Opto-isolator/ optocoupler for electric isolation in IC design.

[Alex_bam]

Hello everyone.

I came across while reading the opto-isolator technique used for electric isolation in embedded systems/ Integrated Circuits design that " Opto-isolator is not recommended for electric isolation in power applications IC's mainly due to its dV/dt sensitivity". So I would like to know:

1)How dV/dt sensitivity affects isolation?

2)Is dV/dt sensitivity the only reason to avoid optocoupler especially in power application IC/embedded systems?

3)Are there other limitations in optocoupler which can compel one to chose other isolation techniques?

Looking forward to your suggestions and guidance. Thanks.

 

[rjenkinsgb]

I'd guess that they are referring to such as high voltage PWM or phase control systems?

eg. If a power control system is switching switching hundreds of volts with "fast" edges, an opto isolator connected to a switched point could be considered as a capacitor between than and low level logic inputs or outputs, which could be rather a bad thing, coupling high voltage spikes across.

Optocouplers are excellent for isolating general "switching" type inputs or outputs. They are not so good for analog signals as they are not linear, though there are special types that can be used for that with additional opamps & feedback.

 

[Alex_bam]

I'd guess that they are referring to such as high voltage PWM or phase control systems?

eg. If a power control system is switching switching hundreds of volts with "fast" edges, an opto isolator connected to a switched point could be considered as a capacitor between than and low level logic inputs or outputs, which could be rather a bad thing, coupling high voltage spikes across.

Optocouplers are excellent for isolating general "switching" type inputs or outputs. They are not so good for analog signals as they are not linear, though there are special types that can be used for that with additional opamps & feedback.

Thanks, rjenkinsgb for your explanation. I was searching optocoupler for power applications. Moreover, I would like to study its limitation in analog signals as you mentioned. I would be thankful if you have /know a specific resource link addressing its limitation. Thanks.

 

[Diver300]

https://www.mouser.ca/datasheet/2/240/LOC111-1547581.pdf

That's an example of an opto isolator that can be linear if used correctly.

The application note may also help:- https://www.ixysic.com/home/pdfs.nsf/www/AN-107.pdf/$file/AN-107.pdf

 

[rjenkinsgb]

If I were to use an optocoupler to isolate the control logic from the power stage in eg. a high voltage servo drive, I'd split it before the power drivers whilst the signals are still at logic levels with minimal or only low-speed variations between the common rails of the two sections.

eg. Between the PWM source and the IR2110 or whatever high voltage drivers.

For analog use, a "simple" circuit does not give linear transfer between input and output, eg.
(I'm just finding on google as I go...)

**broken link removed**

There are opto isolators with two matched outputs. Using one of those, with one output used as feedback at the input side, the signal from the other output (in to a similar load as the first) should very loosely track the input signal.

Example:

 

[schmitt trigger]

Toshiba has many app notes with everything you have always wanted to know about optocouplers, and then some more.
Download them from here;

https://toshiba.semicon-storage.com/ap-en/semiconductor/product/optoelectronics.html

 

[Beau Schwabe]

Just an observation ... if you are using the same Vcc and GND on both sides of the opto as well as the same OP-amp then you really have no isolation. Also on either Op-Amp you might want to provide a method to "fine-tune" the output.