Driving a Triac Directly

A

ACharnley

Member

• Mar 26, 2018

• #1

Hi,

I haven't played with triacs before so want to ensure I've got this right.

Requirement

I need to apply an over-voltage shunt to collapse the input voltage until the AC cycle returns to zero. A triac seems suitable.

Current: constant 500mA.

Chip

I'm looking at https://uk.rs-online.com/web/p/triacs/7958704/ which has a low ~3mA @ 1V turn-on.

Am I right in saying;

1. the triac is latched like a transistor so a gate resistor is required, probably to let 10mA flow as a good margin
2. the heat dissipated will be the voltage drop * current, so 1.2V*0.5A = 0.6W
3. there's no possibility of voltage flowing to the gate (I can't have the mcu destroyed).

Many thanks,

Andrew

 

ronsimpson

Well-Known Member

Most Helpful Member

• Mar 26, 2018

• #2

I think you are OK.
You know the triac will latch on until the voltage/current crosses zero.

 

A

ACharnley

Member

• Mar 26, 2018

• #3

I do, thanks Ron.

 

alec_t

Well-Known Member

Most Helpful Member

• Mar 26, 2018

• #4

ACharnley said:

possibility of voltage flowing to the gate

Click to expand...

There can be slight reverse leakage current through the gate, at under 1V, via gate capacitance, if there is a sudden change in the voltage applied to the main terminals.

 

dr pepper

Well-Known Member

Most Helpful Member

• Mar 27, 2018

• #5

Triacs can have different trigger currents depending on the mains quadrant, so you need to ensure the gate current is high enough to trigger the triac worst case.
You also need the micro to have one power leg referenced to the mains which has safety issues, isolated with an opto is better.

 

A

ACharnley

Member

• Mar 30, 2018

• #6

No mains here, and have chosen a triac to meet those current requirements. Thanks!

 

MaxHeadRoom78

Well-Known Member

Most Helpful Member

• Mar 30, 2018

• #7

If using a MCU typically a zero cross opto is used.
See Fairchild AN-3006.
Max.

 

A

ACharnley

Member

• Mar 30, 2018

• #8

I take it it's to disconnect the high voltage for safety? Without it there's still no chance of that voltage going to the MCU pin?

 

MaxHeadRoom78

Well-Known Member

Most Helpful Member

• Mar 30, 2018

• #9

ACharnley said:

I take it it's to disconnect the high voltage for safety? Without it there's still no chance of that voltage going to the MCU pin?

Click to expand...

Something called Galvanic Isolation.
Max.

 

A

ACharnley

Member

• Mar 30, 2018

• #10

MaxHeadRoom78 said:

Something called Galvanic Isolation.
Max.

Click to expand...

Hopefully given I'm playing with only 36V AC it won't present itself to be an issue. I realise I forgot to mention that in the initial post.

 

MaxHeadRoom78

Well-Known Member

Most Helpful Member

• Mar 30, 2018

• #11

You still preferably require isolation if dealing with MCU control.
Max.

 

S

schmitt trigger

Well-Known Member

Most Helpful Member

• Mar 30, 2018

• #12

By Murphy's law, if something can go wrong, it will.
So without an optoisolator, if the Triac fails for whatever the reason, your MCU could become collateral damage.

In the end, is like driving without a spare tire. As long as you don't get a flat, you can drive many thousands of kilometers and don't miss a beat.
But if you get a flat on a lonely stretch of a road and the night is coming, well, you would really wish that you had one.

 

A

ACharnley

Member

• Mar 31, 2018

• #13

If the triac fails many other components would fail alongside the mcu. The only benefit from using the mcu is variable voltage control, which in itself is a bit pointless, and which could be a liability if the MCU crashes for some reason (i'm prototyping the system to discover reliability). I could use the old zener/resistor route instead.

One thing I didn't verify, a positive current to the gate will turn it on for both AC cycles?