Protect mechanical relays from AC inductive loads 3

[asimpson]

Hi there.

I am running a 20W solenoid valve with 230V AC through a mechanical relay. There is a considerable ammount of electrical noise when the relay switches and this is interfering with other electronics. I would also imagine this is not helping the relay either. I have used diodes in the past with DC loads but what should I do with an AC load?

Many thanks.

 

[itsmoked]

Look up "snubber".

Also, the electrical noise from the relay itself is not usually the problem. It's more about the wires running to it. Try to keep the wire loop small and reroute the lower voltage wires away from the relay's wires. Try also to have them cross at right angles if they do have to get close.
Twisting the 'noise' wires, which automatically 'keeps the wire loop' small, helps too.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[OperaHouse]

Typically these are protected with a RC network consisting of a capacitor .1-.47uf and resistor 47-220 ohms in series. These are available off the shelf in potted assemblies or you can roll your own. This network can be placed parallel to the coil or if that is inconvenient, parallel to switch contacts. RC networks will give the lowest electrical noise. A MOV can also be used but that will begin conducting at a higher voltage.

 

[itsmoked]

Hi Opera; MOVs work well for a little while but they have a relatively low cycle life and will fail in snubbing service eventually.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[asimpson]

Many thanks.

Looks like RC snubber should do the job. Any rule of thumb how to size R and C. One guide recommended R should be 10x resistance of coil. C is a speed of responce variable . I assume C should be chosen on the basis of speed of noise versus speed of responce of switching.

I will also tidy up the wiring harness.

 

[itsmoked]

It's a crapshoot. Really the noise is from the very fast transients and the snubber caps suck them down pretty quickly. Follow your recipe. DO take the time time to calc the steady state current the resistor is going to see. They often need to be 1W/2W sizes to stay cool enough. You want them 2 or 3 times the expected wattage.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[OperaHouse]

If you are driving the solenoid with 60Hz AC, speed isn't an issue. Life issues were brought up. If making your own RC
snubber, Use a capacitor that is rated X2 across the line duty. These are actually two capacitors in one to reduce the
effects of internal corona. These all have agency approvals that assures they have been tested for severe line pulses.
The resistor should be at least one watt or wire wound. Most resistors are film and these have a tendency to open up
with repeated pulses if a smaller wattage is used. I suggest you use a .22uF and 220 resistor combination.

I used to manufacture snubber networks and this reminds me of a funny story. Against my strong objections, the owner wanted
to get CE approval on some legacy RC snubbers. These were grandfathered in under an old UL508 spec and none of them used X2
rated capacitors. Somehow he thought that buying them dinner and a few drinks he could get these approved. Half way into the
test and $40K, 70% of the devices had failed and the test was stopped. I knew they would fail but not that bad. Thought the
resistor would slow down the dv/dt enough to save some of the capacitors.

 

[Skogsgurra]

Very valuable information, OperaHouse!

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[itsmoked]

Definitely. Thanks for the details.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[mikekilroy]

only tidbit to add might be to help pick the wattage rating of the R if building own.... One may assume a little 1/2 watt resistor would be fine....

Xc= 1/(2pi*f*c) ohms....

So assume .22ufd & 220 ohm as suggested on 120vac 60hz contactor for example:

Z=220+ 1/(6.28*60*.00000022)=220+1096=1317 ohms.....

So I=V/Z = 120/1317= .09 AMPS

Since it is a series circuit, the resistor sees V=I*R=.09*220=19.8volts across it.

So it will heat up to W=I*V=.09*19.8= 1.8 watts

NOw if this was a 460v contactor coil..... you do the math

So use at least a 3 watt resistor with these values on 120v, 60 hz circuit

http://www.KilroyWasHere<dot>com

 

[OperaHouse]

I'm not going to fault anyone for a math error, made enough of them myself. As Capt'n Ron says, Anything that's going to happen is
going to happen out there. So I connected a 270 ohm to a cap that measured .175uF and clipped it on a test lead with 123VAC. The
voltage on the resistor was 2.17VAC, which looks like a factor of ten off on the calc.

That leads me into one more comment. We made tens of thousands of 220 ohm .47uF snubbers rated for 250V with half watt resistors.
However, these were made with carbon composition resistors (not to be confused with carbon film). Carbon comp have an amazing capacity
to withstand surges and survive. Carbon film resistors will open a track under the same circumstances. I have a favorite microprocessor
controlled Christmas light timer. I have picked up three of these that have a 1/2W carbon film resistor in the power line. Each of these
resistors had opened up. That is a short season for a life cycle test. I have done stress tests on small caron film resistors heating
them up to a red glow for hours. Resistance changes would hardly be 1%. Yet in pulse situations I have seen many of these CF resistors
fail in line applications.

 

[mikekilroy]

Jeez, thanks for the catch. I am propped up in bed here with back pain going down my leg and sometimes wanting a whiskey bottle more than typing on this dam_ keyboard! Thanks for the catch - don't want to scare off the OP with bad math!

CORRECTED MATH
Z=220+ 1/(6.28*60*.00000022)=220+12063=12283 ohms.....
So I=V/Z = 120/12283= .01 AMPS
Since it is a series circuit, the resistor sees V=I*R=.01*220=2.2 volts across it.
So it will heat up to W=I*V=.01*2.2= .02 watts

Thanks OperaHouse!

http://www.KilroyWasHere<dot>com

 

[asimpson]

Many thanks.

I'll look into that .

 

[mikekilroy]

asimpson, sometimes folks like these gidgets so well they start putting them all over machines. you can buy nice ones in little blue boxes from good companies like rk electric (rke.com).... neat 3ph ones in y configuration with ground....then folks may add a vfd to their machine and promptly smoke them! being designed for this low wattage dissipation @60, they can become the lowest impedance source to ground - for the vfd's 4-10khz noise! the math shows of course that the 12000ohm cap Xc becomes small enough (120ohm@6000hz) to overheat the 1/2 watt 220 ohm resistor! sorta like my first math mistake.

http://www.KilroyWasHere<dot>com

 

[jraef]

This topic made me chuckle. I once had a problem in the field, WAAAY out in the middle of nowhere, with a 480V starter coil causing so much line noise that it caused SCRs in a soft starter to self commutate. It was very brief, but long enough to trigger the Shorted SCR detection in the soft starter. I was stumped for hours on this problem of random SSCR trips, when I happened to notice that a nearby 25HP air compressor came on at the EXACT moment that the Soft Starter gave the false trip. Opened it up and saw the 480V coil and realized what was happening.

So I had to create an RC snubber for that coil from parts I could buy at the nearest Radio Shack store, 40 miles away. Needless to say that the Radio Shack store in Adelanto Ca, a tiny nothing of a town on the edge of the Mojave desert, did not have "industrial grade" components. In fact when I asked the girl (not woman, GIRL) behind the counter where they kept resistors and capacitors, she looked at me like I had two heads. I finally had to string resistors and caps together in a series / parallel network to make something that worked. It was, to put it mildly, ugly.

Oh, and they had a soldering gun, but they were out of solder... This project was at a scrap metal melt furnace and there were lead tire weights lying around on the floor, so I managed to melt them and get the lead to stick by some miracle (no rosin of course). That was one of the worst kluge jobs I have even done in the field when it came to electrical stuff. It likely didn't last long, but I explained to the electrician there that he needed to order an official RC snubber from the starter mfr and replace my handywork ASAP. I seriously doubt he did though...

"Will work for (the memory of) salami"