DPRG List  

[DPRG] Line Following

Subject: [DPRG] Line Following
From: Earl Bollinger earlwbollinger at comcast.net
Date: Sat Sep 13 22:19:00 CDT 2003

Well they aren't very sensitive to white, or red LED's. I tried some
super brights and the photocells don't seem to react well to them.
At Randy's work room they did react really well to the overhead
florescent lighting though. The flood lights at the meeting today in the
Laser room didn't seem to garner much sensitivity either.
Bill Boyer suggested yellow LEDS as these are closer to what a
incandescent light bulb puts out, but I'll have to find some super
bright yellow LED's to test that idea.
Since a flashlight worked pretty wellat home, I am thinking about taking
the lense section and hanging it out over the front of the robot. Then
it can chase the light (carrot?) and follow the line too. 
I think, if I was to hold a flashlight at the contest in front of the
robot, people might think it's following the light and not the line.
Which isn't all that bad actually, it's still showing autonomous light
following behavior and the robot still has no CPU, MCU or transistors on
board. :)

So I can have it start forward, when I shine a light on it, and it stops
when it reaches the white block at the end of the line following course.
Plus it can exhibit light following behavior while chasing a light aimed
just in front of it when it's following the line on the line following
course. :)

-----Original Message-----
>From: rten at new.metronet.com [mailto:rten at new.metronet.com] 
Sent: Saturday, September 13, 2003 9:46 PM
To: Earl Bollinger
Cc: 'DPRG'
Subject: RE: [DPRG] Line Following

The time from off (dark) to on (light) is is critical in that it will
affect how fast the robot can effectively detect a change in line
direction. Besides, cad suflide responds to incident radiation
and, unless you furnish something like white LEDs.

On Mon, 8 Sep 2003, Earl Bollinger wrote:

> Good ol'Dale noted that American Science and Surplus has some big
> cadmium sulfide photocells. Since photocells so fascinate me, I had to
> order some.
> Anyway they arrived, and I had to measure them and categorize them.
> Like all cadmium sulfide photocells, they tend to vary a lot.
> I don't see any patterns, but the lowest resistance measured was 42.1
> ohms (most measure about 60 ohms) and the highest resistance was 355k
> (most measured around 100k ohms).
> I used a simple 40watt incandescent bulb, holding thecell 1 inch from
> the bulb. The other test was in the dark way under my desk (not true
> dark as the cells tend to read really high then.
> 53	41k
> 40	135k
> 54	118k
> 63	103k
> 69	355k
> 42	100k
> 43	72k
> 75	146k
> 53	84k
> 76	172k
> 58	121k
> 63	126k
> 47	100k
> 66	108k
> 54	80k
> Now the trick is to determine what voltage is needed to drive a small
> motor with enough current to work. For example, say I use 3v motors,
> would 9v or 10v or 12v be needed?
> If we back off the bright light, the photocells measure about 100-150
> ohms over a white tape, and 25k to 75k over black paper.
> I almost have my "No-Tran" robot following lines OK, using my earlier
> idea.
> I am working on getting the ambient light more constant for it. Right
> now I have to shine a flashlight on it to get it to work right.
> But a "No-Tran II" with no relays looks very promising too. :)
> -----Original Message-----
> From: Dale Wheat [mailto:dale at dalewheat.com]
> Sent: Sunday, August 31, 2003 9:52 AM
> To: Earl Bollinger; 'Randy M. Dumse'; 'DPRG'
> Subject: Re: [DPRG] Line Following
> Earl,
> But a relay is a switch, just like a transistor is a switch.
> A truly simpler line follower would use the large, 1W CdS photocells
> from
> American Science & Surplus:
> http://www.sciplus.com part number 32155 3/$2.25
> ... with two small (<1W) motors and a bright incandescent light bulb
> (and a
> battery and some wire...).  Each photocell would be used to directly
> govern
> the current going to each motor.  On our course (white line on black
> background), you would need to swap the left and right sensors, so
> as
> the left sensor registered less light (straying into the lefthand
> black),
> less current would flow through the right motor, inducing a right
> Note that this is how the human body works (left side nerves connected
> to
> right side of brain, etc.).
> This model uses no "switches" at all.
> Thanks,
> Dale Wheat
> http://dalewheat.com
> (972) 486-1317
> (800) 330-1915, access code 00
> ----- Original Message -----
> From: "Earl Bollinger" <earlwbollinger at comcast.net>
> To: "'Randy M. Dumse'" <rmd at newmicros.com>; "'DPRG'"
<dprglist at dprg.org>
> Sent: Saturday, August 30, 2003 9:01 PM
> Subject: RE: [DPRG] Line Following
> > No transistor line following robot.
> > I believe it's possible to build a zero transistor line follower
> robot.
> > You use two small motors, and a small low power relay, with a solar
> cell
> > or cadmium sulfide photocell in series with the coil. One motor runs
> at
> > 1/4 to 1/2 speed, the other at full speed through the relay. The
> > coil is controlled by the photocell being on or off. The trick is
> > run a potentiometer in parallel with the photocell, and adjust it so
> > that the relay is almost on, but not quite. This is the ancient
> > technique they used with vacuum tubes in old computers to improve
> their
> > clock speed.
> > Thus there is no transistors involved.
> > You can even run a second "latching" relay with a similar setup too.
> > Except when you trip the relay it opens the motor power circuit and
> the
> > robot stops when it reaches the end.
> > The real trick is starting it without touching it. You guessed it a
> > "second" latching relay with a photocell that you trigger using a
> > flashlite to trigger it on. I'd probably use a shield tube to
> > the start photocell from the daylight and other stray light sources.
> > I know I'll have to find some time to build one.
> >
> > Anyway, I see no reason why one couldn't get a no transistor line
> > following robot to work, except maybe the relay(s) would get stuck.
> > Speed might be critical as it can't go too fast or the relays
> > keep up.
> >
> > On a further note, after my experiments with multi sensor line
> following
> > sensors, I am able to get my little Crystal robot to run the course
> very
> > smoothly, but at a speed price. I think we need to have the robot be
> > able to "anticipate" the upcoming direction changes so it can start
> the
> > turn appropriately at speed.
> > This is the same thing people do, they see a turn coming up so then
> they
> > start taking the appropriate action to negotiate it. Thus the robot
> > needs to know what it's doing now on the line versus what the line
> > like up ahead a short distance.
> > Before anyone goes crazy on that one, remember that a UPS
> > (uninterruptible power supply) can actually anticipate the impending
> > power failure to start up the high voltage switcher, in order to
> it
> > time to ramp up the power before it switches over to battery power
> from
> > the mains. A UPS can do this before the AC mains power actually
> > Now that's a pretty good trick.
> > I've always been fascinated by that one.
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org
> http://nimon.ncc.com/mailman/listinfo/dprglist

More information about the DPRG mailing list