Practical usage of Current Transformers (Simple circuit or dedicated IC)

[ste2006]

I know Current Transformers have been discussed here before but i have a slightly different angle that i cant find the answer too,

I am currently desiging a circuit to measure current from say 0-30A, I have the current transformer, load resistor etc and that is all ok, I intend to feed the variable DC output into a Micro and use AD converters in the Micro to give me the numbers i want,

Regarding though getting from the CT to the Micro, I seem to have two options and cant see the big difference

1) Use a CT, Load Resistor, Peak Detector or Bridge rectifier and feed this into my Micro's ADC??

2) Use a CT and then a dedicated IC like the Maxim 78M6613 or Microchip MCP3911 and then interface this to my micro using SPI or the like,​

I will have approx 8 CT's in a box measuring 8 current sources but cant see the major benefits or drawbacks of using either method mentioned above,

Anyone any experience with this or advice???

Thanks,

Stephen

 

[KerimF]

Do you have an idea about the voltage range on the CT load resistor for 0 to 30A?
Did you mean passive or active peak detector?
Is the shape of the measured current fixed, like being sinusoidal?

 

[ste2006]

Yes CT is 1000:1 and the manufacturers recommend 100 ohm load resistor so about 0-3V.

It is to measure 220V AC so yes it will be a fixed 50Hz AC from the mains.

Have not put a huge amount of thought into the peak detector but probably a passive one if i can get away with it.

Again open to suggestions here, I do not need my system to be highly accurate but at the same time a good representation of the current flowing.

What i fail to understand is the benefit of using an extra IC when the Micro already has multiple ADC's etc built into it,

Thanks,

 

[ste2006]

Anyone any other thoughts on this as i really am at a loss as to which to go with, I am leaning towards cheap and cheerful but i just want to make sure i am not missing out on something obvious,

Thanks,

 

[KerimF]

Sorry for the late reply... my internet connection wasn't fine for a while.
As you know, the passive rectifier cannot detect low level signal below 0.5V for example or 1V if full-bridge.
Even after this low limit, linearity suffers for another 0.5V in the least.
In my opinion using just a diode for 0 to 3V range is not good unless you are not interested in measuring low current (0 to 5A)
Before I go on please remember that the peak dectector gives a value of the current peak and not its RMS. So when you say 0 - 30A you likely mean RMS values and the peak range is actually 0 - 42.4A which is reflected to 0 - 4.24V at the secondary (assuming ideal peak detector).

A quasi ideal rectifier (for peak or average) could be done by using a general purpose opamp like LM324 (quad) or LM358 (dua). Are you interested?

 

[ste2006]

Hi thanks for the reply,

Yes i do mean RMS and yes if you have any info on ideal rectifiers please tell or point me in the right direction,

Analogue was never my strongpoint so the sooner i have it converted to digital the happier ill be

As im using a micro with an ADC lineararity is not a huge issue as i am happy just to use a fixed number of steps, It is more for a general guide on current usage as opposed to having to be super accurate,

thanks,


Thanks,

 

[KerimF]

By using one opamp in LM324 and 4 resistors we can build a quasi ideal half-wave rectifier, say for the positive cycles. Obviously using another opamp in the IC we can build another ideal rectifier but for the negative cycles. As you have already guessed we can use a third opamp as an adder to get them both. The forth opamp can then be used as a simple low pass filter to get the average value (its ripple will be lower if the response is made slower).
It is also possible to build a quasi ideal peak detector by adding a diode in the opamp feedback loop.

I hope I will have time tomorrow to draw these solutions for you on LTspice (a versatile free simulator) so that you will get a clear idea of what I am talking about here. But I hope others will get in and give you the schematics sooner since these circuits are usually known by most engineers.

Kerim

Added_1:
I will assume there is a 5V supply and one of the two terminals of the CT secondary is grounded.

Added_2:
I don't use MCUs with ADC. Would you please let me know the lower/upper voltage limits at the ADC pin?
I guess the relatively low impedance of the opamp output is adequate for the ADC pin.

 

[KerimF]

Hi again,

I didn't sleep early so I draw the circuit so that you can have an idea rather soon if it helps or not.

Perhaps tomorrow I will draw the peak detector circuit, also using LM324.
I guess getting the peaks of one polatity only is good here since the AC current is rather symetrical.

Kerim

 

[ste2006]

thats great, Thanks Kerim,

it can use a 5V supply and will look at that, I can set the limits on the ADC but maybe 0-3V but again i can tweak these things to suit, I just need to get a general idea what i am looking for,

I will spend some time today researching ideal rectifiers and hopefully make your job of explaining a little easier,

Thanks again,

Stephen

---------- Post added at 09:57 ---------- Previous post was at 09:56 ----------

thats great, Thanks Kerim,

it can use a 5V supply and will look at that, I can set the limits on the ADC but maybe 0-3V but again i can tweak these things to suit, I just need to get a general idea what i am looking for,

I will spend some time today researching ideal rectifiers and hopefully make your job of explaining a little easier,

Thanks again,

Stephen

 

[KerimF]

Here are some notes concerning the starting circuit (post #8):

General:
(A1) The inputs of the LM324 can accept a negative voltage if its value is less than 300mV.
(A2) The outputs cannot go below ground.
(A3) The outputs cannot go (linearly) higher than Vcc-1.4 = 5-1.4 = 3.6 V

For U1:
(B1) The voltage divider R1 and R2 ensures that Vin+ won't go below -300mV.
(B2) The negative cycles cannot give an output below ground hence only positive ones are amplified.
(B3) The voltage gain (for positive cycles) is:
G1 = R2 / (R1 + R2) * [ 1 + R4/R2 ]

For U2:
(C1) The resistor R8 could be omitted, I use to include it to better balance the input bias voltages due to the bias input currents.
Its effects could be seen when the resistance values are relatively high above 100K for example.

(C2) The resistor R6 could be omitted, I use including it to get a better stability if the voltage gain is less 1.
Yes, one may think that changing the value of R6 or even removing it can affect the output.
But the output will not be affected at all, practically speaking.
Could you prove it by a simple remark?

(C3) The voltage gain (for negative cycles) is:
G2 = - R7 / R5
where R5 = R5a // R5b

(C4) I tried my best to let G2 = G1
(C5) During the positive cycles, the output stays close to ground.

For U3:
(D1) The voltage gain for input N:
G3n = R10 / (R10 + R9) * [ 1 + R12/R11 ]

(D2) The voltage gain for input P:
G3p = R9 / (R10 + R9) * [ 1 + R12/R11 ]

(D3) Since R9 = R10, and R11 = R12 G3n = G3p = G3 = 1
(D4) The output voltage:
Vpn = V(N) + V(P)
This gives us an output of an ideal full-bridge rectifier.

For U4:
(E1) The voltage gain of the average value of Vpn is:
G4 = 1 + R15/R14
This can be changed to adjust the upper limit (but it is better not to exceed the maximum 3.6V)

(E2) The input RC could be changed mainly C1 to get a smaller ripple or a faster response (a relatively fast response needs a sort of synchronisation with the main input cycles so that the ADC routine can sample the AVG at the same phase).

Practical notes:
(1) the parts of U2 and U3 could be omitted. The draw back will be a slower response by a factor of 2, for the same ripple.

 

[ste2006]

Thanks so much for all your help, Ill mull over all this along with my own research in the coming days and come back if i have any more issues,

Thanks again

 

[ste2006]

Any benefit in using a ideal peak detector vs a ideal rectifier,

I have done lots of study into both and as far as i can see both will do the job??

Thanks,

 

[FvM]

You previously mentioned RMS current as your intended measured quantity. Considering the fact, that AC currents can have considerable distortion, "rectified value" (averaged after rectifier) would be the next best (and commonly used) approximation of RMS and peak value the worst, because it involves the largest waveform dependent deviation from RMS.

 

[KerimF]

In your case here, measuring the average value gives a closer estimation of the RMS than the peak one does.
The main differences between the two detectors is in their response time.
The peak dectector needs only a half cycle to give an estimation of the measured current hence it is good for measuring the inrush current for example.

Kerim

 

[ste2006]

Great, Thanks guys

 

[ste2006]

Have been doing some work on this guys but one thing still puzzles me, The low pass filter. How are you choosing a corner frequency and any reason not to just use a Passive Low Pass Filter??

Thanks,

 

[KerimF]

It is not about the frequency corner but the amplitude of the ripple (between two consecutive cycles).
Using an opamp after RC is just for isolating the passive low pass filter by a buffer (to get relatively low impedance at the ADC pin).

 

[ste2006]

Ok makes sense, Ill do some more work on it and see, The ripple currently is a little larger than i would like, Thanks for your help with this though, You have really pointed me in the right direction,