Building a digital AC ammeter

[Reloadron]

Measure the output of the ACS756 as it should be a DC level proportional to the AC current. I don't get why rectify it?

Ron

 

[crutschow]

You can use an op amp in a precision rectifier circuit.

Alternately you could rectify the current into the ACS756 with a power bridge rectifier. This will give a full-wave rectified output from the ACS756, But you may still need an op amp to amplify the voltage before going to the ICL7107.

You can filter either signal with an RC filter to give the average value of the rectified AC. You then need to adjust the gain of the ICL7107 to display RMS from this average (the difference between RMS and average is .707/.637 for a sine wave).

 

[lilimike]

When I measured the output of ACS756 using DC voltage I got 2.52V on all tests performed.
When I measured the output using AC I got the results posted.

That's why I figured output of ACS756 was AC
Oh and I just realized those ACS756 output results bellow are not mA they are AC mV

50 W = 400 mA / ACS756 output = 16 mA
100 W = 830 mA / ACS756 output = 33 mA
150 W = 1.240 A / ACS756 output = 49 mA
190 W = 1.56 A / ACS756 output = 61 mA

 

[crutschow]

Measure the output of the ACS756 as it should be a DC level proportional to the AC current. I don't get why rectify it?

Don't confuse the poor guy. The data sheet states the output is biased at 2.5V with no input current, and the figure on page 8 clearly shows a plus and minus output voltage around the 2.5V bias for a plus and minus input current.

 

[Reloadron]

Don't confuse the poor guy. The data sheet states the output is biased at 2.5V with no input current, and the figure on page 8 clearly shows a plus and minus output voltage around the 2.5V bias for a plus and minus input current.

Yes, and the output voltage is DC, I am sorry and wasn't trying to confuse him. My point was that the output is not AC but DC proportional to the current. Sorry if that was confusing.

Ron

 

[crutschow]

Yes, and the output voltage is DC, I am sorry and wasn't trying to confuse him. My point was that the output is not AC but DC proportional to the current. Sorry if that was confusing.

That is not correct. The output with an AC input is AC around (riding on) the 2.5V DC pedestal. Look again at the graph on page 8.

 

[lilimike]

Ok I made some more tests this time using my scope and observed these values measuring DC voltage at the output of ACS756.

         [U]0 mA[/U]        [U]400 mA[/U]       [U]830 mA[/U]        [U]1.24 A[/U]        [U]1.56 A[/U]
[U]Vrms [/U]    2.44V       2.44V        2.44V         2.44V         2.44V
[U]Vmax [/U]    2.48V       2.48V        2.54V         2.56V         2.58V
[U]Vpp [/U]     60mV        120mV        160mV         200mV         220-260mV (fluctuating)

I see what you mean by this

The output with an AC input is AC around (riding on) the 2.5V DC pedestal

It looks like AC voltage on top of the 2.5 Volts. So what I understand from this is I should set the ICL7107 to display 0 when it is input-ed 2.48 V and adjust the values from there on?

I would like to test 10 Amp, anything better than 12 x 100 Watts light bulbs?

Mike

 

[Reloadron]

That is not correct. The output with an AC input is AC around (riding on) the 2.5V DC pedestal. Look again at the graph on page 8.

My apologies, I never saw it or caught it. I wrongfully assumed the output of the chip was a DC voltage proportional to the current, be the current AC or DC current. Sorry if I tossed in confusion. My bad.

Ron

 

[crutschow]

It looks like AC voltage on top of the 2.5 Volts. So what I understand from this is I should set the ICL7107 to display 0 when it is input-ed 2.48 V and adjust the values from there on?

I would like to test 10 Amp, anything better than 12 x 100 Watts light bulbs?

Mike

You can use a series capacitor and resistor to ground, to AC couple the ACS756 output, which will block the 2.5V DC before you rectify it. That way to don't have to null that value from the ICL7107. The RC time constant of the coupling circuit should be 25ms or longer for accurate coupling of the 60Hz AC.

Do you have a space heater, toaster, or clothes iron? They all likely take 10A or more.

 

[lilimike]

You can use a series capacitor and resistor to ground, to AC couple the ACS756 output, which will block the 2.5V DC

Ok I will try this.
Thanks

Do you have a space heater, toaster, or clothes iron? They all likely take 10A or more.

I do have a bunch of appliances but I was hoping to find a solution to get exact values from 1 to 10 Amps to test and compare low to high range but I guess I can deal with light bulbs.

Mike

 

[lilimike]

Because I don't really want to add another voltage regulator or change the configuration I have, my only option is the last one on this link (the one based on the OPA2237) as it can operate with 5 Volts.
My circuit runs from a 5 Volts regulator and I am using an ICL7660 to generate the -5V needed for ICL7107.

I am wondering if I will be building a precision rectifier out of this OpAmp, can I not increase the gain at the same time?
Or will I need to add a second OpAmp to increase the voltage in which case I am hoping the OPA2237 is adequate?

Mike

 

[crutschow]

Below is an article on a precision rectifier with adjustable gain. You could use that to get the rectification and gain. Remember that you need a negative supply for the op amp to handle the negative part of the sine-wave.


If you want to use only a single supply, then you could use a precision rectifier such as shown in Figure 5 of this ap note. For that you would need to add another op amp with the desired non-inverting gain. You could do all that with one quad rail-rail op amp.

 

[lilimike]

I will look into this and test as soon as I receive more parts.
In the mean time I am feeding 126 VAC to the ICL7107 but I am getting a lot of fluctuation. It is not possible to fine tune but if I try to get as close to 126 VAC as possible I can observe fluctuation going from 126V to 120V in about 3 seconds and then back to 126V in another 3 seconds and so on...

I understand I should net get accurate measurements but I would like to resolve the fluctuating readings if I could.

Any clues?


Mike