Optotriac resistor selection

Ah, wait, I found down at the bottom of the datasheet the MOC3010 needs max 15mA, min 8 mA to latch the output at 3 volts. Damn, that's a lot for a battery powered app. Time to find a better part...

If you know the zero crossing locations you can drive it with a pulse per cycle, otherwise you can drive it with a pulse train (say 10% duty cycle) to reduce the power consumption.

I seem to recall ~5mA being about the minimum for that class of part, and one would have to be reckless indeed to drive it at the minimum given temperature and aging changes, which are mostly/all negative.

The Vf for the LED vs. current should also be in the datasheet.

--sp

How about an SSR? They often turn on at 1 mA or less.

Or use a bridge rectifier and a mosfet, which would need picowatts of gate drive.

Or a latching relay!

Thanks, I'll try driving it with a pulse train from the uP. Figuring out the zero-crossings is a cool idea, but would require sampling the HV into the uP's ADC, an exercise I don't have time for at the moment...:)

One more question, I know this setup will work fine for low-speed switching, but can the pulse train from the uP be adapted to create a dimming or "fading" effect for the strip as well?

All possibilities, but as I have about 24 hours to complete this one-off, it'll have to wait for Version 2.

What's the sweet spot on el-cheapo SSRs to switch 100V RMS at a few 10s of mA?

With infinite gain?

Not without zero crossing information.

--sp

Gotcha, I guess it would work similarly to a triac incandescent lamp dimmer that reduces the effective RMS drive voltage by "cutting in" only after some delay from the zero-crossing. I wasn't sure if it would work the same with a capacitive load like an EL strip.

The MOC3010 needs 15ma to guarantee a reliable trigger. The diode's forward voltage is 1.5V max. That means the you need to drop 3.7 V - 1.5V = 2.2V by the resistor. So, the resistor's value needs to be 2.2V / 15ma = 147 ohms. A 5% value of

Dave M

You'd have to prowl Digikey for that. We use a couple of 300-volt SSRs, but they're not cheap.

There are small sensitive-gate triacs around, probably very cheap.

Approximately.

The bridge rectifier + mosfet thing could probably do PWM dimming without knowing the power phase. That could get interesting.

The cheapest optotriac at Mouser looks like a good candidate:

The diode trigger current is 15mA typ, 30mA max. Go for a K3023 for 3mA / 5mA max sensitivity.

There are plenty of MOSFETs that can handle that - presumably your EL rope is AC, so the MOSFET will need to be enclosed by a bridge rectifier. That probably means isolation will be unavoidable, a basic transistor output opto would work, but you can also get MOSFET output as well.

Now I think about it - there's a dual MOSFET output optocoupler for telecoms use that can handle AC on its output.

Its non latching - but the triac output only latches for the remainder of any half cycle you trigger it on.

jabut... the trigger only needs to be a few microseconds per period?

.-- inverter --- rope --||-------. ... mosfet could be BJT. | `-|| with antiparallel diode --- || - .-|`--- from microcontroller | --- - it may not be possible to go fully dark.

A MOC3012 is more sensitive but please say what is the frequency from the inverter? If much over 500Hz then any photo-triac will struggle to commutate off after triggering. Check datasheet commutating dV/dt spec.

piglet

I'll check it with my counter when I get a chance, but it's low enough that if I hold my ear close I can hear the inverter whining away when it's loaded.

The 3010 seems to be switching it OK when driven with short pulses through a ~220 ohm resistor from 3.7V.

Silly question: How long does it take for the EL wire to light up when you turn the inverter on?

If the timing is acceptable, how about just switching the inverter power on and off? Heck, if you had a thin piece of double-copper FR4 you could solder wires to it and squeeze it between one of the battery terminals and the holder's contact to get control of the inverter's power.

Jes' curious...

Frank McKenney

It's pretty much instant, but I'm using several lengths of different colors and would like to sequence them independently. The brick is hefty enough to handle several, so I'd prefer to switch the strands rather than use an inverter per.