Car Ammeter Circuit

[barry wilkins]

Hi a few years ago with my previous Honda Car I had a problem with the battery.
I saw an idea where you could estimate the current flowing in the battery by measuring the voltage drop from the negative post of the battery to the chassis earth.

Has any of you seen any circuit ideas to do this ? or would I have to start from scratch ?

Barry Wilkins

[petehall347]

Re: Car Ammeter Circuit

what you are suggesting will only test the integrity of the grounding connections/wire .
ideally shouldn't see any voltage drop there .
you would be better fitting an ammeter inline on the suspect circuit .

[eccerr0r]

Re: Car Ammeter Circuit

Actually you can estimate it, the thing is that the voltage drop will be quite low if the cable is good. I'd guess it'd be no more than 50mV or so during starting, but not knowing the exact resistance of the cable, there's no way to accurately tell how many amps are going through.

You'll need precision op amps with low offset to accurately measure 50mV.

[barry wilkins]

Re: Car Ammeter Circuit

The earth wire has an impedance in the order milliohms ,so when you turn the engine over with the starter motor a current passes through the wire and will generate a voltage .
500 Amps and a resistance of 10 milliohms with equal 5 millivolts.

Barry

[petehall347]

Re: Car Ammeter Circuit

but you have to consider integrity of cables and connections . i regard testing like this is rather useless . it would be better to check voltage drop of the battery . or simply using a tool designed for the job such as an inline ammeter or clamp meter .

[eccerr0r]

Re: Car Ammeter Circuit

Indeed, as in the other thread, a hall effect clamp meter is probably best. A discrete shunt based meter is the old fashioned alternative.

Watch your amps. 500A * 0.01Ω (this is 10mΩ) = 5 volts, not 5 millivolts. I was guessing the cables to be fractions of a milliohm, not anywhere near single digit milliohms.

[redwire]

Re: Car Ammeter Circuit

Using copper (wire or PCB trace) resistance for a current-measuring shunt, the temperature coefficient of resistance is very poor. +0.393 percent/°C, so 10°C is 3.93% error.

A shunt is hard to keep well insulated and not short to the car body, and to take water.

I think DC hall-effect sensors are the best. Might need a bi-polar power supply, as battery current can be +ve or -ve.

Low-cost Winson DC hall sensors are on eBay and Ali, like WCS1800 modules (only 35A).

[petehall347]

Re: Car Ammeter Circuit

auto electrics are fun sometimes . i should know i do it as part of my day job . only time i look at potential differences is looking for bad connections .

[eccerr0r]

Re: Car Ammeter Circuit

When dealing with starter currents, a 5% error is not a big deal, plus you could temperature compensate by measuring ambient. Still not perfect but it's cheap... just that you'll need to make sure you get a quiet, low noise, and low offset op amp.

On the other hand, hall effect sensors also introduce error possibly due to magnetic leakage, wire positioning, etc. that may also generate an error. And these too may require a likewise op amp.

Now I'm curious how much of a voltage drop really occurs along the cable during starting... wish I had a meter that could hold peak voltage..

[petehall347]

Re: Car Ammeter Circuit

smoke and heat applies more than anything with starters . never overlook tight engines . it can happen . not often but can catch us out . a faulty starter will draw too much current if faulty or too little current . cranking voltage drop on a good battery will tell you where to look at next .

[Curious.George]

Re: Car Ammeter Circuit

The earth wire has an impedance in the order milliohms ,so when you turn the engine over with the starter motor a current passes through the wire and will generate a voltage .
500 Amps and a resistance of 10 milliohms with equal 5 millivolts.

No, 500A will develop 5 volts (5000mV) across 0.010 ohms.

Of course, that assumes the vehicle draws 500 cranking amps... (I think warmer climates and the little 4-bangers that are so common, nowadays, means the actual draw is probably much less!)

[Curious.George]

Re: Car Ammeter Circuit

but you have to consider integrity of cables and connections . i regard testing like this is rather useless . it would be better to check voltage drop of the battery . or simply using a tool designed for the job such as an inline ammeter or clamp meter .

Don't connect to the circuit anywhere that would place existing connections (or anything else that could needlessly add variation) in the measurement circuit. E.g., make two dedicated connections to the ground wire and let something else worry about how the connections to battery and chassis degrade, over time. Then, you only have to worry about the characteristics of that wire.

Of course, you can also use high-side sensing (the '+' lead). In either case, you have to be aware that the current flowing out of the battery can (and WILL!) be negative -- regardless of where you measure.

[FWIW, I think most modern cars measure high side]

If you aren't as interested in instantaneous readings, you can use an integrating converter (which inherently works around some of the noise issues you face).

Regardless, be sure your design can handle the sort of load dumps that a car electrical system experiences (far more destructive than the "noise" from plugs, etc.) And, that it is powered off the ignition and not battery (lest it run the battery flat).

[Curious.George]

Re: Car Ammeter Circuit

When dealing with starter currents, a 5% error is not a big deal, plus you could temperature compensate by measuring ambient. Still not perfect but it's cheap... just that you'll need to make sure you get a quiet, low noise, and low offset op amp.

On the other hand, hall effect sensors also introduce error possibly due to magnetic leakage, wire positioning, etc. that may also generate an error. And these too may require a likewise op amp.

The OP may not be concerned with starter currents. So, focusing overly on them may compromise a design. A shunt/HE/current transformer could be sufficient for most of the data of interest. Use a Rogowski coil tuned to the starter "impulses" to cheaply handle the larger magnitude signals associated with starting.

Or, put different sensors in different places: something designed to measure starter current AT the starter lead and something designed to measure "everything else" at another point in the circuit that excludes the starter's load (so the ground strap would not be a good idea)

[redwire]

Re: Car Ammeter Circuit

For a big V8 engine, 1ft. section of #4AWG wire at 500A gives 125mV (I'm using 0.25milliohms/ft).

Now, this 1 ft section of wire is dissipating 62.5W while cranking, it's going to heat up and cause errors. Something to think about.

Charging current of 10A is 2.5mV which is pretty small.

A 4-cylinder engine cranking current is much less, I think under 100A on a warm day.

[stj]

Re: Car Ammeter Circuit

just get a UniTrend UT-210E DC clamp meter.