Build my own CT Sensor

Hi.

I'm think of building my own CT sensor for testing.

I've seen most transformer have the winding around a metal piece that it self goes also around the conductor or the windings go around a toroidal metal and the conductor passes inside the toroid, that in practice seems to be the same.

I would like to know if this model will also be viable as it is more easy to make:
The conductor to be measured passing through a coil like this (this image is just for example, I would roll thin insulated wire on a plastic spindle to obtain something similar):

PS:  I only want to do this because I don't need much accuracy on this project and have scrap parts to do it.

Cheers

calypso_rae's picture

Re: Build my own CT Sensor

I tried this idea some years ago without any success.  Thankfully, these YHDC ones work fine:

http://mk2pvrouter.co.uk/media/f38a139f4c26bb68ffff8175d4355564.JPG

http://mk2pvrouter.co.uk/media/f38a139f4c26bb68ffff8174d4355564.JPG

Because they each have the same number of coils, they have very similar performance.

CidiRome's picture

Re: Build my own CT Sensor

Hi.

That's good to hear that someone has tried it...

Do you have an idea why you didn't have any success?

Is it possible that the physics are not the same when applied without the metal core to transfer the magnetic energy from the primary (live wire) to whe secondary (coil)?

I know about those CTs you posted, just trying to cut on the costs for a testing project where I will need two units.

Cheers.

emjay's picture

Re: Build my own CT Sensor

No, that can't work.  It is not the lack of 'iron' for coupling (though this certainly makes the coupling much more effective) but that the geometry is wrong.  To continue your toroid example, the metal (or in this case air) ring is passing through the hole, so the other winding needs to be a coil in the same plane as the existing multi-turn coil.

By passing along the main coil axis, your degenerate 'one turn' coil is exactly at right angles - no flux change pickup, no energy transfer.

You can successfully use air coupling - it is used when eliminating saturation/frequency limitations are important. Gurgle Rogowski coil if you are interested.  You end up spending money on a sensitive, instrument grade op amp though, so this is not necessarily a cheaper path than just buying a conventional C.T.

 

 

calypso_rae's picture

Re: Build my own CT Sensor

Do you have an idea why you didn't have any success?

Yes, it's because none of the flux lines that are caused by the primary current are going through the turns of the secondary coil.  In the commercial types, the flux is concentrated in a continuous metallic ring that goes around the primary; the secondary is then coiled around this ring.

Without the metallic core, I think this design of CT should still work but very much more weakly.  Opening the clip of a split-core CT by just a tiny amount will wreck the linkage between the primary and secondary circuits.

Gurgle Rogowski coil if you are interested

I tried winding one of these too, also without any success to report.  The signals are just too small.

CidiRome's picture

Re: Build my own CT Sensor

Thank you both for the information.

I will think some more if I can figure a way to cheaply do it with the scraps I have...

Cheers.

dBC's picture

Re: Build my own CT Sensor

Many energy ICs support both CTs and Rogowski coils.  And yes, you need to wind the gain right up on the input, and also enable the digital integrator because the output is proportional to dI/dt, not I.

Robert Wall's picture

Re: Build my own CT Sensor

If you want to make your own ring-core CT, it will be best to follow exactly the same procedure that they use in the factory:

Take a long strip of electrical steel, paint or varnish one side and wind it tightly onto a former. This will be the core of your transformer. The former needs to be a little larger than the centre hole that you need to take the primary winding. Take the coiled strip off the former and bind it with tape to stop it unwinding and to protect the edges. Now wind an even layer of several hundred turns (depending on the ratio that you want) of reasonably fine enamelled copper wire evenly around this doughnut, taking the wire through the centre hole each time. This will be your secondary winding. Finish off by soldering a pair of lead-out wires to the ends of the winding, secure those and protect the whole winding with another layer or two of tape.

And good luck.

Alternatively, you can of course use any mains transformer "backwards", using what was intended as the high current secondary winding as the (multi-turn) primary winding, and the original mains winding becomes your secondary. So a 3 V, 12 VA transformer should give you about 4 A : 50 mA, plenty good enough for an emonTx or Arduino. If you're connecting it into a mains voltage circuit, just make sure that the insulation on the original "secondary" winding is good enough. If you want a higher primary current, you might need to take the "secondary" winding off (it should be on the outside) and rewind it with fewer turns of thicker wire.

CidiRome's picture

Re: Build my own CT Sensor

Hi Robert.

Something along your alternative idea was already on my mind, lets see what I can do...

Cheers.

CidiRome's picture

Re: Build my own CT Sensor

Hi.

I started searching on ebay and found this ones that seem more than enough for my purpose:
http://www.ebay.com/itm/291550421556

And are also cheap.

Cheers

calypso_rae's picture

Re: Build my own CT Sensor

I've used that same model from the same source.  For measuring the flow of real power, the performance of this type of CT is similar to the other ones.  It's good to see the turns ratio shown on the outside.

The central hole for this type is larger than for the YHDC TA17L-04.  This can be helpful when more cables are to be fitted through.  In this system with two loads, one of which has a manual override, diverted power can flow along any of three routes, hence my CT2 needs to go around all three cables.

(This photo must have been taken before the socket & wiring for CT1 was fitted.)

CidiRome's picture

Re: Build my own CT Sensor

And the TA17L-04 seems to be much more expensive.

Cheers.

calypso_rae's picture

Re: Build my own CT Sensor

I thought you were more concerned about accuracy than cost? 

None of these bobbin CTs are more than a fiver.

CidiRome's picture

Re: Build my own CT Sensor

Hi.

In this project accuracy is not very important, unlike the other one where I'm changing my EmonPi firmware.

Sorry, I'm not familiar* with the expression "a fiver", what does it mean?

Cheers.

* English is not my mother language.

calypso_rae's picture

Re: Build my own CT Sensor

A "fiver" is 5 GBP (£)

Bill Thomson's picture

Re: Build my own CT Sensor

Hi CiDi,

A "fiver" is 5 GBP (£)

Or a 5 USD bill. ($)

It could refer to any currency with a denomination of five, but the term is probably used mostly in the UK and USA.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.