hi every one i am a new user of optocoupler pic817 and for my pwm based h-bridge driving circuit i am using it.. is it a right choice for this circuit?...or should i need to replace it with some other package?

one more thing that i want to clear is,currently i am using emiiters output which is being further fed into buffer . so what if i take output from collector?
my circuit diagram is given in attachment

 

What is the purpose of the resistors in parallel to the emitter LEDs?

You can use the collector as the "output" if you move the resistors to the high side. I believe this might give a better signal but the logic will be reversed.

 

hi every one i am a new user of optocoupler pic817 and for my pwm based h-bridge driving circuit i am using it.. is it a right choice for this circuit?...or should i need to replace it with some other package?

one more thing that i want to clear is,currently i am using emiiters output which is being further fed into buffer . so what if i take output from collector?
my circuit diagram is given in attachment

The whole point of an opto-isolator is that the output transistor is floating, you can tie the emitter to ground and the collector via a load resistor to +V, or tie the collector to +V and take the output across an emitter resistor.

You can look at it in the same way as the difference between common emitter and emitter follower - common emitter is inverting, emitter follower isn't. Except the input needs more Vf to turn the LED on, and the input and output are isolated.

 

I use the MOC8106, also go direct from PIC to opto to Ir2110.
Max.

 

Side note I see no need for R7 or R10 (barring serious ground loop issues). Remember, the LED inside still has a Vf. That said, I don't see your Vdd and Vss connections for the MCU or any decoupling.

Your setup may work just fine using the collector, as ian field explained, the transistor is floating. That said, I have generally used them as open collector outputs, at the expense of inverted logic.

 

Side note I see no need for R7 or R10 (barring serious ground loop issues). Remember, the LED inside still has a Vf. That said, I don't see your Vdd and Vss connections for the MCU or any decoupling.

Your setup may work just fine using the collector, as ian field explained, the transistor is floating. That said, I have generally used them as open collector outputs, at the expense of inverted logic.

Most of the safety rated opto-couplers used in the regulation feedback isolation on SMPSUs have no external connection to the base terminal.

To all intents and purposes, the transistor in such units is a two-terminal device - it is polarity concious though.

Years ago I had a magazine project published for a test set for no base opto-couplers.

Since its possible to invert the basic PUT relaxation oscillator and use a TO 92 SCR instead, I made it dual purpose - the pulse generator SCR was also socketed so it could also be used to test the SCRs.

For a while, TO92 SCRs like the 2N5061 were popular in latching shutdown circuits on TV sets.

 

Most of the safety rated opto-couplers used in the regulation feedback isolation on SMPSUs have no external connection to the base terminal.

I actually had no idea there were optocouplers that HAD external base connections. PS2501's have historically been my goto.

 

I actually had no idea there were optocouplers that HAD external base connections. PS2501's have historically been my goto.

There was a rise in the popularity of the ones like a 4-pin DIL chip, the original 6-pin DIL types used to have a base connection, but as new safety regulations came into force, many types without were introduced. A lot of the safety types also have the pins stretched out wider than the usual 0.3".

 

Opto couplers are pretty darn slow, not good for high frequency H-bridge applications.

 

Opto couplers are pretty darn slow, not good for high frequency H-bridge applications.

Not all of them, the digital ones with gate at the end can be pretty fast, and some even contain power drivers like the HCPL3120.

 

thanks all of you for your concern and replies

 

Opto couplers are pretty darn slow, not good for high frequency H-bridge applications.

then what do you suggest? what i should i use instead of an optocoupler?

 

Not all of them, the digital ones with gate at the end can be pretty fast, and some even contain power drivers like the HCPL3120.

@kubeek, is there a way these can be used with a 'high side' mosfet?

 

A standard bootstrapped high side driver configuration will work, similar to how IR2110 and others are configured.

 

Compare the rise and fall times, the one you selected appears to be 18us, the MOC8106 is a little faster.
Max.

 

then what do you suggest? what i should i use instead of an optocoupler?

What pulse frequency? You might just about get away with 100kHz, you may be able to get Schmitt-triger input type MOSFET drivers that clean the pulses up a bit.

Don't use Darlington type opto-couplers - they're even slower than straight bipolar transistor types.

Before giving up - look at the data sheets for some MOSFET output optos, I don't know off hand whether they're faster but it wouldn't hurt to look.

The MOSFET ones are quite popular in telecoms applications.