Options for driving triac

[bruce varley]

Hi, I need to drive triacs from CMOS, the reliable gate current is somewhat
above what the gates can source. There are two obvious options using
bipolars:

OPTION 1. Emitter follower, base to gate output, collector to +Vcc, emitter
via a suitable resistor to triac gate.

OPTION 2. Q1 (NPN) as a switch, with resistor divider from gate output to
base, emitter to gnd, collector resistor divider driving a PNP current
source from Vcc down to the triac. Current limiting resistor between PNP
collector and triac gate. Esentially a complementary switch pair.

It intuitively feels to me that the second option may give better
'protection' to the sensitive CMOS, but in theory Option 1, which uses a lot
less bits, should be OK. I do require a solution that doesn't significantly
degrade MTBF (stated vaguely, I know).

Is there any objective reason why one would go option 2? TIA

 

[Spehro Pefhany]

Hi, I need to drive triacs from CMOS, the reliable gate current is somewhat
above what the gates can source. There are two obvious options using
bipolars:

OPTION 1. Emitter follower, base to gate output, collector to +Vcc, emitter
via a suitable resistor to triac gate.

OPTION 2. Q1 (NPN) as a switch, with resistor divider from gate output to
base, emitter to gnd, collector resistor divider driving a PNP current
source from Vcc down to the triac. Current limiting resistor between PNP
collector and triac gate. Esentially a complementary switch pair.

It intuitively feels to me that the second option may give better
'protection' to the sensitive CMOS, but in theory Option 1, which uses a lot
less bits, should be OK. I do require a solution that doesn't significantly
degrade MTBF (stated vaguely, I know).

Is there any objective reason why one would go option 2? TIA

You'd be better off sucking current out of the gate.

If you tie Vdd to MT1 and use an NPN (collector to gate through
resistor, emitter to Vss, base resistor to input...



Best regards,
Spehro Pefhany

 

[legg]

Hi, I need to drive triacs from CMOS, the reliable gate current is somewhat
above what the gates can source. There are two obvious options using
bipolars:

OPTION 1. Emitter follower, base to gate output, collector to +Vcc, emitter
via a suitable resistor to triac gate.

OPTION 2. Q1 (NPN) as a switch, with resistor divider from gate output to
base, emitter to gnd, collector resistor divider driving a PNP current
source from Vcc down to the triac. Current limiting resistor between PNP
collector and triac gate. Esentially a complementary switch pair.

It intuitively feels to me that the second option may give better
'protection' to the sensitive CMOS, but in theory Option 1, which uses a lot
less bits, should be OK. I do require a solution that doesn't significantly
degrade MTBF (stated vaguely, I know).

Is there any objective reason why one would go option 2? TIA

There aren't too many practical situations where the triac and control
circuit are safely DC-coupled.

Driver matching and islolation are fairly easily obtained using pulse
transformers, with little or no circuit mtbf degradation.

As triac conduction is most reliably maintained using pulse trains, a
low power flyback configuration is an obvious choice for low-frequency
control - pulse trains being gated by the controler.

RL

 

[Pooh Bear]

Hi, I need to drive triacs from CMOS, the reliable gate current is somewhat
above what the gates can source. There are two obvious options using
bipolars:

OPTION 1. Emitter follower, base to gate output, collector to +Vcc, emitter
via a suitable resistor to triac gate.

OPTION 2. Q1 (NPN) as a switch, with resistor divider from gate output to
base, emitter to gnd, collector resistor divider driving a PNP current
source from Vcc down to the triac. Current limiting resistor between PNP
collector and triac gate. Esentially a complementary switch pair.

It intuitively feels to me that the second option may give better
'protection' to the sensitive CMOS, but in theory Option 1, which uses a lot
less bits, should be OK. I do require a solution that doesn't significantly
degrade MTBF (stated vaguely, I know).

Is there any objective reason why one would go option 2? TIA

Have you considered paralleling some CMOS gates to provide more drive current ?
I've never done this myself but see no practical reason why it can't be done. I
have a feeling the technique was mentioned in application notes of old.

Graham

 

[Spehro Pefhany]

Have you considered paralleling some CMOS gates to provide more drive current ?
I've never done this myself but see no practical reason why it can't be done. I
have a feeling the technique was mentioned in application notes of old.

Graham

Enough voltage and current comes out of the triac gate to potentially
latch up the CMOS chip under some oddball conditions. I think the
pennies for the transistor are very well spent, in fact for a real
product I'd add more protection except for the cheapest applications.


Best regards,
Spehro Pefhany

 

[Bob Monsen]

Enough voltage and current comes out of the triac gate to potentially
latch up the CMOS chip under some oddball conditions. I think the
pennies for the transistor are very well spent, in fact for a real
product I'd add more protection except for the cheapest applications.

There are also those nice opto-coupled triac triggers, like the MOC3021
and it's bretheren. However, I'm guessing a standard cmos output won't
be able to generate enough current to trigger them (they require
something like 10mA, as I recall). Way more expensive than the
transistor, but good isolation.

--
Regards,
Bob Monsen

If a little knowledge is dangerous, where is the man who has
so much as to be out of danger?
Thomas Henry Huxley, 1877

 

[[email protected]]

I've always liked to use an opto isolator. They are cheap and I get to
sleep at night.

Dick Hagerman

 

[Rich Grise]

There are also those nice opto-coupled triac triggers, like the MOC3021
and it's bretheren. However, I'm guessing a standard cmos output won't
be able to generate enough current to trigger them (they require
something like 10mA, as I recall). Way more expensive than the
transistor, but good isolation.

Be careful with those MOC optos if you're looking for phase control -
the 3030's, that I know of, have a zero-crossing detector built-in,
so they switch at V=0.

But I was going to suggest optoisolators of some kind, yes.

Cheers!
Rich

 

[Spehro Pefhany]

There are also those nice opto-coupled triac triggers, like the MOC3021
and it's bretheren. However, I'm guessing a standard cmos output won't
be able to generate enough current to trigger them (they require
something like 10mA, as I recall). Way more expensive than the
transistor, but good isolation.

Of course if your power supply is non-isolated, the isolation isn't
all that valuable. Transistors don't age badly either. But, horses for
courses, the isolation can be very useful, and is rather inexpensive
since the AC line supplies the current for the triac gate.


Best regards,
Spehro Pefhany

 

[Bob Monsen]

Of course if your power supply is non-isolated, the isolation isn't
all that valuable. Transistors don't age badly either. But, horses for
courses, the isolation can be very useful, and is rather inexpensive
since the AC line supplies the current for the triac gate.

Another thing is that they trigger in both quadrants with a positive
pulse. I seem to recall that some triacs take more current when
triggering on the negative side with a positive pulse, or something like
that.

The OP may not know that these are actually triacs themselves, which are
turned on by the optical connection. Thus, the positive control pulse
generates a negative pulse into the actual triac, which may make it
trigger more reliably.

Also, the zero-crossing detector variants that Rich mentions are useful
if you are using them for a switch, because it only triggers when the AC
voltage is near zero (even if you power the LED before that), thus
preventing a big transient current if you trigger it at the top of the
waveform into lots of capacitance. If you are using it for phase
control, use the ones that don't have zero-crossing detection.

--
Regards,
Bob Monsen

If a little knowledge is dangerous, where is the man who has
so much as to be out of danger?
Thomas Henry Huxley, 1877