Larry said:
Part of the challenge of doing this properly will be to get the
sensitivity of the indicator to vary with frequency in a way
that mimics average hearing sensitivity. If you do not do
this, kids whose voices are near the peak in that sensitivity
function will have an easier time satisfying your device, while
those whose voices are off the peak may not realize that
they need to speak up, having been misled by your device.
Have the level adjusted by a listener with average hearing to just start
to flash when the trainee speaks properly. Much easier than trying to
get it right for everybody.
Here is a circuit that works for me:
(courier, please. Otherwise, it looks like the gibberish it may well be)
1MEG 10k
___ ___
.--|___|-o--|___|--.
| | |
VCC | | |
| LED 1.5k | /| | Vcc |
--- 9V ___ | /+|-' | |
- GND ----|<--|___|---o---< | .-. |
| \-|--->| | |
GND \| | | |
10k 1MEG U1D '-' |
___ ___ | |
VCC .---------o--|___|-o-|___|---. GND |
| | | | | |
..-. .-. | VCC | | |
| |10k | |330k | + | | |
| | | | | | | | .-------o
'-' '-' | | | | | |
| | | | | | | U1C |
| | U1A | | | U1B | | |\ |
| Cf | |\ | .-. | |\ | '--|-\ |
| o--|-\ | | |<-)---|-\ | | >-'
| || | | >---' | | | | >--o--->|-o----o--|+/
o----||-----o----)--|+/ '-' '---|+/ | | |/
| || | | |/ | |/ | |
| | | RSens | | |
| | | | | |
| | ' | | |
| Rf | | | | .-.
| _ .-. .-. | 1uF --- | | 1MEG
'-/ \ | | | | |1k | --- | |
(Mic)| | | | | | | '-'
.-\_/ | '-' '-' | | |
| | | | | |
-o-----------o----o--------------o--------------------o----o---- GND
(created by AACircuit v1.28 beta 10/06/04
www.tech-chat.de)
U1A is a gain stage. U1B is a comparator, which feeds into a trivial
peak detector (the diode) which decays with a tau of 1s. U1C is a
buffer, so the decay of the peak detector isn't affected by U1D, which
is a comparator that controls how much time the light is on.
The opamps are A-D on a single LM324 chip.
The pot on the V- input to U1B controls level sensitivity.
The pot on the V- input to U1D controls how long the LED stays on after
it's triggered.
The MIC is a cheapo electret microphone
Cf and Rf adjust the sensitivity of the input. A 1k resistor and 1uF
capacitor will give a reasonably OK filter for voice, I think. However,
playing around with different values will give you a different profile.
the function of attenuation with frequency is
a = Rf/sqrt(Rf^2+(1/(2*PI*f*Cf))^2)
--
Regards,
Robert Monsen
"Your Highness, I have no need of this hypothesis."
- Pierre Laplace (1749-1827), to Napoleon,
on why his works on celestial mechanics make no mention of God.