Figuring out RC snubber

I’ve killed a relay that was switching 120vac mains. I suspect it running a lamp with 3 compact fluorescent bulbs was just asking for trouble with the startup spike that those bulbs have X3.
It stayed on, so my guess is the contacts arc welded closed.

Anywho, I’ve ditched the CFLs and threw in LEDs, but I’d still like to CFL-proof the setup with some kind of snubber circuit.

The relay is a latching DPDT, but I’m only using a single pole and one throw (other being not connected as “off”).

I’m looking at pictures of RC snubber circuits, and since this is AC, I’m wondering how being paralleled across the ‘switch’ would work if the switch is open. Wouldn’t AC current still pass the resistor and capacitor?

What values of resistor and capacitor are used for 120vac @ 60Hz, <2A running (but unknown spike levels to attenuate)?

Yes a snubber will still pass current with the contacts open, but way less than the load and almost purely capacitive (reactive). Snubbers need to be roughly matched with the inductance of the load, so that's the first thing to find out. Basically the energy stored in any inductance is pushed into the capacitor as the contacts open, on a timescale much shorter than the mains period, but much longer than the typical inductive spike.

If the capacitor is too small it will allow the voltage to rise very high, and the contacts will still arc, if its too big the quiescent current is high and wasteful.

The resistor is needed to prevent accidental resonance (its an RLC circuit) with the inductance in the load and prevent high spike currents when the contacts close.

Another approach is a voltage clamping device like a MOV or TVS that can handle the inductive energy comfortably - and is UL listed, mains rated, etc.

Ok, so incandescent lamps are resistive, compact fluorescents are inductive, what are LED bulbs?

Above their turn on voltage, very low resistive - hence should be driven by a current source...

regards

Allan

allanhurst: Above their turn on voltage, very low resistive - hence should be driven by a current source...

regards

Allan

LED bulbs, the ones shaped as light bulbs. They're getting mains power and have internal transformer circuitry. Get with the program.

I think you meant that the contacts welded shut, not open?

It seems to me that a capacitor across the contacts is going to cause a current spike when the relay contacts are closed.

The big problem with CFLs and LED bulbs when they turn on may be capacitor charging current. CFLs are not inductive, they have a bridge rectifier followed by a smoothing capacitor. So not really inductive, capacitive, or resistive as it only has current flow on the peaks. I took apart a bunch of CFLs and found resistors in all of them, presumably to limit this inrush current.

I would expect to find something very similar in LED bulbs.

Perhaps your relay is not rated for a high enough current. Never use a 3A relay in a 3A circuit.

An old style fluorescent tube with a coil ballast and a starter will have some inductance. It is a complex impedance.

The relay is a latching DPDT, but I'm only using a single pole and one throw (other being not connected as "off").

Dual coil or single coil controlled? (may need several flyback diodes). Wouldn't mind seeing your control circuit, code and part# for the relay. I'm quite familiar with latching type relays.

Ok, so incandescent lamps are resistive, compact fluorescents are inductive, what are LED bulbs?

Incandescent lamps have the highest derating factor for relays. Definitely its inductive component is very high.

From here ...

and here ...

Anywho, I've ditched the CFLs and threw in LEDs, but I'd still like to CFL-proof the setup with some kind of snubber circuit.

Snubber circuits by design would pass a small current ... enough to make CFLs flicker. Not sure about LED bulbs, but adding a snubber would probably make them impossible to turn off

dlloyd: Dual coil or single coil controlled? (may need several flyback diodes). Wouldn't mind seeing your control circuit, code and part# for the relay. I'm quite familiar with latching type relays. Incandescent lamps have the highest derating factor for relays. Definitely its inductive component is very high.

From here ...

and here ... Snubber circuits by design would pass a small current ... enough to make CFLs flicker. Not sure about LED bulbs, but adding a snubber would probably make them impossible to turn off

Hi,

That does not sound right, unless i am not understanding you correctly.

Incandescents are resistive, not inductive, however their resistance swings roughly 10 to 1. So during startup they might draw 10 times the normal current. An inductive load would do just the opposite: draw less current during start and more later. The inductive kick back is usually a problem however when the device is switched off.

The relay 'welded' contacts can come from a high start up current or perhaps arcing during turn off. To resolve the problem it has to be determined which is causing the problem, or possibly both.

A snubber would be the idea if the relay contacts arc over during turn off.

If the problem is high startup current though then some kind of current limiter would be needed.

A typical snubber for DC would be a diode, bleeder resistor, and capacitor. The capacitor absorbs energy when the contacts open, and the energy slowly dissipates in the resistor. The cap and resistor and diode have to be sized according to the amount of kick back energy it has to handle and how frequently the switching action occurs. A snubber for AC could be similar, but instead of a diode a bridge rectifier. The rectifier allows energy to flow only into the cap, and the resistor bleeds the energy off at a given rate. If the contacts switch fast then the values have to be sized to handle the energy at the rate needed. Some efficiency is lost because of the bleeder resistor and it conducts when the relay contacts are open so the sizing is a tradeoff between fast action and low power dissipation. Perhaps a TVS instead.

If the problem is more about a high start up current, then some sort of current limiting is needed if the relay contacts can not be oversized. Of course going to a heavy duty relay would be a good idea because current limiting isnt that easy. For example, a second relay to keep a resistance in the circuit for a short time and then switch the resistance out by shorting it out and thus allowing full current. An inductor would work here too, but sizing and secondary behaviors are not that easy to deal with. Another idea is an inrush thermistor. These are made for the very purpose of limiting current during startup of various devices like motors and capacitor filters. They are available in a variety of sizes, but they do get hot during normal operation, so that has to be dealt with.

So really the first thing to do is find out if it is an inrush current problem or an arc over problem, or both.

Dual coil latching, DPDT, 2A, 5V coil, 28mA coil current As I said, the relay coil side is not a problem.

After it was stuck open, I took off the casing and the insides were moving as normal again. So it wasn't a permanent melted-metal-welding, frankly seems like I could've just flicked it loose like the ol' hammer on the fuel pump relay trick in trucks.

I have a 3V version of the relay wired up now and it's doing just fine with the LED bulbs, being turned on and off throughout the day, hasn't gotten stuck.

The LED bulbs say 130mA on them, so unless that rating is purely just for the LED (post-transformer rating) and the transformer is horribly inefficient, I think a 2A relay should be able to handle the job of two such bulbs. The CFLs were in the park of 300mA each and there were 3 of them, which is still only 900mA during normal operation.

I have two switching modules, one based on a PT2272-M4 RF receiver (writes a pin HIGH when signal received and directly drives the relay set/reset, works dandy when tested for hundreds of cycles with no load on the relay switch side), and another based on an ATTiny85 that I made to work with a wall switch as a button input.

The relay that froze was controlled by the PT2272-M4 RF receiver module.

I've used really high current dual-coil latching relays (100A and 200A). Experienced the same issues with the internal magnetic mechanism. It was the mechanism that gets stuck, not the contacts.

It mostly comes down to the pulse you use to control them (it should not be a steady state signal). Your relay is specified at >= 10ms, but I wouldn't go much more than 200ms. Heating of the coil reduces performance.

Also the voltage of the pulse needs to be within 5% of the nominal coil voltage.

Another thing that affects performance is how closely the relays placed next to each other. Placing them to close together can also freeze the mechanism.

Are you using 2 diodes - one for each coil?

EDIT: @MrAl Yes, I agree with your comments (I striked-out my error) ... thanks.

Hi,

Oh so this isnt really a welded contacts problem then?

If the mechanism gets stuck then, do you know if it is because of the magnetic part itself or the mechanical mechanism itself (or maybe both) ?

If it is the magnetic part where the iron becomes permanently magnetized, then a fix might be to introduce a gap in the magnetic path, which could simply be a very thin material like a small piece of mylar film. This will help keep the metal from staying magnetized once the current is removed. I dont know if that modification is possible with the actual relay being used but if it is then that would help significantly. Of course follow up testing is required to make sure the thing still operates at the expected voltages and that the coil does not over heat (separator too thick AC models only).

If it is the actual mechanical part (like it gets stuck) then the mechanism would have to be examined carefully to find out what is sticking, and whether or not lubrication would help or not or some sort of machining would be needed.

I can say one thing though, if the relay gets stuck for any of these two reasons then i would get another relay of a different type. I have worked with high power relays in the past and never had this sort of problem, and that was with relays that had to connect power converters directly to the line (photovoltaic energy systems). The relays used for that were called “contactors” which have high current ratings. The only problems encountered were slow speed and inductive arc over.

Designing RC SNUBBERS

RC SNUBBERS

Off the shelf