My goal is to replace the analog battery gauge in the dashboard - with a more accurate, digital gauge.

My first idea was to look for a specialized IC. I'm familiar with battery gauges for LiPo batteries like the STC3100IST, so I thought "probably there's something similar for lead-acid batteries, since they are used in nobreaks".

And indeed there was, like the bq78412 - but all hard to find, overwhelmingly complex. Also, they are designed to measure multiple cells, with current measuring features which for a car are of course not useful (think about installing a current sensing resistor...)

Then, my second choice was just to use an ADC after a voltage divider made with resistors to reduce from the 12-14V to hundreds of milivolts - giving a 10x margin to survive voltage spikes.

But this second choice seems almost too simplistic. Will it work? Will it survive the harsh environment a car is? Are there any feasible alternatives?

Thanks in advance.

  • \$\begingroup\$ Are you trying to measure battery state of charge (SOC), or just measure terminal voltage in a different way? \$\endgroup\$ – HikeOnPast Nov 17 '12 at 0:19
  • 1
    \$\begingroup\$ That presumes a linear or near-linear relationship between SOC and terminal voltage, which you may find lacking when monitoring Pb-acid batteries. (en.wikipedia.org/wiki/Peukert%27s_law) What is of more interest to you: voltage or SOC? \$\endgroup\$ – HikeOnPast Nov 17 '12 at 2:33
  • 1
    \$\begingroup\$ Most accurate lead acid battery SoC testers do an Interrogate and Response load test. This is related to the worst case cell in series and it's specific gravity and then scaling by the number of cells and capacity of each cell. \$\endgroup\$ – Tony Stewart EE75 Nov 17 '12 at 3:19
  • 1
    \$\begingroup\$ @DeanB SOC is more important, but it'd be nice if I could have both. Mainly, I want to detect two situations: (1) when the battery is approaching a too low charge while the car engine is off (for whatever reason - sitting idle for too long in the garage or the headlights were left on) and (2) when the battery is not recharging correctly with the engine on (if the alternator isn't working perfectly or the battery is old/defective) preferably before the battery gets too much discharged, so the driver could be warned in advance. \$\endgroup\$ – fceconel Nov 17 '12 at 21:23
  • 1
    \$\begingroup\$ @NickAlexeev my answer to dean also answers some of the points you asked, for the others: yes, I want to measure the standard car battery; and no, I'm doing nothing more than is regularly done (i.e. supply the engine electric system, power the car lights, the car stereo, etc.). \$\endgroup\$ – fceconel Nov 17 '12 at 21:25

You actually don't need a sophisticated setup (like bq78412) for measuring the SoC of a starter battery in a car. Your second approach with measuring the battery through a voltage is adequate.

The lead-acid batteries can be split into two (2) groups:

  1. Starter batteries, such as the ones found in common cars and motorcycles. These always operate high SoC under normal conditions. They are discharged by the starter for just a few seconds. That takes may be 5% of the charge. Afterwards, they are recharged by the alternator in just a few minutes. An important requirement for starter batteries is the ability to deliver high current.
  2. Deep cycle batteries are designed to discharge down to 20%. Probably, bq78412 was made with deep cycle batteries in mind.

If you know the battery capacity within 20%, you can still detect the faults, which you're describing (lights left on, defective alternator or battery).

P.S. This question have been up for a day, there are no responses, we've got use cases from @fceconel in his comments. I feel that I can provide a systems-engineering answer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.