$2 Carabiner AC amp sensor (aka current transducer, CT sensor, amp meter, split core clamp-on ammeter)

 Impressed by the usefulness of AC current sensors but put off by the prices? Me too! So I decided to make one.  Purchasing a new split-core current transducers costs around $25. Here is a $2 option.

Here are some creative uses of current sensors:
Real-time Web Based Household Power Usage Monitor
Tweet-a-watt


What you need:
-cheap carabiner (what other use for those fake caribiners!)
-or use a $1 c-clamp for better results
-3' of small gauge wire
-a store bought clamp-on current sensor

Once complete, read your sensor with a millivolt meter ; such as, a multi meter, micro controller, arduino, etc 

Step 1: Wrap the carabiner and calibrate

 Wrap your carabiner with enough turns to achieve the required resolutions. I initially used 30 turns with no resolution so I doubled to 60 turns as shown below.

Next, slice/separate the wires of an extension cord in order to test/calibrate your CT sensor.  If you try to sense both the load line and neutral line, the magnetic fields cancel and no current is induced.  

Next, plug something into the extension cord.  I used a space heater. Measure the amperage with a store bought clamp-on meter.  

Calibration
I measured 11.8 amps with my Field Piece meter.

The field piece has been calibrated so that 1amp AC is equal to 1milivolt AC (1AAC/1mVAC).  My Caribiner sensor measured 0.3 milivolt.  

11.8 amps / 0.3 millivolts = 39 AAC/mVAC

We're done! 

If you are considering putting this around any high voltage wiring, consider wrapping all exposed metal parts in electrical tape.

Comments

[joablen]

ok i'm starting to get angry with this thing... i haven't been able to get a reading... when I use my multimeter I get 0. using the arduino with this circuit (http://openenergymonitor.org/emon/sites/default/files/current.png) Burden = 27 ohm, C1 = 10 uf, Rvd = 10k ohm
the analog reading is always 510.

any ideas?

[Punkguyta]

 I really want to do this, but how would I attach it to my computer for real-time analysis of current use??

[hydronics]

 yea... its been a month... sorry i have no good excuse for you...

probably the easiest way to plug this into your computer is to use the above link: 
www.instructables.com/id/Real_time_Web_Based_Household_Power_Usage_Monitor/

and use the op-amp described (AD8820, AD627 or equivalent) but instead of plugging into the ioBridge you can plug it into an Arduino or something like this.. and use that to connect to the PC. hows that sound? fun... you go! :) Let me know.

[Punkguyta]

 Is there a simpler way? I know just like controlling a relay from your computer, it's easy to control 1 or 2 relays from the parallel port of your computer, but it gets infinately more complicated as you add more. I was wondering if the same goes for these AC clamps, would it be possible to hook even just one up without too major of circuitry?

I have lots of spare/junk components laying around but nothing elaborate like apics or arduinos or even any useful I/C's.

[E_MAN]

try wiring to the tip and sleeve of a mono 3.5mm male plug, then plug that into your microphone port. this should give you a nice input, just a matter of writing of finding some software to graph it.... i used to have some......

[chrwei]

Would magnet wire work better than the hookup wire you are using?

[hydronics]

I'm not familiar with magnetic wire.... I tried bare copper wire and it seemed it did not function as well as insulated wire.

[chrwei]

right, bare copper would short on the clamp and not work any better than just a plain-old clamp.

magnet wire is an insulted wire used in electric motors and electromagnets, so it's designed for inductance. I'm no EE, but it seems like this may improve the resolution of your readings, or make reading smaller currents easier. I could be wrong.

[paulstich]

Have you tried a ferrite choke?  Digikey has a bunch and Radio Shack sells a snap-together choke.  Or you could hack up an old monitor cable.  My results from a small steel clamp were less than stellar.

[hydronics]

hey thats interesting stuff thanks for the links.

the digikey looks like ~$11 for a 1" diameter ring.  I couldn't see the dimensions on the radio shack item but half the price. I'd really like to see how the performance matches up so I'll probably try one out eventually.... so you say there's ferrous in a monitor cable too? can you add to that, cause I didn't see any last time I checked.

You should get better results with iron than steel. whats the difference? from what I can remember, iron has more iron in it.  Impurities are added to steel to make it stronger (~ 1% carbon for instance).  Iron often appears rougher, darker, and rusts a lot easier.

regardless, I think you should get good results with the steel. what kind of AAC/mVAC are you getting?

[paulstich]

most pc monitor cables have a ferrite choke on one end (see this).  You would have to cut open the cable to get the ring.

1" is probably bigger than necessary for most applications.  I was thinking about 495-3856-ND which has a inner diameter just under 1cm, which should be plenty to go around most any power cable I would measure, even with the thickness of the windings.

Ferrite is very brittle, so it is probably not feasible to cut it to make something that snaps together.  However, for my application it is probably okay to have a solid ring (I would have to break the circuit to put the ring around one wire.

The Radio Scrap snap together choke would be a pain to wrap, because of its length.  But, it does snap together - and is easily available. (Hard to justify a $1 digikey order).

I saw no deflection on my analog voltmeter set to a 300 mA range. I may not have given it enough load, though - I think I was drawing less than an amp.

[uberdum05]

I think you can cut it with a hacksaw but you have to go very,very slowly and gently

[hydronics]

 Awww I see.  I was pretty happy with the resolution reading my space heater in the range of 5-12 amps plugged into 120 volt outlet.  

At ~1/2 amp you would only see 1 mVAC on your voltmeter using an arrangement similar to the C-clamp.  As you play around with materials you may also want to try more windings. Also, I tried once with uninsulated wires and it didn't work so don't go that route.

thanks for the lead on the monitor cables.... Sense I've been looking for ferrite it is kind of hilarious that it turned out to be so close.

[paulstich]

what are you using to measure mVAC?  Even when I upped the current to 10A, the voltage was still too small for any of my cheapo multimeters.  I exchanged my c-clamp for a steel snap clip (something like this) with about 5 feet of wire wrapped around it - enough so the windings were snug against each other.  I think I need to build an op-amp based full wave rectifier that averages (so I get a steady DC voltage) with a capacitor, and also tweak it to amplify, so my voltages come out between 0 and 2.5 VDC.

By the way, Radio Shack has some toroids too, along with a different snap-together choke that might be much better for this application, although at $8, it alone breaks the $5 price point.  www.radioshack.com/family/index.jsp?categoryId=2032273 at the bottom of the page.

[hydronics]

 I'm using a Fieldpiece.  It measure VAC in the ranges of 600VAC, 200VAC and 200mVAC (the mVAC range is made for a c-clamp plug in). It is necessary to have resolution in the mVAC range, say around 1-20 mVAC based on the type of clamp you use.  The above setup sounds like you have plenty of windings and steel to get resolution in this range.

[paulstich]

I got it working this weekend.  I used JasonT's schematic with a few modifications to amplify & rectify the signal to 0-2.5VDC.  I tested a bunch of core materials, including a c-clamp like yours, the stainless steel snap, the solid toroid ferrite, and the radio shack snap-together ferrite.  The solid toroid ferrite was vastly superior to all the others.  It had over 10x the signal of the snap-together ferrite, which was about 10% better than the c-clamp, which had about twice the signal of the snap.  I noticed that the results of the snap-together ferrite are greatly reduced if the two parts are not firmly snapped together, which isn't surprising.

Getting it working with the hobo was a little tricky. I had hoped to use the Hobo's 2.5VDC supply, but it is pulsed, so it only provides power during the sampling (plus about 20ms beforehand), which is once per second (or less frequently), so I ended up powering it with a little dc transformer I had kicking around.

Given that the solid toroid ferrite is so much better and cheaper than anything else, I think I'll go ahead with making a bunch of them.  I don't really need a snap-together, since I plan on measuring circuits at the breaker box, and it is easy enough to turn off the breaker and remove the wire to slide the toroid around it.  My goal is to have 8 of these running simultaneously (two 4-channel Hobo U12's).

[hydronics]

 nice job. thanks for adding this information about the HOBO....

[Nerdz]

You can improve inductance by making your wrappings tighter around the core. You also should be able to just wrap a coil of wire around the current you  want to measure and still be able to measure current. It would still work the same since a electron moving in a wire will produce a magnetic field, and since that magnetic field changes, it will induce a voltage in the coil of wire you wrapped around it, proportional to the amount of current flowing in that measured peice of wire.

iirc, electrical steel > ferrite > steel > Aluminum in terms of permeability

[Broom]

"You can improve inductance by making your wrappings tighter around the core."

 

Why would that be?

 

Inductance = (DiameterOfCore*NumberOfTurns^2)/(Length/Diameter OfCore  + 0.43)

 

Nothing in the equation mentions the diameter of the wrapping, nor the distance between the wraps and the core.

[struckbyanarrow]

if you wrap it tighter you get more number of turns

[Broom]

Um, no. We're talking about tighter to the core, not tighter as in more wraps per inch.

[struckbyanarrow]

oh, I'm sorry, my bad.

[Nerdz]

Because you lower the amount of Empty Space between the windings and the core. By doing so you Lower the Reluctance of the inductor.  Air is poor when it comes to concentrating magnetic fields (come to think of it, Air is poor for Anything except cooling purposes :P) .


Hmm Maybe Im using the wrong terminology here. But Basically, The closer your windings are to your core, The less empty space there is, the magnetic field lines can penetrate deeper and are more concentrated than if you wound a coil all willy-nilly. Since we are making a transformer (in essence anyway), There is mutual inductance (inductance shared by the 2 coils). It is given by the formula:

M = k*sqrt(L1*L2)

M is the Mutual Inductance
K is the Coupling Coefficient (<than 1)
and L1, L2 are the Inductors

Now For examples Sake, if L1,L2 = 5H, then our formula becomes

M = k*sqrt(25)
M = 5*k

We can  easily see that If k was some Really small number, it would effect our mutual inductance. A small k would represent A very Horribly wound core, while a large K near 1 Would be A very tight core. You can look at k Like efficiency, ie the transfer of magnetic fields from one core to another.

My point in saying this is His cores look Poorly wound, so He may get better results by winding the coil tighter around the core.

[Broom]

Thanks. My EE courses on this are pretty rusty (2+ decades), and haven't been used since. Forgot that k is dependent on the constituent materials (and the equation I grabbed above apparently assumes an air core, since k isn't even mentioned).

[hydronics]

 So do you know if you can get a good reading using one of these square ferrites 
upload.ecvv.com/upload/Info/200801/China_Product_C200835144238683422_ferrite_core_toroidal_core_transformer_bobbin_inductor_magnet.jpg
In other words, not have to 'clamp' or circle the metal core completely around the wire being sensed? 

[hydronics]

 yea, that sounds right.   'electrical steel'... looks like pretty interesting stuff... For currents over an amp, I think the steel would probably give you the resolution you needed.  Do you know were you can source cheap 'electrical steel' or 'ferrite'? 

[Nerdz]

What do you mean resolution? Electrical Steel and Ferrite Just have better properties more suited for inductors. They have a higher permeability which allows for more inductance per turn of wire.

I dont of any place that sells Ferrite or Electrical steel, you can probably disassemble a transformer and take out the laminations from there.

You could probably use a square ferrite core, but how would you get it around the wire that you want to measure current from?

[Broom]

Resolution = smallest incremental change that can be read from the meter.

However, I think he actually means "sensitivity", which is the amount of change in output per change in input. He has expressed this as 1/sensitivity in his comparisons, so that smaller numbers are better.

(The difference between resolution and sensitivity is that a tiny resolution may be readable from the meter, but "in the noise", or below the noise level, so that changes that small may not actually represent changes in the measured signal.)

[dstahlke]

The proper way to use a current transformer is to measure the current across the secondary, not the voltage.  When used in that way, the current across the secondary (your windings) is equal to the current across the primary (the power cord) divided by the number of windings, with no need for calibration as long as the losses are not significant (depends of the material the C clamp is made of).

Measuring the voltage on the other hand gives you something like the rate of change of current times inductance.  Even when calibrated for one load this will give wrong results with a different kind of load because the rate of change of current depends on frequency.  Frequency is normally 60Hz or 50Hz depending on what country you live in, but the tricky thing is that electronic devices generate significant harmonics at higher frequencies.  So, if you calibrate it using an incandescent light bulb it will not give proper results when measuring the power draw of a computer.

If you have an ammeter (i.e. the "amps" setting on a multimeter) then just use that.  If you can't measure current and need to measure voltage (eg. for a microcontroller input) then put a 1 ohm or smaller resistor across the coil and measure the voltage across that.  Use ohms law to convert the voltage to current, then multiply by the number of windings, and you end up with the current running through the power cable.  Keep in mind that the waveform is typically not sinusoidal so many multimeters will not give you the proper reading.

[hydronics]

 I'm trying to build a $5 power meter.
Real-time Web Based Household Power Usage Monitor
The idea is to interface the meter with an online usage analyzer like Google's Powermeter.  This device would serve organization or communities trying to understand reduce their energy use.  This could be 3rd world country application or your local neighborhood communities.  Let me know if you would like to collaborate. I'm really close... I've got half the PCB board designed and fabricated. -thomas

[hydronics]

www.instructables.com/id/5-Power-Meter/