Da-Doo Ron RONJA
February 7, 2007 8:45 AM Subscribe
RONJA is an optical networking device that can be built by nearly everyone, using readily available components and using only free software.
Pretty cool. These things are a pretty obvious idea, but I didn't know anyone had done it... is there any advantage to doing this rather then using a directional wifi antenna?
posted by delmoi at 9:34 AM on February 7, 2007
posted by delmoi at 9:34 AM on February 7, 2007
very reminiscent of Lichtsprechgerät.
quite cool, I may have to build one...
posted by dorian at 9:58 AM on February 7, 2007 [1 favorite]
quite cool, I may have to build one...
posted by dorian at 9:58 AM on February 7, 2007 [1 favorite]
is there any advantage to doing this rather then using a directional wifi antenna?
For one, the radio spectrum has a ridiculous amount of governmental oversight.
posted by cowbellemoo at 10:06 AM on February 7, 2007
For one, the radio spectrum has a ridiculous amount of governmental oversight.
posted by cowbellemoo at 10:06 AM on February 7, 2007
It'd be a lot harder to intercept or break into, as well... it would require specialized equipment that most people won't have, AND line of sight. Put solid encryption on that link and it would be nearly as secure as running a physical wire.
You could also do numerous point-to-point applications in a very small area, since each carrier beam is so narrow.
posted by Malor at 10:11 AM on February 7, 2007
You could also do numerous point-to-point applications in a very small area, since each carrier beam is so narrow.
posted by Malor at 10:11 AM on February 7, 2007
is there any advantage to doing this rather then using a directional wifi antenna?
If someone is going to be sniffing your packets, you've got a pretty good idea who -- or at least where -- they are. Looks like the range might be a little better as well: Twibright are claiming 1.4KM on a standard setup. Not a lot of hard claims as to cantenna distances but I found one saying they got some at what they figured to be a bit under 2KM with significant degredation in signal. Presumably one could boost the laser wattage and get some significant range improvements.
posted by Ogre Lawless at 10:20 AM on February 7, 2007
If someone is going to be sniffing your packets, you've got a pretty good idea who -- or at least where -- they are. Looks like the range might be a little better as well: Twibright are claiming 1.4KM on a standard setup. Not a lot of hard claims as to cantenna distances but I found one saying they got some at what they figured to be a bit under 2KM with significant degredation in signal. Presumably one could boost the laser wattage and get some significant range improvements.
posted by Ogre Lawless at 10:20 AM on February 7, 2007
is there any advantage to doing this rather then using a directional wifi antenna?
1) Gain. It's vastly easier to focus 875nm light than 12cm Microwave, so getting high-gain beams is easier. The power used to modulate the signal is tiny -- the LED drops 1.7V, and forward current is 100mA. They get the full 10Mbps at range, where Microwave based WiFi drops speed rapidly.
2) Security -- the beam path is far narrower, and you'll know the moment something intercepts the beam.
3) Interference. LEDs aren't known for causing EMI issues.
The thing I'm impressed with is how they're driving the LED to generate the signal. They've come up with a very clever way to get 16MHz out of an LED that's specified as having a 40ns rise/fall time, which implies 9MHz bandwidth. Most LED implementations have ranges in the tens of feet and bandwidth in the 100kbps range. They've gotten a full order of magnitude more bandwidth, and two orders of distance. Nice, nice hack.
The big win is if you can see it, you can make the link. The bad this is there's lots of things that can change what you can see. Deeper into the IR would help with fog and rain, but that's the realm of lasers. Lasers are the right answer for this on technical grounds, but there's real safety issue once the power starts climbing. However, certain diode lasers switch vastly faster than IR leds (VCSELs leap to mind,) which is why 1GB Ethernet is pretty much the *slowest* fiber optic link out there, so a laser could vastly increase the bandwidth.
Free space is hard, though -- thing move in free space, and you need a really bright light if something like a truck gets in the way....
posted by eriko at 10:25 AM on February 7, 2007
1) Gain. It's vastly easier to focus 875nm light than 12cm Microwave, so getting high-gain beams is easier. The power used to modulate the signal is tiny -- the LED drops 1.7V, and forward current is 100mA. They get the full 10Mbps at range, where Microwave based WiFi drops speed rapidly.
2) Security -- the beam path is far narrower, and you'll know the moment something intercepts the beam.
3) Interference. LEDs aren't known for causing EMI issues.
The thing I'm impressed with is how they're driving the LED to generate the signal. They've come up with a very clever way to get 16MHz out of an LED that's specified as having a 40ns rise/fall time, which implies 9MHz bandwidth. Most LED implementations have ranges in the tens of feet and bandwidth in the 100kbps range. They've gotten a full order of magnitude more bandwidth, and two orders of distance. Nice, nice hack.
The big win is if you can see it, you can make the link. The bad this is there's lots of things that can change what you can see. Deeper into the IR would help with fog and rain, but that's the realm of lasers. Lasers are the right answer for this on technical grounds, but there's real safety issue once the power starts climbing. However, certain diode lasers switch vastly faster than IR leds (VCSELs leap to mind,) which is why 1GB Ethernet is pretty much the *slowest* fiber optic link out there, so a laser could vastly increase the bandwidth.
Free space is hard, though -- thing move in free space, and you need a really bright light if something like a truck gets in the way....
posted by eriko at 10:25 AM on February 7, 2007
RONJA is one letter removed from being my real name. Weird.
posted by grapefruitmoon at 10:27 AM on February 7, 2007
posted by grapefruitmoon at 10:27 AM on February 7, 2007
This looks like it could be cool, but I'll settle for building a DIY Pringles can antenna if it will allow my shit wireless hub to project a decent signal as far as my living room. (I don't know what the hell the problem is, but my Tivo can barely see the wireless, and we're talking on the order of 50 feet or less away. I was half tempted to just run an ethernet cable; god knows I have enough hubs to keep the signal strong on the way...)
posted by caution live frogs at 10:36 AM on February 7, 2007
posted by caution live frogs at 10:36 AM on February 7, 2007
Now there's an idea. A 45kW beam link.
Pros: Guarantees line-of-sight.
Cons: Massive power costs, potential military intervention. May remove landmarks.
posted by Skorgu at 10:41 AM on February 7, 2007 [2 favorites]
Pros: Guarantees line-of-sight.
Cons: Massive power costs, potential military intervention. May remove landmarks.
posted by Skorgu at 10:41 AM on February 7, 2007 [2 favorites]
Damn, I have the perfect place to use one of these, but not the skills or time to build it. Any nobody's selling them. Is there a commercial equivalent?
posted by cosmicbandito at 10:44 AM on February 7, 2007
posted by cosmicbandito at 10:44 AM on February 7, 2007
Thanks, dorian, for the link to the German WW2 voice-over-optical field telephone. Exactly the sort of thing I like about Mefi.
posted by gdav at 12:56 PM on February 7, 2007
posted by gdav at 12:56 PM on February 7, 2007
"Now there's an idea. A 45kW beam link."
Hey, we've got a special on roast pigeon today!
posted by zoogleplex at 1:12 PM on February 7, 2007
Hey, we've got a special on roast pigeon today!
posted by zoogleplex at 1:12 PM on February 7, 2007
Ah, I've been looking for a project. Gettin' me a bandwidth upgrade!
posted by quite unimportant at 3:41 PM on February 7, 2007
posted by quite unimportant at 3:41 PM on February 7, 2007
cosmicbandito - yeah, but they're nowhere near as cheap... google or google
posted by russm at 10:29 PM on February 7, 2007
posted by russm at 10:29 PM on February 7, 2007
which is why 1GB Ethernet is pretty much the *slowest* fiber optic link out there
Eh, 100BASE-FX, dude.
posted by ZenMasterThis at 11:26 PM on February 7, 2007
Eh, 100BASE-FX, dude.
posted by ZenMasterThis at 11:26 PM on February 7, 2007
I'm fairly certain that you need an FCC Amateur Tech license to build and operate homebrew equipment on the "experimental unregulated" VHF and UHF bands (which also include many WiFi and cordless telephones). Also even if you have the license your power is gimped, and there is only so much you can do by just building bigger antennas.
posted by KirkJobSluder at 11:37 PM on February 7, 2007
posted by KirkJobSluder at 11:37 PM on February 7, 2007
A "device that can be built by nearly everyone"? Most people can't solder. Or understand the schematics well enough to build the thing.
Sounds awesome, though, and if I had a compound, that's how I would network it.
posted by Monday at 2:39 AM on February 9, 2007
Sounds awesome, though, and if I had a compound, that's how I would network it.
posted by Monday at 2:39 AM on February 9, 2007
which is why 1GB Ethernet is pretty much the *slowest* fiber optic link out there
Eh, 100BASE-FX, dude.
Eh, 10BASE-F, dude.
posted by kjs3 at 12:52 PM on February 9, 2007
eriko: where did you find the description of how they achieved the bandwidth? I can't even find schematics on the twibright website, though I did find PCB layout files.
posted by polyglot at 1:19 AM on February 10, 2007
posted by polyglot at 1:19 AM on February 10, 2007
« Older Beatlemania | Aftermath of another war Newer »
This thread has been archived and is closed to new comments
Official site
RONJA wiki
Talk at WSFII London (audio)
posted by goodnewsfortheinsane at 8:45 AM on February 7, 2007