Optocupler question.

[hapetter]

If I have a sine AC signal (50Hz) signal going into a optocupler. How will the output look like?

 

[VEBill]

It depends on if and how the devices are biased, how the signal is coupled, and the input and output amplitudes compard to the linear ranges.

With care, you can force optocouplers to be linear. I've used them to carry audio to good success.

But if you've not made any effort to keep things linear, then the output would look like a half-rectified near squarewave.

 

[hapetter]

I'v found something that I find interesting, a zero-cross optically isolated triac driver seems to take an AC signal into smal square pulses TTL signal, right?

 

[marginal]

but triacs need there current to be above their holding current.

if you are after zero cross detector yuo can do it with especially ac coupled opto coupler (i think its HA11A1 or similar).......the transistor can then be used provide an input to a microcontroller with say a pull up resistor.

or use it to turn a transistor on or something

 

[BobM3]

There are a whole bunch of different types of optos. Are you looking for a particular output for a particular input?

 

[VEBill]

My response (above) was addressing the simplest type of optocouplers: an LED on one side, and a phototransistor on the other.

 

[itsmoked]

Yes, there are probably 20 different families of optos. If you want a valid answer you need to be more specific.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[hapetter]

Thanks.

This is what I want exactly:

I want to make a cleen square puls from a 0 - 220VAC 0-120HZ. The signal must be noise free. The output (Square pulse) is going into a AVR microcontroller. Then I like to calculate the frequence of the AC signal.

The positiv part of the AC sine signal must be 5VDC and the negative part 0VDC.

Any opto component for this?

 

[benta]

You are not going to get any signal at 0 V or 0 Hz. You need to make your specifications realistic.
5...230 VAC is realistic.
0.5...120 Hz is realistic.

Benta.

 

[hapetter]

Ok I agree.

But the range must be as wide as possible.

1. So will an zero-cross optically isolated triac driver work in, lets say 5-230 VAC , 0.5-120 Hz?

2. Do I need to filter the signal with other components to get a clean signal?

Hans

 

[benta]

1: No. The Triac will turn on and stay turned on. You need a completely standard transistor output optocoupler.
2: Probably not. Choose a slow coupler (Darlington), that should be filtering enough.

You REAL challenge here is to build a circuit that will supply a constant current of, say, 20 mA to the optocoupler LED regardless of whether the input peak voltage is 5 V or 340 V (peak value of 240 VAC).

Benta.

 

[BobM3]

Take a look at the H11A as Marginal suggested. It's used to do what (I think) you are looking for:

http://www.vishay.com/docs/83608/83608.pdf

 

[Johnee2]

Hi,

The above referenced part is nice. It has back-to-back LEDs, however, so it will turn on twice per cycle, equivalent to a full-wave rectifier. One LED for the positive half-cycle, the other LED for the negative, or 100 Hz for 50 Hz main AC. It also directly exposes the LEDs on the input wiring so significant circuitry will be needed to drop the AC Mains voltage to provide the maximum (not to exceed) 60 mA LED current.

Unless you're keen on routing 230 V on your PC Board, a small transformer may be a nice option to drop the 230 V to something safe like 15 before hitting the PC Board. Otherwise some sort of safety barrier covering the tracks and opto-input makes a lot of sense.

Then some current limiting resistors. If you want to detect a wide range of AC voltage, clamp diodes will prevent overloading the LEDs. I think that part of the prior post might have been misinterpreted.

The optocoupler output is just the transistor response to the analog input voltage/current. For noise rejection and edge sharpening, a comparator or schmidt trigger is probably needed.

http://en.wikipedia.org/wiki/Schmitt_trigger

It sounds like someone probably should have all this in a module.

 

[Johnee2]

Here's a reference circuit from a clock chip supplier:

http://www.maxim-ic.com/appnotes.cfm/an_pk/1994

You can be pretty sure that the edge shaping and noise rejection is built inside the chip.

 

[itsmoked]

Just use any junker single LED opto.

Put in a series dropping resistor calculated to limit the LED current at the peak line voltage to 20% of Maximum Allowed Forward Current.

Put in a silicon diode in series with the opto LED to block the reverse voltage. This cuts the power dissipation of the dropping resistor in half.(a good thing!).

So you now have a 230V AC source, running to a dropping resistor, running to a diode, running to the opto LED, returning to the 230V source.

Now add another reversed diode in parallel with the the LED. This protects the LED from any destructive reverse voltage.


Monitor the opto's output (the logic levels). The point at which it drops low is going to be the same place on the 230VAC waveform every cycle. Use that as your synchronous point.

I have done this many times for SCR and TRIAC firing control. It is pretty much the standard for this function.
You don't need to worry about the trailing edge of the logic signal. Set up the AVR to count from the falling edge to the next falling edge. That will give you the AC frequency to the accuracy of your counter. If you need the frequency faster than a power line cycle set the aforementioned circuit up in both directions with two opto's. Measure the period between both falling edges. This will get you the frequency in half a line cycle.

Make sure you calculate the power dissipation of the dropping resistor/s. It won't be ignorable.

Keith Cress
kcress -

http://www.flaminsystems.com