Ammeter for PWM-controlled Motor

[calzap]

I'm interested in installing an ammeter for a large PWM-controlled motor in a small electrical utility vehicle.  Power is from 6 large 6-volt lead-acid batteries in series.  Motor is series wound, 8 HP,  6 kW at 1600 RPM.  Not sure what locked-rotor power consumption might be.  Controller is a PWM type, 400 A rated, made by/for the vehicle manufacturer (Textron),  and I haven't been able to find additional specs on it.  However, a replacement would be a Curtis M series: http://curtisinstruments.com/index.cfm?fuseaction=cDataSheets.dspListDS&CatID=67&siteID=1&langid=1.  Curtis PWM frequency is 16 kHz.  Suspect Curtis made the original for Textron.

So, I should use a 500 A shunt resistor, which are available with a voltage drop of 75 millivolts.   But then what?  An ordinary digital millivolt meter won't like pulses though.  I'm thinking of using an  RC filter, but what should the specs and design be and is there a better way?  The ammeter will be permanently installed in the vehicle, so needs to be relatively inexpensive.  There are some  cheap RMS multimeters on eBay and elsewhere that could theoretically do the job, but I have my doubts about them handling 16 kHz square waves.  The ammeter does not need to be highly accurate.  So, maybe an old analog millivolt meter would work by mechanically averaging the pulses.

Any ideas will be appreciated.

Mike in California

[SeanB]

place the sense resistor in the power supply line to the inverter, and then use a RC on the sense lines, 1k and 1uF will work well enough for most DVM modules, though you will need to use an isolated power supply ( simple enough to get as a block DC-DC converter) to provide power to the display, as it likely is using a 7106 or variant with a DC path through to the supply pins from the input lines.

Often the sense resistor is done by using the battery lead to act as the sense resistor, just placing the sense wires at each end gives a usable if needing calibration voltage which can be simply scaled for the meter to read from 0-100% of current.

[calzap]

Thanks for the advice Sean.  The vehicle has 12 VDC available for accessories, like lights, horn, etc., which is created from taps off an adjacent pair of the 6 V batteries.   Couldn't this be used to power the display or is there some noise or short circuit consideration than I'm missing?  One nice thing about the electrical system in the vehicle is that the chassis is not part of it; every device has + and - wires to it.  So no worries about shorts to the chassis,  corroded ground connections, etc.    Shielded, twisted wires will be used to carry the sense signal from the shunt resistor to the meter.  The batteries are close to the motor and PWM controller, so the shunt resistor will be in an area likely to have EMI.  Meter will be in the cab, 2-3 meters away.  RC filter will be at the meter, not the shunt resistor.

I'd like to use my scope to look at the signal at the shunt and at the meter, but it's mains powered.  I could jack the wheels off the ground, but that would be an unrealistic, low torque test.  So a hefty 12V battery and inverter will be used to power the scope.

Mike in California

[johansen]

The motor inductance at 16 khz keeps the current flowing. i would estimate less than 10% current ripple, so an average reading digital meter will be fairly accurate, to about  +-1%.

[calzap]

You are right about ripple being 10% or so.  Attached is a figure from EDN in a 2001 article:  http://www.edn.com/design/test-and-measurement/4379582/Measure-PWM-motor-efficiency

Main point of the article though is how difficult it is to measure current in PWM applications.

Mike in California

[David Hess]

A voltmeter reading DC voltage averages the input and will reject a *lot* of noise.

Given the current levels involved, I might use a DC current clamp instead of a current shunt resistor for safety and galvanic isolation.

How much accuracy do you need?

[calzap]

+/- 5% accuracy would do; maybe even +/- 10%.  It's primarily to keep tabs on the motor and controller for changes that might indicate need for service or impending failure.  I thought about a clamp-type ammeter but have several concerns.  First is cost.  The clamp would have to be 2-3 meters from the readout.  It might be possible to shorten this though.  The forward/reverse switch is in the cab and handles the full load.  I find this surprising and would have thought a relay would have been used instead of long runs of heavy copper cable.  A "solenoid" (Textron's term; really a contactor) is used to apply power to the controller anyway; it would have been possible to include forward/reverse into it with a few more contacts.  The meter/clamp should be capable of handling 400 A.  Second is how well a clamp-type ammeter can handle the ripple.

Mike in California

[David Hess]

I was thinking something like a Fluke i410 when used with any standard multimeter.  I do not think it will have any problems measuring the average DC current in the presence of ripple and the leads to the multimeter can be extended.  There are similar but less expensive non-Fluke versions of the same DC clamp of course or you could pick one up used.

Battery life is only 60 hours but with care it could be externally powered.  A current shunt has the advantage of not requiring any power to operate but I might want to avoid breaking a 400 amp circuit even at low voltages if I can avoid it.

DC accuracy is suppose to be 3.5% + 0.5 amps over temperature with additional drift outside the range of 18C to 28C but there are more accurate clamps like the Fluke i1010.

I have the older Fluke Y8100 which uses the same design.  It is very handy for measuring things like starter and alternator DC currents or even solar panel currents.

[calzap]

David, thanks for the info, but this is to be a permanent installation in the vehicle and dedicating a Fluke 410 for that purpose is a little pricey.  I can get a digital millivolt meter for US$10, and capacitor, resistor and cable cost very little, so that's what I'll try first.

Mike in California

[David Hess]

You could also use an analog meter with an external current shunt.  Then no power connection is required.

[calzap]

Will use a 500 A shunt that has a drop of 75 mV.  Run sense wires via shielded cable to digital voltmeter and have RC filter at the meter.

Mike in California

[Zero999]

How about using a Hall effect current sensor?

[calzap]

I'm looking at using a Hall-effect transducer, which is a neat idea.  However, it will be the back-up plan if using a shunt resistor, RC filter and ordinary digital voltmeter doesn't work.

Mike in California