Power over-limit detector, or can a current transformer power a buzzer?

[Artlav]

I want to make a device that would clamp onto a mains wire and make a sound if the current in the wire exceeds a certain limit (~20A).
It's going to be something fairly straightforward - a CT, rectifier and a comparator driving a buzzer.

What i'm not sure about is powering it. I can just add a couple wires from the mains and put an AC-DC block on it to power the whole thing, but i wonder if there is a more elegant way of doing it?
Specifically, can a rectified output from a current transformer be used, in practice, to power a comparator and a buzzer?
I suspect there could be problems:
-The load of the comparator skewing the reading in unpredictable ways
-Various spikes and transients burning stuff out
-The resistor needed to provide adequate voltage would overheat

So, the question is - how would you make such a detector, and is there a way to power it in a simpler way than an extra line and AC-DC?

[john_c]

That's an interesting question. It seems to me, that if you take a typical CT of 1000 turns, then given your 20A condition we have a 20 mA AC current source.

Although it seems like the needed power may be available when the 20A condition is met, one possibly simpler approach that occurs to me is to rectify the CT output to DC, and use that to charge a small battery.

As for your question about the detector, I think you're on the right track that the output is rectified and then you can just use a comparator.

[Seekonk]

We built a self powered device like that with a comparator built in.  It tool a lot of turns on the CT. The resistive load on the had to be constant so we had to switch between between a resistor and the load which in this case was a relay. Line powered it is much easier.  Self powered with just a relay is easier, but a pull in of about 20 would have a drop out of about 3.

[David Hess]

The switching power supply in Tektronix 7000 series and 485 oscilloscopes work this way; the switching regulator controller is entirely powered by the primary side current sense transformer through a bridge rectifier which produces DC into a capacitor and shunt regulator.  A series resistor between the bridge rectifier and filter capacitor allows the voltage at the output of the bridge rectifier to increase as the current increases so the current can be measured.

The trick is that the current transformer has a maximum volt*second product on the secondary so if the secondary voltage is too high, the transformer will saturate.  Any current transformer you find should have this specification.

[john_c]

The trick is that the current transformer has a maximum volt*second product on the secondary so if the secondary voltage is too high, the transformer will saturate.  Any current transformer you find should have this specification.

I'm trying to understand this part. There's a maximum of voltage, integrated over time? For what time period?

What if we were to use this CT; how does it apply? http://cdn.sparkfun.com/datasheets/Sensors/Current/ECS1030-L72-SPEC.pdf

[David Hess]

The trick is that the current transformer has a maximum volt*second product on the secondary so if the secondary voltage is too high, the transformer will saturate.  Any current transformer you find should have this specification.

I'm trying to understand this part. There's a maximum of voltage, integrated over time? For what time period?

Exactly, the maximum volt*time product that the transformer will support before saturation is a constant.  Transformers intended for power line use specify this differently; they include a maximum secondary voltage or load resistance and assume a 50 or 60 Hz AC only input.

The Tektronix design I mentioned used a roughly 9 volt shunt regulator on the secondary side of the current transformer which limited the secondary voltage preventing saturation.  But they also included a 15 volt shunt zener diode as protection to prevent the secondary from going open circuit which is very bad in a current transformer.  So their design supported more than 9 volts across the secondary but this was pretty easy for them to do; they built their own custom transformer and it was operating at 10's of kHz lowering the volt*time product compared to a 60 Hz transformer.  I used their design as an example of what can be done.

What if we were to use this CT; how does it apply? http://cdn.sparkfun.com/datasheets/Sensors/Current/ECS1030-L72-SPEC.pdf

In this case the numbers to look at are the specified secondary DC resistance of 250 ohms, the 10 ohm load resistance, and the 300mV output voltage.  Nothing more is needed in aa normal 60 Hz application.

Assuming it could be rectified, 300mV across 10 ohms only yields 30 milliamps and 9mW.  Of course we should know that because 60 amps divided by the turns ratio of 2000 is 30 milliamps.  The actual secondary voltage which we cannot access is 30 milliamps across 260 ohms or 7.8 volts.  Divided by the winding ratio of 2000, this burdens the primary side with 3.9 millivolts.

At some point as we raise the load resistance, the transformer will saturate and the output will drop but the specifications do not say anything about that; it will have to be measured to see if the transformer is suitable.

Note that there are lots of low voltage energy harvesting ASICs now which could directly operate off of a low voltage AC source to charge a capacitor or battery.  It should be possible to use these with a standard AC current transformer and it would probably be less expensive than procuring a better transformer.