# AC realy driver ?

Discussion in 'Electronic Design' started by Rich, Mar 3, 2007.

1. ### RichGuest

Hello,

I have some AC relay questions. I have a generic relay driver control
signal that drives a small RC to gate a triac. The triac provides a ground
to a 24VAC off board relay. Here's the circuit:

24VAC
|
C|
o---------o G C| relay coil
| \ C|
0-5V control ___ | MT1 |<| MT2 |
signal -|___|o---o o--o|\|o---------o-------o
330 | | |>| |
o | o
--- | ---
0.47 50V --- | 0.1 100V ---
o | |
| | o
| | |
o------o----------------o
|
GND

Now, MT2 is greater than MT1 so the "left" pointing gate will conduct when
gate (G) has a negative or positive voltage. Thus, I "source" a ground to
the 24VAC relay coil and can activate the relay. My questions are as
follows:

1) What is the proper method to size and rate the 0.1 cap? I surmise this
cap is used to absorb the energy created when the current through the
inductor (relay coil) is suddenly stopped. So the energy stored is
(1/2)*L*I^2. Is the 0.1 cap really needed? Won't the triac alone route
any remaining sine wave from the AC signal until the current drops below
the holding current thereby absorbing the energy. I ask this because I'm
concerned if the 0.1 cap fails shorted (improper voltage rating) then the
relay coil is always energized! Maybe the analogy is a zener/diode
combination in DC circuits to speed up the energy absorption especially if
the relay is driving an inductive load, is that what the 0.1 does in this
case? The RC would be the coil resistance and 0.1uF so this would be a
small time constant (not sure if its faster than the holding current method
though!).

2) I surmise the triac must be selected with heavy consideration to ensuring
that the relay coil current is >> than the holding current of the triac,
right?

3) I'm curious about the RC. The control signal comes from a 5891 driver.
Do triac gates require slower slew rates? If not, why the RC? If so, why?

Thanks!

2. ### RichGuest

After reading a bit more, I think the 0.1uF cap is to help ensure triac turn
off. With an inductive load (ac relay coil), the current and voltage will
be out of phase so as the current falls below the triac holding current,
the voltage will be near a peak (positive or negative) due to the inductor.

I see two issues with this. 1) If the voltage is high enough, the triac can
stay on regardless of the gate (static high forward bias scenario is the
way I'm thinking about that). In my case at 24Vac I think this is a
non-issue with a 2N6073 with a 400 peak repetitive off state voltage. 2)
More importantly, the quick stopping of current in the inductor will create
a spike in voltage that I'm guessing will have a significant slew rate to
it (???). If this slew rate exceeds dV/dt for the part the off cycle
command will fail regardless of the current dropping below the holding
current. What's odd is that I think in most cases this would be handled
with an RC not just a C!

So I'm not sure what rate to expect the voltage profile to have when the
inductor is shut off (HELP!). And once that is defined,

1) I presume you target the RC to have a time constant much slower than
dV/dt. Any guidelines? Should the RC time constant be 5x > than dV/dt,
etc.?

2) I'm still uncertain how to size the cap voltage rating. If I'm on track
that this cap is for triac off state protection, then it seems the most
critical voltage is the spike during inductor current flow cut off. I'm
trying to estimate that by:

0.5*L*I^2 = 0.5*C*V^2 <<< Solve this for V
V = ((L*I^2)/C)^0.5

So I'm using V as my max voltage experienced by the cap, L as the ac relay
coil inductance and C the value of the cap. Is this the correct approach?
If so, any good estimates for L of an ac relay coil? Of course, I have I
which is the relay coil current.

I'm still very interested in the questions from my this and my first post so
if you can help me out it would be must appreciated!

Thanks

3. ### jasenGuest

Quite the opposite, it's there to absorb the voltage step when the triac
switches off. those things don't turn of during current peaks.

sizing is done by observing the dv/dt limit for the triac the voltage of the
supply and the inductance of the relay.
yeah, but unless it's masively oversized for the task that's unlikely to be
a problem.

Bye.
Jasen

4. ### RichGuest

I realize the triac won't cut off until up to 1/2 cycle later and that the
voltage is out of phase with the current on an inductive load (relay coil).
But with the gate removed and no inductive load (V and I in
phase...resistive load), the triac should cut off without the cap near the
next zero crossing, right? That is to say, any triac without an inductive
load doesn't need the cap? Or is the cap present for the fact that the
relay coil is an inductive load causing the out of phase relationship I
just mentioned and a voltage spike on shut off that could violate dV/dt for
the triac and perhaps the traic forward breakdown voltage (I'm not worried
about the latter as the triac is rated to 400V repetitive off-state
voltage)? I think it's the inductive load that requires the cap but was
confused by your statement "quite the opposite." The voltage step you are
referring to is from the inductive load, right? Also, wouldn't it normally
be an RC not just a C to control the slew of the inductive voltage spike?

So when sizing the cap, I think a more typical approach would be to size an
RC to protect against violating dV/dt. Is there a way to estimate the
voltage rating of the cap? Or does this really need to be done on the
bench top? I've been trying to estimate worst case voltage the cap will
experience based on the energy stored in the magnetic field of the inductor
(relay coil) and assuming the cap must withstand this.
Yeah, for my part Ih is 15-35mA and the coil voltage is in the neighborhood
of 125mA. Good point!

5. ### jasenGuest

right, if it was a non inductive load the capacitor isn't needed.
Because it shuts off with no current flowing through the inductor there's no
inductive spike, so forwards breakdown won't be exceeded.
it's caused by the current crossing zero near the voltage peak, not by
a sudden collapse of a magnetic field in the inductor, to me that's a
different effect.
Actually it's an LC
coil current? hmm, that seems kind of close.
most of the above I learned from a document called

I think this is the URL

http://www.onsemi.com/pub/Collateral/HBD855-D.PDF

Bye.
Jasen

6. ### John FieldsGuest

---
It seems to me that your circuit's overly complicated.

Here's what I'd doView in Courier)

24VAC
|
+--[<Z50V]--+
| |
[COIL] |
| |
+--[<Z50V]--+
|
MT2
VIN>---[R]---G
MT1
|
GND

You may not even need the Zeners; it's going to depend on how much
current is in the coil when the TRIAC turns off.