DPRG mailing list archives: [DPRG] Wheel encoder theory

Mon Oct 9 12:46:31 CDT 2006

Vimal,

I would not recommend using a CDS visible light photocell for encoders
because they are slow to react to light. Also, they typically have a large
surface area, which means that they will not be saturated by the light
coming through the encoder slot. You will end up seeing a sine wave increase
and decrease in drop across the photocell. Then you have to try and detect a
threshold that indicates the edge of the encoder slot.

There are a number of easy to use transmissive type optical encoders out
there which are already amplified and will produce a digital output. Check
Mouser or Digikey for these. They will have 3 wires on the receiver for
power, ground, and output instead of 2 wires of a phototransistor. When
selecting a device, make sure that the aperture is as small or smaller than
the smallest encoder slot in your wheel.

You don't need an absolute encoder for wheel encoding, because you can count
the increments as they happen with the micro. Absolute encoders are also a
lot more complex.

Incidentally, I also built a differential drive robot with a dedicated micro
for each motor control. It worked ok, but the problem was that a
differential drive robot needs to constantly make adjustments between the
two wheels in order to go in a straight line. Even if you have a PID
algorithm, one motor will inevitably go faster than the other for a moment,
and this will cause the robot to turn. (Not to mention the wheels not being
the same circumference.) To correct for this, a good controller will need to
track the difference between the motor speeds and compute a direction
vector. Then it can compensate for the erroneous turning. This is
accomplished much more easily with a single controller for 2 wheels, rather
than 2 separate controllers.

On a separate note, I have also used this technique of a separate controller
for each motor in a 6 axis robotic arm and it works quite well. In this
application though, there is not the critical need for correlation between
the motor speeds.

PS: If your encoders are connected directly to the wheels, you don't need to
worry about gear slop in the motors, because it will not affect the encoder
count.

I would be interested to see this VB script for designing encoder wheels.
Can you provide the link?

If you are interested in designing quadrature type encoders, I can make a
quick tutorial. It's not as simple as it might seem. 

Good Luck, 
Rick

-----Original Message-----
From: dprglist-bounces at dprg.org [mailto:dprglist-bounces at dprg.org] On Behalf
Of Vimal Shankar
Sent: Monday, October 09, 2006 10:45 AM
To: dprglist at dprg.org
Subject: [DPRG] Wheel encoder theory

I did some searching on wheel/shaft encoders on Google and Yahoo, and I got
a VB Program to design wheel encoders for me.
I also got a few documents on designing my own shaft encoders using old mice
parts.

But what I want to do, is design a encoder system which may use IR/Visible
light (CdS/Phototransistor) but NOT the readymade reflectance sensors. (Is
this a bad idea ?)

I also want a few articles/links which teaches me the theory behind
designing absolute/angular encoders and relative encoder measurements, so
that I can deisgn my own encoder wheels and ALSO learn how to make sense of
them/read them.

Altough I have a readymade program to design the encoders for me, I will not
be comfortable being ignorant of the messy details "behind the scenes".

I have a feeling that a good wheel encoder is integral to designing a good
micromouse which can take diagonal/angular turns.

Specifically, I am thinking of designing a differential drive mechanism
which I can use to navigate ANY angular turn.

My idea is this :

I shall have a differential drive which has a dedicated microcontoller to
it.

I just instruct it to turn 'x' degrees from the current position, and it
gives me feedback about how much it really did turn.

I hope I could describe my goals clearly. Looking forward to some great help
!

PS : Are there any method to handle backlash for Geared DC motors and
slippage ? miscounted steps in stepper motors ?

Regards

Vimal

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around
http://mail.yahoo.com _______________________________________________
DPRGlist mailing list
DPRGlist at dprg.org
http://list.dprg.org/mailman/listinfo/dprglist