Measuring AC line voltage with OptoCoupler

[Piet de Pad]

Dear Reader,
I recently simulated the circuit shown below using MultiSim and obtained some interesting results.
The simulation suggests that you can accurately reproduce the AC line voltage from the input to the output. Given that optocouplers are typically not linear, this outcome is surprising to me.

Has anyone else conducted a similar experiment or observed this on an oscilloscope? I looked at the Spice model and saw various sub. models used to make the Opto model. To me, it looked as a normal diode transtor modeling. Could the result be due to the simulation model of the ptocoupler used by the simulator (SPICE)?

I look forward to your responses why this simulation is within margins of what you can expect or that this simulation way of from reality. Two heads are better than one.
Regards,

 

[std_match]

The way to do this in the real world is to have two identical optocouplers. The first driven from the measured signal, and the second driven from a feedback loop adjusted so it gives the same output signal as the first optocoupler. The A/D-converter measures the input to the second optocoupler.

 

[FvM]

The circuit gives absolute value output for AC as expectable according to ACPL-214 AC function. The symmetric output is however strongly idealized, ACPL-214-500E CTR can vary in a range between 20 and 400 % (20:1).

The optocoupler is intended to isolate "digital" (on/off) signals, not analog measurements. std_match mentioned a possible solution for analog optocoupler operation, you can refer to Vishay IL300 as an example https://www.vishay.com/en/product/83622/

 

[Piet de Pad]

The circuit gives absolute value output for AC as expectable according to ACPL-214 AC function. The symmetric output is however strongly idealized, ACPL-214-500E CTR can vary in a range between 20 and 400 % (20:1).

The optocoupler is intended to isolate "digital" (on/off) signals, not analog measurements. std_match mentioned a possible solution for analog optocoupler operation, you can refer to Vishay IL300 as an example https://www.vishay.com/en/product/83622/

Hi FvM thanks for responding. You say something interesting. "The circuit gives absolute value output for AC as expectable according to ACPL-214 AC function." What do you mean exactly with "according to ACPL-214 AC function." have you seen documentation that says that this ACPL has a reasonable linearity in the I-forward current area as I use it?.

 

[thannara123]

So by using two optocoupler ,Can we get the same input signal to the output ?

 

[betwixt]

NO!
Read the IL300 data sheet, it has one LED and two sensors inside one package. One sensor is used to carry the signal, the other is used via a feedback circuit to regulate the LED current to keep it stable and in the linear region. If you use two opto-couplers you would be monitoring the current of the wrong emitting LED.

Brian.

 

[Easy peasy]

IL300 is the go

 

[Piet de Pad]

So by using two optocoupler ,Can we get the same input signal to the output ?

So by using two optocoupler ,Can we get the same input signal to the output ?

Not so fast, My simulation says yes you can, but that is exactly my question. Is my simulation close to the reality. I don't know that, and that is my question.

--- Updated Jun 15, 2024 ---

NO!
Read the IL300 data sheet, it has one LED and two sensors inside one package. One sensor is used to carry the signal, the other is used via a feedback circuit to regulate the LED current to keep it stable and in the linear region. If you use two opto-couplers you would be monitoring the current of the wrong emitting LED.

Brian.

I know the IL300 but my question is my simulation correct because if so I don't need an Il300.

 

[FvM]

Two separate optocouplers don't work well because it's effectively impossible to get identical characteristics. Problem in simulation is that gain variation isn't modelled unless you introduce in explicitely. IL300 is solving the problem by using two matched photodiodes in package. The matching isn't perfect however and still needs individual gain adjustment. Analog optocouplers are a last centuries technology if you're looking for precision, today's analog isolators are digital (AD/DA).

Once again regarding post #2 question. To transmit AC with analog isolator, you need to bias the LED current to midrange. An AC input optocoupler like ACPL-214 is no solution.

 

[betwixt]

The simulation will be correct for the given component parameters, the issue is the spread of parameters to expect in real life. If the CTR can spread over a ratio of 20:1, so can the results. You could be lucky and get the expected results or you could be 20 times out and if you are building many of these you should expect varying results. That is why the IL300 was suggested, because it monitors its own LED luminance, the CTR can be taken out of the equation.

Brian.

 

[Piet de Pad]

The simulation will be correct for the given component parameters, the issue is the spread of parameters to expect in real life. If the CTR can spread over a ratio of 20:1, so can the results. You could be lucky and get the expected results or you could be 20 times out and if you are building many of these you should expect varying results. That is why the IL300 was suggested, because it monitors its own LED luminance, the CTR can be taken out of the equation.

Brian.

Thanks, I'll keep that in mind