Picking correct level of resistance for a voltage divider

Question

I want to measure my DC voltage source (aka batteries), vs my known Dc voltage from my DC-DC regulator. To do this I'll need a voltage divider. I can calculate the resistances I need - but what I'm not sure about is the over all level of resistance I should go for.

If the voltage divider resistance is to low, I'll be burning current for no effect. If it's excessively high.... is there any downside if the current is near zero? Should I just got for 100K ohms and leave it there?

The math is easy. But picking the right target? That's hard.

Answer 1

When you factor in the input impedance of the measurement device, your options will reduce but yes, there will be a range of values that can be chosen and you will tend to choose the option with the highest impedance that still produces acceptably low errors in measurement.

For instance, the ADC inbuilt in such devices as a PIC micro will not want to see a signal impedance more than a few kohm. As you go to greater values than a few kohm, the measurement error will increase and this can be alleviated, to some extent, by adding a capacitor from input to 0V.

So, read the data sheet for your measurement device, restrict your range of impedances and choose the highest set of resistors that give you an acceptable measurement error.

Sometimes you find that there is no overlap and, in cases like this you would use a spare IO line to activate a MOSFET to connect your battery to a potential divider. Then you can use quite low values and get good measurement accuracy and the battery energy spent is only for a short period of time in the bigger picture of things: -

Taken from this stack exchange answer (by me).

Answer 2

The impedance of a voltage divider depends critically on
a) what's driving it and
b) what's loading it

Given your driver is either batteries or a PSU, then you can go down to very low values successfully.

You don't say what your load is, but I will assume it's a DVM. Now most of these have a 10M input impedance, but I did buy a cheap one recently that was 1Mohm input. If you drove that from a 100k divider, you would have up to 10% voltage error, not very good. Be warned that some ADCs have lower input impedances than that.

Keep your divider impedance

If we assume you have a worst case 1Mohm DVM, and want the loading effects to be negligible compared to the DVM error and the resistor tolerances, so 0.1% from loading, then you would want 1kohm impedance divider. That won't embarrass your PSU, and is probably OK for your batteries.

If we assume a 10Mohm DVM, and a 1% error, then your 100k is OK.

Answer 3

The answer also depends on your local EMI environment. Say you have a bunch of 12V car batteries for some electric car conversion project. Then you’re probably gonna have some heavy switching noise from your power electronics. Even the short distance (10 cm) between battery poles to a battery-mounted voltage-measurement chip can attract a lot of EMI. So your best bet then is to draw some measurement current, say 50mA (240 ohm resistor in this case) into your shielded measurement circuit which could be based on the other peoples suggestions here. To decrease your used energy, you can draw current only when measuring, say during 100 us per second (implying power loss of 60uW in my example, excluding circuit losses).

Answer 4

Consider just using a slidewire. All it takes is a meter stick and a length of wire to make a divider with 1:1000 resolution. Typical voltage-measuring bridges used a zero-center meter and momentary key so the source (or reference battery) only need be connected for a few milliseconds.

Nowadays, of course, a 1M to 10M voltmeter is inexpensive and convenient. The slidewire has an advantage if you want to measure microvolts, I suppose.