Help needed. Zero crossing with RC snubber problem

t w o

o 0

nts no

to

of

Hi, Michael. First off, you should spec a 12V relay that's made to handle inductive loads (you can see a HP rating). This type of relay has contacts which open more quickly, and are farther apart when open. A small standard relay might not even open up far enough to stop an inductive arc at line voltage.

Next, when using a relay to drive a relay, you have to be aware of the delay-on-make, which can be several milliseconds, especially for larger relays. That might help explain some of the curious results you're getting. Turn-on delay can be affected by wear and aging. It also varies from unit to unit, even in relays with the same part and lot number. Trying to get this kind of timing accuracy might be the wrong way to go.

It might be better to take a look at suppressing the arc, which you've already started to do. Here's an intuitive way to start. First off, remember your basic goal: you want the voltage across the contacts to rise just slowly enough so the contacts can pull away without sustaining an arc. That's all.

Remove the cover from the driving relay so you can see the contacts. Next, find the current rating of those contacts, and use Ohm's law to find a resistor which will result in about half that current.

For example, if you have a 220VAC load, and your relay can handle

10A,:

R = 220V / 5A = 44 ohms

Choose a 47 ohm, 1 watt resistor here (carbon comp is best). Now get a selection of self-healing AC line-rated capacitors, and switch the inductive load with the 47 ohm and C snubber directly across the load, repeating and increasing the cap value until the contact arcing disappears, or at least is minimized. Without knowing anything about your 220VAC relay, it sounds like your 0.027uF cap might be on the small side.

Note that ITW Paktron makes a series of Quencharc snubbers that you can just plug in, which makes selection a snap. They're one-piece, epoxy-encapsulated units, and are very easy to use.

If all else fails, remember that physical distance is also helpful. Do what you can to place the arcing contact as far away as possible from sensitive circuitry.

Good luck Chris

Very true!

There also is a relay-off delay. And the RC snubber makes the turn-off delay even worse as the current will run a little longer. Expect something here between 5 to 20 milliseconds.

What you need to know is the current running through the contacts, and the voltage spike you want to allow. What a snubber does is store the energy of the coil in the capacitor. The resistors "eats" this up as the capacitor discharges over the load. The voltagespike at turnoff (aasuming an empty capacitor at that moment) is the load current * resistor. So if you have R= 100, C= 47n, I=2A, you get a voltage spike of 100*2 = 200 volt even before your capacitor charges. The load inductance with the current give the stored energy: Q=L*I=C*U. So the capacitor size should match the load inductance, otherwise you get a high voltage there too. The capacitor will for example be loaded to

100 volts with a Q = 27n*100V = 2,7 uCoulomb. An inductance that contains this would be L=Q/I= 2,7uC/2A = 1,35 uH. But the resistor already "eats" up a lot, so the voltage will be lower. When you turn off larger motors, transformers, inductors (the coil of a large relay!), the capacitor must match the load to keep voltages limited.

The prevent oscillations, resistance values must not be too low, ususally between 30 to 100 Ohm is ok. For large inductive loads I would not take 100 Ohm but 47 Ohm as normal value (see below what Chris wrote), you contact must ve able to handle that. The capacitance could be anything you want, for larger devices 220n, 470n etc.

Also very true: the RC snubber also gives a current peak when you turn ON the relay. Nice thing about inductive loads is that the current does not rise fast, so the load and RC current do not occure exactly at the same time (also depending on RC, where a smaller R gives a higher rush-in current but also a better timing).

The peak where the snubber works is at turning the relay OFF again.

What I often use is 47 Ohm to 100 Ohm with 100 nF.

You need TWO RC networks: one across the driving relay, one across the large relay, and do not forget the diode (or also RC) across the coil of the driving relay. So you have 3 things that need attention.

Do not forget that long cables (also with resistive loads like lamps) act as inductance. 10 meters or more may also need a RC!

Regards, Pieter

I have used a triac assisted relay before.

The triac shorts out the relay contacts just before the relay is switched on and off. Because the relay is no longer swicthing current they pretty much last forever. You can get away with a small triac as its not on for long.

0 e

ero

y

ak

=A0to

resents

lace

of

so

Right. If you have a resistive load for the final 220VAC load, a plain old solid state relay will work fine (be sure to heat-sink the SSR at about 1.5 watts per amp load). If you have an inductive load, you may want to spec a SSR made to switch these loads, which have back- to-back SCRs to eliminate the possibility of not being able to turn off the SSR).

If this is a class project, you won't lose any points by going for the simple solution, as long as it also happens to be the best one.

Good luck Chris