Author Topic: Resolve Under Sampling  (Read 4826 times)

Offline Seán McT

  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
on: February 07, 2010, 10:33:13 AM
Hi when building a source of infrared what voltage is needed in order to produce a smooth stream on the Z axis? Currently I have two Infrared bulbs hooked up to two AA batteries, but the Z axis samples are jumping from high two low without a stream of data in between, Does anybody have any ideas that would help me? Would save me a lot of trial and error research. Thanks

Seán



Offline bobster23

  • *
  • Posts: 19
  • Karma: +2/-0
    • View Profile
Reply #1 on: February 07, 2010, 03:39:58 PM
Hi Sean,
I'd like to help you but am not completely sure on what you mean so I'll do a bit of guess work (and try to answer your question).

First off I don't really know what you mean on the z axis. The wiimote camera is able to follow up to 4 points and for each point give the following information: x axis location, y axis location, size. So i'm guessing what you mean when you talk about z axis is either your using a program that looks at the size of a point, or your doing what the wii is doing which is looking at 2 bulbs that are at a fixed distance from each other and measuring their distance from one another with the wiimote camera (and calculating the z axis using triangulation). If you say you have 2 infrared bulbs I'm guessing your doing the later in which case the cause of your problem may be that one bulb is flickering or your not holding the bulbs at a fixed location.
You can easily check if a bulb is flickering in one of 2 ways. If you have the wiimote hooked up to your computer then download wiinremote from this site: http://onakasuita.org/wii/index-e.html which it lets you see the raw feeds from the wiimote (that way you can see the points the wiimote is seeing). Another method is to take a camera on a cellphone and look at the infrared bulbs through it. The trick is that some cellphone camera (not all) are able to detect infrared light, so you should be able to see if light coming out of your bulbs that should look like a bright white light. That way you can check if the bulbs work properly.

Secondly what do you mean by infrared bulbs. Most people use infrared LEDs so I'm guessing you mean IR LED (plus they look a bit like a bulbs). This is important for the following reasons.
1. In a clear (non defused) led the light is brightest from the top, like in the picture on this site http://www.instructables.com/id/Wiimote-Whiteboard-IR-Pen/step4/Final-Assembly/
(the 3rd picture)
so you have to aim the led correctly at the wiimote or it will not see properly (and maybe cause some sort of flickering on the wiimote camera)
2. LED stands for Light Emitting Diode, meaning it is a type of diode so the direction of the current is important, so you have to connect the LED in the correct direction. (If you get something in the z axis I'm guessing you did but its worth a check anyway)
3. Not all IR LEDs are the same. The wiimote has a filter (the black thing in the front) where the maximum sensitivity is at this wavelength 940nm. By all means that doesn't mean other LEDs won't work. Its just that if you can see the LED with a cellphone camera, and not with the wiimote it could be that the LED is not in the correct wavelength.

I also came across this site which may be helpful to you:
http://doctabu.livejournal.com/64758.html

I hope I was helpful 

Bobster



Offline Seán McT

  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Reply #2 on: February 07, 2010, 06:37:25 PM
Wow thanks man, thats a great help sorry I have been working on this project alone and made up some of the technical terms myself so all of what you assumed I was trying to say was correct. I have been using the phone camera to see the source of the LED's ive been using. See the photo attached for the circuit I have been using. I am sending the data to my system via a program called OSCulator, this reads all of the data sent via the wiimote.


What you mentioned about triangulation, Is there a set distance that is required in order to get an accurate reading of how far a user is distance-wise from the screen (this is what I was referring to by the z - axis sorry!).
The OSCuator program is reading the infrared that isn't a problem as I can see this on the X and Y axis, What you referred to as triangulation may be the answer, any Ideas on how the reading could be more accurate? The image attached has the LED's about 7 cm apart this is just a prototype circuit but it is what I have been using.



Offline Seán McT

  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Reply #3 on: February 07, 2010, 06:40:27 PM
Sorry Just saw the link you attached thanks a lot man, thats a huge help!