Allegro ACS758LCB-050B-P not working as expected

Automation and Control Sensors Transducers
, , ,
1

Hi,
We are designing automation of our electrical panels and to read current we have gotten couple of Allegro ACS758 hall effect current sensors. According to the datasheet and search on the web alongwith user reviews this device produces 2.5 V DC output at NO LOAD and increments the voltage by 40 mV / 1 A. We are using this to measure AC current and we have tried couple of the devices and gotten same results. We passing 6 A of current through the IP+ and IP- terminals but we our DC voltage out is still 2.50 V and not even bulging even by a mV. Even though it is AC if IP+ and IP- terminals are reversed we should see lower voltage but it is not. We even tried changing the polarity of these terminals but no avail. We are using the capacitors as suggested in the datasheet. We have seen some reviews online and users are seeing the change in DC out based on current being pulled but we are not. I could imagine if one of them could be bad but we have tried 3 of them and all are giving same results.

Please help if anyone knows what is going on.

Thanks.
Regards,
Jatin.

2

Hi Jatin,

I will take a look and see if I can find a possible solution. However, you should note that in June of 2017, Allegro changed the status of this part from “Active” to “Not for New Designs”. This generally means that they plan to obsolete the part some time in the near future, and/or that they have created an improved replacement. They recommend the ACS770LCB-050B-PFF-T as the next generation part.

3

Hi Jatin,

What is the voltage and frequency of your AC input signal? What type of device are you using to measure the output voltage (pin 3) from the ACS758? Have you tried passing a DC current through to see how it reacts to that?

4

Thanks for looking into this. We are measuring current through 110V 60 Hz circuit. We are using a calibrated DMM to measure the current passing through the wire and an output dc voltage. We haven’t tried passing DC current since it is not our application and we do not have something that would pull such dc current.

5

Are you using the DMM to measure the output voltage or an oscilloscope? The output should be cycling at a 60Hz rate, based on your input.

6

Hi David,
We are using DMM. According to the data sheet we should be getting proportional dc voltage out for AC or DC current input. Isn’t that correct?

7

This is a fast-reacting current sensor. It is giving you the nearly instantaneous current reading throughout the sine wave. If you are feeding it a 6A RMS current, the actual signal is a +/-8.5A peak-to-peak waveform. The output should then be a voltage that varies from 2.84V to 2.16V with a 60Hz frequency.

8

We did measure the output voltage in AC mode and at no load it was reading zero and 0.3 V at 6A. It measures RMS value.

9

Based on that, it sounds like it is working properly. Most DMMs aren’t real good at the mV level in AC readings. You will probably need an oscilloscope to properly evaluate that signal. The intent of this part is to feed into an analog-to-digital converter (ADC) for waveform analysis (motor control and such). It is not typically used for RMS measurement. To do that, you would have to feed into an ADC and use a processor to calculate RMS from the digital data coming in.

10

Hi David, out primary purpose is to feed this output voltage into ADC but that requires a DC input. In this case, AC current produces an AC signal which can’t be read by ADC and interpreted to read the actual current. If there is a way you think it is possible then please advice. Again, thanks for all the prompt feedback.

11

When you say “actual current” are you referring to RMS current or instantaneous current? The output will always be a signal that varies, plus or minus, around 2.5V (0.5V to 4.5V for 50A peak-to-peak signal), with a 120Khz resolution reflecting the “actual” instantaneous current. This is easily read by an ADC (as long as its range and bandwidth is sufficient).

This device will not output an RMS value directly. You must either calculate the RMS value with your processor using the incoming data from the ACS758, or use a different device which can output an RMS value based on an AC current.

12

Hi David,
Yes, I mean an RMS current as an ultimate goal to be able to achieve after post processing the data fed to the ADC. In this case it seems like we will have to acquire data at a fast rate and then post process it to get the rms current. This is not something we intended. We are looking for a device that will provide us pure DC voltage proportional to the change in AC current. We can also use AC to DC converters and I will look into this. If you know of any device(s) then please advice.
We are able to achieve this with one of the current sensors I have attached the data sheet here. But, we are looking for sensors that are little cheaper to measure 50A or less. If you know of any other current sensor in this range of current meausurement that will provide proportional DC output then please let me know.

Again, thanks for your help.

h722hc_i0a2.pdf (720 KB)

13

OK, so I believe you need a sensor that outputs a DC voltage which is proportional to (i.e. changes with) the magnitude of the RMS AC input current.

I’ll see what I can find.

14

It could be dc value proportional to peak ac input current which cna be converted to rms with easy math.

15

With the ACS758, assuming your current is nearly sinusoidal, you could approximate this by examining incoming values and keeping only the max and min values, but this still requires comparing previous values to current values and looking for the peaks and troughs of the sine wave. That is a bit more processing intensive (comparing consecutive values) than just getting the RMS value directly, but simpler than doing a true root-mean-square calculation on a bunch of data points.

I am still looking for a simple, lower cost solution than your H722HC.

16

Hi Jatin,

The lowest cost item I can find that will natively output a DC voltage which is proportional to the RMS AC current of the input is the CR9580-50. It appears to be somewhat lower cost than the H722HC, based on what I found online. The only caveat to using this part is that for maintaining its stated accuracy of 0.5% at full scale, it requires that the device measuring the output voltage have an input impedance of at least 1M Ohm.

Typical ADCs I have seen have input impedances in the 50K to 150K Ohm range, which is well below the 1M Ohm level specified. If you can live with a lower accuracy, you can probably interface this directly to your ADC. If not, a simple unity gain buffer circuit with a single op-amp should increase your input impedance to well above 1M Ohm.

For lower cost alternatives, I could not find any which give an analog DC voltage proportional to the RMS of the input current. Those which output an analog voltage all seemed to give instantaneous, rather than RMS values. As described above, you can determine RMS current from these, but it is more processor-intensive.

If you could handle a serial interface, such as I2C or SPI, we do have some IC’s which CAN calculate the RMS current. One example of this is the ACS71020KMABTR-090B3-I2C, which outputs I2C, but which unfortunately is out of stock at the moment.

We also have a number of energy metering IC’s which can do this as well. Unfortunately, the level of circuit board complexity and software required to use these is significantly more advanced than the other options, so this may not be the way you would want to go.

One other possibility that I found is to use a current sense transformer along with some external circuitry which, together, can approximate an RMS current sensor. I found an application note titled Precision Rectifier Circuit for CT Signal Conditioning on the CR Magnetics website which describes how to do this. They give the following circuit example:

image
image789×231 10.9 KB

This solution would be somewhat higher priced than the ACS71020KMABTR-090B3-I2C, but somewhat less than the more integrated CR9580-50.