Mains voltage zero crossing detector with optical isolated output

Maybe build it with this?

I like this:

It's a dual optoisolator, with the LEDs antiparallel and the phototransistors totem-poled. It needs very little line-side current since the phototransistor load will be a cmos port pin. Software deglitch maybe, if it's necessary.

That circuit lacks precision and has too much dead time, also aging and temperature issues with the LEDs..

A plagiarism

Hmmm- your circuit

How much dead time do you think it has? How much can the OP tolerate?

also aging

Please elaborate.

Not exactly the same... not totem pole outputs.

Yikes, 19 parts!

IBIS modeling courtesy of Yours Truly... at least of the TAOS chip inside ;-) ...Jim Thompson

I put a zero-crossing detector in my laminator hack, using a standard AC-input opto-isolator, a cap, and a resistor. It produces a signal that's high for about 50% of the time, when the AC voltage magnitude is highest. The trick here is, the MCU needs to measure the "off time" of the signal it gets, and set a timer to trigger half that time from the next pulse, for each pulse. That puts the timeout right at the zero crossing. Many MCUs can do this kind of measurement in hardware.

(detector is in the upper right corner)

(before anyone picks on my circuit - yes, I know it's a hack, but it's MY hack and it works for me ;)

THen you have not looked hard enough.

Jamie

It's not going to degrade much in your lifetime at, say, 0.25 mA average LED current. And the totem pole output need an approximate CTR (...scribbles mathematical equations furiously...) of zero.

Maybe because they never thought of it?

ac

so

pe

om

..

a...

Thank you for your reply Fred. The two EDN links you provided have some very good mains voltage, optical isolated output, zero crossing detector circuits. I will be using LTspice over the next few days to obtain a better understanding of how the circuits work. The Dego circuit is the circuit that initiated this hunt to see if it could be replaced by a single IC. Evidently there isn't sufficient demand to justify manufacture an IC for this application.

Howard

is

iac

e so

ype

com

...

ra...

Nah- that circuit is just a noise amplifier, you will have ultra-long storage times and a bunch of pulses when the off transistor finally comes out of saturation, with a lot current shootthrough in between. It looks good on paper but you will spend a lot of effort cleaning up that messy output. Also, it would not be unusual for something like that to take half an hour to finally settle down to steady state behavior.

Thank you for your reply Joerg As far as I can tell the HCPL-3700 is a better fit for an application that needs to know whether 207 Volts Ac or 240 volts AC is being applied to a circuit than for use as zero crossing detector.

Howard

Thank you for your reply John. Your suggestion looks like a nice low parts count approach. I will simulate the circuit to see how it compares with the other suggestions.

Howard

Thank you for your comments on accuracy Robert. However, the zero crossing detector will be used in application like D. J,. Delorie's laminator, phase fired power controller, where 1 to

Howard

How does an "off transistor come out of saturation" ? What does that mean?

Where do "ultra-long storage times" come from? Makes no sense.

with a lot current shootthrough in between.

All insane. There is no shoot-through; one transistor turns off, then the other turns on. Where is there a half hour of time constant? This will start working nicely in 1/2 of a line cycle.

I mentioned that the uP could deglitch it in software if necessary. Or if you don't want to do that, and are worried about glitches, add a tiny-logic Schmitt trigger thing. That works it up to 4 parts total.

The antiparallel coupler/totem pole is far better than an optocoupler with a pullup resistor, and much cooler besides.

They'll all be like that because, as you hinted in another post there is not enough market for a isolated zero-crosser chip. The question is, how precise do you need to detect the zero crossing? 1-2% as you have hinted elsewhere should be feasible with an optocoupler device.

If you are willing to install a tiny transformer it gets even easier: Run the other side into clipping diodes, via a resistor. Now place a window comparator (can be bought in a chip) around zero and watch its output.

While it's not impossible to make such a device, I don't think they exist. Consumer electronics (the main customer) is very cost sensitive, and this sort of thing would not likely be cost-effective.

I suggest that you should design something using a standard optoisolator that is safety-agency approved, along with a few inexpensive appropriately rated discrete parts, just like everyone else does.

type

com

What the hell are you talking about? How does high impedance loading - essentially no loading - slow down a phototransistor turning on? You don't understand this stuff.

The period of the AC line is 16 milliseconds!

That's crazy. When there's current in one of the LEDs, why would its phototransistor wait a half hour to turn on?

It's a totem pole because two NPN transistors are stacked, one above the other. They can't conduct simultaneously because only one LED can be driven at a time. The drive frequency is 60 Hz, for Pete's sake. I've used this totem pole opto circuit at 100 KHz, and it worked fine.

ftp://jjlarkin.lmi.net/Opto_Totem.JPG

I've also use it in linear mode, as the output stage of an amplifier that swings 400 volts p-p.

ftp://jjlarkin.lmi.net/HVamp.JPG

All this stuff works.

For accuracy, just use a capacitive dropper type power supply to power a comparator that drives an opto-isolator. Then you can get as accurate as you desire.

A straight opto-isolator, with no helper gain, is simply a "straddle-the-zero-crossing" indicator :-) ...Jim Thompson

that is

ing triac

But

ve come so

ill

this type

current

are

ology.com

rate?

kinda

ncy

CTR

tector...

.jpg

rrentdra...

I know how to read a datasheet and also understand transistor saturated operation better than you do apparently. You have a combination of two phenomena working to turn the detector transistor OFF when it comes out of saturation: recombination of excess base charge and collector current discharge. Since you have no collector current to speak of, your circuit must rely on recombination alone. My sampling of some popular Toshiba optos list storage times of 100us as typical with even as much as 4.7K loading to 24V! This means the transistor you will be trying to turn on will be loaded by the other transistor which is still a fully conducting saturated switch.

Give us some part numbers and your input drive resistor values.

This is a totem pole

Do you see that little branch there you have labeled 'FB'? Think that would make any difference compared to your open loop drive for a zero crossing app?