Lowering signal voltage

[exuvo]

I have 4-cell LiPo batteries that are between 12V and 16.8V (initialy thought max would be 15V) depending on charge and I'm currently measuring this on a 3.3V 12bit ADC using a voltage divider followed by a sallen-key op-amp filter. It currently seems to work fine but I'm not sure if the voltage divider disturbs the filter characteristics, maybe somone could enlighten me on that?

Because I know that the input voltage won't go below 12V i was thinking that i might be able to somehow subtract 12V from the input and instead of measuring 0-17V i could measure 0-5V with higher resolution. My first idea was to use a diode in series but then i'm unsure how that would affect the measurement as there is a capacitor in the filter, does it need a bleed resistor? My second idea was to use a negative supply (with fig 4 from here) and mix that in but I don't understand how to get a negative voltage from my 0V, 3.3V, 5V or ~16V supplies. What is the best way to accomplish this?

The op-amp is a LMC6482IN running on 5V and the micro with the ADC is a LCP1768-H running on 3.3V (5V tolerant IO). This is for a robot I am building picture, more pictures, build log in swedish.

[pmbrunelle]

The voltage divider formed by R2 and R7 is equivalent to an ideal voltage source in series with a resistor. "Thévenin's Theorem" will allow you to transform your voltage divider into a voltage source with series resistance. Then, you can analyze your filter like a standard Sallen-Key circuit.

I would try to work with a modified difference amplifier:
http://en.wikipedia.org/wiki/Operational_amplifier_applications#Differential_amplifier_.28difference_amplifier.29

With this configuration, you can connect V2 to the battery, and you can subtract a precise positive offset voltage V1. V1 can be generated from something like a TL431.

Connecting a capacitor in parallel with Rg will provide some low-pass filtering. To analyze this filter like a normal RC low-pass, you can use Thévenin's Theorem again.

[tszaboo]

Why do you need the sallen key? For all the AC signal coming from a battery?
If you want to lower the voltage, you can use a difference amplifier. Make a reference voltage, make 12V from this, and connect it to the negative input, and indeed, your resolution will be 3-4 times higher. If you properly
Or take a 5V reference, divide your input by 2,4 and adjust your gain for full scale (again, a difference amplifier). If the difference amplifier is properly protected, you dont need higher supply voltage than the battery.
Also, be aware, that batteries have the inherent capability to power your circuit even when it is off.

[exuvo]

Thanks, the information was very useful. I used the sallen key filter beacuse I didn't want to measure any noise coming from the main motors, tower rotation motor or switching buck converters. I have also added oversampling (1kHz) and averaging (latest 100 values @10Hz) in software so the filter might not be neccessary anymore.

I updated the schematic with the op-amp in differential mode and component values that i have calculated and verified in simulation. Signal goes through voltage divider with ratio of 0.22 which makes the 5V swing into 1.1V swing. Then the op-amp, with a gain of 3, subtracts 2 * reference 3.96V and the result is 0-3.3V. Anything else i need to think about?

Another question is in regards to my hall effect current sensor ACS756 (top half of attached schematic), it needs a minimum load resistance of 4.7k Ohms and has a maximum load capacitance of 10 nF. I currently have a 68K resistor (not sure what my thinking was for that value) but I'm wondering what the optimal value might be and what difference, if any, it makes?