Splitting a 5v power source into 0.125v X 40

[cobra1]

Im working on a water tank that needs to be filled to a certain level and then release its contents into another tank.

This happens every 10 days or so. The dispenser tank holds 1 Litre.

I would like the ability to change the desired dose if required. In order to do this i need to know how much water is going into my dispenser. The way i want to do this is to have 40 conductive sensors up the side of the dispenser.

This will give me 1 sensor every 25ml, problem is this would not work on my 16f877 chip that i plan to use for the entire project. After looking about i find that using a big voltage divider (40 outputs) might be the answer, at each point where the resistors meet there will be a tap point, this point i would like to read at 0.125v, every sensor that is underwater will allow its coresponinding output to be added in series, this will allow the voltage to rise in increments of 0.125v and this can then be read using ADC.

I think that using 40 resistors between a power source might be a little ott. So basically im wondering if there is a better solution here.
If possible i would like to still use the ADC on the chip.

Hope this makes sense.

Cheers

 

[dougy83]

Perhaps just 2 submersible conductive [sense] rods and an additional reference rod or two (with all but the tip insulated). Use the reference rod(s) to provide the conductivity of the water and the sense rods to give the amount of liquid (resistance changes with level).

 

[iamsuperbleeder]

could just go very low tech and perhaps use some sort of float attached to a potentiometer? and the varying water levels would be reflected as the varying resistance?


excuse the crude drawing every finished project starts out as a sketch, lol




**broken link removed**



obviously that would have to be refined, but you get the basic idea

 

[dougy83]

Maybe a capacitive sensor?
capacitive water level - Google Search

 

[Reloadron]

Given a choice I would lean towards what iamsuperbleeder suggested. The reason being that using a float method eliminates changes in the conductivity of the water screwing up the level measuring process. The mhos can change.

When using the solution conductivity to measure the level you are betting that the conductivity of the solution will always be the same and never change. Considering 40 points that is a pretty open gamble. Also, when basing the level on conductivity you need to worry about the sensing points being always clean, remember electrolysis can figure into this and become a nightmare with the sensing. Thus I would look at a float system driving a potentiometer and let the pot drive your PIC. That would allow for easy calibration and I believe reliability with fewer worries. Simple can be a "good thing".

Ron

 

[cobra1]

just looking at this idea now, there appears to be 2 types of level sensor. One is a shaft with a magnet floating magnet, the othe rlooks like what iamsuperbleeder has drawn.

I am inclined to go for the Floating magnet type as it can all be enclosed easily. I will look into it a bit more.

Thanks for the ideas.

 

[Reloadron]

just looking at this idea now, there appears to be 2 types of level sensor. One is a shaft with a magnet floating magnet, the othe rlooks like what iamsuperbleeder has drawn.

I am inclined to go for the Floating magnet type as it can all be enclosed easily. I will look into it a bit more.

Thanks for the ideas.

The magnet flavor generally uses a series of reed switches with resistors across them. The magnet is configured to float, generally a ring around the tube containing the series of reed switches. As the level changes different switches close, in some cases two switches may be closed at once denoting a "half point". The switches are spaced equally. The problem with this method is the overall length of the reed switches. If each switch is 1 inch (25.4 mm) and you want 40 points you will have a very long tube for your water container. Additionally each switch would have a resistor placed across it. What you are really doing is creating a very long voltage divider that will have a voltage output proportional to the level of the floating magnet in the container. The method does work and work well depending on the resolution you want and how the distance between switches is configured with respect to the strength of the magnet in the floating ring. I am not trying to dissuade you from this method but in your case I just don't see it as viable. However, it is something else to think about and consider.

<EDIT> I should point out that a series of hall effect switches could replace the reed switches but that can get more complex. </EDIT>

Ron

 

[cobra1]

so the potentiometer with a float connected as per the drawing above would potentially be the best idea...
Im guessing that using one of these type sensors the proportional output will be much smoother.

I have found this.

**broken link removed**

Im thinking these are small enough to do the job required.

 

[Reloadron]

so the potentiometer with a float connected as per the drawing above would potentially be the best idea...
Im guessing that using one of these type sensors the proportional output will be much smoother.

I have found this.

**broken link removed**

Im thinking these are small enough to do the job required.

In my opinion using a float on a pivot arm is a good choice. The float will give you a nice analog output that won't be plagued by some of what has been covered thus far as to level sensing. I see it as a viable solution. Well, for your application anyway. I liked iamsuperbleeder's suggestion for the reasons I mentioned. Your call on how you do it.

Ron

 

[iamsuperbleeder]

wow... nuB here had a decent answer

lol, 3 years in drafting and design paid off

 

[Reloadron]

wow... nuB here had a decent answer

lol, 3 years in drafting and design paid off

Yeah, a very good suggestion with a real nice cartoon.

Ron

 

[cobra1]

Hi Ron,

Going back to your previous post about the reeds switches with the resistors, i like the idea of this and could possibly make something like this work well for me.

Trouble is i dont know how to work out the resistor values

I have looked around on the net and i have found examples of how they work and how there wired, but no resistor values to create my dividers.

Is this something you could help me on??

 

[jrz126]

Why not just use something like this:
**broken link removed**
and manually adjust the position as needed?

 

[cobra1]

that would be too much hassle, the computer has to calculate its dosage itself.
the unit you have shown will require human intervetion.

by the way i like your sinature, im an engineer lol

 

[Reloadron]

OK, attached is a resistor string. R1 through R10 can be 10 K in the example. I am not nuts about going this route but as long as you asked. When using a circuit like this or for that matter a R2R ladder the resistors should be .1% tolerance if you want it to be accurate. Magnetic reed switches like the ones needed typically can cost $7.00 USD EA. and those are inexpensive ones. Keep in mind your switches are stacked vertically unlike the way they are drawn and need to be spaced equidistant. Also, this assumes the circuit it is driving is a very high input impedance or some impedance matching (a buffer) would likely be needed. There are other oladder logic methods but this is simple compared to them. Also, this only covers 10 points. 0 to 5 Volts in .5 Volt steps.

Personally I like the other way but your call. This can easily become an expensive affair.

Ron

 

[cobra1]

Hi Ron

Thanks for the explanation, it has helped a lot. There is a reason for me looking a bit closer into this method. Basically space is too limited and the tank is very deep it also has a lid on it so any mechanical methods for using a wiper arm get a little complex.

i was looking at this reed switch method as it may be a little easier for me to impliment into my design.

not sure what sort of reed switches your using but i can get them for about 45p each so shouldnt cost very much to build one.

 

[Reloadron]

Give it a try then. I would build a basic version and see if it gives you what you want. The switches come in many flavors so find some and experiment. Let me know how it goes.

Ron

 

[colin55]

Get an old water level indicator out of a washing machine. It has a pressure sensor that can be adapted to what you want.