembly and installation instructions.

Also, I'm not sure how easy it is to download code to a micrprocessor btu this could potentially be a difficult task for a DIY'r that wants to take this project on.

Cheers
Derrick
Breamer
 
Posts: 4
Joined: Sat Sep 22, 2007 6:16 am
Location: Sacramento

Postby Rudeofus » Mon Sep 24, 2007 11:58 am

On the frequency band, my personal feeling (not backed up by calcs) is that 2400 MHz is too high for a 100m distance requirement - the extra path loss and limited allowable transmit power are big trade offs in the link budget. As I said I haven't calculated this though but can do.


Just FYI: Bluetooth class 1 does 100m, older wireless LANs stretched up to 300m. It can be done and has been done 10 years ago, but only with spread spectrum (you need to do spread spectrum so you are allowed to broadcast @ 100mW)

The 2400 MHZ band is the lowest frequency band which can be used world wide (France and Japan only with restrictions) and the lowest frequency band where you can implement a meaningful form of frequency hopping (which is the best way to combat fading effects). Also the antennas for 2.4 GHz are neat and small.

The biggest draw back for 2.4 GHz is that it's pretty much impossible to handle at amateur level, you'd require modules which can be simply plugged into your circuit and are already FCC/ETSI/ARIB conformant.

Another draw back is the common availability of wireless LAN networks which will cause considerable interference ...

Just my 2 Eurocents ...
Rudeofus
 
Posts: 27
Joined: Sun Sep 23, 2007 11:28 pm

Postby chrisb » Mon Sep 24, 2007 1:33 pm

Jon wrote:
FCC approval. Is it the whole thing that needs approval or is not necessary as long as an approved tx/rx is used.


We're not talking about mass producing these are we? I thought the point was to create an easy to follow parts list and set of instructions for people to build them themselves.

If people are building them themselves, we don't need to get FCC testing. We need to be sure that we're not interfering with anything, but if we're not marketing it or producing it ourselves, we're not responsible for testing. That's the point of opensource/homebrew.


This is wrong, the unit will still need FCC Part 15 testing and you are responsible for testing it even if you provide it as a kit - you are providing a product to an end user.

If the RF transmitter/receiver is already approved, then the process is easier as you don't have to apply for an FCC ID and obtain radio approval for the RF portion.

Regardless of the approval on the RF side, you are still required to get the whole unit tested to FCC Part 15 (this would be a Part 15 Class B device).

-chris
chrisb
 
Posts: 7
Joined: Mon Sep 24, 2007 1:04 pm

Postby strogg » Mon Sep 24, 2007 1:38 pm

What if you provide it as a partlist, a PCB layout diagram, a schematic an dsome instructions? Is FCC testing still required if there's not a physical product?
strogg
 
Posts: 19
Joined: Thu Sep 20, 2007 8:31 pm

Postby chrisb » Mon Sep 24, 2007 2:09 pm

strogg wrote:What if you provide it as a partlist, a PCB layout diagram, a schematic an dsome instructions? Is FCC testing still required if there's not a physical product?


If you provide it this way, then technically, each person is responsible for getting it tested - although practically speaking no one will (and the FCC realizes this).

However, lets stop for a second and think about your end product and the size. Providing it as suggested above will mean anyone wanting to build this will have to obtain everything themselves including PCB, components and enclosures. I don't believe people can build this for the costs being proposed if they have to purchase these items by themselves.
A second thing to consider is the soldering ability of the people who you are expecting to build this. You will be looking at a small footprint, how many people will build this when they realize they are going to have to solder 201 chip resistors?

-chris
chrisb
 
Posts: 7
Joined: Mon Sep 24, 2007 1:04 pm

New Nordic USB dongle

Postby Thonord » Mon Sep 24, 2007 2:56 pm

Naturally I'm a supporter of using a Nordic chip, but even if I wasn't I'd look closely at this.
http://www.nordicsemi.com/files/nRF24LU ... f_v1_0.pdf
If I'm making a fool of my self, its because I know nothing about ASM 2.4GHz and dongles.

I'll give them a call tomorrow, our time.

Tom
Ppl who agree need normally not reply, those who disagree or have questions do.
Or - just ignore me.
Thonord
 
Posts: 50
Joined: Fri Sep 21, 2007 2:32 pm
Location: Norway

Postby JonSenior » Mon Sep 24, 2007 3:08 pm

chrisb said:
I don't believe people can build this for the costs being proposed if they have to purchase these items by themselves.
A second thing to consider is the soldering ability of the people who you are expecting to build this. You will be looking at a small footprint, how many people will build this when they realize they are going to have to solder 201 chip resistors?


I disagree. The per-unit price of the xBee module is around $17. The Nordic one is a similar price (I believe). Simple 433 or 868(?) Mhz receiver / transmitter units can be had for less but put more load onto the chip. A micro with enough ports and power to fulfill our requirements will be pocket money.

Modern micros are designed to be highly self-sufficient, often providing internal pull-up or pull-down resistors and containing reset circuitry and even clock generators. The actual "glue" in many cases is just wire.

My suggestion would be that we provide code, design and a proven working PCB layout. If enterprising individuals in various countries decide to produce a small run of PCBs, or if a band club together to get a discount on parts then so be it, but the aim should not be (In my opinion) to form a not-for-profit competitor to the PW. It should even be possible to provide separate SMT and DIL designs so that those who feel confident soldering surface mount can, and those that don't can just plug in the big chips.

I see no reason why the enclosure has to be complex. A hotshoe to PC (or whatever) adaptor will get a trigger signal off the camera and a small plastic box with a trigger-in socket can contain the electronics. The cost of small scale tooling of a robust custom shoe mount box would probably exceed the one-off component costs of this project.

Search the strobist forum for "mod" or "modding" to see how far a large number of people are prepared to go.

And remember that with the right licence it remains open for a commercial entity to use the design to produce a final product on the shelf with all necessary legal coverage.

And remember that the FCC is a USA thing. While I appreciate that a large proportion of the target audience will be in the United States, FCC compliance is not required in Europe... or (I presume) Asia.

Jon
JonSenior
 
Posts: 60
Joined: Fri Sep 21, 2007 7:27 pm
Location: Paris, France

Postby Firebird » Mon Sep 24, 2007 3:23 pm

chrisb wrote:

A second thing to consider is the soldering ability of the people who you are expecting to build this. You will be looking at a small footprint, how many people will build this when they realize they are going to have to solder 201 chip resistors?

-chris


This is absolutely true. 99 percent of the people here won't be able to build a surface mount PC board, and probably 90 percent won't be able to build a thru-hole PC board. I think the TTL thing is most Pie in the Sky. I'm going to try out the TX/RX modules I purchased to see if it's possible to do the IR/VL to RF, RF to IR/VL thing to transfer the manual power control pre-flashes of the Nikon CLS system. It this works with the hardware I've ordered a simple kit may be possible. If this doesn't work out, I'm done. I've already got 2 TX and 5 RX ghetto trigger units and where I operate they are 99.999999 percent effective.
Firebird
 
Posts: 6
Joined: Sat Sep 22, 2007 7:38 pm
Location: East Coast

Postby chrisb » Mon Sep 24, 2007 3:35 pm

JonSenior wrote:And remember that the FCC is a USA thing. While I appreciate that a large proportion of the target audience will be in the United States, FCC compliance is not required in Europe... or (I presume) Asia.

Jon



FCC compliance is a USA thing agreed, but there are requirements for Europe and Asia, European compliance is mandated by the EMC and RTTE Directive, Asia is regulated by individual countries, Australia - ACMA, Japan - VCCI.

-chris
chrisb
 
Posts: 7
Joined: Mon Sep 24, 2007 1:04 pm

Postby brittonphotography » Mon Sep 24, 2007 7:49 pm

i have quotes for xbee stuff coming, sales rep called me today
xbee= indoor range 40m, outdoor range 100m
xbee pro = indoor range 100m outdoor range i forget maybe 300m

xbee and xbee pro have same pin out. so we could use either and swap for a higher price for longer range in the same housing

coming out with a 900mhz soon with even longer range same pinout...

will update later with more info.
brittonphotography
 
Posts: 25
Joined: Fri Sep 21, 2007 8:52 pm

PreviousNext

Return to Wireless

Who is online

Users browsing this forum: No registered users and 1 guest

cron