James Berg

3 March 2005

TURBIDITY METER

PROGRESS REPORT 1

I have spent the last month attempting to figure out how to work two sensors, one 90° and one 180°, into the turbidity meter. This poses a much greater challenge then a simple one sensor device because math is now involved, specifically division. While this is not complex math for a person to do circuitry has to do it in a rather round about away, I believe by doing many multiplications. There are two different ways to the divide the voltage from the 90° sensor by the voltage from the 180° sensor, an analogue method anda digital method.

The sensors themselves output an analogue signal, a voltage that varies continuously with light intensity. Op amps set up in varying configurations can be used to perform certain math functions with voltage. Specifically, given two voltages V90 and V180, take the Ln(V90) & the Ln(V180) => Ln(V90) – Ln(V180) = Ln(V90/ V180) => exp(Ln(V90/ V180)) = V90/ V180. It is fairly easy to us a logarithmic amplifier to take the Ln of a voltage but, despite the fact that I have been told that there are inverse-logarithmic amplifiers, I have been unable to find one. It is true that this method would use a lot of separate amplifiers but they are fairly cheep. The major disadvantage of the analogue method is that it would require a fair amount of soldering, and it still does not solve the problem of an output device.

The other option, and the one that looks most promising, is to convert the analogue signal into a digital one and then run it through a device called a micro controller. A micro controller is an integrated circuit that has a little bit of rewritable memory, allowing the user to program specific functions into it, using C. This would mean that the micro controller could not only be programmed to divide the voltage from the 90° sensor by the voltage from the 180° sensor, but also to convert the final signal into NTU's rather then Volts. Additionally the micro controller has an analogue to digital converter built into it and can be set up to output directly to a digital readout. The major disadvantage is that the analogue to digital conversion is limited. An analogue signal varies continuously, the conversion process divides it up into a set number of possible values; so for an 8-bit chip there are 254 possible digital output values evenly divided between a high and low value. Some converters are preset with a high and low value, say 0V to 5V; other chips allow the user to set the high and low values. Because the digital signal can only be one of 254 different values the resolution of the whole sensor assembly would be limited. If your minimum turbidity is 0 NTU and your maximum is 254 NTU then your resolution is going to be limited to 1 NTU, the solution to this problem is to get a higher bit chip, but that costs more.

Another issue with the micro controller is that the initial cost can be rather high, anywhere between $80 and $150. The initial cost is so high because there is equipment that is needed to program the micro controller, however once a micro controller has been programmed it can be removed from the equipment and installed in any device you wish. Subsequent micro controllers would only cost $10 or $12 as the programming equipment can be reused. It is also possible that I could use equipment in another lab to program the micro controller. Selecting a micro controller poses another problem, they come with a dizzying array of options and I do not have the technical expertise to understand all of them. A friend has recommended I talk to a professor (whose name escapes me at the moment) who is knowledgeable about these devices; however I have not met with him yet.

Component / Price ($)
Light Sensor (2) / 8
Op Amp (4) / 4
Micro Controller / 12
Digital Readout / 3
Total / 27

The final design using a micro controller would look something like this. Two light sensors at 90° and 180°; they would each connect to an op amp, a very simple device that multiplies the input voltage by a specific, easily definable value; if necessary at this stage the signals can be put through another op amp with a slightly different configuration to average the value over a given time period, 2 or 3 sec (the 90° and 180° sensor signals are still separate); the two signals are fed into a micro controller where they are first converted into a digital signal; the 90° sensor value is then divided by the 180° sensor value, giving you Vdivided; Vdivided is then put through a function that converts it from volts to NTU; the signal is then turned into a signal appropriate for a digital readout; then the signal shows up as number on a digital readout.

The resolution of a micro controller may end up not being an issue depending on the desired NTU range and the desired resolution. Unfortunately I do not have exact specifications for desired final product, this is large oversight. Last semester the goal was just to see if a cheep turbidity meter was even remotely feasible and the goals were somewhat flexible. Some ranges I found for commercial turbidity meters were from 1-199 NTU or 0-1000 NTU. Off the top of my head 1-199 NTU sounds like a reasonable range but of course the necessary range varies depending on where you are. This turbidity meter does not need to read exceptionally high values because at some point the water becomes so dirty that one would only need to look at it and declare it undrinkable. Again though I do not know what these values are, nor do I know what kind of resolution is needed for a turbidity meter to be useful in determining the necessary level of chemicals that should be added. Determining the required range and resolution, finding an appropriate micro controller, and starting to program the micro controller are my primary concerns for the next month.