Re: Tasks Accomplished Jan31

[Tim Vanderhoek <vanderh@ecf.utoronto.ca>]

On Tue, Feb 01, 2000 at 08:41:20PM -0500, Joyce Poon wrote:
> 
> I'm not THAT dumb.  I added it to the last stage. 

Just checking.  :)


> > I think you _may_ have better luck adding another stage to the amplifier if
> > you put a small cap across the input to the extra stage and the ground
> > line.
> 
> That kind of creates a low pass filter -- hence the circuit at that last
> stage will have a band pass filter.  I think I tried that last time but
> it didn't work.  One wonders whether adding that low pass filter will do
> anything since there is, in theory, nothing oscillating at that freq range
> other than our signal.

You mean: "There are no IR light sources oscillating at that freq.
range other than our signal."  :)

What I'm thinking is a very very small cap.  TBH, if the noise we're
seeing is just random electrical noise that's being amplified, then I
don't think there's much that can be done.  We were seeing some type of
high-frequency noise.  I learned in today's physics lab that random
white noise results in equal voltages at all frequency levels (roughly
speaking).

The idea is just to try many different configurations...  This is
(from experience) time consuming...


> Actually -- band pass filters might not work very well at all --remember
> when I changed the time scale on the scope, we saw that (at least I
> saw that) the output was a squarish/sinusoidal wave modulating a bunch of
> sinusoids oscillating at a much higher freq.  Using a bandpass filter may

This is the "sinusoid in a sinusoid" you were talking about?  You
weren't able to make the scope show it again.  However, depending on
how the scope was triggering, this might be the white electrical
noise that I mention above.  You are correct --- it is not very easy
to filter that out.  I can propose things that would possibly deal
with it, but you wouldn't like any of them.  :-)

What level of amplification is actually being applied?  If it is white
noise, then it should be in 1x10^-10 Volt magnitude range.  Possibly
higher.

[I did choose that lab for a reason!  :-]

Alternatively, it's possible that the modulating squarish/sinusoidal
wave is actually some multiple of the power supply's 60Hz wave.
Hmmm...  I didn't see the "sinusoid in sinusoid", and I don't know
what time scales we're talking here, but it is _very_ hard for the
power supply to supply a clean voltage level.  I very strongly suggest
you try running the sensor circuit from a battery power supply at
least once to see if it goes away.

[The more I think about it, the more I think that this must be the
 explanation and that the white noise explanation only applies to the
 lower-voltage sign waves you saw that were being modulated _by_ the
 larger one].

If you want to save batteries . . .

It is possible to simulate a clean voltage supply using a battery pack
loaded over a voltage divider that feeds an opamp, where the opamp is
powered by the main power supply.  Common ground, of course.  If this
works well, it may be a good idea to solder a permanent one together
that uses a pot for the voltage divider.


> Besides band pass filters make the sensors less flexible (ie it's harder
> for us to change freq if we want to later).

Yes, I know.


> Good thing that Tim doesn't care for band pass filters anymore, as stated
> in his earlier email.

Band pass filters were a response to your concern for reflected
signals.  That was changed by the video.  :)


> I drifted off topic.  I was going to boost sensor range by increasing the
> voltage to the emitter.  The 555 timer can have a max Vcc of 18V.  I ran

If that works, that's a _far_ easier solution to increasing sensor
sensitivity.

If the resistors get too hot, you can wire them in a square matrix to
spread the power dissapation around.  Simply buying bigger resistors
is the preferable solution, but the ones you had were pretty big
already...

  *..%..*..%..*
  .     .     .
  .     .     .
  *..%..*..%..*
  .     .     .
  .     .     .
  *..%..*..%..*

The standard square matrix: "%" are resistors all of the same value,
"." are wires, and "*" are nodes.  Overall resistance equals the
resistance of one "%".  Pretty straightforward.  :)


-- 
Signature withheld by request of author.