Scope trace of ZC?

[djulien]

Does anyone have a scope trace of an H11A11-based ZC signal from one of the dimming controllers? I'm wanting to see the actual pulse width, and the shape of the signal edges.

I was trying to calculate the ZC pulse width by doing some calculations and got a large number (something like 25% of the AC cycle), which seems too large. I'm also trying to get a PIC to tell me that info, but if someone already had a scope trace that would be less work. :)

thanks

don

[P. Short]

The pullup resistor on the output of the H11AA1 is usually 27K, so you need about 160 μA to pull the output down to 1V or so. The transfer ratio of that opto is nominally 0.2, so that translates to about 800 μA on the input side of the opto to obtain that current on the output side. The input resistors to the opto total 30K (for 110VAC), so the AC voltage must be more than 24V (30K * 800 μA).

Since 24V is a lot less than 170V (the peak voltage for 120VAC RMS), we can use a linear approximation (instead of trigonometry). The approximation comes out to V=170 * (2πft), where f is the frequency and t is time in seconds, or t = 24/(170 * 2 * π * 60), or .38 mS. You need to double this to calculate the duration of the ZC signal, because this 0.38 mS is the time from the exact zero-crossing until the ZC signal goes low, so the duration of the ZC should be around 800 μS.

This is an approximation, of course. The circuit is not all that precise, and my calculations ignored the voltage drop of the opto input diodes.

FWIW...

[RavingLunatic]

Don,

Now that you have the hard facts behind the ZC calculation from Phil, here are some pics to visualize it.

(This is off of a Renard SS8)

[djulien]

Thanks Phil and Raving!

For comparison purposes, I'm trying to adjust Phil's calculations to the Renard SS8 and then compare them to the scope trace from Raving so I have a better understanding on how these can apply in other cases ... (please bear with me).

According to the Renard SS8 circuit diagram, the H11AA1 is connected to the transformer output. The 3FS-312 transformer description says it's a 6.3V transformer, which would give it a peak output of ~ 9V. Since the resistors on the input side of the H11AA1 in the Renard SS8 are 750 instead of 15K, then it should be triggered when it saw 1.2V rather than 24V if using 120 VAC line voltage and 15K resistors. Is that correct?

If yes ... Then I can take the 1.2V and calculate when that would occur on a perfectly-shaped sine curve, which I realize doesn't quite match with an AC dimming environment. Using Phil's approximation (since 1.2V relative to 9V is small like 24V was to 170V), it would be 1.2/(9 * 2 * π * 60) =~ .35 msec, as opposed to .38 msec for the calculation at 120VAC. So it looks like the ZC scope trace for 120VAC would be slightly wider than the one for the Renard SS8 - is that correct?


The transfer ratio of that opto is nominally 0.2

Is that the CTRce symbol on the Fairchild spec sheet?


and my calculations ignored the voltage drop of the opto input diodes.

The Fairchild spec sheet says the emitter forward voltage is 1.17V typical, and 1.5V max, so in the case of the Renard SS8, should I calculate the point in the AC curve when it reaches 1.2 + 1.17V ~= 2.4V, rather than just 1.2V? If I use 2.4V rather than 1.2V, that would make the pulse width double what it appears to be on the scope trace, but if I just use the 1.2V then it matches okay. Does that mean I can always ignore the H11AA1 emitter voltage drop in the calculations?

don

[dirknerkle]

Don, RL and Phil,

You don't know how much stoobs like me appreciate the academic and sensible way you approach things. It's just delightful to sit back, read the postings, questions, replies, and try to absorb it. To be able to witness some of the logic and consideration that's been put into designing DIYC gear is tremendously gratifying and helpful.

Thanks.

-dave

[P. Short]

That spec for the diode forward drop is for 10 mA of current through the input diodes, but the current that we are concerned with is for 800 μA of current. The voltage across the diodes should be smaller at that lower current level, although the CTR is also a lot lower. The data sheet for Vf vs If is unfortunately not usable at those low current values (the graph is unreadable for currents at the sub-mA level).

[djulien]

Don, RL and Phil,

You don't know how much stoobs like me appreciate the academic and sensible way you approach things. It's just delightful to sit back, read the postings, questions, replies, and try to absorb it. To be able to witness some of the logic and consideration that's been put into designing DIYC gear is tremendously gratifying and helpful.

(I didn't help with the design, so I shouldn't really be included in that comment - I was just trying to learn some of the "secrets from the masters" so I could apply it to other scenarios).

don