I was thinking of using 1 of my spare boards to record voltages for battery load tests. The problem is the AIP's are only 10 volt. The batteries I need to test are 24V and 36V. A voltage divider will change voltage as the battery voltage drops and create a slow error that will exist all the way thru the test. So what are some thoughts on the best way to accomplish this?
Battery voltage monitor
- Thread starter todster
- Start date
BraveSirRobbin
Moderator
Not sure what you mean by create a slow error using a voltage divider.
I don't see why you cant use two resistors as a divider and monitor voltage across one of the resistors. You would have to develop an equation to convert the voltage read to battery voltage (the CAI unit should be able to do this right?).
I don't see why you cant use two resistors as a divider and monitor voltage across one of the resistors. You would have to develop an equation to convert the voltage read to battery voltage (the CAI unit should be able to do this right?).
If my resistors are 4k7 and 14k1 and I use the 4k7 to measure the drop across. At 38v the drop will be 8.1v and at 32v the drop will be 6.9v. To say that the14k1 would always be fixed would also be wrong. So the question is how to deal with this. Create a table or calculation that acounts for this and figures it into the calculared voltage?
BraveSirRobbin
Moderator
I agree with az1324, select resistors values so you are only drawing milliamps through them. With 1% you can calculate the actual swing you could expect, but you could even possibly reduce this (not taking into account any drift of course) by using actual measured values, or better yet, use a variable supply in place of the battery and perform your own check points (supply voltage vs resistor measured voltage) and ensure that it is indeed linear as well as develop the conversion using those numbers.
What kind of accuarcy are you looking for?
What kind of accuarcy are you looking for?
CAI_Support
Senior Member
If you have a stable reference voltage, you can measure the difference between your battery and the reference voltage, so that you can have better reading. For example, if your 24V battery is fluctuate between 23V and 29V, the difference is six volt. You can setup a reference voltage at minimum voltage 23V to measure how much the battery voltage above the reference voltage.
The only problem is that since DC are all ground together, the AIP inputs can only be used as one ground, so that you can only measure batteries in one range, not both 24V and 36V, in this method.
The only problem is that since DC are all ground together, the AIP inputs can only be used as one ground, so that you can only measure batteries in one range, not both 24V and 36V, in this method.
BraveSirRobbin
Moderator
BTW, you can get some pretty precise and stable (wire wound) resistors (will pay more).
For instance here is a Digikey example of one with +/_ 0.01% with a thermal stability rating of +/_ 10ppm/Deg C.
For instance here is a Digikey example of one with +/_ 0.01% with a thermal stability rating of +/_ 10ppm/Deg C.
By far the "simplest" way is a zener diode and a small load resistor.
I use this for monitoring my own 48V battery bank and to monitor the nominal 105V input from my solar panels.
If you have a 24V battery, you will probably be interested in the range from about 22 to 30V
A 21V zener diode from your battery to the WC board input, and perhaps a 10K resistor from the zener to ground to supply some bias current, will make your WC board read about 21V low. So 24V will show as 3V, 30V will show as 9V etc.
This gives you far more resolution with your 10 bits than a purely resistive divider.
If you are using my terminal board and amplifiers, there is a jumper block for each input where you can install such a zener, and the 20-turn trimpots will let you adjust the gain for an easy to work with calibration.
Edit: You can use two diodes in series, one low voltage and one higher voltage - to get the temperature coefficient "close to" zero.
I use this for monitoring my own 48V battery bank and to monitor the nominal 105V input from my solar panels.
If you have a 24V battery, you will probably be interested in the range from about 22 to 30V
A 21V zener diode from your battery to the WC board input, and perhaps a 10K resistor from the zener to ground to supply some bias current, will make your WC board read about 21V low. So 24V will show as 3V, 30V will show as 9V etc.
This gives you far more resolution with your 10 bits than a purely resistive divider.
If you are using my terminal board and amplifiers, there is a jumper block for each input where you can install such a zener, and the 20-turn trimpots will let you adjust the gain for an easy to work with calibration.
Edit: You can use two diodes in series, one low voltage and one higher voltage - to get the temperature coefficient "close to" zero.
CAI_Support
Senior Member
Ross,
You mentioned this zener diode trick before, it is a wonderful idea!
You mentioned this zener diode trick before, it is a wonderful idea!
I'm getting closer to actually doing something with this. So with the zener idea in mind, my thinking is to use a 10v for the 1st unmonitored drop out. Then to auto select the battery voltage ie 24 or 36 I will monitor the 2nd and 3rd zener and determine if I need to use analog 2 or analog 3 for my voltage and add it to the 10's from the previous analog tests. IE if the voltage is 34 volts analog 1 will show full scale of 10v as well as analog2 so we add 30 to whatever analog 3 is. For a 24v if analog 1 is full scale and analog 2 is not then we add analog 2 to 20V.
CAI_Support
Senior Member
If you have a relay board to switch between different battery connections and zener diodes, you could have that measured in one AIP by switching the inputs.
CAI_Support
Senior Member
Your program knows which voltage source it measuring, so that in your code you could adjust that.
