Hi Paul,

You may not obtain very good precision if you use an optocoupler for this sort of voltage sensing. The sensitivity too may not be that great for rapid voltage variations; and that could be critical for the controllers that you are contemplating.

I suggest you consider using a suitable LEM-Hall Effect Voltage Sensor.

There is a certain distributor of LEM sensors in Bengaluru: one Globetek India.

Just contact them, and get the sensor of voltage range that you need. They distribute the entire range of LEM current sensors as well, by the way. And you get decent discounts if you buy sensors in bulk.

-Sanjay

Sanjoy is quite right - of course. Optocouplers aren't really very linear. If you absolutely need a high voltage isolation and speed ( and cost!) isn't too much of an issue you could digitise the voltage at the source and transmit it digitally via the coupler. It needn't be binary - it could be a simple mark/space signal and couplers are designed to handle these signals very well.

But I would be the first to admist that's getting a bit difficult for a simple circuit.

Thanks for your answers.

i am working on a approx 30V and 10-20W converter.

Going for a hall voltage sensor may be a bit expensive.

I have seen many hardware models using 4N25 optocoupler as voltage feedback unit. But, I am not convinced how this variable voltage is being sensed by an optocoupler.

the data sheet show a current transfer rate of 20%

I havent practically tested 4N25 for its linearity with voltage applied; which I am looking forward to.

the applications specified in datasheet is mainly of logic type either a HIGH or LOW will do. But for SMPS or voltage feedback applications, this logic type outputs will not do, some sort of linearity in output is required.

Being seen hardware implemented with 4N25 in voltage feedback for converter (5-20W), I am bit confused.

Hi Paul,

Personally, I would never have risked the performance of a controller on something as basic as sensor performance.

But you say that you have actually seen hardware implementations with the 4N25 in the feedback. If that is so, the most straightforward thing to do will be to configure the hardware yourself, and take a decision as to whether the performance is acceptable or not ! 

With best wishes.

-Sanjay

One solution is to add voltage to frequency converter in the DC circuit you are measuring and transmit the pulses through the optocoupler.  Then on  the other side convert the pulses from the optocoupler to analog voltage or use a counter if you want to have the measurement in digital form. Naturally the speed of measurement depends on the frequency used and maximum frequency of 4N25 may not be enough for switch mode power supplies with small filter capacitors.

Paul, you are working at 30V but must the output 30V be electrically isolated from the input - which is the reason for the optocoupler? If it isn't isolated you don't need to isolate the feedback path at all. If it has to be isolated then you need to either:

1. Transmit a very accurate measure of the output voltage via an isolation device like a transformer or optocoupler, or

2. Generate your reference voltage on the isolated side of the convertor when you can then transmit only a difference signal through the isolator which is a bit easier. Doing this can make the system harder to self-start.

Best wishes!