Analog isolation from Micro controller

[khansaab21]

Hello
I want to measure a analog voltage (as high as 220V) from micro controller (PIC). Yes i will use the built in ADC of micro controller and a potential divider to bring down the voltages within measurable limits, but, the problem is, how can this analog voltage be measured with the micro controller. I can not mix analog and digital grounds for obvious reasons.
I read somewhere to use opto-isolators, but can analog voltages be transferred over the isolator. I have only used opto isolators to isolate two digital circuits. The isolator i used was ILQ74. Can analog voltages be transferred over this isolator??

 

[nandhu015]

There are linear optocouplers like IL300 which suits your appliction.

Google "linear optocoupler" to find more parts

Nandhu

 

[kender]

I'll second Nandhu on the linear optocoupler suggestion. LOC110 is another linear optocoupler chip.

Consider also a different approach - put a separate A/D converter chip on the isolated side, run digital signals across the isolation barrier.

 

[bobcat1]

Hi

Simply use 1:1 isolation transformer to isolate the 220V from main
and then use the 1:20 resistor divider

Also you can use measurement IC like ADE7753 from analog and optocupler to measure voltage and current from 220V rail

Please remember to act with caution when using high voltage

All the best

Bobi

The microcontroller specialist

 

[khansaab21]

After a bit of research on the posted replies, i found that if too high values of resistances are used on such high voltages then, noise will be generated, and if too low, then divider network will start loading the main line un-necessarily. I think that following solution will be ok

220V will be stepped down on a separate line to about 6V (This procedure is just for measurement), then it will be passed through a voltage divider, following which is the isolator chip. Across the isolator, ADC will be waiting for the voltages.

Any suggestions are most welcome

 

[FvM]

I see the problem, that you didn't yet give a single specification of your measurement (execpt for a voltage range),
that would allow to select a suitable or optimal isolation method.

You didn't even tell about AC or DC. For AC measurement, a transformer as suggested by bobcat1 is a simple and
commonly used technique, bus it has frequency dependant gain and phase errors. For this reason, a precise energy and
power factor measurement should be better use an A/D frontend before the isolation barrier, as the said energy
measurement chips mostly do.

Also analog isolators have limited accuracy, so it's difficult to discuss about their usage without a specification.

 

[khansaab21]

FvM, basically at this point I am unable to detail the specifics because I am still in the stage of selection of methodology for the solution. And question I raised was about one of the methods that crossed my mind and also seemed reasonable. Until this point following are the specifics

AC mains (220V) -> Rectification -> Filter -> DC bus (313V)

This DC bus has to be monitored for over voltage and over current.

And I have the following in mind for the specified tasks

For over voltage: A connection from AC mains will be taken out -> stepped down to 6V -> filtered -> analog isolator -> ADC

For over current: A resistor in the reference ground connection in negative DC bus - voltage across it will be input to a op-amp working in common mode -> signal conditioning will be attached with op-amp -> analog isolator -> ADC

Any way to improve this scheme or any faults in it?? Or any suggestions are most welcome.

 

[Siniša]

Specifics would include accuracy, resolution, bandwidth,... As FvM said those would greatly affect advice someone could give you.

I will venture to say that you might be satisfied with average 10bit ADC included in your MCU and under that assumption might be easiest to do what kender already suggested. Use opamp to amplify signal on current sense resistor, resistor voltage divider for voltage and MCU with ADC on your line voltage side and transfer result as digital signal to second MCU on isolated side through simple OC.

 

[khansaab21]

But the issue is, I am trying to avoid using any sensitive digital electronics on the high power side board as it will require a separate digital supply on that side + another issues will have to be taken into consideration like, management of power and ground traces, health of clocking signals etc.

And a 10-bit ADC is good enough.