Shunt vs Hall Effect

Korishan

Korishan

Administrator
Joined
Jan 7, 2017
Messages
7,542
I'm watching Julian Ilett's videos on the Hall Effect sensor ACS712. I was surprised that there was so much noise in that sensor. I know it's a fairly cheap sensor, but still. He did find a way to correct it, reduce the noise at the loss of loosing sampling times, but even at a 1000 times per millisecond, that's pretty good :p (i think that's what he figured, I might be off by a fraction)

But what I was really wondering, has anyone used Hall Effect sensors to measure their amps being used? The pros of using an HE sensor is that it's truly bi-directional. So you can use 1 sensor to determine charge into the batteries, and discharge from the batteries. I'm not sure how easy this would be able to be done with a shunt.

I'm assuming that you could read the voltage difference on either side (as normal) (assuming left-right arrangement, batteries left & load right) and when the voltage is higher on the left, amps are discharging, and if the voltage is higher on the right, amps are charging. Is this correct thinking?

If that is so, then I would gather that a shunt would be far superior over an HE sensor as the HE sensors have a very limited amperage threshold; whereas the shunt, could be made at home and use a known resistance to get the calculations (far cheaper)

Which is more accurate,shunt oran HE sensor? Again, I would assume that a shunt would be because it is less likely to be subject to external interference (ie. a magnetic object gets close to an unshielded HE sensor can through it's readings off easily, as Julian showed in his first video).



Part1:
Part2:
Part3:
 
Ok this is quite a topic and to my knowledge i will just bring up a couple of things: (I will leave out details to why and if so it can be googled if needed or if someone want to answer)

Shunt is more accurate in general. (This depends on how its setup but if you buy ebay stuff....)
Shunt is hooked up into the system so you need to take in account the voltage.
Shunt will also dissipate some of the current going through in heat....
Shunt can meassure both ways without problem. the value just goes negative....

HE does not affect the system at all since its just clamped around
HE does not care about the voltage and is isolated
HE doe pick in every little noise very easily if not done correctly
HE can measure very very very big currents without issues.

I think the Wikipedia do have some good info about it. I did not check it for accuracy though.
https://en.wikipedia.org/wiki/Current_sensing_techniques


Personally i would always use shunt on circuits i could attach it easily and dont break anything. If i just need to see something quick like current going into 1 phase and dont care about logging i add a HE instead. It takes 2 min to install and you dont have to care about voltages or such.

I shunt can easily measure from some mA to several 100A if correct hardware is used :)
 
daromer said:
HE does not care about the voltage and is isolated
HE can measure very very very big currents without issues.
Click to expand...

I'm kinda confused on these points. In Julian videos and some others, they say the voltage is limited. Is that because of the microcontroller on the board and not the actual HE chip itself, or is that because the device is designed to run on an arduino-like setup?

In another video (can't recall the author right off), he had to use voltage dividers to be able to measure higher voltages. In his example it happened to be 35V system. But then he had to actually put that into code (which could be a variable that is read before hand, I know), but the voltage dividers are hard set. So it's not an easy change. Higher the voltage, the more the dividers have to work, and the more inaccurate the reading (as more voltage/ampere being put into the same graph range; 0-1024, or 0-5V)

So in this case, it'd be better to go with the shunt since it is a whole lot easier to adapt to.

Correct? Incorrect? Misunderstood by myself? Just trying to do brain wrapping here :p
 
The HE is a coil so the output voltage depends on what goes through it. You need to pick the coil/transformer or what you use depending on the current going through.

Frankly i have not got that much knowledge about HE in practice since i mostly been working with shunts.
 
I'm working with two shunt (200A/75mV). One for Ibat (positive= charging, negative= descharging) and the other for Ipv (0 or positive). But I'm thinking to put other outside of inversor: STC-013, to calculate consumption AC.

I'll post my experience with it asap.



image_knidhi.jpg



image_etnvzh.jpg
 
There are three types of current sensors:
a) Shunt Resistor + opamp (of some sort) - Measure voltage drop across resistor
b) Current Transformer - 1 turn from cable, multiple turns for sensor output
c) Hall Effect Current Sensor - 1 turn from cable, silicon based magnetic sensor to sense current.

The Current Transformer and Hall effect Sensor both work on the basic principle that current flowing though a cable generates a magnet field, which is then picked up and amplified. As they are magnetic sensors that respond to low level signals, they will pick up noise. You can either filter the noise on the output, say get rid of AC signals if your only measuring DC, or shield the sensor from the noise in the first place.

Accuracy of all three methods of sensing current depend entirely on what you buy. I've got hall effect current sensors good for 0.1%, you'll struggle to find current shunt resistors that accurate.
 
A proper shunt should have very little resistance. Just enough to make a voltage differential. If the shunt is getting too hot, then it's too small for the job. The power loss should be in the milli-watts or less.
 
The option b) current transformer above only work with AC signals.
Shunts & hall effect types can measure either.
 
There's another major concern with HE sensors. They will stop functioning correctly after awhile. There is a youtuber that uses(d) the HE sensor for monitoring the solar input into his batteries. He had to replace them several times in a year. I would imagine the sensor gets burnt out somehow. At least, the cheap ones we can get on eBay anyways.
 
station240 said:
There are three types of current sensors:
a) Shunt Resistor + opamp (of some sort) - Measure voltage drop across resistor
b) Current Transformer - 1 turn from cable, multiple turns for sensor output
c) Hall Effect Current Sensor - 1 turn from cable, silicon based magnetic sensor to sense current.

The Current Transformer and Hall effect Sensor both work on the basic principle that current flowing though a cable generates a magnet field, which is then picked up and amplified. As they are magnetic sensors that respond to low level signals, they will pick up noise. You can either filter the noise on the output, say get rid of AC signals if your only measuring DC, or shield the sensor from the noise in the first place.

Accuracy of all three methods of sensing current depend entirely on what you buy. I've got hall effect current sensors good for 0.1%, you'll struggle to find current shunt resistors that accurate.
Click to expand...

When im allowed i would like to add a 4th type of current sensor similar to HE but more accurate (especially the zero offset) working with a different physical principle. Its a current transformer working with electronic as a compensation technique providing directly voltage per Amp, from DC up to 100 kHz. Of course isolated, accuracy close to the 0.1 % region possible. Different models with internal or external electronic.
There are no clip around models (by principle), you have to feed through.

Producer is VAC (Vacuumschmelze) in Germany. There may be others I dont know.
 
Cherry67 said:
When im allowed i would like to add a 4th type of current sensor similar to HE but more accurate (especially the zero offset) working with a different physical principle. Its a current transformer working with electronic as a compensation technique providing directly voltage per Amp, from DC up to 100 kHz. Of course isolated, accuracy close to the 0.1 % region possible. Different models with internal or external electronic.
There are no clip around models (by principle), you have to feed through.

Producer is VAC (Vacuumschmelze) in Germany. There may be others I dont know.
Click to expand...

Do you have a link to this device? Sounds interesting. Always up to learning more about this stuff :)
 
 
Ahhh, that's an interesting device. Thanks for added info. I'll be looking at more of these to get a better grasp on how they work and how to implement them.
 
In general, i can say the accuracy is better than HE.
Better linearity, better accuracy, lower zero Offset.

The technical prinziple is different.
While HE basically uses a transformer, and measures the flux in the core with an HE, the VAC current sensor have a compensation principle. The sensor does not measure the flux, but controls a compensation winding to compensate the flux in the core to zero. Then the current in the compensation winding equals the Measurement Signal.


image_rfepvp.jpg


(the picture is taken from public info from VAC)


Due to the accuracy, even higher currents can be measured with acceptable low current resolution, and without shunt losses.
 
The "losses" in a decent shunt are negligibly small, the resistance is often less than your battery cables & typically develp only eg 75mV across them at rated max current.
 
Redpacket said:
The "losses" in a decent shunt are negligibly small, the resistance is often less than your battery cables & typically develp only eg 75mV across them at rated max current.
Click to expand...

You are correct. yet, at peak power of some 100 Amps that still makes a few watts. Not too nagging in form of losses, but a heat to dissipate of several watts. The losses in cables, and switches, fuses, and connectors are present as well.
Its just another 75 mV, as you say. In an 48 Volts system still different to look at, as in a 24V.
 
Hi....i am a new user here. As per my knowledge the Current Transformer and Hall effect Sensor both work on the basic principle that current flowing though a cable generates a magnet field, which is then picked up and amplified. As they are magnetic sensors that respond to low level signals, they will pick up noise. You can either filter the noise on the output, say get rid of AC signals if your only measuring DC, or shield the sensor from the noise in the first place.

circuit board assembly
 
Yeah, I agree with the noise issue. The hall effect uses a tiny magnetic sensor to detect the current. The transformer on the other hand, converts the magnetic field generated by AC current into a voltage. Basically the AC wire is 1 turn on the transformer, and then the secondary sensor wire is wound several hundred times. Usually the transformer sensor has a max rating of amps it can pull from the primary wire (AC line) as the secondary is usually wound for a max of 50mA. (not looking at mine, so going off memory here). Then use a set of resistors to convert that to +/- 2.5V (for 5V max reference) or +/- 1.65V (for 3.3V max reference)

The transformer is a bit more work to get it working properly and accurately. The current shunt is a little easier and straight forward as the IC usually outputs a reference voltage based on it's input (5V or 3.3V).
I'm probably going to wire up a bunch of current transformers to do my monitoring as I can clamp them on as I expand. Current shunts have to be connected in series to function properly. Not exactly idea for house mains lines ;)
 
To get back to the first post about the ACS712:
I have the 5A version here and tinkered a bit with an arduino. I do have some noise in the signal, mainly from the USB-power of my computer. With an external supply, I could come down to +-12mA. when trying to measure mAh with a 4.2V cell but no load, the result stayed at 0mAh even after 30 minutes. I didn't bother displaying a more precise number because: Who cares?
The only problem is the voltage offset at startup. You need to "zero" the sensor before you can get good results.
 
And every time the unit has a change in power state (say, replacing the battery, or using a different usb wall-wort), this would need to be done. Altho, that's fairly easy to do. I have the code in one of my other projects that does just that. Fixes the internal reference so it gets an accurate reading of the voltages on the analog pins.

I've come to the conclusion that the ACS712 is just not good enough for my tastes. I want to have as little noise as possible and easy to work with. Depending on the project, I will be using some sort of transformer based sensor instead of a hall effect. For more accurate, and a lot more stable, imho
 
You must log in or register to reply here.
Back
Top