Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Torroid Core and Hall Effect device.

Status
Not open for further replies.

kinarfi

Well-Known Member
I was trying to use a torroid core with a slot cut in it and a hall effect device placed in the slot to detect the current through a wire. (Image 1 & 2) I used a A3144 (On-Off) and an Honeywell SS495A (ratiometric). With the A3144, with the current one way, it would turn on and stay on, I had to reverse the current to get it to turn off, with the SS495A, it started with one voltage reading, I would add current, and remove it and then have a different voltage reading, add and remove current again and the reading would change some more. It finally dawned on me that the torroid core was magnetizing. My next attempt at a solution was to use a ferrite core induction (#3) and things worked like I thought they should, the A3144 turned on and off at about the same current level each time and the output of the SS495A returned to it's static level after current was applied and remove. I haven't tried inductor #4 yet.
Jeff
 

Attachments

  • POST.png
    POST.png
    234.6 KB · Views: 290
What you are seeing is magnetic remanence.

If you want to use a toroid, you have to use a very low-perm material, which usually also posses very low remanence.

Micrometals #2 material could be a suitable option.
 
The problem I'm having is that most of the inductors I have are salvaged so I don't know what the core's material consists of. I didn't get a chance to try the air core yes, I hope to do that tomorrow. The one that is working well is just like the #3 with 6 turns of wire
Jeff
 
I have done a similar thing, a current clamp for automotive use, the hall sensor was analogue not digital the circuit I used ended up being capacitively coupled, however you could make the system work if you had an auto zero circuit coupled to a analogue hall device.
 
Made a mess of this post, I'll try again later
 
Last edited:
upload_2016-6-26_7-59-22.png

Here's the coils I tried and the results I got using the Honeywell SS495A, I believe there is some potential for this set up.
Jeff
 
Try the test again, if the 0 reading is caused by remanence then the reading will be completely different next time.
 
I was trying to use a torroid core with a slot cut in it and a hall effect device placed in the slot to detect the current through a wire. (Image 1 & 2) I used a A3144 (On-Off) and an Honeywell SS495A (ratiometric). With the A3144, with the current one way, it would turn on and stay on, I had to reverse the current to get it to turn off, with the SS495A, it started with one voltage reading, I would add current, and remove it and then have a different voltage reading, add and remove current again and the reading would change some more. It finally dawned on me that the torroid core was magnetizing. My next attempt at a solution was to use a ferrite core induction (#3) and things worked like I thought they should, the A3144 turned on and off at about the same current level each time and the output of the SS495A returned to it's static level after current was applied and remove. I haven't tried inductor #4 yet.
Jeff


Hi there,

I dont know where you got that drawing from but that's not the way to do it. It looks like a very basic way to construct a current measuring device, but only in the most basic sense. A practical design needs magnetic feedback.

In the more practical design, the op amp you see in that drawing is used with a driver to energize a SECOND winding on the same core. This second winding and current in it cancel the magnetic field effect from the current in the primary winding (current to be measured). Because of the turns ratio, when the current in the secondary is a certain level that exactly cancels the current in the primary, the secondary current is related by a constant to the current in the primary. So for an ideal example, if the primary has 1 turn and secondary 10 turns, if the secondary current is 0.1 ams then the primary current is 1 amp, when the "magnetic balance" is satisfied. This is again a take on the ancient dual pan weight balance, where when the weights match the unknown weight is related by a constant to the other weight and the constant is related to the placement of the fulcrum, and the other weight is the calibrated standard weight used to compare.

So to be practical, you need a second winding, driven by an appropriate circuit, and using the Hall Effect sensor to detect when the field has been brought up to the original intensity before any current was applied. To get a zero, you can use an adjustable offset circuit.

Note that once you do it that way the core will always be at the same level. If it does get magnetized however, you may have to use a degaussing circuit too or else just try adjusting the offset.

Even with the single winding though, as they say, "Kudos" for trying :)
 
Last edited:
Try the test again, if the 0 reading is caused by remanence then the reading will be completely different next time.

Will try my test again, as soon as I get caught up.

Hi there,

I dont know where you got that drawing from but that's not the way to do it. It looks like a very basic way to construct a current measuring device, but only in the most basic sense. A practical design needs magnetic feedback.

In the more practical design, the op amp you see in that drawing is used with a driver to energize a SECOND winding on the same core. This second winding and current in it cancel the magnetic field effect from the current in the primary winding (current to be measured). Because of the turns ratio, when the current in the secondary is a certain level that exactly cancels the current in the primary, the secondary current is related by a constant to the current in the primary. So for an ideal example, if the primary has 1 turn and secondary 10 turns, if the secondary current is 0.1 ams then the primary current is 1 amp, when the "magnetic balance" is satisfied. This is again a take on the ancient dual pan weight balance, where when the weights match the unknown weight is related by a constant to the other weight and the constant is related to the placement of the fulcrum, and the other weight is the calibrated standard weight used to compare.

So to be practical, you need a second winding, driven by an appropriate circuit, and using the Hall Effect sensor to detect when the field has been brought up to the original intensity before any current was applied. To get a zero, you can use an adjustable offset circuit.

Note that once you do it that way the core will always be at the same level. If it does get magnetized however, you may have to use a degaussing circuit too or else just try adjusting the offset.

Even with the single winding though, as they say, "Kudos" for trying :)

Got the drawings from googling "hall effect current sensor" , there were a few that showed feed backUntitled.png but most did not, your suggested method and it's comparison has got me thinking, using feed back is like balanced scales, and the lack of is like measuring with scales using springs. That probably why the measure gold with balance scales.
Jeff
 
In the pic you posted the hall effect measures the magnetic field and the op amp applies a current in the coil to counter the magnetic field produced by the red wire, the current from the op amp being used as the measured signal.
I have a couple made by Abb, however they are for Ac, I suspect Dc ones are arranged such that the offset produced by remanence is small compared to the lowest measured signal.
 
Will try my test again, as soon as I get caught up.



Got the drawings from googling "hall effect current sensor" , there were a few that showed feed backView attachment 100170 but most did not, your suggested method and it's comparison has got me thinking, using feed back is like balanced scales, and the lack of is like measuring with scales using springs. That probably why the measure gold with balance scales.
Jeff

Hi there,

Yes :)
The balance scale can be very accurate because the position of the lever is exactly the same before and after the measurement. The only thing that can prevent that is a little sticking friction, which can cause the scale to look balanced when it's very slightly off. A very thin shaft helps with that and with other things like lubricant it keeps the error down.
This is the way a lot of things in electronics works these days and in days past, including the Wheatstone bridge, any balanced bridge really, and of course the best example of all: the Op Amp itself.
I am sure there are articles on Hall Effect current sensors on the web which probably describe this stuff in detail.
 
I did it again
Untitled.png



Seems to remain fairly stable, not as stable or accurate as a LEM current transducer #HX 03...50.P/SP2, <1% Accurascy, Out put is 2.5 v +/- 12.5mv per amp.
Jeff
 
Hi,

It's not apparent what you are doing there.
Normally the hall device is inserted into a slot in the toroid.
 
Agreed!! I did that with a toroid by "Dremelling" a slot in a ring and discovered that on DC, the remanence was high enough to make an A3144 device act like a latching device, on and latched with forward current even when current was removed, off after current was reversed. So I'm just testing different methods of using the hall device and the effect of remanence.
I'll update this with a photo or 2 after I get going and take them.
Jeff
 
Hi,

Oh ok great, i look forward to seeing them.

Also, the core should be a low remanence type.
A closed loop system is shown in the picture.
 

Attachments

  • HallSensor_ClosedLoop-1.GIF
    HallSensor_ClosedLoop-1.GIF
    12.5 KB · Views: 172
Last edited:
Yes that was repeatable, wasnt expecting that.
You used to be able to get high Remanence or square ferrite, it was used for data memory a long time ago, you can get low Remanence factor ferrite especially for current transformers, something I've not played with, yet.
Would your application lend itself to a degaussing circuit like in old Tv's?
 
IMG_20160629_161508.jpg Here's my test bed, One Fluke 87 for current, one for voltage, switch for measuring coil 1 or 2. the output lead on coil 2 looks like it contacts the coil, but it doesn't. The toroid has an a3144 in it and if I send current through the toroid and turn on the LED, the LED stays on until I reverse the current, I have something similar that came off the control board for a spa that did the same thing, but it wasn't a toroid, just an open loop of metal in a plastic case to house the hall device, that's the one that got me started on this project, I originally blamed the supplier for sending me latching halls, but they didn't latch when triggered with a magnet, I had to know why! It was remanence.
The whole idea, for me, behind this is to shut down the power to a 12 volt power steering controller when the steering reaches the end of travel and the current goes locked rotor, fried many a FET by holding it turned too sharp. The first unit I built had a chain drive and I could mount vanes on the links to break the magnetic field and shut it down, Then I got a more sophisticated drive train and had no where to put the vane and I've been trying to figure out a current sensor for some time. I think this may be it. I can use an air core winding and adjust the sensitivity by how many turns on the coil, I think. No remanence with air cores. :)
Jeff
 
Yes that was repeatable, wasnt expecting that.
You used to be able to get high Remanence or square ferrite, it was used for data memory a long time ago, you can get low Remanence factor ferrite especially for current transformers, something I've not played with, yet.
Would your application lend itself to a degaussing circuit like in old Tv's?

Hi,

As you know, when you apply a current to a winding that goes through a magnetically active core, the field is the ampere turns times a constant, so we have:
B=N*Idc*K
where N is the number of turns and Idc is the DC current.
After we remove the current, we end up with an equivalent B which we can call simply B1:
B1=N*Idc1*K
where B1 is the remanence flux and Idc1 is an 'equivalent' DC current that can be thought of as still flowing.
In order to get that flux back down to zero, we would apply a negative current -Idc1, and we would think that would lead to zero:
0=N*(Idc1-Idc1)*K=N*0*K=0
but it actually doesnt work because there is time involved too, so we end up with a different level:
B2=N*(Idc1(t1)-Idc1(t2))*K
and this level may be slightly less than the previous level, although perhaps negative.

It turns out it's not that easy to determine the exact level or time, so what we do is just apply a sine wave that starts out at some higher level and gradually dies down to zero. So the current then would look something like:
I=k(t)*sin(w*t)
where k(t) could be something like e^(-a*t) which would cause the sine wave to tend to zero over time.

So the idea then is to run a current in a winding that is sinusoidal, and gradually decrease the value. That's the basic way to degauss a magnetic core. It may in fact work with square waves too but i've never tried that, nor with triangle waves.

But hey, we already have a device that can sense static fields inside the core, so why not use that? What we could do is apply a reverse current and see if we could get the field down to zero with that. If it went past zero, we could always reverse again. The core would be degaussed when the sensor read zero again.
However, when in the feedback loop, this should happen automatically because the circuit will not rest until the flux is again zero. So probably all we need is a 'rest' period between measurements. Once it is zeroed, it should stay that way because it will be forced back down to zero again anyway. There might be some stabilization mechanism required too though, so this wuold have to be tried out.
 
View attachment 100202 Here's my test bed, One Fluke 87 for current, one for voltage, switch for measuring coil 1 or 2. the output lead on coil 2 looks like it contacts the coil, but it doesn't. The toroid has an a3144 in it and if I send current through the toroid and turn on the LED, the LED stays on until I reverse the current, I have something similar that came off the control board for a spa that did the same thing, but it wasn't a toroid, just an open loop of metal in a plastic case to house the hall device, that's the one that got me started on this project, I originally blamed the supplier for sending me latching halls, but they didn't latch when triggered with a magnet, I had to know why! It was remanence.
The whole idea, for me, behind this is to shut down the power to a 12 volt power steering controller when the steering reaches the end of travel and the current goes locked rotor, fried many a FET by holding it turned too sharp. The first unit I built had a chain drive and I could mount vanes on the links to break the magnetic field and shut it down, Then I got a more sophisticated drive train and had no where to put the vane and I've been trying to figure out a current sensor for some time. I think this may be it. I can use an air core winding and adjust the sensitivity by how many turns on the coil, I think. No remanence with air cores. :)
Jeff

Hi,

Oh yes, very nice. The air core has the disadvantage of not being able to 'amplify' the magnetic signal. Back when i experimented with this i found that almost any metal like even steel would increase the field and thus give a higher reading. Even a black binder clip would work to some degree because the spring part is made of steel.
The best cores are the ones made for magnetic applications, but yes there are other things that have to be considered. With an air core you'd just need more turns i guess as that makes the field stronger:
B=N*Idc*Kair

The main basic advantage a core has to offer is a higher field intensity per winding turn. Since a single winding turn has a certain resistance, then N winding turns has resistance that is N times one winding turn, and so the fewer winding turns we can get away with the lower the resistance of the entire winding.
Since the multiplication factor for a core can be easily 100, using even a low mu core leads to 100 times the field intensity and that means 100 times the sensitivity. So for a core that needs 10 winding turns to read 1v on the meter with a certain current, that same size coil with an air core would require 1000 turns. Pretty big difference right? Even a common 10 penny steel nail might have a mu of 10 to 100.
That's only the tip of the iceberg too, because many cores have mu factor of 1000 not just 100, and so that always leads to less turns.

I'd be interested in hearing more about this too as you go.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top