Trouble with current sensor

I'm trying to use a honeywell cslh3a45 open loop current sensor (data sheet at http://sccatalog.honeywell.com/pdbdownload/images/cslh3a45.pdf) to measure AC current. Hopefully I can use this to build a mediocre energy meter.

I know the sensor is working because when I hook it up it gives me half of the the supply voltage as it's supposed to and the output varies when I move a magnet near it. I chopped up a power cable and ran one lead through the hole, but when power goes through the cable (I know power goes through the cable because the device turns on) the output voltage doesn't change from supply/2. I remember seeing something about needing a coil, so I tried wrapping some magnet wire around the frame but still no cigar. Any idea what I could be doing wrong?

 

Hi sigfrid3141,

Perhaps you could provide some more information. The transducer data sheet is quite useful.

How are you measuring the change in output signal when the monitored current is running through the window - with a multimeter (AC, DC - what range?), oscilloscope .....?

 

Hello again,

If you are passing an AC current through the window you need some idea of the actual current value - the device appears capable of responding linearly to currents of +/- 45 Ampere-Turns. The sensitivity is quoted as a max of 22mV/NI. So the maximum output variation is only likely to be +/- 990 mV at best. If your AC source can only supply a couple of amps you can increase the sensitivity by passing the wire several times through the window - I think this is what you were referring to anyway. So if you can source up to 5 Amps you might pass the wire through the window 8 or 9 times.

If using a multimeter on AC range you will need to remove the DC offset (Vs/2) from the measurement. Presumably you are using AC current at mains frequency.

You could use a DC supply for the source as the current through the window - again you could wind several loops through the window to increase sensitivity. With DC current you can take a multimeter DC measurement and subtract the zero current offset voltage (~Vs/2). The change in Vout above or below Vs/2 will be the actual output voltage change for the given input current [ampere-turns].

 

Put a cap in series with an AC meter to read the AC component of the output from your sensor. (my old Simpson 260 has one built it )

 

Thanks all for your advice.

You are correct, I am running US mains current through the window. I was able to squeeze four turns through the window. Unfortunately my multimeter is terrible and its lowest AC range is 200V.

I thought that it might work to put the sensor output through a bridge rectifier. I put one of the AC inputs of the rectifier to ground, one to the sensor output and then measured the DC voltage across the two outputs. It was a bit lower than Vs/2, but didn't respond to current going through. I tried putting a .1uF cap in series with the sensor output but that made no difference.

When a proper multimeter arrives later this week I'll be able to get a useful AC reading.

 

Also, just to be sure I'm not crazy I supplied the sensor with 5VDC and wound a conductor 9 times through the window, for a sensitivity of 135mV/amp. Then I connected a 100 ohm resistor to said conductor and passed 5VDC through that. In theory that means there should be .05 amps going through the conductor and the sensor should output about 2.508 V. My multimeter isn't very accurate, but it did register an increase of around 10mV whenever current went through the conductor.

Long story short, I have it working for DC but not AC. I'm hoping the problem is just that my multimeter is too lousy to register the tiny changes in AC voltage that I would expect.

 

With the DC setup proposed you would expect to see a change in output of 6.8mV to 9.9mV. Based on your observation it looks like you are in the "ball park".

Hopefully, your new multimeter will resolve this more accurately. If it comes with a relative measurement option that will make the task a lot easier.

In your test setup you are trying to measure quite small currents with a device designed for a much higher range - at 0.45A effective current in the loop you are still 2 orders of magnitude below full scale range!

Adding a series capacitor is a great idea for the AC case - again you'll need to inject a reasonable amount of current through the window to obtain a measurable response. A 0.1 uF (26.5KΩ @ 60Hz) shoud be OK with a multimeter with 10MΩ input resistance. Again your meter sensitivity on AC range may be the limiting factor at the moment. If you have access to an oscilloscope this might make life a lot easier for you.