Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Parasitic Drain Wattmeter?

Status
Not open for further replies.

Thender21

New Member
Hello,

I work in automotive. It is getting very hard to diagnose an intermittently dead battery.

Sometimes there are very small short circuits or switch problems that cause a small parasitic draw of lets say 100-200 mA. Its the fact that they are intermittent that is a problem.

Sometimes a module has a bad input that causes it to wake up repeatedly or to wake the others and this can cause battery drain.

In most all cases there is no diagnostic information stored. Zero.

And if the problem is intermittent there is no cost effective repair. Nothing to check test or inspect except the entire car. For a problem that is not even there.

I am wondering if I could make a small device to record an indication of current flow and put it inline with a fuse or a ground connection or connector so that there is a smoking gun vaguely in the direction of the problem.

Whole cars are sometimes bought back by the manufacturer because of this issue. Othertimes a $5000 vehicle wide wiring harness is replaced and may or may not fix it.

Meanwhile, the car is useless. Who knows if it will start? Only a manual battery disconnect and reconnect every day could work around that. Even in that case, a lot of unwanted things can happen as a result of losing battery voltage...
 
One thought is to set the vehicle up with small low wattage resistors in place of certain control/backup power circuit fuses.
If a current draw goes over a specific limit for more than a few seconds the tiny resistor burns up showing that circuit tried to take more power than it should.

If you know what the maximum current a circuit should take on standby the resistance for that resistor and its rated wattage can be calculated to allow that without burning up but not much more for very long.
 
I like the idea of translating current flow into a physical change.

There could be 150 different fuses in the 5-20 amp range staggered hierarchially from 100 and 50 amp fuses downward.

The goal is to set the car up and let it be driven.

One thing I remember thknking of is using a small dial like an analogue clock that receives power whenever the key Off power mode is indicated.

And then if current flows on a circuit still the clock counts the "overtime".

Maybe something like a voltage drop across the fuse could be used to indicate clock should be started. Via a transistor perhaps.

The clock itself would consume little power if connected to the vehicle battery or alternatively a separate battery could be feasibly installed to power it. It could be mechanical or digital

Circuit activation time during key off is sufficient. Amp draw is probably not needed since intermittent large amp draws are rare and easier I think.

It only has to help narrow down from 400 possible circuits to 40. I cant analyze a whole car but I can use heiristics, experience, inspections and circuit analysis to get close to home from there.

Thanks,
 
You pose some tough problems. The general place to start is here: **broken link removed**

But, it doesn't help an intermittent problem at all.

I'd probably be looking at aftermarket stuff, Dome fuse, power source and voltage regulator first.
 
I have one "off the wall" idea after sleeping on it for a little bit.

1. Some gizmo that operates off of an SLA battery.

2. Charger that somehow knows the car is running. Alternator ripple is one turn-on method. Cigarette lighters are either always on or IGN only. Kinda wish they were selectable. Monitoring system needs to know when engine is running.

3. The negative side of the battery cable can act as a current shunt. So, voltage measuring across this is surely possible.
(It may or may not find negative battery cable connection problems)

4.The +Bat to alternator terminal should take care of alternator type issues like diodes, winding breakdown and regulator.
(Measure voltage there too)

5. Maybe design some sort of power line disturbance monitor kind of software with a threshold which is a trigger.

6. I'd also monitor the dome circuit, It keeps track of car doors, trunk light possibly etc.

7. In other words 3, 4 and 6 are triggers. They wake the processor up.

8. One could then record the "power" and/or currents for those circuits and tally 2 and 4 into watt-hours or watt-minutes. This is some sort of threshold that the car goes in for analysis. Something like watt-mins accumulated with engine off.

9. If you want to be able to somehow find the problem, I might suggest a custom bed of nails that uses spring loaded pogo PCB test probes that would contact the test points of the fuses in one or more fuse blocks.

10. You would then multiplex the inputs and try to read differentially the voltage across the fuses. You would have already have told the computer the type and amperage of the fuses ahead of time, so hat voltages can be converted to approximate currents,

Re-iterating: Te dome, alternator to battery and bat neg to ground voltages measured across the respective cables becomes the microprocessor trigger to wake up.
it does the total watt-min and the watt-min on the alternator side of things.

Now it tries to measure the individual fuse drops using a bed of nails that's attached - kapton Tape?

This **broken link removed** is one such company that makes probes. A bed of nails can be made on a polycarbonate sheet.

What creates issues is multiple fuse boxes. At least two. One inside and one in the engine compartment.

The bed of nails creation could come from a photo, the turned into a drawing and then CNCed.
 
I have one "off the wall" idea after sleeping on it for a little bit.

1. Some gizmo that operates off of an SLA battery.

2. Charger that somehow knows the car is running. Alternator ripple is one turn-on method. Cigarette lighters are either always on or IGN only. Kinda wish they were selectable. Monitoring system needs to know when engine is running.

It's actually the opposite, we need to know when the engine is off and the key is out. At that time, more or less no current or very little current should be flowing. It should certainly be less than the amount that would drain a battery overnight.

It should be easy as any number of things change with key position.

3. The negative side of the battery cable can act as a current shunt. So, voltage measuring across this is surely possible.
(It may or may not find negative battery cable connection problems)

4.The +Bat to alternator terminal should take care of alternator type issues like diodes, winding breakdown and regulator.
(Measure voltage there too)

5. Maybe design some sort of power line disturbance monitor kind of software with a threshold which is a trigger.

6. I'd also monitor the dome circuit, It keeps track of car doors, trunk light possibly etc.

7. In other words 3, 4 and 6 are triggers. They wake the processor up.

Right, a wakeup / detection function is necessary.

8. One could then record the "power" and/or currents for those circuits and tally 2 and 4 into watt-hours or watt-minutes. This is some sort of threshold that the car goes in for analysis. Something like watt-mins accumulated with engine off.

Because the batteries are rated in Amp Hours, as one system, there is only so much time they can provide current at rated voltage...

9. If you want to be able to somehow find the problem, I might suggest a custom bed of nails that uses spring loaded pogo PCB test probes that would contact the test points of the fuses in one or more fuse blocks.

10. You would then multiplex the inputs and try to read differentially the voltage across the fuses. You would have already have told the computer the type and amperage of the fuses ahead of time, so hat voltages can be converted to approximate currents,

Re-iterating: Te dome, alternator to battery and bat neg to ground voltages measured across the respective cables becomes the microprocessor trigger to wake up.
it does the total watt-min and the watt-min on the alternator side of things.

Now it tries to measure the individual fuse drops using a bed of nails that's attached - kapton Tape?

This **broken link removed** is one such company that makes probes. A bed of nails can be made on a polycarbonate sheet.

What creates issues is multiple fuse boxes. At least two. One inside and one in the engine compartment.

The bed of nails creation could come from a photo, the turned into a drawing and then CNCed.


The problem can be explained simply and generalized from there.

Suppose we have a circuit with a switch and a bulb and a battery. On/Off, it works well. But the switch wears out and instead of OPENing the circuit, it has some resistance in the OFF position, and allows a very small current to pass through it.

The next day the battery is dead and the switch doesn't work.

In a vehicle there are some hundreds of circuits distributed across the car. When a hard short occurs, the fuse blows. And that provides diagnostic direction.

If a resistive short circuit occurs, it just slowly drains the battery. And these we diagnose by finding which fuse is providing power and going from there (while monitoring the net discharge rate in milliamps of the battery).

The type that is likely to bankrupt a car is one that occurs intermittently and leaves no evidence of where the fault occurred.

Starting from the battery we have 100 amp fuses, then 50 amp fuses, these feed power to fuse boxes around the car, in the dash, trunk, engine compartment, behind seats. From there power is fed directly to actuators and modules, and from modules to other actuators. There are 10-30 modules.

The engineers have different designs, so current flow often does not have a logical direction. I just dealt with a car that had one 20a fuse feeding 25 modules from the engine to the gas tank and within the dashboard and steering column, under the carpet. It had a bad splice. And several junctions where it joins six or eight circuits together.

That car had a lot of shunt-wiring that drove me nuts. A lot of circuits would have logical flow but have some kind of shunt going elsewhere that acted as a signal that the circuit activated.

So in a nut shell, I am wondering if there is a way to run a clock for example when the engine is off but current is passing through a circuit. I would apply this device to larger fuses first and then work down to Narrow the Field.

The whole body/chassis/engine is at ground potential so that means this device could be planted at a ground connection as long as there is enough voltage drop right there to indicate current flow. There are only a handful of groundsand a handful of large fuses, so a divide and conquer approach is how it could go.

Think of it like if you have ever seen the water meter for your house, they often have a numerical indication of consumption as well as a little black or red triangle. The triangle moves even if there is a tiny bit of wate flowing, like a broken sprinkler pipe in the ground.

it provides an indication of whether the system has a leak somewhere, or something is using water that shouldn't. It is only useful with all faucets closed, dishwasher and laundry machines off, etc. Much like my device would need to count mA so to speak with the engne off.

Regards.
 
I'm into automtive too, and fork lift trucks.
To trace currents I put togther a dc clamp meter, with one of these you could clamp it around the battery lead and connect it up to a data aquisition unit, and trace out current flow over hours/days/weeks.
You can buy dc clamp meters, dunno if you can buy one that will connect to an acquisition unit.
I've had stuck relays before and a diesel burner that refused to shut off, most annoying when you want to go out in the car and it doesnt start.
Another thing if you have indepth knowledge of the vehicle is to monitor the bus activity and see whats happening, assuming the dash/engine control unit is aware of whats going on.
 
Now, there is something you can make/buy easily, BUT you would need access to 120 VAC. The nanovoltmeter consumes 30 VA max.

You can use the resistance of the fuseholder an/or wire as an uncalibrated ammeter.

The keithley 181 Nanovoltmter ends up on ebay for < $300.00 USD. The connector will run you about $40.00 w/ postage if not included.

That meter has an analog output, so it would not be too difficult to use that to trigger a simple timer. Difficulty - The analog output +-2V, I believe, is only good for one range. It can withstand 120 VDC on the mV ranges which is good.

The mV ranges uses the connector and the volts ranges uses the binding posts.

So, now with a simple comparitor and voltage source (10 turn potentiometer), and the right "Hour meter", you could totalize the amount of time the voltage was above a certain value which means the current was above a certain value.

It becomes an "easy" way to get "something" that would help you, but It does come with restrictions.

==

If you look here: https://www.alliedelec.com/flir-com...=30980760979&gclid=CO3PmN77iMYCFUUXHwodrbsAFg

Which are Kelvin clips for a milli-ohmmeter.

and look at the datasheet on the page, it shows the corresponding Milli-ohmmeter.

So, if you could measure the resistance of the fuse removed at the fuse test points using Kelvin leads (two leads to each terminal), you would be able to convert the voltage measured by the nanovoltmeter to a current because you know R.

That milli-ohmeter can also help you find connection breaks in an unconnected harness. Some other milliohmeters are here: https://www.alliedelec.com/catalog/pf.aspx?fn=1595.pdf
The list isn't meant to be exhaustive.

==

As "food for thought: This **broken link removed** acts much the same way as the nanovoltmeter. Therefore, with a LOT more work, you might be able to get the device you want.
 
These, https://www.zoro.com/enm-hour-meter...gclid=CP2a04Sbi8YCFY4WHwodvCcAkw&gclsrc=aw.ds or similar, resettable hour meters exist. So, two; One for elapsed time and one for time over a certain current.

I do agree that a direct reading energy meter would be more appropriate and I'm not saying it cannot be done.

A signal for "Engine not running" is still difficult to do although IGN on (e.g. not in accessory) may suffice. e.g. ECM has power.
 
Most vehicles report a stream of data through the obd port on startup and switch off, and sending a pid request for rpm or any other engine data wouldnt get a reply without the engine running or at least key on, however an actual on/off signal may not exist.
You could do something like leave something switched on that is only on with the ignition and use that as a ignition on indication, the radio might be a good start.
 
100 to 200 mA leakage should not discharge a healthy 60 Ah car battery overnight, or over the weekend. 200 mA @ 24 hours = 4.8 Ah, say over a 50 hours weekend 10 Ah.
Ensure everything is switched off, I take the negative lead off and check if there is any spark, or not. Then as others already said, try to sectionalize the main feeds to alternator, fuse boxes. and fit a small 12 Ohm resistor or 200 mA fuse in series with one off the battery leads, and see if it blows overnight, it will give an indication of what type currents are flowing.
I know it all takes time but a systematic approach is needed.
Alternators sometimes have faulty diodes and can be intermittand.
 
Just been thinking about this one again. Have you tested the actual battery ? or fitted a known good one.
A proper discharge test and internal resistance may have to be done, to prove the battery hasn't got a internal fault ?
Sometimes the links between cells can go intermittand as well.
 
RODALCO
the OP said:
The type that is likely to bankrupt a car is one that occurs intermittently and leaves no evidence of where the fault occurred.

Remember, he's looking for an INTERMITTENT fault, not continuous.
He already has a handle on how to find continuous faults. Battery and alternator would be in that list.

So, I think he would want to build or buy something that doesn't seem to exist yet. It's OK if it works on one circuit at a time.

A fuse can be used as a shunt, so one of the difficulties is access. Some of the automotive fuses have two test points on them and MIKE suggested a plug. In one of the vehicles I work on, that plug would not work because one of the fuse boxes is between the door and the dash.

Access to the fuse/wire drop is mechanical.

Another difficulty is determining when the car is "not running". One can make that "manual" though. Some cars don't have ignition switches, so what does "IGN" really mean?

A "Nanovoltmeter" is a specific instrument which can measure low voltages, but it does have to be protected against overloads and transient dumps. An Instrumentation amplifier is also likely to work. Expected drops have been tabulated elsewhere on the web.

A "milliohmeter" is capable of measuring low resistances.

Once you know the shunt and the voltage across the shunt, you have current.

Voltage of the system could probably be ignored and just use mA/hr since reset or something.

1. Assure power is removed
2. Mechanically fasten (to fuse or wire)
3. Measure R
4. Attach to gizmo which hopefully could do #4
5. Normalize to measure current
6. Stick it to a power IC
7. Measure energy or just an average current drain/time when the car isn't running and maybe some other useful stuff like an alarm when some current is exceeded (used when actively troubleshooting).
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top