Flickering zone of dark activated switch and triggering sensitive gate TRIAC

[Willen]

Hi,
I am doing a solid state AC switch project triggered by dark activated switch. The dark activated switch part is designed by MikeMI, member here. The solid state part is from random search.

Maybe the dark switch is a nice design with hysteresis control. It looks like is has digital output instead of linear. However when I completed the project, it flickers between ON and OFF state (let's say during evening time).

Test 1 with filament lamp:
-turns off in light,
-turns on in dark,
-flickers in evening.
Test 2 with LED lamp:
-turns off in light but flashes twice in a second,
-turns on in dark,
-flickers in evening.

LED lamp flickers even in OFF state. It that because of "Sensitive type of TRIAC" and leaking current?
Both flickers in evening. Do it need to use comparator based circuit instead of the circuit to minimize/remove flickering zone?

 

[Pommie]

Does the lamp illuminate the sensor?
With LED lights the leakage current causes the flicker. A resistor can be used across the bulb to absorb the leakage current.

Mike.

 

[Willen]

Does the lamp illuminate the sensor?
With LED lights the leakage current causes the flicker. A resistor can be used across the bulb to absorb the leakage current.

Mike.

No, I have blocked the light of the lamp to the sensor. Ok, maybe placing a bleeder resistor across the LED supply will solve the OFF flickering.

Now problem is flickering zone in evening.

 

[gophert]

I thought sensitive gate trials could be driven directly by microcontroller.

also, you have way too much current into the gate (check datasheet). I think they only need 5mA (less) to latch on.
PS - are those the triacs I sent you some years ago?

 

[rjenkinsgb]

Try reducing R5, to give the circuit a bit more hysteresis?

Gophert - the MOC3021 is a zero-crossing driver, the resistors are purely transient protection as the triac should turn on at a very early, low voltage point in each half cycle.

 

[Willen]

Hi all,
I found the dark detector circuit designed by MikeMI is awesome! Just an idiot mistake is, there is a tiny red LED near the sensor making the 'evening flickers'. It is very dim and not focusing to the sensor, however its dim light affecting it. I covered the LED and worked well. Even better result with lower value of R5. I also increased 3.3k opto feeder to 30k, maybe the sensitive gate getting around 5 to 8mA now.

Remaining problem is, LED lamp still flickering (twice in a second) in off state. I used bleeder resistor of 330k across input filter high voltage electrolytic capacitor, used 82k across output to the series of LED. Not solved. Any other modification? Filament lamp does not flicker in off state. I would be happy to block leakage current from triac, instead modifying LED lamp.

PS - are those the triacs I sent you some years ago?

Yours are genuine things made in west, so kept them safe. I used BT134 used commonly here in fan dimmer and available 5 pcs in USD 1. China made, once bouht 3 and found shorted all of them. Once bought dimmer, found same shorted dimmer inside.

 

[Pommie]

I think I used a 47K 3W resistor across the LED bulb. I was using an SSR and the leakage current was 5mA which was enough to charge the power supply capacitor in the LED and make it flash at around once per second. I'm guessing this is your problem too so try different resistors.

Mike.

 

[Willen]

[/QUOTE]

I think I used a 47K 3W resistor across the LED bulb. I was using an SSR and the leakage current was 5mA which was enough to charge the power supply capacitor in the LED and make it flash at around once per second. I'm guessing this is your problem too so try different resistors.

Mike.

Hi,
I will do that too. But instead of modifying every LED lamp, it would be better to solve the triac leakage problem, isn't it if we can. There is a 47nF cap which is bypassing the triac. Which is maybe bypassing/leaking AC too. Maybe it is used here to supress EMF from inductive load. I am not going to drive inductive load, So can I remove this one to decrease/remove the leakage current?

 

[gophert]

Yes, that should be ok. And, I don't think you can call the cap a leak. AC power can transfer through a cap - it should be expected.

 

[rjenkinsgb]

There is a 47nF cap which is bypassing the triac.

That cap provides the initial gate current to latch the triac; it may not work reliably without it?

 

[Willen]

That cap provides the initial gate current to latch the triac; it may not work reliably without it?

Same as you said, when I removed the capacitor, I cannot turn on the triac. Replaced with 4.7nF, failed. Used 33nF, turned ON but leaked as before. Still the LED lamp is blinking. Still I want to solve the problem around triac side instead of lamp side.

I used a 330R resistor from gate to MT2 to pull down the gate leakage but it turned ON the triac. Tried different opto feeder resistor from 820 ohms to 30 k ohms with 4.7nF capacitor, and got nothing. Strange to me.

I have an experience, once I used another sensitive gate TRIAC to dim light. Dimmer was common found around internet. With a sensitive triac, I was not able to fully turn off the light with pot turn down. Then I changed BT136 (Non sensitive) then I was able to turn off the light with pot turn down.
Do I need to change BT136 instead of sensitive BT134 to reduce the leakage?

 

[rjenkinsgb]

As long as there is a capacitor (and or resistor) providing an AC path through the switch circuit, the LED lamp internal power circuit will be able to gradually charge then flash and discharge itself.

Just remove R8 and C1, so the opto and a low value resistor are the only things in the gate circuit.
Change R9 ro 100 Ohms, or also remove R7.

A resistor alone is a valid configuration according to the MOC3021 datasheet - that shows just 180 Ohms total in series feeding the gate,.
See figure 9 in this datasheet:

https://www.onsemi.com/pub/Collateral/MOC3023M-D.PDF

That should hopefully get rid of the leakage?

 

[gophert]

Change the 220k resistor in the hysteresis loop to 100k. This "should" take care of the dawn/dusk blinking. If it is not easy to remove the 220k, just add a second 220k in parallel with the first to make 110k. If you still have blinking issues, bring that resistance down to 68k or so.

 

[ChrisP58]

This application circuit shows a resistor from the gate to the lower MT, which is different than what's in your schematic. You do have two resistors, R7 and R9, but they're effectively in series with each other. Should R9 be reconnected as shown here?

https://optoelectronics.liteon.com/upload/download/DS-70-99-0019/MOC302X series 201606.pdf

 

[Willen]

It worked!
As all of your suggestion, I removed capacitor there and just used 220R opto feeder, worked. To protect sensitive gate TRIAC from heavy firing, I increased feeder resistor value till 30k and still worked! Again to protect sensitive gate I used a 680R pulldown resistor and still worked. Thank you for the more alternatives! I also reduced the hysteresis resistor down to around 100k to improve flicker zone.

Filament lamp working perfectly. LED lamp is also working without leakage flickers. I just have noticed chattering/firing around switch/socket during LED lamp ON and OFF, I guessed of its SMPS design. All of SMPS makes such powerful spark in socket, I don't know why!

I used to think I can place the load in MT1 or MT2 anywhere. But it did not work. Just worked when wired as in the schematic.

 

[ChrisP58]

. . .
All of SMPS makes such powerful spark in socket, I don't know why!
. . .

Glad to hear of your success.

The spark when connected or turned on is because of the inrush current into the front end filter capacitors in the power supply. It could be 10 or ?? times what the operating current is. But only for a very brief moment.

 

[Willen]

It is so fun, working with basic thing!
After solving all the problem, I modifed the circuit which uses tiny 3V solar cell, scrapped from calculators. And it works! I added a pulldown transistor!