Measuring high constant current

[stoyanoff]

I need chart for measuring high constant current(3A and above). I need the high voltage circuit to stay separated from the low voltage circuit.I`ll use a microcontroller for measuring. I want to ensure the controller wouldn`t be estroyed from a current pick.
Any ideas???

 

[walkura]

Personaly i would use an ACS from allegro for this.
It gives a outputsignal You can connect straight to the micro and it has no electrical contact between the measured signal and the micro.
(Its based on hall sensors)
I used this setup from a few Ampere's upto 150 Ampere's, depending on the model used.

 

[stoyanoff]

Can you give me some more info about this ACS?? I know teoreticly what is Hall`s sensor. I think I can buy some, but I think I need and ferit hart.

 

[walkura]

I have a link for You so that You can have a look at how it works and where You can find datasheets.
**broken link removed**
You don't need any ferrite's its all intergrated and it gives a output voltage depending on the currentflow through the power sensing line.
You power it out of the same 5 Volt going to your processor and on the output pin You will get 2,5 Volt +- your sensed current (depends on the type You use how much millivolts You get for every Ampere)
It basicly senses how much current flows through the conductor by the amount of magnetic field it creates, so there is no galvanic connection between the sensed line and the output.

 

[goldsmith]

Hello my friend
You have many ways . and the difference between them is : 1- price 2- precision
So what do you need clearly ? what is your maximum current . and is precision important for you?
Best Wishes
Goldsmith

 

[stoyanoff]

The precision is not so important. I think the maximum current should be around 15A. I need to measure the current just for orientation. I just want to know what`s happening with the load and if the current goes above determined value to stop the device. I saw this current sensors, but there is a problem. In my country there isn`t supplier. In their site are written 2 addresses but there are in Germany. I looked in the shops around me and I can buy Holl sensors like UGN3119U or UGN3177U. I can also buy a magnetic ferrite. Is it a good idea to construct my own sensor and are there any rules which I should follow???
My idea is to make a hole into the ferrite and to put there the sensor. Next to connect a operational amplifier and to read the voltage with a microcontroller. Is this posible??

 

[walkura]

You need a lineair hallsensor (so that the output voltage correlates with the amount of magnetic field the current generates in a wire) if You want to measure current.
The ones You showed are to measure for instance the RPM of a engine or anything rotating like for instance a speedmeter for a bicycle.
I did see clamp meters to make at home but i don't have a clue where those schema's are or who and when published them.
What country are You from?
I still have a few ACS712-30A and maybe i can send You a few.
Otherwise You will have to find a lineair hallsensor type, i did use them in the past for measuring current (but only for currents exceeding the standard ACS sensors upto 200 Ampere)
About those ferrites You would have to cut it on one side to fit the hall sensor in the gap (probably there are examples on the allegro site in the white papers concerning lineair hall sensors)

 

[goldsmith]

Again Hi
For current measuring , we have many methods. i want to talk about some of them. 1- with a very low value series resistor ( series with main line ) . 2- CT ( current transformer ) . 3- hall effect sensor . 4- Clamp method .
I think , according to the specifications of your project , clamp method is ok.
Good luck
Goldsmith

 

[stoyanoff]

But I don`t need the measure the current manually. I must use a microcontroller which to dispaly the data in LCD screen. So I think I need a sensor, not a clamp

 

[walkura]

@Goldsmith.
Ofcourse You could use a simple sensing resistor, but to provide the isolation barrier to the micro You would have to design an optical barrier with for instance an IL300 (plus the extra powersupply to run it)
Current transformers could be used for AC but for DC they are of no use unless You go invent some nice trick to prevent it from saturating.
Measurement clamp would be possible but You would still need a lineair hall sensor.
I just proposed what was in my eyes the most simple way to measure current and provide the isolation barrier.

 

[goldsmith]

Dear walkura
Hi
My mean by CT , is not on DC line . each DC line provided from AC line . you can measure AC line current before rectifying .
Best Regards
Goldsmith

 

[walkura]

Good evening Goldsmith.
That way would make the use of a current transformer possible.
In a way if the math isn't to hard (assuming You would calculate and make it yourself) it would even be a prefered option when You consider precision.
(Hall sensor's like the ACS have a tolerance of a couple of mV which can come down to 250 mA or even more as measurement uncertainty)

 

[goldsmith]

Dear walkura
Good night
Yes . i'm completely agree with you
Respectfully
Goldsmith

 

[stoyanoff]

Hi, again! I found this Allegro Hall sensors ACS713. Where should the input pins be connected??? Directly to the power circuit or to a shunt resistor???
Thanks!

 

[emontllo]

ACS713 directly to the power circuit. No current sense resistor is needed. Just check that the isolation voltage matches your requirements and also current range.

 

[stoyanoff]

I want to use ACS713ELCTR-30A-T. I think this means 30A max through the sensor. I`m not sure this SIOC 8 package can hold such current. If I have undestood you right you mean I should connect it like the chart in the attached file?!
In this case I could have some problems. For example my pcb leads must be quite big to hold this big current?! My first idea was to use a cabels which can easily hold big currents.....

 

[emontllo]

The connection is correct, provide a decoupling capacitor to Vcc too.
The package can hold the current if appropriate connections are provided. Ensure a perfect solder between leads and pad. Avoid thin tracks and thermal reliefs to the pad.
However, if you prefer you can also use other packages from allegro such as ACS756KCA with higher current ratings.
There are other similar options (Hall effect current sensor) such as LEM USA LTS 25-NP.

 

[stoyanoff]

I`m starting to think about creating a current devider with 2 etalone resistors in the power circuit. If I use 0.1 ohm and 0.01 ohm I`ll decrease the current through the sensor 10 times.

 

[emontllo]

If you assume a loss of accuracy this is a good option. Note that the solder joints have a certain resistance which is difficult to evaluate and will add error (arround some mohm or less).
However, the circuit power dissipation will increase and efficiency decrease (at 30 A P = 30^2 * 0.01 = 9 W). One of the advantages of hall effect current sensors is that you avoid using resistors.

 

[mtwieg]

Depending on how much isolation you need, a high impedance differential amplifier may suit you needs. Then you can just use some precision resistors (assuming using a shunt resistor for the current is acceptable).

For hall effect sensors, you should also look at LEM products, though they may not be available to you either.

Building a custom hall effect sensor is pretty difficult... it's incredibly difficult to machine ferrite, as it's very brittle, so unless you can find ferrite pieces already in a suitable shape, it's probably not feasible. It would likely be more feasible to use a shunt based current monitor, even if it involves more circuitry.