Non-contact current measurement - HALL EFFECT SENSOR

Hello, i am doing a project for non-contact current measurement using a Hall Effect Sensor built into an air gaped toroid core such as the Honeywell CSLA1CD( +- 57A) - https://sensing.honeywell.com/index.php?ci_id=49794 ; this is the data sheet. My understanding is that the hall effect sensor is a transducer that outputs mV in relation to the flux density in a linear fashion. An idea i have come up with is too have a current carrying conductor pass through the toroid core, inducing a magnetic field onto the core, which will then concentrate the flux density onto the Hall Effect Sensor, which in return will output a mV value depended on the current passing through. Now i assume that this mV output is going to have to be amplified ( i was thinking an LM358DT amplifier ) before going into the ADC converter of a micro-controller ( Arduino UNO ) but my issue lies in the fact that i am not too sure on what the gains of my amplifier should be, if or how to range the output/arduino so that the displayed reading is accurate, or how to convert the signal going into the micro-controller from mV to mA to be read off of a LCD display. Any insight on my problem would be greatly appreciated, and if you see anything that you see that you think might not work please let me know so we can further discuss the issue. Thank you.

Hi,

Maybe, keep things simple: e.g. 10mV in reads as 10 or 0010 or 10.00, etc., out on the LCD display. Not up on Arduino workings, sorry, anyway. If an ADC has an input range of 200mV or 2V, for example, that would display at most 1999 on a 3 and a 1/2 digit display, so if current range/sensor voltage were 0 to 50mV, you'd need a gain of 4 to make the most of the ADC/display, so long as you take into account the min and max current to be sensed. If the sensor has a min output/offset of midsupply, then you need to bias the op amp to the same level so its output looks like 0A at 0A current. 0mV to 250mV for a 200mV ADC, need to put a divider on the input, that divides the sensor voltage down to 0.8, e.g. 2K4 and 9K6, to fit the ADC input range, then you wouldn't need an op amp to amplify, unless you wanted to add one as a unity gain buffer.

As you know the numbers/supply and device and sensed voltages and currents involved, it's not so hard fitting them together, and reducing them to ratios first to see how they need to fit together for 10mV to look like 1.000A on the LCD. Hope that helps a bit.

Hello, thanks for the quick reply. For this particular sensor the output is 49.6 mV/A, that is if 1 ampere of current is being carried through the conductor going through the toroid core, 49.6 mV will the output from the sensor, some confusion lies in the fact that if no current is being passed, the output of the sensor will be 1/2vcc which vcc being 5v would be 2.5v. For this project i want to be able to pass a known current say .5A through the sensor and be able, from the magnetic flux induced into the core, and the mV output of the hall effect transducer, have a LCD display this current. Not worrying too much about temperature compensation or residual magnetism drifts but would like it to be reasonable accurate, i am not going to be pushing a lot of juice through this circuit as it is only for a thesis project.

There are a whole range of commercial Hall effect/gapped toroid current sensors available, but they work in a rather different way.

They place a Hall sensor into a gap, just as you are doing, and the differential output from the Hall sensor goes directly into the inverting and non inverting inputs of an op amp. The output of the op amp drives current through a multi turn coil wound on the ferrite toroid.

The idea is that the op amp drives current in such a direction through the coil as to always null out the flux in the toroid. This is an extremely linear process, because the full gain of the op amp seeks the flux null. Doing it your way the flux in the core must change, and due to non linearities in the BH curve of the core, your millivolt output will be highly non linear as well as temperature dependant.

This is a typical example of this type of sensor:
https://www.alldatasheet.com/datasheet-pdf/pdf/115009/LEM/LA55P.html

This one has a 1,000 turn coil, so coil current is 1mA for each primary amp, up to 50 amps.

The return current path through the coil is through a resistor of your own choice for scaling the output voltage. This is a much better and more accurate solution. For a 50 Amp range sensor, the zero error might typically be only a few mA.

Hey Tony, i see where you are coming from, due to the non-linear nature of the saturation curve you believe that the accuracy will be greatly affected? are there any ways around this with the core without the internal windings nulling the flux? i looked into the sensor you posted the link too and see a few problems, the main being the time to receive the sensor ~22 weeks i believe it said. Also i would like the sensor to be able to work off of AC or DC. There is really not much information regarding this sensor for this application available online; i am not extremely knowledgeable in this aspect as i am only a student.

Many cheap DC current sensors on the market are not using the compensation method, either sensors without a core like Allegro AC7xx or "direct measuring" sensors like LEM HTFS. They still achieve linearity down to 0.5 percent.

Yea, i checked out all of these options already but the "direct measurement" does not meet my given criteria as this is for my thesis project, and i was assigned a "non-contact", non intrusive means to measure current with a hall effect sensor.. i don't necessarily have to use a ferrite toroid but that is what was recommended by my instructor.

**broken link removed**

It really depends on what you plan to use this for. If you want pretty good stable measurement accuracy for dc or ac then the servo type sensor with the feedback coil is definitely the best way to go. They are widely available, even on e-bay, and ex stock from just about all of the major component suppliers.

The direct reading Hall sensors certainly have an application, especially for current limit monitoring where speed of response and low cost take priority over absolute accuracy.

but the "direct measurement" does not meet my given criteria as this is for my thesis project

Click to expand...

Suppose you didn't yet understand what a "direct measuring" or "direct reading" hall sensor in contrast to a compensating sensor is, in any case you didn't yet specify your requirements.

Thanks for the responses guys, sorry for the lack of understanding and not being specific. My goal is, to be able to measure current off of a hall effect sensor.. i been looking the LM55p servo type sensors with the feedback coil around the core. so your saying this https://www.digikey.ca/products/en/sensors-transducers/current-transducers/525?k=la55p would perform better then my original idea to use this https://www.digikey.ca/products/en?keywords=480-4818-ND&from=home sensor? again i apologize for my lack of experience but my goal is too be able to place a wire through the core, hall effect sensor pick up the flux density and output a mV output.. from there i figured the signal would have to get amplified be more readable for a microcontroller, from the controller code convert that given mV signal into a mA signal.. to be read off of a LCD. That was my vision, but as i look into it deeper i see complications which is why i came to this fourm to ask you guys what your opinions were figuring you guys would have a greater understanding.

This is a typical example of this type of sensor:
https://www.alldatasheet.com/datasheet-pdf/pdf/115009/LEM/LA55P.html

So, Tony, i ended up going with the LEM-55P transducer, thanks for your input regarding the difference of open and closed loop. I plan on having a 12VDC power source going into the sensor, then a LM358 amplifier (with a 10X gain?) on the output, with that going into an arduino for signal processing to an LCD to display, do you see any issues with this circuit?

[/QUOTE]

The LEM sensos, as with all these types of servo sensors are specified with split supply rails so that the full ac current waveform can be reproduced, usually +/-12v or +/-15v supplies.

If you are only interested in dc measurement in one direction, you will still require split supplies, but the negative supply requirement can be greatly reduced in both voltage and current requirement.
So if there is only a +12v dc supply available, a negative supply will need to generated somehow.

ok, i understand what you mean.. where vcc is actually bipolar the +12v by itself wont work, what would the be the best, and in my case cheapest, way to achieve this? some sort of a voltage divider? or a split rail?

a cheap rail splitter i came across is the https://www.digikey.ca/products/en?keywords=NMK2412SC Data sheet -- https://power.murata.com/data/power/ncl/kdc_nmk.pdf , this takes 24VDC source and turns it into +12v and -12v, just wondering though, it has 2 seperate outputs, one for the +12 and one for the -12.. would both of these outputs be tied together going into the VCC of the LEM 55-p?

sorry about that last post, i realized that there is actually a positive pin, and negative pin on the LEM sensor, what about the 0v pin.. is that ground? or the output and there is no ground pin? can i use this sensor interfaced with an arduino ? any insight would be appreciated, thanks.

I suggest you don't use a rail splitter if you only have +12 volts, but a voltage pump circuit to generate a not too critical negative supply.
A rail splitter will leave your output with reference to half the dc rail, which is far less convenient than a ground referenced output.

Check out the Intersil ICL7662. This is a very venerable chip.
That chip although still more or less readily available is now officially listed as obsolete by Intersil.
But now has has several different higher performing replacement alternatives offered by other manufacturers.
http://www.allxref.com/search.htm?part=ICL7662

This will generate about roughly -11v from a +12v input without requiring any inductors. Its very simple and easy to use.

The servo part of the Hall sensor (with a 1,000 turn coil) requires up to +70mA with a +70 amp input, but the negative supply only needs to supply the negative side of the op amp used in the servo loop. That might only me 1mA or even less, but am not absolutely sure of the actual current.

But I can tell you for sure, the negative supply is not voltage critical or heavily loaded, so a fairly primitive non voltage regulated capacitor voltage pump circuit will work fine. I have done this many times myself.

i am going to be using a 24vDC source into NMK2412SC to get the +-12v required for the LEM sensor according to the spec sheet, im just not 100% sure of the 0v pin can you elaborate on that please?

Hi,

The "0V" pin appears to be the output pin, and it says in "Remarks" "IS is positive when IP flows in the direction of the arrow".

(Would that mean it is negative if current flows in the opposite direction?)

I looked at this datasheet: LA 55-P/SP1