21/11/06

Ordered two IR LED's as well as a load of samples from Microchip including temperature sensors as I can't get hold of the high accuracy LM335AZ/H model here. These should have the same sort of accuracy. Also ordered voltage regulators, ADCs and DACs. I've also looked into controlling the fans. PWM is OK, however this will mess up the feedback from the tachometer as we are effectively switching the fan on and off, with no tacho readings when the fan is off. This is quite a problem as we can't tell how fast the fan turning to give an accurate indication of the air flow through it. Possible solutions include putting a capacitor across the fan to keep it going when PWM is off. Use of a frequency to voltage converter to drive the fan rather than PWM alone. PWM will drive the converter which will output a voltage proportional to the input frequency. - use passive filter to average waveform.

07/11/06

A very scientific test of the smoke sensor prototype...

As you can see the voltage output changes by approximately 0.05V when Mr Sheen is sprayed in front of the light source and thereby scattering the light directed at the receiver, resulting in less light being detected.

The other option for optical detection is to have the receiver shielded from the light source, and scattering would increase the light it receives. The former method is easier in implementation.

A more rigorous test will be done once I get hold of all the necessary components. Unfortunately only the phototransistors were in stock.

To measure 0.01V changes over a 5V range this would require 500 levels, corresponding to a minimum resolution of 9 bits. This won't be a problem, as the microcontrollers I'm looking at using have 10bit minimum.