[SOLVED] Triac light dimmer - triac not opening

[bmandl]

Greetings!
I have a home project in which the microcontroller light dimmer is involved. I built custom PCB with AVR and triac controlled dimmer, but it doesn't work. When I was in testing phase, with protoboard circuit, everything worked perfectly, but now on PCB it does not. I made PCB exactly as I connected everything on protoboard. I used BT137 triac and MOC3021 as optocoupler with diac output. I meassured voltages in circuit with oscilloscope and I get about 1.5V drop on current limiting resistor for gate and about 1.5V drop on output diac, but I get almost full 220V drop on gate of triac. This is kind of wierd. Triac just won't open. I have zero cross detector and I am sending right signals to optocoupler input, so this is not the problem.
I don't understand why circuit worked, exactly the same one on protoboard, but it does not now on PCB. I already replaced triac in case it was fault, but it was not.

My trigger circuit:

 

[Audioguru]

A door opens or closes. A triac turns on or turns off.
Does the triac not turn on?
Maybe you have the main terminals of the triac reversed.

 

[bmandl]

TriAC is alternating device right? It conducts both ways. I really don't understand your suggestion. No disrespect, but do I sense a bit of sarcasm in your answer?

PS.: Please, don't resent me, but English is not my first language, so it can happen, that I miss a meaning of a fraze sometimes, so with triac open or close, I acctually ment turned on or off, like you already indicated.

 

[betwixt]

The most likely reason for failure is lack of voltage at the input. The LED side has a recommended current of 10mA giving a Vf of 1.15V, the transistor needs an additional ~0.7V Vbe so the base needs to be elevated to at least 1.85V for the opto-coupler to reliably turn the triac on. Can your AVR circuit do that, considering the 10K series resistor in the transistor base pin?

It might be better to put the opto-coupler input and series resistor in the transistor collector and ground the emitter pin. It will reverse the driving logic which can be fixed in software but also reduce the voltage requirement to turn the switch on.

Brian.

 

[KlausST]

Hi,

Open/close:
The problem is that not clear what you mean when you say "a triac is not opening".
When a door is open, then you can walk through it..
A valve is open when water can go through it..
Is a switch open when current can flow through it?
Is a triac open when current can go through it?

I find it difficult to answer, maybe because english isn't my native language...

Klaus

 

[Audioguru]

When I used a triac more than 50 years ago it conducted better when one of its main terminals was at the supply voltage. Maybe your gate current is too low for the triac to conduct when wired both ways.

 

[ard]

HI,

If you mean that the triac comes on, and *stays* on, continuously lighting the load at full intensity, try a pull-down resistor between your opto pin 6 and the triac pin connected to mains neutral in your schematic.

Also, do you really need a transistor to drive the opto input? Your AVR should be perfectly capable of driving the opto LED directly, thru a 470 ohm resistor.

Check out this thread, its quite insightful. https://www.edaboard.com/threads/181497/

Regards!

 

[bmandl]

ard, my problem is not understanding how triac works. As I said, the circuit worked fine on protoboard with exactly same components. Problem is, that I am getting some strange signals on AC side of a circuit.
I will try to lower 10k resistor on base of transistor.
Thank you for your answers.

 

[bmandl]

I just noticed one interesting thing. When I touch gate of triac with oscilloscope probe, bulb flicker. What could that possibly mean??

 

[betwixt]

Most likely that you are not using an isolation transformer with your oscilloscope or you have a ground connection to it.
It does no harm, but may be beneficial, to add a resistor between the triac gate and 'N' to discharge leakage currents through the opto-coupler.

Brian.

 

[bmandl]

I am testing my circuit with isolation variak, because I have FID switches in house, so that's not the case.

- - - Updated - - -

But still, this is the way to turn triac on. Something most not be right. Why triac does not turns on. I swear that I have exact same circuit as I have it on breadboard. I even tryed to take triac from breadboard and test it here and no difference. Everything else is same.

 

[j33pn]

I agree with Betwixt that you should be switching the GND side of the opto, not the Vdd side.

See page 4: https://www.ti.com/lit/ds/symlink/moc3021.pdf

 

[bmandl]

This is not the option anymore. I already have my PCB and I assembled everything (all the TSSOP stuff and I really don't want to do that again). Besides, why would that be a problem just suddenly on the circuit, that is exactly the same as the other one, except that one is on PCB and the other is on protoboard? The opto side is ok, believe me, I was using this circuit for months without problems. Only thing that is differend here is microcontroller, which is not the case of the problem (please don't ask me, I know it is not...).
So, to put it all together:
- LED side of optocoupler is ok (exactly the same as the working one)
- the other side of the optocoupler is ok (exactly the same as the working one)
- Bulb is ok
- Components are ok (replaced each of them already)
- Microcontroller is ok (test program with 100% brightnes works - i meassured output with oscilloscope. I am getting nice 5V signal every start of half periode)

Now, I can't think of anything more. Nothing is reasonable for me now, why this circuit doesn't work. So, do you have any suggestions of what could really cause triac, not to turn on. What could be a major issue, not just like that maybe it's reversed (exactly the same triac - phisicaly, in the other circuit works exactly the same polarized and here does not...I don't think so...)?

 

[KlausST]

Hi,

You say it's all the same..
Except the microcontroller, but you promise it is not the problem..

Now lean back and ask yourself: how can we help you? What do you expect?

****

I'd say there must be a difference.

Either a bad solder joint, two accidentally shorted pins, a defective part, a wrong resistor value...

My hint:
Check the signal coming from microcontroller, through transistor, through resistor to MOC.
Is there really current, enough voltage drop across the resistor? (I'd rather use a PNP instead the NPN emitter flower with it's large voltage drop)
Then MOC output, through resistor to triac. Mains voltage, load connection triac connections..

There is no way around this. It won't heal itself..

Klaus

 

[DanyR]

hi bmandl,
Perhaps some small tests you can easily execute:
-short the emitter (3) and collector (1) of the driving transistor and the triac should always be on (lamp on).
- short the base (2) of the driving transistor to ground and the triac should always be turned off (lamp dark).
- short pins 4 and 6 of the MOC and the triac should always be turned on (lamp on)

 

[FvM]

Good you finally revealed that "not opening" means not turning on, the usual open/close switch terminology would suggest the opposite.

According to datasheet, the worst case MOC3021 trigger current is as high as 15 mA, 8 mA is a typical value. 1.5 V across 330 ohm results in only 4.5 mA, even with 5V supply you never achieve 15 mA. So you have at least one possible problem cause.

Suggestion: Use ohms law when designing a circuit.

 

[bmandl]

I apologize to everyone. It was not all the same. Triac main terminals were backwards and I didn't even think about this as a cause. I always thought that it doesn't matter how you turn triac mains around. @Audioguru, special thanks to you, that you mention that, because I wouldn't dig around this if you wouldn't mention it. And I apologize to you to, that I didn't believe you.
How I found this out? I measured conductivity from triac gate to each of main terminals. One was almost shorted and it was the wrong one on bulb. So I modificated PCB with knife and now it works. Too bad thou that PCB looks modificated now. I will use another triac if I will build one more PCB (I have 6 more ).

 

[Audioguru]

The gate works with both main terminals but one terminal is more sensitive than the other. So if you increase the gate current then the triac will work fine when the terminals are reversed.

 

[FvM]

The gate works with both main terminals but one terminal is more sensitive than the other. So if you increase the gate current then the triac will work fine when the terminals are reversed.

No. The optotriac has to be connected between T2 and G. Connecting it between T1 and G results in no operation.

As reported, there's an internal low ohmic resistance between G and T1 to prevent self triggering by leakage currents. The triac is triggered by a voltage of typically 0.7 V between G and T1.

 

[Audioguru]

No. The optotriac has to be connected between T2 and G. Connecting it between T1 and G results in no operation.

As reported, there's an internal low ohmic resistance between G and T1 to prevent self triggering by leakage currents. The triac is triggered by a voltage of typically 0.7 V between G and T1.

Maybe you are thinking about an SCR used for DC, not a Triac used for AC.
The Philips datasheet of the BT137 shows an ordinary Triac without "an internal low ohmic resistance between G and T1" and the spec's show that all four of its quadrants work but one of its four quadrants has half the sensitivity of the others. Its maximum trigger current for all of them to work wired any old way is 100mA!

The four quadrants:
Ι + Mode = MT2 current positive (+ve), Gate current positive (+ve)
Ι – Mode = MT2 current positive (+ve), Gate current negative (-ve)
ΙΙΙ + Mode = MT2 current negative (-ve), Gate current positive (+ve)
ΙΙΙ – Mode = MT2 current negative (-ve), Gate current negative (-ve)