help understanding the monster I created :-)

[Tom Horsley]

I've been working on a silly gadget using the cut & paste bits
of other circuits technique combined with a large does of
ignorance :-).

The gadget details can be found here:

http://home.att.net/~Tom.Horsley/phonetale/markII.html

With my breadboarded circuit, I have been observing a
phenomena that my tiny little brain can't figure out.

When I expose the cadmium sulphide photoresistor to light,
my relay triggers once (briefly) as expected and desired.

However, fairly often (not all the time), when I put the
lens cap back over the resistor and block off the light,
the circuit will trigger the relay again.

The only theory I have (with no access to a scope that
can record whats happening at different points and show
me) is that the photoresistor resistance actually bounces
around some as it gets dark rather than simply smoothly
going back up. Could that be it? Or has the ignorance part
of my design process merely led me to do something
silly along the way?

Any ideas for improvements to my design or explainations
of this puzzling behavior gratefully accepted. (I doubt
it is gravity waves though :-).

(Perhaps chuck the whole thing and detect incoming calls
via the headset interface is a better plan - but I'd have
to find out about interfacing to cell phone headsets
for that to work...).

[Andrew Holme]

Tom Horsley wrote:
> I've been working on a silly gadget using the cut & paste bits
> of other circuits technique combined with a large does of
> ignorance :-).
>
> The gadget details can be found here:
>
> http://home.att.net/~Tom.Horsley/phonetale/markII.html
>
> With my breadboarded circuit, I have been observing a
> phenomena that my tiny little brain can't figure out.
>
> When I expose the cadmium sulphide photoresistor to light,
> my relay triggers once (briefly) as expected and desired.
>
> However, fairly often (not all the time), when I put the
> lens cap back over the resistor and block off the light,
> the circuit will trigger the relay again.
>
> The only theory I have (with no access to a scope that
> can record whats happening at different points and show
> me) is that the photoresistor resistance actually bounces
> around some as it gets dark rather than simply smoothly
> going back up. Could that be it?

Yes, that's correct. The LM339 is very sensitive: it responds to very
small differences between the inputs. As the photocell resistance
passes through the critical threshold, the slightest "bounce" is enough
to momentarily push it back below the threshold. To overcome this, you
should add a little hysteresis by connecting a very large resistor (at
least 100k, maybe even 1M) between the output and the non-inverting (+)
input. The on-off and off-on transitions will then occur at slightly
different points e.g. like a thermostat.

[Jamie]

i just looked at the circuit you have there. from what i can
see, puting a small cap from the output of the OP-amp to the
- input could increase the response time thus absorbing the
quick pulses you may be getting.

[Thomas A. Horsley]

>To overcome this, you should add a little hysteresis by connecting a very
>large resistor (at least 100k, maybe even 1M) between the output and the
>non-inverting (+) input.

Thanks! That did indeed do the trick. In another Doh! moment, I notice that
the web page I point to with info on comparators even describes the reasons
I'd want to add hysteresis, but I never made the connection :-).

the comparator page I mention is at:

http://home.cogeco.ca/~rpaisley4/Comparators.html
--
>>==>> The *Best* political site <URL:http://www.vote-smart.org/> >>==+
email: Tom.H...@worldnet.att.net icbm: Delray Beach, FL |
<URL:http://home.att.net/~Tom.Horsley> Free Software and Politics <<==+