Wireless needs a Brainstorm!

Wireless Design and Protocol Discussion

Moderators: seaton, strogg

Wireless needs a Brainstorm!

Postby Thonord » Fri Sep 21, 2007 3:58 pm

The new recruits in an armored division were given a tour of the base.
Propulsion said: We're the most important. If the engine or tracks fail you ain't no good to nobody!!
Comms said: We're the most important. If you can't communicate and coordinate, you ain't no good to nobody!!
Armor said: If the cannon fails, all you are is a 40 ton WalkyTalky.

So, in this case, with all due repect to the others - without us, what will they have.

We need to get cracking, because we are the ones that tell the others what they can do

And I have just shot myself in my foot with my signature. Ignore it :D

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 mattsteg » Fri Sep 21, 2007 4:20 pm

So far I like the idea of using a serial protocol for control and using wireless transport that's pretty much transparent. That way you can swap in different sorts of radios for people with different needs etc if necessary. The radios that I like the looks of are the XBee pros. They've got good power and range, along with performance that looks sufficient. They do a lot of the work for us, so that we can concentrate on the flash side of things. They can send messages to single recipients or send broadcast messages (so the ability to address individual receivers to make settings adjustments iplus the abiility to trigger everything at once are built-in). That way, the commander-receiver protocol can be kept simple and hassle-free (can basically write code and protocol for sending messages to one flash and get the multiflash features for nearly free). Sending 1 byte at 56kbps (half the rated interface speed) gives a delay of around 1/3500s - way quicker than sync speeds.

We can talk about rolling our own protocols and hardware etc, but at least read the XBee pro datasheet first. There's a lot of very convenient capability in that module for modest cost. It might be a lot of effort for little gain or even some loss imo. I think that such a solution is worth looking into. Yes, we get a lot of stuff that we'll never use, but most or all of the stuff that would slow things down significantly vs. a custom implementation looks to be disablable.

Also, regarding a standard serial-based protocol over some simpler, stripped-down bit-banging - for triggering serial is clearly fast enough (as long as you make sure nothing gets stuck in buffers or such) given the delay listed above (and you'd want some multi-bit signal anyway for reliability) any sync problems are going to be from processing, not transmission of the signal.
mattsteg
 
Posts: 26
Joined: Fri Sep 21, 2007 5:21 am

Micro controller

Postby Thonord » Fri Sep 21, 2007 4:57 pm

I just had a look at the xbee. Do you know the cost, say at 100 units?

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

Re: Micro controller

Postby mattsteg » Fri Sep 21, 2007 5:43 pm

Thonord wrote:I just had a look at the xbee. Do you know the cost, say at 100 units?

Tom
I've only seen the single-unit price of $32 for the pro and $19 for the lower powered non-pro ones. No idea how much they discount for larger orders.
mattsteg
 
Posts: 26
Joined: Fri Sep 21, 2007 5:21 am

Postby onelovephoto » Sat Sep 22, 2007 2:20 am

im no wireless expert, but has anyone thought of 802.11g or b technology? i brought this up in the thread but received no response.
onelovephoto
 
Posts: 3
Joined: Fri Sep 21, 2007 9:00 am

Postby seaton » Sat Sep 22, 2007 4:11 am

I like the look of the Xbee and we can offload quite a bit to them, i.e. we don't need to develop that lower level, all that is abstracted for us, Oh yes they are already FCC approved in a heap of countries.

But on the flipside they are fairly expensive and will add a fair chunk to the final unit costs.

It would seem we have some ppl with RF experience in the role call. so we may be able to roll our own hardware and implement something like the zigbee stack or even our own for that matter in firmware.
seaton
Site Admin
 
Posts: 134
Joined: Mon Jan 23, 2006 8:48 am
Location: Bunbury, Western Australia

Postby Breamer » Sat Sep 22, 2007 6:47 am

I thought about this a few months ago and did some research on it. Whilst iTTL,eTTL functionality would be nice I would design a simple trigger first and forget about iTTL,eTTL for now. Why? Well I don't know iTTL and eTTL protocols. eTTL I believe has 5 pins, so 5 data lines. Implementing this serially would make it hard to get a decent sync speed. Implementing in parallel makes everything more complicated.
KISS sounds good to me.

After a spec. is agreed upon I think the next step is to pick a frequency range. Each has advantages/disadvantages.

I would then suggest sourcing RF chips. There are plenty around and they can be cheap (e.g. 5$). THe final chip selection will be dependent on many things including chosen modulation type etc etc.

But to get the ball rolling we really need to have a basic idea of what we need to achieve and that seems to be being trashed out in the other groups.

FCC and European licensing is also an issue. I believe it costs $5000US for a device to be tested.
Breamer
 
Posts: 4
Joined: Sat Sep 22, 2007 6:16 am
Location: Sacramento

Postby brittonphotography » Sat Sep 22, 2007 8:38 am

need solution that is already FCC approved
lead time and investment are too much to go through testing for us.

i have email into xbee for pricing on volume discount
brittonphotography
 
Posts: 25
Joined: Fri Sep 21, 2007 8:52 pm

Different approaches

Postby Thonord » Sat Sep 22, 2007 4:09 pm

1.
Robustness in a wirelesslink is taken care of in layers 2 and 3. Layer 1 can be pretty dumb and simple.
I asume that the xBee and the Nordic chip contain atleast those three layers.

I don't think we can equal the quality of those products.

However, ours is the question of the "right" quality, an this is where cost is part of the equation.

Our thingy will contain a AT91SAM7X or equivalent. If we move layers 2 & 3 into it, we might use the same tx (R433A) and rx (don't remember) as the eBay trigger. We will have a significant saving that way, ofsett by additional lines of code.
Is this approach worth consideration.

2.
FCC approval. Is it the whole thing that needs approval or is not necessary as long as an approved tx/rx is used.

3.
802.11. I dont know enough about it, but very interesting.
Should be investigated.

My two bits worth (as in US currency)

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 Jon » Sat Sep 22, 2007 10:55 pm

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.
Jon
 
Posts: 1
Joined: Sat Sep 22, 2007 10:51 pm

Postby brittonphotography » Sat Sep 22, 2007 11:45 pm

we may not need to produce these ourself, but it could be very helpful to people to have completed parts such as the logic board and housing sold as a kit.

Could we sell a mostly completed kit where the consumer would only have to solder a couple points to make it work and we would get out of fcc testing?

still good to pick parts we know are fcc cleared already though so we don't go interfering with stuff.
brittonphotography
 
Posts: 25
Joined: Fri Sep 21, 2007 8:52 pm

Postby seaton » Sun Sep 23, 2007 5:11 am

I agree in that the RF side should already be FCC approved, that way we can just not have to worry about it, I'm aware that the final product would also need to be FCC approved if sold as a completed product, however as a kit it I don't think it will require it.

There are plenty of kits on the market that have no approval.
seaton
Site Admin
 
Posts: 134
Joined: Mon Jan 23, 2006 8:48 am
Location: Bunbury, Western Australia

Postby seaton » Sun Sep 23, 2007 5:48 am

Breamer wrote:I thought about this a few months ago and did some research on it. Whilst iTTL,eTTL functionality would be nice I would design a simple trigger first and forget about iTTL,eTTL for now. Why? Well I don't know iTTL and eTTL protocols. eTTL I believe has 5 pins, so 5 data lines. Implementing this serially would make it hard to get a decent sync speed. Implementing in parallel makes everything more complicated.
KISS sounds good to me.


I think that out of the 5 pins, only 2 are really used for digital I/O, one is the strobe trigger other is data clock. not too sure of the last, but looks lke beam assist enable. so in reality only the commands and data need to be handled by the network, all the rest are handled by our units. Our protocol could be fairly simple and leave the units personality to bridge to specific protocol. I understand that there will processing latency but I've not taken that into account.

e.g. our protocol could be made up of 3 bytes, command, data, checksum, this can be uused to set up power levels etc, while the trigger could be a magic byte that is a combination to keep overheads low, a four bit ID and four bit checksum.

Assuming the most units we can have is 15

IDIDXXXX

IDID = dest unit id FF=all devices
XXXX = checksum

Isn't the most important command the trigger command, so yes this one needs to be kept within sync timing, but the others are setup before the shot so don't necessarilarly need to be within max sync timing. I know that if shooting at a large FPS then this may be effected, but inreality lighting would not probably change that much between frames so would keep the same lighting data as far as the strobes are concerned. Say you are shooting at 12fps then that is about 82mSec between frames, I would think that woudl be plenty of time to send data over the wireless as a broadcast to all the strobes somewhere within the, 1/1000th window, maybe 1/1000th may be out of our reach and we may need to relax this?

I guess this is the type of specs we need to nail down.
Max Sync Speed and Max frame rate we want to achieve easily and effectivly, not to mention as cheap as possible.

Or do we look at say the XbeePro or whatever off the shelf device we choose and accept the Max Sync speed and frame rate that it will allow us, then live with that as long as they are acceptable?

Q: What are the max specs in terms of sync speed and frame rates for the pocket wizards? I have not looked at these devices so I don't really know, but seem to be popular, can anyone shed some light?

Stephen...
seaton
Site Admin
 
Posts: 134
Joined: Mon Jan 23, 2006 8:48 am
Location: Bunbury, Western Australia

Postby TwoLeftFeet » Sun Sep 23, 2007 6:17 am

So, should wireless layer mean just the physical layer? Or more?

I think 802.11 links would be "cool". However, what would those chipsets do to the cost? Also, is it overkill for this type of application?

Anyway, I think a discussion of what the wireless layer entails is important.
TwoLeftFeet
 
Posts: 12
Joined: Sun Sep 23, 2007 5:52 am
Location: Toronto, ON

Wireless layer

Postby Thonord » Sun Sep 23, 2007 9:37 am

TwoLeftFeet
The RF part (actual tx and rx devices) are the physical layer and are " fire and forget".
An intelligent device, like xBee is like TCP, it covers more than 1 layer.
I know,I know - bad analogy!
Yes discussion is required, but I think we are in the Brainstorming mode right now.
Throw your thoughts int as they arrive.

Then we compile and digest and start to make descissions.

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

Next

Return to Wireless

Who is online

Users browsing this forum: No registered users and 2 guests

cron