Go Down

Topic: measure AC ampere and monitoring/signal (Read 586 times) previous topic - next topic

Wawa

...(even if it DID measure AC current.. which it does not).
A CT must be biased mid-voltage, and the the AC current sinewave must be sampled many times to get the shape or peak of the wave.
Same difference with a bipolar ACS712/723, except that the biasing is already done for you (idle out = VCC/2).
Not a problem to measure AC with these chips.
Leo..

krupski

RTFM !!!!!!
It DOES measure AC.
I checked DigiKey just now. The -AB suffix parts do AC and DC. I stand corrected. However, in my defense, only "ACS723" was mentioned. I've used these before, but only the DC parts and I never noticed that there was also an AC version.
Gentlemen may prefer Blondes, but Real Men prefer Redheads!

krupski

A CT must be biased mid-voltage, and the the AC current sinewave must be sampled many times to get the shape or peak of the wave.
Same difference with a bipolar ACS712/723, except that the biasing is already done for you (idle out = VCC/2).
Not a problem to measure AC with these chips.
Leo..
A CT requires a "burden" (a load resistor). The AC voltage across the burden can be rectified and filtered, then fed to an ADC.

There will be a small intercept error at low currents (where the diode is just beginning to conduct). A schottky diode or better an op-amp rectifier can minimize this problem.

I still think those little hall effect current sensors are not safe to use directly on the mains.  Maybe a CT and an ACS723 together would be ideal.
Gentlemen may prefer Blondes, but Real Men prefer Redheads!

raschemmel

#18
Dec 11, 2019, 04:06 pm Last Edit: Dec 11, 2019, 06:20 pm by raschemmel
I linked the Sparkfun ACS723 breakout bd
which specificAlly designed with the AC/DC version to reach a a larger market.


The CRMAGETICS CTs include the burden resistor.
This is the one I have:


CR8459-2000N DATASHEET


I might be wrong about it including the burden resistor.
The diagram shows an included resistor DCR (74 ohms for this model) in series .
It also shows a resistor "R" in PARALLEL with the CR. (This must be the BURDEN resistor since it is in
PARALLEL) but no value is given for this.

For this part:
Ir = 200 (this is the MAX CURRENT SPEC)
VmaxRMS = 11.5
Te = 2011
DCR = 74
Frequency(f) = 20Hz - 1 kHz
R = UNKNOWN

Which yields :
Let Ir = 200 (for example purposes)

VL = Vmax - [(I * DCR)/Te]
                        11.5 -[(200 * 74)/2011
        VL = 4.14 V

SPEC sheet says (in fine print):

"For best linearity, choose R such that V < 0.8 VL"

Let:
V = 0.8*VL
V = 3.13V
I = 200A
Te = 2011

Thus,

3.13V = I*R/Te

3.13V  =  200*R/2011
(cross multiply),
(devide both sides by 200A)

R = (3.13V)*(2011)/200A

Burden Resistor R = 33.3 ohms

(AFAIK)

Wawa

I never noticed that there was also an AC version.
The bi-directional version is common (± DC and AC). The uni-directional one is not.

A CT requires a "burden" (a load resistor). The AC voltage across the burden can be rectified and filtered, then fed to an ADC...
But it's usually not done like that, because of that diode threshold.
The AC signal is usually presented mid-voltage to the analogue pin, and peak/peak value is extracted by continuous sampling at least 3/4 of the sine wave.
Leo..

krupski

"For best linearity, choose R such that V < 0.8 VL"
Years ago I was an engineer at a company that made current transformers including the "split core" type.

The reason that the voltage across the resistor at max rated current is specified is to keep the core flux density in the most linear region that the particular core supports.

At the low end (low R, low voltage) the volts per turn is very low and most of the energy goes into overcoming the magnetising force of the core (making the output voltage much lower than it should be).

At the high end (large R, high voltage) the volts per turn exceeds the capacity of the core and it saturates.  Look at a B-H curve and you will clearly see the linear region and the terribly non-linear ends.

Interestingly, with no resistor the output current is theoretically zero and the output voltage theoretically infinite.

A well insulated CT can produce spectacular hot, blue/white arcs at the output terminals with no resistor. I'm talking 3, 4 or more inch long hot, wicked arcs.

Amazingly, if the CT is well made and not phase compensated, it can actually survive this abuse unharmed!
Gentlemen may prefer Blondes, but Real Men prefer Redheads!

krupski

The bi-directional version is common (± DC and AC). The uni-directional one is not.

Are you thinking about sensors that can measure DC current in both directions (such as -30 to +30 amperes) or sensors that can handle actual AC current?

I've seen all kinds of DC sensors, but only heard about the AC ones yesterday!
Gentlemen may prefer Blondes, but Real Men prefer Redheads!

raschemmel


Go Up