SCR switching 120 VAC

[batman]

Is there a way to use an SCR to turn on a 120 VAC light fixture? thanks

 

[john1]

Hi batman,

Normally a triac would be used for AC,
but if you must use an SCR then you can put it in
a diode bridge.

This makes other problems,
the control to its gate is not referenced to a
stable point.

This can be overcome by having all the circuitwork
associated with the gate completely independent of
any thing else,

Or by having a small dedicated transformer,

Or by using an opto isolator.

Best of luck with it,
John

 

[batman]

I dont understand how this schematic would work, because whe the AC supply reaches negative, wouldnt the SCR commutate?
thanks

 

[john1]

Hi batman,

Hmm...

If you replace the SCR in the diagram above,
with a short piece of wire ...

Then you can see that whichever way up the supply is,
the current will always flow the same way,
in the short piece of wire in place of the SCR.

As the current here is one-directional an SCR can be
used to control it.

Will it commutate?
Dunno.
It might.
Wots commutate?

If you mean the voltages on the SCR would vary with
the supply, then yes.
Thats why you would have to isolate the input to
the gate.

Regards, John

 

[Sebi]

Many years ago all dimmers contains same solution with UJT trigger circuit.SCR comutate on all zerocross, with trigger impulse delay You can control the AC voltage effective value. In static application (as switch) apply DC voltage between cathode and gate. It will comutate on zerocross, but fired always immediately when the voltage rising about 2.5V after zerocross.

 

[batman]

john1, in your schematic, will the voltage across the scr ever reach zero? If it does, the scr will shut off. this is commutation. i dont want it to shut off.
thanks

 

[john1]

Hi batman,

I am very sorry, i had thought you wanted to control the
lamp.
I have re-read your first post, and you only want to turn it on.
Yes you can just turn it on, and make it stay on if thats what
you want.

As you say, if its forward current drops below its hold-on value
the SCR will stop conducting, until triggered again.

you could do this by having a continuous current through its gate,
Or by making sure that the SCR carries on passing some current
during the zero portions of the supply.

Maybe a small capacitor across the SCR, to effectively 'carry'
it past the zero portions of the supply, or the gate could be fed
from a DC source, so as the SCR is continually in the triggered
state.

I am guessing you want to do this by switching the gate on or
off and allowing the SCR to take the load, the lamp ?

Simple circuit coming ... give me a little while ...
unless sebi wants to draw one ... ?

 

[john1]

Hi batman, Sebi,

Heres a simple circuit with a switch on the SCR gate,
i imagine sebi would put forward a similar one.

The values are only guessed, i would start with them
and see how it goes.

I would be aiming for about 10 volts or a bit more on
the 5 mfd. It doesnt have to hold much, or for long, it
doesnt want to be so big as to cause a delay when switching
off. If it does make a small delay when switching off,
just reduce the value a bit till its ok.

Sebi, if you feel these values are within reason or
maybe too far out, please say so.
They are only guessed.

I imagine 3k would work the gate ok, they are pretty high
impedance, but some are lower than others.

30k at 120 volts sounds like about half a watt to me,
so that shouldnt get too hot.

Sebi ... comment ?

Regards, John

 

[batman]

In response to john1's original schematic, I tried the capacitor thing, it didn't work. I also tried connected the gate to the anode so that when I want the light switched on, the anode wil never fall to zero volts. This killed the pin that was connected to the gate. Is there any way to use an SCR to turn on a 120 VAC light that actually works for real?

 

[john1]

Hi batman,

So was it correct to assume that you want to switch the
light on and off, using an SCR ?

I am sorry to hear that you have damaged your SCR.
The gate is quite a sensitive connection.
Dont really see why you thought that would prevent the
anode voltage from going to zero.

Yes the circuit i have offered is incorrect,
although it should have done no damage.
Please accept my apologies, i have submitted another
circuit which should work satifactorily.

It was incorrect in that there would be very little
voltage across the SCR when conducting.

This next should be ok.

John

 

[batman]

I don't understand this schematic because, it uses a mechanical switch. What I want to do is turn on the gate with a signal from a CMOS chip. Thanks[/code]

 

[Roff]

Batman, you need to go back and read "Rules for Posting".

John, you need to be a mindreader as well as a circuit designer.

 

[Sebi]

John,
no problem with Your circuits, cannot damage for SCR. But no need serial gate-resistor, and - avoid triggering when the switch is off - need a low-ohmic gate-cathode resistor (0.1....1k)

 

[john1]

Thank you Ron, Thank you sebi,

Sebi, i have never had an SCR conduct with an open circuit gate,
but i suppose if it had lengths of wire over to a switch, then
yes i can see that a spurious spike could easily trigger the SCR,
so i have to agree, a resistor between cathode and gate would be
a wise inclusion to make sure it stays off when its supposed to.

Without the series resistor, you would effectively be discharging
the capacitor using the gate, which may be alright, but i am a
bit wary of damage, i have always thought the gate to be a bit
sensitive.

Ron, yes sometimes its not easy to know what is required, and
sometimes its easy to just get things wrong.

Batman, if you're using cmos stuff on a home made assembly, i
think you should make sure that its on a stable footing, and not
flapping around at supply voltages, because cmos stuff is well
known to be extra easy to damage from external discharges that
normally would go un-noticed. This type of circuit has the
potential to damage cmos stuff very easily, because its running
at supply voltages from ground.
I would suggest you consider TTL equivalents for the cmos you
have in mind,
Or more sensibly consider using a triac instead of an SCR,
which would put the circuitry on a much more traditional footing,
that is one side could be neutral, which makes for a more
easily understood circuit anyway.

Best of luck with it,
John

 

[Sebi]

Batman,
see my circuit in previous post, this is exactly what You want...