Maybe this question has already been asked before, but after 4 days of searching the internet I give up and have to ask it here... I'm getting desperate so I hope somebody can help me. For information, I am a computer programmer with very basic knowledge of electricitiy so please keep you answers easy lol
I'm using a little arduino nano to monitor all kinds of sensors (or that was what I was planning). The sensors and arduino are powered by the same 12V source. I programmed an Android tablet interface to visualise all the sensor values and everything is working.
To start I wanted to monitor the battery health. So I made a voltage devider (R1 = 1000 ohm and R2 = 477 ohm) over the battery and connected it to analog A0 to read out the voltage on a 0 - 5 V scale and reconvert it. When I measure the voltage off the battery with my multimeter it is exactly the same voltage as I read on my tablet
Being able to check correctly the battery voltage I wanted to add an extra sensor to my setup I tought it would be a piece of cake adding an extra voltage devider and this is where things go wrong. The sensor is a 12v temperature sensor with a resistance varying from 1200 to 0 ohm. So I connected it on the existingcircuit.
I tried to read out A0 and A1 and I noticed my readings of A0 where less accurate.
After some measuring I noticed that when I add the second voltage devider the restance of R1 is altered. Not being 1000 ohm anymore, but only 990 ohm. For a reason I don't understand R3 and R4 are influencing the values of the resistor R1 and faulting my measurments.
Things are only getting more worse when I also tried to add another 12V sensor with a variable range between 3 and 160ohm. Now R1 was down to 980 ohm...
So long story short, how can I use multiple sensors ond multiple voltage deviders without them affecting eachother. I still have some other sensors to add and if each time they affect eachother it is impossible to calculte the right correspondance for my output
Any idea is welcome because I can't figure it out.
Thank you
Firstly your dividers are way too low an impedance for healthy battery life. That 300 ohm resistor in
the last circuit would be burning 0.5W and would need to be a power resistor - that's very wasteful.
Secondly how did you determine that R1 changed resistance? You known you have to measure
resistance out of circuit?
Generally for Arduino use arrange that dividers have roughly 10k impedance unless there's a reason not
too, which in simple terms mean the smallest resistor of the pair should be about 10k. So for instance
the battery voltage divider could be 22k:10k. This would drain only 0.4mA, rather than 8mA for your 1k:477R
divider. The 300 ohm divider could be draining as much as 40mA.
Your high current loads may be causing the battery voltage to droop, this could explain your accuracy problem.
Thank you for helping me out. I'll try to follow up your advice, but how can I pump up my impedance?
I calculated the impedance for the resistors according to the sensor impedance, to bring it down from 15V to 5V readings. Do I add an extra resistor in series to the sensor and recalculte the first resistance?
The arduino is also fed by a 14V car battery attached to a 75A alternator. Would the 40mA really matter (just asking. I'm a complete noob on this)
For the measurment. I did the first voltage devider on an onpowered board and measured the impedance of R1 (= 1000 ohm) Than I plugged in the second voltage devider and remeasured over R1 and it dropped ...
"For the measurment. I did the first voltage devider on an onpowered board and measured the impedance of R1 (= 1000 ohm) Than I plugged in the second voltage devider and remeasured over R1 and it dropped ..."
It will and each time you add another resistor sensor series circuit you are adding another parallel changing resistance. The lower Rtotal the higher Itotal since resistance and current are inversely proportional. As mentioned you really don't want low resistances and again each parallel combination or resistor sensor lowers the Rtotal. So what you are seeing is normal and in this case an undesired effect.
Ok Ron_Blain, that was already a post that makes sense. So is there a walkaround or another way to wire up these sensors without having any interference.
How would you wire this up?
Just for reference:
A0 needs to receive a voltage I can measure to measure the battery voltage 0 -15v)
A1 needs to receive voltage from the temp sensor (variable from 0 ohm to 1200 ohm)
A2 needs to receive voltage from the pressure sensor (variable from 3ohm to 160 ohm)
Thanks in advance guys. Your help is really appreciated
A0 you want a 3:1 divider circuit, we can get that later. A1 and A2 it would be good to know the sensor types. We know for example for A1 and A2 we will want a voltage proportional to whatever the sensor is measuring. The reason it is important to know what the sensors are is for example they are temp sensors like a thermistor then we need to watch how much current we pass through to avoid what we call "self heating effect". So there are considerations. This is when a temp sensor begins adding its own heat to what it should be measuring, it gets ugly.
OK here is what I came up with. The problem goes back to the low sensor resistances. We want to keep the current down but have usable signal (voltage).
There are other ways to do this and use some amplification but that makes things more complicated. I just wish it were easier to get a greater voltage swing. The total current worst case is about 6 mA if I did my homework right. R2 and R4 are your sensors, R1 and R3 are resistors you need to add. Your 3 - 160 Ohm will give you about 4.3 volts to 5.0 volts out and the 1200 - 0 Ohm (really can't have zero) will give you about 4.0 to 4.8 volts.
Will that much work for you?
Got to thinking about this and remembered you mentioned 12 Volts and I can change that but are we talking about automotive 12 volts which can be anywhere between 12.6 VDC and 14.0 VDC so any divider outputs will swing along with the voltage. This is why important sensors in automotive run on a regulated 5 volt supply like the MAP sensor for example.
Your is right. The voltage can be between 12.6v and 14.7v
This is why in my first schema I used a voltage devider to calculate the real tension of the circuit.
According to this value I could than calcultate the impedance of the sensors according to the divider voltage and the fixed pre-resistance. At least that was my initial idea.
Is that the right way to do? I would like to continue on a 12v -15v power source as I have a constant source of power for this, and not havint to bring it back to 5v ...
Your idea ? And how would the schema look like with 12v-15v ?
Well we know automotive voltage will vary depending in several variables like engine speed. You really want a constant voltage and more important a stable voltage. Using a LM7805 is inexpensive and there is always a step down (buck) converter option. Both are inexpensive and easy to use. Not like we have a high current demand. Amazon offers a wide range of buck converters which would be my first choice.
You only need one. There won't be that much current. Getting a LM7805 in a TO220 case will make for easier mounting on something and they also make TO220 case heat sinks a Google will bring up dozens of flavors I would just get a simple slip on type. For a few cents it can't hurt. Also look at a LM7805 Data Sheet as you will want a few caps for input and output. I would likely put maybe a 1,000 uF 50 Volt rated on the input along with a 0.1 uF just to help clean up automotive noise. They do make 5 volt regulators for automotive applications but really a 7805 will do fine.
Ron is correct- your resistors are WAY too small. The analog inputs draw practically no current with an input impedance specification of 100 megOhms, so your voltage divider resistors should be at least 50k and up.
But given the low resistance of your sensors, you measurements may be less precise.
This could be an application for a Wheatstone Bridge.
