Kilian's Robot Shop of Horrors - 04/06/2001

Click thumbnails for a larger view.

Subject: [TCRG] Rangefinder experiments.
Date: Sun, 8 Apr 2001 12:50:20 -0500 (CDT)
From: Alan Kilian
To: tcrobots@orbis.net (Twin Cities Robots group)

We had a really good time last Friday experimenting with light.

My 5 Year old son helped me clean up the RSOH before anyone came over, so we had a real nice place to work.

You all know that Brynn has designed a laser range finder for his Tryclops robot, and he is getting it put together probably as I write this. His is kind of like a ring oscillator using the time the light takes to go out to the target, and bounce back to set the frequency of the oscillator, and then he counts the frequency to figure out the range. Pretty slick, but it uses some stuff I don't have the tools for, and don't really want to learn about like FPGA design tools for a bazillaHertz counter and things like that, so I am taking a different approach. I waited until now to start so we could talk about Brynn's version while he was designing it. So here goes my attempt.

Brynn (Or someone) found a nice article by a guy named Joshua Andrew Strickon from MIT called _Design and HCI Application of a Low-Cost Scanning Laser Rangefinder_ (I'll add a URL whenever I find it)

This is a nice paper he wrote for his Masters in EE/CS which describes not one, not two but THREE laser rangefinder concepts.

I chose to try the system he calls a "Continuous-Wave phase measurement device". The idea is to turn a light on and off real fast (Like 25,000,000 times-per-second) let the light go out to an object, bounce off the object and reflect back to a photo detector.

Now if you look at the signal that is sent to the light source, and the signal you get back at the photo detector, they will be a little bit different. Let's say everything is perfect. You send out a sine wave with the light source, and you get back a sine wave. (Yeah, RIGHT) You will see that if the object reflecting back the light is really close to the lightsource/detector pair that the signals will be almost identical. They are "in phase" or their phase difference is zero.

As the light gets brighter, whammo the light bounces off the target, and the signal at the photo detector starts to get brighter at the same time.

Now at 25,000,000 Hertz, and using 3*10^8 meters-per-second (300,000,000 Meters-per-second) as the speed of light, the light will be able to travel (out and back) 12 meters while we change the brightness of the light source through one sine wave.

So, if we move the target 1.5 meters away from the rangefinder, the light needs to travel 3 meters before it gets back to the photodetector, and if we compare those two sine waves, we'll see that the signal running to the light source is a little bit "ahead" of the signal we get at the photo detector.

How much ahead?

Well, the light traveled 3 meters. 1.5 Meters out, and 1.5 meters back. That travel took 3 Meters divided by 3*10^8 Meters-per-second = 1 * 10^-8 Seconds. (Huh?) OK. That is 0.00000001 Seconds (10 nanoseconds) And at 25,000,000 Hertz (2.5*10^7 Hz) the light bulb takes 40 nanosecond to go from off-to-on-to-off So, 10/40 = 0.25 of the full cycle, so that is a 360 degree * 0.25 = 90 degree phase shift. (PI/2 radians for those who are trying to keep up using their physics books)

So, what do we know? If we measure the phase difference between the outgoing signal and the returning signal, and it is 90 Degrees lagging at the detector, the object reflecting the light is 1.5 Meters away. (or (1.5 + (n * 6)) Meters where N is a integer. You figure THAT one out)

So what did we DO LAST FRIDAY ALREADY???

1) Need a light source that can go on and off real fast.

a) Laser diode. I started looking up laser diode driver circuits. People make them real complicated. Feedback photo diodes and all.
b) Wait for Jeff to say something. This usually works better. This night, he said "Let's start with an LED" HEY, great. an LED
So I hooked a red LED up to my variable frequency signal generator, turned it on to 1,000 Hz added a current limiting resistor, and lit it up. Using a real nice photo detector I got from Global Surplus for $75, we could look at the LED going on and off at You guessed it, 1,000 HZ.
Well, that's not fast. Let's turn it up. How fast will an LED go on and off? 1kHz? 10kHz, 100kHz? What would you pay for such a thing? 1000 kHz? Well, we got it going 5 megahertz!!! I couldn't believe it, but there is was. an LED blinking 5 million times per second. Thank goodness for those (theories) of electromagnetism.
So, we got a light source.

2) Need a detector. I already have one. We put a lens in front of the detector to capture as much of the reflected light as we could and get a nice large signal, but we still used a mirror as the target, and aimed the outgoing beam right back into the photo detector.
We got good results as shown in several of Jeff's photographs.

Lots of fooling around with lenses. We just don't know any optics, so we didn't know how to get good images etc.

Then we pulled the LED out, and plugged in a red laser diode from a laser pointer Brynn left here, and IT BLINKS AT 5,000,000 Hz also!

Now we didn't use any lenses at all, and things were even better. As you can see in the later images. Much better, cleaner signals.

Ron hooked up a 20 MHz "can" oscillator a 74ACT244 buffer and we drove the laser diode directly from that, but it blew 20 MHz signals all over the house, so I'll build a 20 MHz laser diode driver in a shielded can for further experiments.

Well, we learned a lot, and could actually see a phase shift, so now I'll continue to build a receiver amplifier, filter and see if I can get it to work reliably.

--
Alan Kilian

Alan's range finder experiments

Setting up a mirror as an optical target.	An LED connected to the signal generator. A lens in front of the optical receiver.	The LED running. You can see an image through the lens.	The target's view of the photo diode in the receiver.
Distance of ? (Close?)	Distance of ? (Farther?)	Distance of ? (Farther?)	About 15 feet.