High speed control for AC switching.

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
Im trying to control a 220VAC , ~10uA load with a signal from a uc.

I need it to switch faster than 100us either way.

I've tried with an ssr ,and an optoisolator (moc3041 & bt136), and they are both too slow, especially when going from on to off. The opto was faster than the ssr, but still too slow :(.

Any suggestions are highly appreciated.
 

Ian0

Joined Aug 7, 2020
9,677
Both BT136 and SSR are triacs, when switched on, they remain on until the end of the mains cycle.
Please post a diagram of what you are trying to achieve.
 

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
So, what im trying to do is, get the rising time and falling time of a signal that is made by the load.
1618584755684.pngThis is from the datasheet. Vcc 5V


1618583734223.pngThis is without the relay/opto(manual switching): t_r=~171us and t_f=~66ms

1618583936927.pngThis is with the relay controlling the load. Much slower, and i think like you said, it waits for the end of the cycle to turn off.

What can i use that is in the ns range switching time that won't affect the rising and falling times?
 

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
Sorry, its 50-100VAC @50-60Hz. The signal is from a photodiode, and the load is a switchable glass or smartglass. Im trying to automate the measurement of how fast the glass changes states.
 

Ian0

Joined Aug 7, 2020
9,677
How are you generating your trigger signal? Is it synchronised to the mains zero-crossing?
As @MaxHeadRoom says, the MOC3041 is a zero-crossing opto-triac. It will only switch the triac on at the beginning of the mains cycle. A triac will only switch off at the end of the cycle.
You will get faster switching with a MOSFET, but you have the problem of isolating the trigger signal from the mains.
 

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
The trigger is a pull down from the uc. I actually want the zero-crossing for the on switch, but not for the off, because i think it would add more time to the measurement for t_r.
 

Ian0

Joined Aug 7, 2020
9,677
Won't the switching times depend on the instantaneous mains voltage?
How are you synchronising your microprocessor to the mains?
 

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
Yes, the switching times depend on the instant v.
Im not synching, i just analyze when the pdiode signal isn't changing anymore, i send the trigger.


btw, this is how the glass behaves(top) with the mains(bottom)
|
\/
1618590861307.png
 

Ian0

Joined Aug 7, 2020
9,677
Does the glass have to be driven with AC? Will it work with DC? It looks as though it behaves identically with either positive or negative voltage.
 

Thread Starter

manonpajamas

Joined Apr 15, 2021
12
Yes, it needs to be ac.
1618597467967.pngThis is a crude representation of how they work. The green bars are the electrodes. In a way its similar to a cap.
 

BobaMosfet

Joined Jul 1, 2009
2,110
Im trying to control a 220VAC , ~10uA load with a signal from a uc.

I need it to switch faster than 100us either way.

I've tried with an ssr ,and an optoisolator (moc3041 & bt136), and they are both too slow, especially when going from on to off. The opto was faster than the ssr, but still too slow :(.

Any suggestions are highly appreciated.
You need a zero-cross detection circuit. Elsewhere in this forum, the topic has been covered, and a fantastic zero-cross detector, with isolation schematic was provided. Temperature independent, it neatly provides the middle of the pulse.

Not sure where you get a 100uS requirement from- your signal is 50-60Hz. That's SLOW (an eternity) in electronics.

Look at Post #19 in this thread:

https://forum.allaboutcircuits.com/threads/zero-crossing-detector.168250/
 

BobaMosfet

Joined Jul 1, 2009
2,110
Its not from the mains, its from the speed of change of the glass, which has a off-on time of around100us and on-off around 100ms.
Understood, thx. One thing at a time- reliable zero-cross detect (which can also be used to tell if MAINs is even generating a signal), and then a method to control it via the window of time you need to switch in (MosFET will do this).
 

Ian0

Joined Aug 7, 2020
9,677
Apologies for the pencil drawing, but I don't have an opto-triac in my copy of Spice.

Your load is such that it below the holding current of any opto-triac, so it's a bit of guesswork as to how it will behave, see: http://www.farnell.com/datasheets/97984.pdf
Use the MOC3021, because it does not have a zero crossing detector. It should switch on in about 1us.
It won't latch, being below the holding current (100uA) so should switch off again when the LED current switches off; but it might switch off slowly.
There's no need for a main triac at the current you are using, and certainly no need for a snubber as that just slows the switching speed down.
If you add a resistor across the load, it will make the triac latch. It will turn off at the end of the mains cycle.
If you want a rapid switch-off use a series resistor and a parallel opto-triac. The panel will be on until you trigger the opto-triac, at which point it will switch off (in about 1us) and remain off until the end of the cycle. You can then measure the switch-off time.BE86C11C-5D5B-43A3-8497-F1428602B291.jpeg
 
Top