Lego Sound Sensor
Lego Sound Sensor
van Mattijs, Stef en Robbert Mientki
This was our first project, to test our knowledge of the electronics
of the RCX and to get a feeling of how to put the components into a lego brick.
For convenience we used standard components and no smd-components.
We designed it straight forward and of course this leads almost exact to the design of Michael Gasperi.
His design was also published (without reference) in a Dutch magazine,
Elektuur july/august 2000, chematic 30. Of course this is not a clean electronic design,
but more "getting aximum unctionality from a minimal number of components".
We have tried to optimize the design (see schematic below),
which resulted in the following differences with respect to the design of Michael Gasperi:
we used an electret microphone, which is much smaller and has a higher sensitivity
the power supply capacitor C3 is much smaller,
even with 1uF the sensor works perfect (see my arguments)
we saved 1 diode, by combining the creation of the positive power supply
and enerating the output signal we amplified the signal extra with U1B (5*),
one of the resistors R6 or R7 is mounted in way it can easily be changed we increased
the value of R8, saving a lot of energy (minus 5 mA).
This sensor uses less than 1.5 mA of current,
which can even be reduced by a factor 2 by using a TLC272 (CMOS opamp) instead of the LM358.
The RC-time of C2/R5 seems to be optimal,
we've tried other values but they peformed worse.
Increasing R8 from 1 kOhm to 10 kOhm gives a great reduction of power needs.
This is quit simpel to understand, if you assume that most of the time there is no sound,
then output 7 of U1 will be 0 Volt.
The RCX drives its terminals 97% of the time (hard) to +8 Volt (see measurements).
So the power consumption of R8 will be roughly 8 Volt / R8, which in the case of 1 kOhm is 8 mA.
There is a little drawback, the effective signal range is only 2.5V ... 5V, instead of 0.5 V ... 5V.
Because the amplifier isn't linear at all, and even has a memory effect,
the decreased range doesn't affect the performance of the cirquit.
Combining the positive power supply and the signal generation,
results in a power supply for the opamp which is 0.7 Volt less than normally.
This is no problem because the opamps will work correctly downto 3 Volt,
it will even save some (very little) current.
It was a hard job to get the last 5 components in it. Because the electronics just fits
in the height we have to mount a 1/3 height plate as a bottom cover (this wasn't planned).
The largest mistake we made, was the use of the wrong resistor type.
We used 0.25 Watt metal film resistors, in the future we will use 0.125 Watt carbon film resistors,
these are the smallest resistor components normally available.
Another improvment may be the use of smd-ic's, which still can be handled
without the use of a printed cirquit board.
|Proximity Detector CS209A/D|
|12C508 IR Proximity Detector|