CMOS buffer hysteresis?

[Mark Haase]

Hi all--

I'm building a drivetrain for an autonomous robot, and I bought a
photointerrupter to use as part of a homebrew shaft encoder. With
tweaking of the circuit, the photo interrupter outputs about .5V when
the encoder disk is black, and about 4.6V when the encoder disk is
white. These vales are probably OK for talking to my HC11, but I figured
that to "do it right" meant using a CMOS buffer. (The exact part I
bought is here:
http://rocky.digikey.com/WebLib/Texas%20Instruments/Web%20data/CD74HC4049
,4050.pdf).

I used the output of my circuit to send signals to an input compare pin
on my MCU. The problem is that the CMOS buffer's output apparently
oscillates rapidly, and so when the optical encoder moves forward one
segment, my MCU gets dozens or hundreds of input pulses.

Without a scope I can't tell exactly what's happening, but AFAIK buffers
usually have about a .5V hysteresis, true? So I guess what's happening
is that the CMOS buffer is oscillating between 0 and +5V very rapidly.
So my question is, how on earth should I smooth out the input signal to
my MCU? The pulses I'm interested in will be at least a few MS long (the
optical encoding disk has 16 segments, and the motor's output is about
120RPM). The MCU is 2Mhz, but its juggling a bunch of other tasks and
there's no FPU, so I can't do the smoothing in software--unless there's
a real efficient way I don't know about. (Ie. I'm familiar with FIR and
IIR but they seem to be too expensive for my application).

Thanks for any hints--I don't even know the terms for which I should
google

Mark

--
|\/| /| |2 |<
mehaase(at)sas(dot)upenn(dot)edu

 

[CWatters]

Try adding an RC filter to the output of the photo interrupter and change
the HC4049 for something like a Schmitt Inverter (eg like an 74HC14 but
perhaps there is one with the same pin out as the 4049?).

 

[John Fields]

On Fri, 20 Aug 2004 04:18:53 -0400, Mark Haase <mehaase@earthlink.net>
wrote:

Hi all--

I'm building a drivetrain for an autonomous robot, and I bought a
photointerrupter to use as part of a homebrew shaft encoder. With
tweaking of the circuit, the photo interrupter outputs about .5V when
the encoder disk is black, and about 4.6V when the encoder disk is
white. These vales are probably OK for talking to my HC11, but I figured
that to "do it right" meant using a CMOS buffer. (The exact part I
bought is here:
http://rocky.digikey.com/WebLib/Texas%20Instruments/Web%20data/CD74HC4049
,4050.pdf).

I used the output of my circuit to send signals to an input compare pin
on my MCU. The problem is that the CMOS buffer's output apparently
oscillates rapidly, and so when the optical encoder moves forward one
segment, my MCU gets dozens or hundreds of input pulses.

Without a scope I can't tell exactly what's happening, but AFAIK buffers
usually have about a .5V hysteresis, true? So I guess what's happening
is that the CMOS buffer is oscillating between 0 and +5V very rapidly.
So my question is, how on earth should I smooth out the input signal to
my MCU? The pulses I'm interested in will be at least a few MS long (the
optical encoding disk has 16 segments, and the motor's output is about
120RPM). The MCU is 2Mhz, but its juggling a bunch of other tasks and
there's no FPU, so I can't do the smoothing in software--unless there's
a real efficient way I don't know about. (Ie. I'm familiar with FIR and
IIR but they seem to be too expensive for my application).

Thanks for any hints--I don't even know the terms for which I should
google

---
The easiest way out of the mess is to dump the 4049 and use a
comparator with some hysteresis, like this:



+5v>-----+-------------+----------+
| | |
| | [2K]
| | |
[100K] +--|--[100k]--+
| | | |
OPTO>----|---[10K]--+-|+\ |
| | >--------+---->OUT TO HC11
+------------|-/
| |
[100K] |
| |
GND>-----+-------------+--------------->GND

You could use something like half of an LM393, and if you do, make
sure you connect the inputs and outputs of the unused comparator to
ground.

--
John Fields