about current transformers

[gabriel]

hello:
I need to measure the RMS of a big current, so my idea is to use a current transformer with a true RMS converter like AD736.
the current may be pulsed/switched one , so non-sinusoidal. My question is: does the current transformer gives a true image on the secondary in case of pulsed/non-sinusoidal currents??
anyone have experienced that already?

thanks in advance

[solutions]

Define "big current", nature of pulse/switched, how the measurement will be used, etc

Generally, no.

[pickit2]

you would have to model it, you may need to sample&hold over a set period. what your asking is like" if I monitor with a meter, I want the value,when meter leads, are not making contact.
you need to ask your self, "is it repeatable" & "what is the end result, I want to see"

I like your "big current" for some projects 100MA is way too much.

[gabriel]

my question is simply: can a current transformer be used for non-sinusoidal currents? (like in thyristors bridges). I mean also is the shape of the current preserved on the secondary side?
I know we can use hall effect devices, but they need supply, I was wondering if there is something simpler and cheaper.
the current is about 100A.

[Gallymimu]

my question is simply: can a current transformer be used for non-sinusoidal currents? (like in thyristors bridges). I mean also is the shape of the current preserved on the secondary side?
I know we can use hall effect devices, but they need supply, I was wondering if there is something simpler and cheaper.
the current is about 100A.

Yes it can work, It must be properly designed for the desired bandwidth and peak flux density.

Pearson, for instance is a manufacturing of precise measurement transformers like you are looking for.

Also, be careful using the AD736.  There are VERY significant crest factor limitations that you must consider if you want a decent RMS measurement of a pulsed signal.

[Finzi]

For this level of current, I think you should use a Hall Effect hall sensor.
 The first advantage is that you'll have a linear response over the limit frequency ==> Vo(t) = k*I(t).
 The second advantage is that you won't have additional losses due to the inclusion oa a series element. I can suggest you the LA55-P. Here is the data sheet
  http://igor.chudov.com/manuals/LEM_LA55-P-Current-Transducer.pdf
It can handles 50A RMS, 70A max, bandwidth 0 - 200kHz. There are hall effect sensors for 100A and 150A (rms) too.
To get the RMS value you can use the same AD736 mentioned in above posts.
Good luck.

[Gallymimu]

For this level of current, I think you should use a Hall Effect hall sensor.
 The first advantage is that you'll have a linear response over the limit frequency ==> Vo(t) = k*I(t).
 The second advantage is that you won't have additional losses due to the inclusion oa a series element. I can suggest you the LA55-P. Here is the data sheet
  http://igor.chudov.com/manuals/LEM_LA55-P-Current-Transducer.pdf
It can handles 50A RMS, 70A max, bandwidth 0 - 200kHz. There are hall effect sensors for 100A and 150A (rms) too.
To get the RMS value you can use the same AD736 mentioned in above posts.
Good luck.

Most current transformers don't have a series element.  What kind of current transformer has a series element that you are describing?

I'm not sure what device in these categories wouldn't be "hall effect" it's the fundamental principle.  Maybe you are thinking of something I am not aware of.

Your suggestion uses an active sensor architecture which is fine and may work great but is expensive compared to passive current transformers such as the following:

http://www.digikey.com/product-detail/en/CST206-1A/237-1096-ND/242539

One nice thing about the LEM sensor is functionality down to DC.

[Finzi]

Hi, Gallymimu. Well,when I considered the current transformer as "series connected element" I considered only topologies where at least one of the windings are electrically connected in series with the circuit. But, you are right. If you consider a topology where the electromagnetic field, generated by a wire conducting an AC current, is used to induce a voltage across the terminal of a electromagnetically coupled transformer (current transformer) we will have almost the same Hall effect. But, there is a small difference. Current transformers, as described before, cannot be used to monitor DC current components in a complex waveform. Gabriel has stated that his main concern was to reproduce a sample of a non-sinusoidal  current with no distortion. Well, DC and low frequency components will be lost using a current transformer.

Here is a link where we can find a hall effect sensor (-100A  to  100A). It is not that expensive.
http://www.newark.com/honeywell-s-c/csla1cf/current-sensor/dp/16M2414?in_merch=Popular%20Products

Anyway, I was mistaken considering only old topologies of current transformers. My apologies.
Cheers.

[Gallymimu]

Hi, Gallymimu. Well,when I considered the current transformer as "series connected element" I considered only topologies where at least one of the windings are electrically connected in series with the circuit. But, you are right. If you consider a topology where the electromagnetic field, generated by a wire conducting an AC current, is used to induce a voltage across the terminal of a electromagnetically coupled transformer (current transformer) we will have almost the same Hall effect. But, there is a small difference. Current transformers, as described before, cannot be used to monitor DC current components in a complex waveform. Gabriel has stated that his main concern was to reproduce a sample of a non-sinusoidal  current with no distortion. Well, DC and low frequency components will be lost using a current transformer.

Here is a link where we can find a hall effect sensor (-100A  to  100A). It is not that expensive.
http://www.newark.com/honeywell-s-c/csla1cf/current-sensor/dp/16M2414?in_merch=Popular%20Products

Anyway, I was mistaken considering only old topologies of current transformers. My apologies.
Cheers.

Oh, okay,

I wasn't really trying to beat you up.  I was more interested in seeing if there were things I wasn't understanding or was missing.  Thanks for the clarity.

You bring up a good point.  I am not sure if Gabriel made it clear whether there was a DC component to his measurement or not.  That has a pretty significant impact on the selection doesn't it!  Gabriel, you should share what the bandwidth requirements for your measurement are.

[alphaville]

Try this: http://www.allegromicro.com/en/Products/Current-Sensor-ICs/Fifty-To-Two-Hundred-Amp-Integrated-Conductor-Sensor-ICs.aspx

Cheaper, very good linearity and high bandwidth

Regards

[gabriel]

only AC currents, 50Hz pulsed ones.

[solutions]

You didn't answer my questions. I didn't make a list of the basic information needed for you to only fill in what you felt like filling in. You also have a very poor design practice of having picked the solution, then wrapping it in bandaids until maybe it works.

Use a resistor and a thermocouple....since you can't seem to be able to describe, or show, the waveforms, or state how the measurement will be used, everybody here will be guessing.... the R and thermocouple will work no matter what you have.

[Finzi]

Oh, okay,

I wasn't really trying to beat you up.  I was more interested in seeing if there were things I wasn't understanding or was missing.  Thanks for the clarity.

You bring up a good point.  I am not sure if Gabriel made it clear whether there was a DC component to his measurement or not.  That has a pretty significant impact on the selection doesn't it!  Gabriel, you should share what the bandwidth requirements for your measurement are.

No problem man.

Pulses of 50Hz are a bit strange. What is the base frequency?
 "solutions" has a point. Whe need more information to help you, Gabriel.

[enzine]

only AC currents, 50Hz pulsed ones.

Can you post a picture of  this current waveform?

[ahmetuyanik]

Dear Gabriel

As you know the current transformer can measure AC current. If the current contains DC part, the CT will be saturated or its output will be distorted. Thus, you cannot measure correct value. If the current is purely AC, you can measure it with the CT. However CT has a limited bandwith. Pulsed waveform has all harmonics and Current transformer will filter out some harmonics. If you would like to measure excat current waveform you must use hall effect sensor. Hall effect sensor need external supply and I do not like. Also, Hall effect sensor is more expensive than CTs. Thanks to Harting started to manufacture hall effect sensors. LEM and ABB's prices are expensive but Harting's price is reasonable. Please check following link : http://www.harting.com/fileadmin/harting/documents/lg/hartingconnectivitynetworks/news/newproducts/2012/stromsensoren/98_42_938_0201_harting_hall_effect_current_sensors_gb.pdf

Regards,

[solutions]

You guys are not reading the question completely and are off on all kinds of tangents about measuring current.

The ONLY way to measure true RMS, without digitizing and computing it, is with a resistor and a thermometer.

[Finzi]

You guys are not reading the question completely and are off on all kinds of tangents about measuring current.

The ONLY way to measure true RMS, without digitizing and computing it, is with a resistor and a thermometer.

Hi solution.
Well, thinking about the linearity of the measured signal, over a very wide frequency spectrum, you are completely right. But, there are a few problems.  The main one is the lack of isolation between the circuit where you are taking the measures from and your control/instrumentation system. Since, in most cases, we have a limited frequency spectrum we may use other methods that provide electrical isolation and good linearity. As a matter of fact this is mandatory for some applications like monitoring of High Voltage Transmission Lines and Power Transformers. But, anyway, if linearity is the main concern, and isolation is not a problem, your solution is the cheapest, reliable and more precise one.

By the way, where is Gabriel? I think he should be participating on this debate since he started it. Or else, i think there is no point in continuing this "brain storm". But, again, I might be wrong and some useful ideas may emerge from all of this.

Peace!!

[egstudio]

The best literature..the best (modern) ic's..

and the best AC RMS (analog !) millivoltmeter (I've got one years ago via eBay)...HP3400a

http://fp.io/5d9dcmfe/

Posted on: December 16, 2012, 03:44:23 03:44 - Automerged
..and this:

http://fp.io/e237884c/

LT1088..is  the best RMS (10 Mhz !!) RMS analog (thermal) converter in the world.
In this (legendary)  app. note, they explain the way to convert an 3400a to a modern RMS millivoltmeter.
But, unfortunately, it's phased out...Brokers that sell the LT1088 are mainly in China.

Modern (not thermal but RMS)  IC's ranging to 10 khz...2 mhz maximum

[kreutz]

Do you need to monitor the r.m.s. value of your current constantly?, or is it only during development/troubleshooting? If it is the last one you could use a h.f. current probe and oscilloscope, some models allow you to measure AC +DC.

[ahmetuyanik]

By the way, where is Gabriel? I think he should be participating on this debate since he started it. Or else, i think there is no point in continuing this "brain storm". But, again, I might be wrong and some useful ideas may emerge from all of this.

Peace!!

I agree with Finzi. However Gabriel question was  "My question is: does the current transformer gives a true image on the secondary in case of pulsed/non-sinusoidal currents??" and I could say Gabriel should be careful about saturation (if the current has DC components) and bandwith (should check CT bandwith and do not forget pulse or square wave signal has many harmonics) .

[gabriel]

It's clear to me now, either I can use a good CT like Pearson ones, suggested by Gallymimu (specially made for that) or a hall effect device (which need a power supply, not a big problem). It's all that simple. Again the current is AC pulsed 50Hz (no DC average).

thank you all!

[Gallymimu]

It's clear to me now, either I can use a good CT like Pearson ones, suggested by Gallymimu (specially made for that) or a hall effect device (which need a power supply, not a big problem). It's all that simple. Again the current is AC pulsed 50Hz (no DC average).

thank you all!

Glad we helped!  Good luck with the project.