# snubber in contactor coil

Discussion in 'Electronic Design' started by a-bike, Oct 14, 2006.

1. ### a-bikeGuest

I have a problem of noise caused by the conmutation of a contactor that is
hunging the dsp in the control board of my system. Some one told me the
posibility of connecting a snubber or varistor in parallel with the
contactor coil in order to reduce that noise. Does anybody knows about the
method used to dimensioning the snubber?

2. ### John PopelishGuest

A common AC (you didn't say whether the contactor coil is AC
or DC operated, so I guessed) coil snubber consists of a
resistor in series with a capacitor, connected across the
coil. When the driving contacts opens, the coil current
detours to the capacitor, which limits the rate of change of
voltage as it absorbs the energy that was stored in the
coil. The lower the value of the resistor, the better this
works. Except that when the driving contact closes at peak
voltage, there is a large inrush current as the cap is
popped up to that peak voltage, while the contact is
bouncing. Bad for both RFI and contact life. So a
compromise must be struck on the resistor value. Since the
coil draws essentially no current during the bounce time,
because of its inductance, I usually size the resistor to
limit the peak current to no more than the contact current
rating. The capacitor value and voltage rating must be
large enough to contain the peak stored coil energy without
exceeding its voltage rating. But way too much capacitance
causes the series resistor to get hot while the coil is
energized.

In some cases, I have used a smaller capacitor than could
safely store the peak coil energy, and limited its peak
voltage by paralleling it with an MOV.

So to design an effective snubber that doesn't get too hot,
about the contact rating (that is driving the coil) the peak
driving voltage, the coil's peak current and its inductance
(to come up with the (I^2*L)/2 energy that gets transfered
to (V^2/C)/2 in the capacitor.

Or you but an off the shelf RC snubber rated for the coil
voltage and hope it works and survives.

3. ### John PopelishGuest

John Popelish wrote:
(snip)
Sorry. The energy stored in the capacitor is (V^2*C)/2

4. ### Fred BloggsGuest

It's hard to tell whether you're talking about snubbing the coil or the
contacts or both? And the OP is classically conclusory: has no idea in
hell what he's doing but still *knows* a snubber is the answer. You are
a glutton for punishment it seems.

5. ### Homer J SimpsonGuest

0.1 uF cap in series with 100 ohm resistor.

6. ### ChrisGuest

Either that or John's just a natural teacher.

Cheers
Chris

7. ### Frithiof Andreas JensenGuest

Step 1) Look up contactor in catalogue.
Step 2) Order the matching supressor/snubber/whatever it is called in catalogue.

Then fun starts:

3a) Find out that the leads, the mechanical fixing e.t.c. do not match
contactor.
3b) Find out that contacts need snubber after having got one for coil.
3c) Find out that coil needs snubber after having got one for contacts.
3d) Find out that this is after all not the real problem having got all snubbers
available.

8. ### Tom BruhnsGuest

A snubber cross the coil will keep the flyback energy from the
inductance from exceeding the breakdown voltage of the driving device,
if properly designed, but the coil side should not be causing the sort
of noise that would hang the DSP. I'd guess more likely the very fast
risetime transients from the contacts opening and/or closing are the
culprit. You can also apply a snubber to them. The snubber design
depends on the load they are driving: the voltage, the current, and
whether the load is resistive or inductive (or possibly capacitive).
Beware that optimal design of a snubber for contacts is different than
optimal snubber design for a solid-state switch driving a relay or
contactor coil. You should be able to find web references for both; I
have in the past, and some are quite good.

Cheers,
Tom

9. ### Mike MonettGuest

Hi Tom,

Back in the old days, a laser company moved in next door and all our
sensitive equipment went belly up. It turns out they used 10KW heaters to
soften the glass, and the temperature control was a simple thermostat.

Every time the contacts opened the arc resonated with stray inductance and
capacitance in the wiring, and generated a burst of 40MHz noise, measured
with my HP spectrum analyzer. The RFI went everywhere and was impossible to
filter or shield against.

Hertz would have been proud

I talked to the owner and he allowed me to add simple snubbers across all
the contacts. I believe it was something like 47 ohms in series with 0.1uF,
but this was a very long time ago. Anyway, it was pure guesswork.

Since there was no way of figuring out the wire inductance in the ratsnest,
there was no calculation involved, and the first try solved the problem.

I guess the trick was to prevent the arc from starting in the first place.

Regards,

Mike Monett

Antiviral, Antibacterial Silver Solution:
http://silversol.freewebpage.org/index.htm
SPICE Analysis of Crystal Oscillators:
http://silversol.freewebpage.org/spice/xtal/clapp.htm
Noise-Rejecting Wideband Sampler:

10. ### Tom BruhnsGuest

Yep, exactly: for opening contacts, keep dv/dt low enough that the
contacts spread fast enough that no significant arc develops. The arc
itself should not be the problem; it's the very fast dv/dt when the arc
finally quits. And to do the job, you need capacitance, with just
enough series resistance to keep the current pulse in the capacitor
when the contacts close under control. But the snubber across a
contactor coil driven by a solid-state switch such as a triac will
typically use a much larger resistance, too large to work well for a
contact snubber.

Arcs/sparks have a bad reputation for causing EMI, but it's not the arc
itself, it's that the arc can act like a very fast switch. Spark
transmitters used the spark to connect a capacitor in parallel with a
coil, with the capacitor pre-charged to typically a rather high
voltage. Though the spectrum was fairly broad, most of the energy was
concentrated at the resonance of the capacitor and inductor. Spark
transmitters generating up to a significant fraction of a MEGAwatt
output were built and put in service to communicate roughly half-way
around the world at what we'd now call VLF frequencies: in the tens of
kHz.

Cheers,
Tom

11. ### Rich GriseGuest

I heard or read somewhere once something to the effect that the resistance
should equal the inductive reactance, anthough they didn't say at what
frequency, and as you've said, there was no way to measure the inductance,
so a WAG, if it worked, was obviously the right answer. ;-) Also, the guy
that made that reactance remark just said to use a large enough capacitor
that its reactance is negligible; I'd think "critically damped" would be
the value that would resonate at the inductance you used for the resistor
calc, at the frequency of interest.

At least that's a starting point - I tried to learn about snubbers once,
but it turned out to be as much black magic as science, much like ferro-
resonant transformers.

Cheers!
Rich