[duff] found this and sent it in. The video demonstrates that the Wii ‘sensor bar’ is just an array of IR leds. The actual sensor is in the remote control – which probably sends data to the Wii via RF. These guys faked the ir signal using a pair of standard remotes. This’ll probably open the door to some controller cheats as things progress. [Better than using a belt sander on a trackball.]
[Hey – If you want to get on the podcast, email some questions or comments as mp3s to podcast at hackaday.]
It could be that the gyroscopic sensor is sending coordinates to the array of ir leds.
But because the device is moving so much, direct line of sending signals maybe discussable. Therefore an array is in place, so the whole field of reception is widened.
– Unomi –
QUOTE:It could be that the gyroscopic sensor is sending coordinates to the array of ir leds.
Actually, if you watch the video, that is clearly not the case! See how they DISCONNECTED the infrared panel? There is no information being transmitted back to the Wii through infrared. I assume that all these infrared lights do is to disable any remotes that are not facing the sensor bar (which we can assume would be on the top of your tv) to keep signal inteferance down. Think about it; If you and all of your buddies jump on over to your place, and you are all totin your favorite wii controller, and every last one of those is just blaring out signal, a great way to minimize wireless signal interferance would be to simply add a switch triggered by infrared light emanating from the top of your television set.
Having worked on a similar project dealing with accelerometers and gyroscopes on a moving platform, I have a pretty good guess at why those IR transmitters are there:
It is very hard to estimate orientation and position of the Wii controllers just by inertial sensors (accelerometers and gyroscopes).
Both have drift problems, and if you just integrate their values, the error on position and orientation becomes huge after some seconds. There are ways to solve this problem, mainly using a Kalman filter to fuse the data from both sensors and produce a reasonable estimate. But the problem is never solved completely, and the algorithm needs an external reference to cancel the drift. This reference doesn’t need to be available constantly, but at least from time to time.
Sensors like the Xsens (http://www.xsens.com/), use the magnetic field of the earth as a reference, but this method suffers greatly indoors if you are close to distorted magnetic fields. Nintendo probably can’t tell people that they would need to calibrate their controller every time you move your TV or couch, so this is out of the question.
The only other solution is a couple of IR transmitters located in a fixed position, and that is the function of the IR bar on the Wii.
This is great, now those who have projector setups can make DIY IR bars because the cord length of the originals are too short… or they could just make the cables longer.
Maybe we could figure out how to place the IR field to improve accuracy?
Interesting. Maybe a unit that sticks on the end of the remote with IR LEDs and a small battery pack, that would stop you having to have the sensor bar.
Or you could just leave the sensor bar in.
so if you really wanted to, you could line your TV with a bunch of IR LEDs, so you’d have the ultimate sensor for your remote?///
Um, some of you guys seem to be confused about what the sensor bar does. It has a few IR leds and thats all. it really isn’t a sensor, the sensor is on the wiimote. Its so that the controller can see where its pointing, like amaldo said.
The wiimote has an optical sensor in it, like a webcam, without an IR filter (like the IR webcam hacks) but a “standard light” (for lack of a better term) filter, the sensor bar itself is just IR leds and could just be ran off a battery pack, data from the wiimote is sent via bluetooth back to the wii. It’s the same reason why bright light can mess the wiimote up as it’s getting through the filter. I’m surprised it wasn’t worked out before. Check this out http://psyark.jp/?entry=20051208003209
Alex ~
Im surprised nobody figured this out earlier with a cameraphone (phones don’t have the IR filter).
I wonder what would happen with just one, or more than one IR transmitter? Would accuracy be related to the amount of LEDs, or would more than 2 not make much difference?
seems like the sensor bar is included for the sole purpose of keeping people pointed at the TV. IE. if you turn around and move you may get confused and mess up so Nintendo just solved this by disabling the wiimote if it dosent have the signal.
just a thought…
this also leads to an interesting problem i just thought of… what if you play with a LCD projector where the Wii is behind you where the projector is. how long is the cord for the bar?
Isn’t “don’t + no” a double negation?
Anyways. We can all wait for TVs with built-in IR leds. Not just for these applications, but it could work for information distribution to handhelds without sacrificing image space.
Most of you guys are wrong. The sensor bar is just IR LEDs that the Wii Remote uses to triangulate the position of the remote… I don’t know why they’re trying to “hack” the sensor bar, they should focus of hooking the remote up to a computer via bluetooth, it’s already recognized, it just needs software.
The bar really does do something. Just go into you sensor bar sensitivity settings. Those two dots you see are the IR sources. By figuring out their slant and distance from each other, the remote senses where it is pointing and how far from the TV it is. Pretty cool and elegant way of doing it if you ask me.
The one thing that I think can’t be ruled out yet is if the IR LED’s are always on or not. It looks like they have a regular blink or something like that at least.
The IR LEDs probably pulse to help the Wii Remote figure out how far it is from the bar.
yes, the sensor bar just emmitts IR light, however, this is very important to sense the position of the remote, when it is used to move a courser espectialy. sinply all the remote doers is find the locations of the lights in the camera. the lights are a fixed distance apart in the sensor bar, so if the appear closer together in the camera that means the remote is further away and vise versa. in addition, if the two ir lights do not appear vertical then it means that the remote is tilted. a very simple and accurate way to calculate the remotes position. so making a ton of ir lights would probably just confuse the thing, and changing the distance between the lights would make the remote believe it is closer or further away than it is. the lights do have to be in a fixed position, cannot move, and must be horizontal for the remote to sense its position properly. as for why it wasn’t thought of before now, it is simply because the system has only been in the hands of gamers for 48 hours. not late at all really.
Just read about how it works here:
http://en.wikipedia.org/wiki/Light_gun at the “Infrared emitters” section.
@10
It’s a neat idea, but no. Several games – including some (or all, I’ve only played a couple) of the Wii Sports games, and Excite Truck, require that the wiimote be fully mobile, not always pointed at the screen.
answering #3, I agree with you about the led bar as a refrence point, but i don’t think the wiimote even has an accelerometer, accelerometers don’t “drift”, rather their used to detect gravity as a refrence point with which to correct drift that acumulates in the gyro’s readings, (thats the principle on which self stabalizing vehicles such as the segway are based) the led bar most likely serves a dual purpose, also serving as a method of identifying controlers currently in use.
Guide to download video from youtube at http://www.soft29.com/how-to/download-video-from-youtube.htm
I am stoked to learn this due to an issue i’ve run into with playing the wii on a projector: games which use the “sensor” bar (such as link and call of duty, which have a cursor on the screen) get sorta wonky when played on a projector. When playing on a normal tv, the place where the wiimote is pointed sights exactly where the cursor is, however with the projector, there is a discrepancy between the actual and detected location the wiimote is pointing.
It seems to me if one built a sensor bar which has a larger space between the ir led’s (or allows the distance to be adjusted) this might offset these calibration issues.
for a better explanation of the calibration problems, check out this joystiq article, specifically the part where they connect a laser to the wiimote to show the difference between where it is pointed and where the cursor shows up: http://www.joystiq.com/2006/11/10/wii-en-la-casa-de-joystiq/
This doesn’t take too much to figure out — the Wii has a calibration screen in the options that literally displays a modified camera output from the remote (heavily contrast filtered so that sources are white spheres and the surrounding area is black). When a lamp with an incandescant bulb showed up as a bright spot next to the screen, I figured the Wii remote just had a cheap CCD in the tip. This video clip helps prove it, and should lead to some nice hacks.
http://doctabu.livejournal.com/64758.html
They did build their own ;)
i don’t know why everyone is acting as if it were some big secret that the sensor bar was just an IR emitter. ever since E3 when people were taking video with digital camcorders we could see that it was an IR light source, and the cord for it is so thin that the only thing it could run through there is a couple power lines.
as far as telling if it’s leds are always on or not, you could just keep a digital camcorder on it (say one of those cvs camcorders from a while back) and look through the screen to see if the IR shows up.
what i’m interested in though is making a stronger sensor bar so i can play from across the living room since the stock bar is just weak enough to not quite make it to my couch. also since having the light source being too intense might confuse the wiimote, i’d bet a potentiometer in any aftermarket or hacked sensor bars would be good to have so you can tone it down some for when you’re playing closer.
maybe someone can make a couple IR spotlights, get a jumbotron, and play wii sports in a stadium from the feild.
okay, like amaldo said, the sensor has NOTHING TO DO with in-game tilt/pointing mechanisms. all this thing does is orient the wiimote, tell it which way the tv is. once it figures that out, the gyroscopes take care of everything else.
this is just another example of how nintendo is staying low-tech in order to save a few dollars. Buy a ps3. and don’t even think about making a comment like “but the ps3 costs 3 times as much. wah wah wah…” because if you are really a video game enthusiast, you would know that the price of games adds up to a LOT more than the price of the console.
Wow i gotta try this out on my friends wii, did u guys hear about the wii grabbing pictures off bluetooth phones without asking to pair?
http://forums.nintendo.com/nintendo/board/message?board.id=revolution&message.id=2549265
“if you are really a video game enthusiast, you would know that the price of games adds up to a LOT more than the price of the console”
I totally agree with you. Add up the price of ten Wii games and ten PS3 games and compare.
Oh wait, TEN PS3 games…. wii might be in for a bit of a wait for that.
Before you reply, it’s pretty obvious that you’re not too hot on maths here, so I’ll elaborate on my point… PS3 games are a LOT more expensive.
People, people, people! Calm the fudge down, and listen:
The sensor bar is simply a point of referance (POR) for the wiimote to figure out what ‘up’ is. That’s all. There is no data being sent through the Sensor Bar.
The gyroscopic sensors in the wiimote only sense changes in G forces. Without a POR the wiimote dosen’t know which way is up, and therefore it woulden’t be able to translate your movements into the game.
There is nothing magical about this. There is nothing hidden anywhere. Calm down, and take a breather.
Given this information, one could make a sort of invisible sensor bar in space for the projector people by pointing two collimated beams of properly-modulated infrared light at the screen, to the immediate left and right of the center. I imagine this would give better calibration, as mentioned in #20.
Does anyone know the voltage required to run the sensor bar? I am running the WII on a projector and will just mount the sensor bar on the pull down screen with rechargable battery power and a switch..
I just cut open my sensor cord and tested it at 12v. The same as the powersupply…. Likely the IR led’s are run in series to obtain the lower voltage they need. Perhaps this is part of the reason for the array…
Now testing a 12v wall plug adapter
Sweet… I tried with a 12v wall plug and it worked beautifully. Got to thinking an AA battery pack would need to be pretty big, so I tried a 9v. Works like a charm. Now I head to Radioshack to get a 9v box with a swich and I am done… No more wires… This was my biggest complaint with the WII because I am using a projector and the sensor wire was not long enough… Now I just pull down my screen, switch the battery on, and don’t have to worry about running a wire around the house…. Sweet..
I just measured the current and it is drawing about 50mah. This means on an alkaline battery, I should get about 10 hours of use per battery., or if I use a rechargable battery, I would only get about 4-5 hours.
Here is a link showing it works
http://www.youtube.com/watch?v=ULZ8Gd06Kqg
Surely how the wii works is obvious to most hackers, it was clear from the first day they mentioned the need for the sensor (IR) bar – cheap yaw/horizontal movement stability.
It is exciting when you discover how something works though so im with them on that!
Extending the cable on a sensor bar:
I modified the sensor bar for my kids’ Wii so the wire was long enough for the sensor bar to reach our theater’s screen to the Wii at the back of the room. I ordered a “tripoint” tool from the web to unscrew those stupid 3 point screws on Nintendos, opened up the sensor bar, de-soldered the existing wire, and soldered in a short length of telephone wire. I then put 3.5 mm (1/8″) audio plugs on both the new cable from the sensor bar and the old wire that connects to the Wii. Finally, I created a long (~10m, or 30 ft) extension cable that had a male 3.5 mm audio plug on one end and a female plug on the other end.
I was careful to wire everything “tip positive” so I didn’t reverse the polarity on the sensor bar.
I tucked the long cable between the molding and the carpet along the wall of the theater, out of sight and out of the way.
If my kids want to move the Wii into another room, they can just unplug both ends from the extension cable and plug the short cable back together.
I also bought a 12 volt power supply from Radio Shack, and fitted it with a 3.5 mm plug. I can plug that directly into the sensor bar (making sure the polarity is “tip positive” like all the connections I made) and run the sensor bar from the power supply, without any connections to the Wii.
Here’s a small photo gallery showing the sensor bar and my finished rig:
http://www.pbase.com/duncanc/wii_sensor_bar_hack
Guys–You realize that some Wii games don’t really use the sensor bar, right? Excite truck sends all its signals right from the remote. It detects tilt left and right and doesn’t care about the sensor. Games like boxing, and the other Wii sports use the sensor bar. I can’t watch the video currently, (at work), but game are they using to demonstrate this concept?
HA!!!!!
i came up with a better way to play the wii w/out the two remote controls!!! simply change the sensitivity setting to 5,then light two candles and place them on top or below your tv. now u have your homemade, wireless, nonelectric, battery-free sensors!!!!
It works with Candles,
I suck, I posted without reading…
the sensor bar is to see where u point at the screen. the remote has the sensor in it. i tried it without the sensor barand it works fine. u just cant point at the screen and click on things. the buttons use bluetooth. so does the sensor.
ok so im not a techie but i am more confused then ever i dont have a sensor bar and my wii dont work any ideas? it turns on but it doesnt respond to the remote
Dude you are a fuckin genius. You just saved me from an entire week of boredom since I’m at my gf’s house over Christmas
Cheers!
If this is true, why when you point the wiimote at the screen, the aim is offset and you have to compensate. Like you can’t look down the scope of a wiimote and have it aim at that area of the sceen.
The remote for the tv isnt even connected to the wii
This is a pretty interesting find and come open up some hacks with the Wii remote.
The wiimote contains a gyroscopic sensor it detects its angle. While the sensor bar just detects the distance of the wiimote and if its angle is pointing towards the sensor’s “screen” area (“screen” because its what the sensor thinks is the screen). The wii is pretty fail the sensor bar isn’t all that necessary, were i the programmer behind the wiimote mechanics i’d just make a configuration setting that asks the player to point towards the screen to find where the screen is, instead of making some cost inefficient infrared bar. However the wii is lame and the sensor bar is required to point at the screen therefore you cannot navigate the menu without the sensor bar.
Quick story about making an emergency replacement for the Wii IR Sensor bar.
While on vacation, we forgot to bring our Wii IR sensor bar with us. All we needed the Wii for was the Netflix movie app this time … so, we tried to have one person set up two Sony remote controls on top of the TV pointing back at the viewers. We had one person press “Volume Down” simultaneously on both Sony remotes. This allowed us to “roughly” navigate the Wii Menu to get to the Netflix Application launcher using the controller. The pointer was very jumpy, but it worked.
So in conclusion, two IR sources do work as a sensor bar replacement – but more specific definition needs to be given to signal, frequency, distance apart, angle apart, etc to really nail a perfect IR Bar replacement.
Just thought I’d look around the net after running
My wii off of a lamp for a couple days.. Have a “dead”battery
Powered sensor bar and it appears that, as long as a
Light is shinning into the led in the front of the sensor bar
You can aim at that light source.. A little bit ago just to see
I fired up the flash on my iPhone 5 and aimed the remote at that light
And spun the phone toward the sensor bar and viola !
Cool thread . Just thought I would share that