Controlling activation of a circuit with a moisture meter

[lostminty]

Hi,

I'm building a make shift spectrophotometer. It's only concerned with a narrow bandwidth of visible light. I'm using two LEDs, one which is lit, the other which is co-incident with a sample in between absorbing light from the lit LED.

I pass the voltage of the receving LED to an LM358 opamp which then gets output to an LM3914 bar dot driver.

This gives use-able readings. So that works.

I want to only turn on the sensing circuit when water is passing through the apparatus. So I've been investigating moisture sensing circuits.

http://www.instructables.com/file/FGO5G9DFH9ZFLVQ

Someone on yahoo answers suggested using a 555 timer in astable mode instead of DC which I have so far tried to some success (see circuit above). This was because the electrodes will get eaten away quickly with that set up.

Someone on reddit suggested the arduino moisture sensor also, really not sure what to use to take the results from that sensor, apparently it outputs resistance?

http://www.seeedstudio.com/depot/grove-moisture-sensor-p-955.html

The point of this is to improve battery life. The current end design I'm hoping to power by a 9v battery. I'm thinking it would be good to use a "low dropout" 5V regulator.

 

[Baluncore]

If you use electrodes in contact with the water you should use AC if possible. If you use DC then keep the voltage below 100mV. A very low power CMOS analogue gate such as CD4066 could be used in a bridge configuration to alternate the polarity and so reduce electrolysis.

Liquid water has a dielectric constant of 80 so it is easy to detect with an oscillator having a timing capacitor made from part of the flow's path. A CMOS micro-controller could wake up once per second, measure the time constant of the cell and perform a reading if water is there.

Power consumption will be reduced if you can make less measurements and so keep the LEDs off for say 99.99% of the time. Also low power CMOS op-amps will consume much less power than an LM358, they will plug straight in as a replacement.

How often does water flow ?

When not flowing is there air in the cell ?

While flowing, how often do you need to make a measurement ?

How long do you expect it to run on a single 9V battery ?

 

[lostminty]

How often does water flow ?

every 10 mins

When not flowing is there air in the cell ?

Yes

While flowing, how often do you need to make a measurement ?

Once

How long do you expect it to run on a single 9V battery ?

As long as practical...hard one to answer. maybe a month?

that is my current design...I think.

 

[Baluncore]

Two x LM555 = 10mA, Dual amp LM358 = 1mA, Total = 11 ma
9V battery, say 200mA.hr
200(mA.hr) / 11(mA) = 18 hours
So the water detection must be done using very low power rather than a couple of LM555s.

LM3914 bar dot driver with one LED will need about 15mA in dot mode when displaying the transmission measurement. You really need to use an LCD display.

A micro-controller in sleep mode that detects the flow, make the measurement and then displays it on the LCD display would meet the power requirements.

When water flows, what is the minimum time it will flow for ?
Can you use the water flow to generate the power needed to make the measurement ?

 

[lostminty]

Well there you have it. Would I get the functionality I desire out of an arduino?

minimum time..about 5 seconds?

potentially could generate power from the flow. doubt it though

 

[Baluncore]

I have no experience with the Arduino so I cannot say if it has the low standby power you need. If you want only one, then by all means use Arduino, with a bigger battery if needed. If you will want to manufacture many, avoid buying more hardware than you need by designing a single chip solution.

I would use a single chip that had the ability to sleep then wake every second to check for flow, or that was triggered by the flow. It would have an ADC to digitise the signal when needed, with sufficient outputs to support a slow LCD display. I would program it at the assembly level. Everything unnecessary needs to be eliminated. It should easily run for a month on one 9V battery.

LED emissivity changes with temperature, maybe it can calibrate itself without the water present. Soap or oil can form a film, (with a thickness of less than a few wavelengths), on the optical surfaces that can destroy any calibration. It can be very difficult to identify problems like that.

I have no idea of your application so I cannot guess at the specifications, constraints, size, or volume of the flow involved. I guess you are building a turbidity meter of some sort.