solar charger voltage sense trouble

Thread Starter

jdelaney

Joined Aug 2, 2012
3
I 'm charging a 12v lead acid battery with a solar panel and reading the battery voltage with a microcontroller(the micro uses a mosfet on the high side for constant voltage). the problem i'm having is when the battery is charging at arount 4-6 amps the micro thinks the battery voltage is about 0.3v higher than at the actual battery terminals and gets worse with larger currents. when not charging the micro reads the battery voltage exactly. Can someone give me some idea how to remotely sense the battery voltage at the battery terminals that wont be affected by the charge current?

Thanks for any help!

Jamie
 

#12

Joined Nov 30, 2010
18,224
It is in the nature of most things to develop a voltage when current is passing through them. You can't stop that from happening. You can either develop a look-up table to convert to the correct voltage or check intermittently by turning off the charger.

I recommend that a look up table will be specific to the batteries and their age and not be very versatile. I think the "stop and measure" idea will work.
 

Thread Starter

jdelaney

Joined Aug 2, 2012
3
I'm aware the current is causing a voltage in the current carrying lines, what I was hoping for was a seperate set of small wires that carry no current that would hook directly to the battery and the micro can use a different a/d channel to read the voltage. what i can't figure out is how to connect this up. my battery gnd and solar panel ground connect together on the same board the micro is on, so my ground plane potential is moving around during charging. Is there a way to connect seperate sense wires to the battery that carry no current and are used just to measure the voltage? can I use an op amp to sort this out?

Thanks again.
 

#12

Joined Nov 30, 2010
18,224
You're talking about a "Kelvin" connection. Yes. It exists. It works. But it will not cause the battery to have a valid "resting voltage" measurement while you're pushing current through it.

The ground voltage problems are a different problem. For that, you need a "star" connection. No high current paths are combined with the sensor path to "ground".

If you know this stuff, it seems simple. If you don't, it's major frustrating and expensive. Can you post a drawing of what you have, so far? There are a dozen ways to wire this and dozens of places to make a mistake.
 
Top