Question about relays..

[Deac_1]

Fist a little introduction. I was on this forum a LONG time ago. Forgot my PW, life got in the way etc. The end result is that it's been a long time since my last visit. I have a rudimentary background in electronics and have just enough knowledge to be truly dangerous..

I have two questions and I will pose them in separate posts to keep things straight.

First: I put up a fairly large outdoor Xmas light display and, among other things I use a Mr Christmas Sounds and Lights dedicated controller to add some interest to my light display. In a nut shell it sets lights to music in a fashion similar to a LOR but in a simpler manner. It has 6 ac outlets for a six channel display.

This year I added a 6 channel mega-tree, 20 strings of 100 lights for a total of 2000 lights in 5 color channels and one channel to flash 18 LED stars. The tree is located 75 feet from the controller and, not wanting to run 6, 75 foot, extension cords (and also not exceed the controllers 2 amp per channel limit) I had to come up with something else. Simplicity was the watchword. At the controller I used a box with six small 12vac 300ma transformers. sending that low voltage out to the tree via category 5e, rectifying that and driving 6 relays. (switching 120vac sourced at the tree) Each color channel draws 680ma so initially I used 1 amp relays. The first relay failed within three days! An autopsy reveled badly burnt contacts. Thinking the relays were under rated I opted for 10amp versions. They lasted a bit longer but failed also. Next I added a snubber circuit to each relay (100 ohm resistor in series with a .47 cap) That helped a bit but relay failures continued!

What I can't understand is why 10 amp relays are failing under a 680ma, non-inductive, load? Since we are dealing with AC I thought any arc would be self extinguishing due to the zero crossing of the ac voltage, and at the very least the snubber circuit would suppress anything else...


Any thoughts??

P.S. I guess something using solid state relays will be next though exact implementation/design, has yet to be decided....

Kerry

[CeaSaR]

Hi Deac_1 (Kerry),

I remember you from the night-light project you asked about. How'd that come out?

As for the relays, the only thing I can come up with is that the way you are using them,
pulsed to music, is the cause for the burned contacts. They don't get enough time to settle
and they arc due to all the activity. I am only an enthusiast, so I may be off, but that is
what I can deduce. Mayhap one of the more "experienced" (read :engineer, tech, etc.) guys
will chime in.

CeaSaR

[jimmy101]

Are there incandescent bulbs in the strings? If there are, the startup current will be much higher than the rating of the string. IIRC, the resistance of an incandescent bulb can swing by a factor of 10. For example, a typical 100 watt 120V light bulb is about 10~20 ohms at room temperature but ~150 Ohms at operating temperature.

[dyarker]

Let's see; relays in Mr Christmas (probably solid state) controlling relays near Mr Christmas, which control relays near the tree which turn the lights on-off. WOW!!

I can't claim to be a relay expert, or engineer; but have some ideas that may help.

I looked up Mr Christmas. The easiest solution is to move Mr Christmas closer to the tree, and plug in the lights directly. I'm guessing the problem with that is it moves the music 75 ft away from the audience.

The 12VAC sent to the relays near the tree is rectified you said. Do NOT put filter caps after the rectifiers. Caps would slow the relay switching speed, making contact chatter mentioned by CeaSaR worse.

DO put a resistor in series with a reverse diode across the coils of the relays near the tree. This will speed up contact release by dissipating the magnetic field faster. Resistor value depends coil current and how high you want the reverse spike to go; higher = faster.

Put a resistor across the contacts of the relays near the tree so that the lights just barely glow when viewed up close. This keeps the lamps warm, so their resistance is higher, reducing in-rush current when the relay turns on. Side benefit of reducing thermal shock on the lamps so they last longer too.

Cheers,

[Deac_1]

Hi,

Ok a little clarification is in order. The reason the tree is 75 feet from the Mr Christmas (heretofore referred to as "MC") is that the MC controls lights on the house and front yard. Tree is in the lot next door.

As for implementation, (read hack design!) The MC has six AC receptacles (driven internally by what is basically 6 SSRs). I have a box with 6 transformers plugged into the MC, dropping the voltage down to 12VAC and sent out 75' to the tree on CAT5, where it is rectified and drives 6, 12vdc relays, which in turn switch AC sourced at the tree to drive 4 light strings per channel X 6. Each string draws ~170ma.

I initially tried using a small value electrolytic cap, BUT the relay(s) chattered badly, not liking the unfiltered DC at all. Smallest value that kept the relays happy, while allowing a fast response. was around 100uf. 12VAC relays would have been helpful but budget (non-existent) and availability made the use of 12VDC relays necessary. Didn't try the diode/resistor cross the coils but will. There was/is a snubber circuit, consisting of a 100ohm resistor and .47 cap in series, across the contacts.

Interesting idea regarding the resistor across the contacts to limit inrush current. I will give that a try too.

I was under the impression (mistaken??) that contact damage was occurring when the relay contacts opened, rather than upon initial contact.

Thanks for all the input!

Kerry

(If you have too much time on your hands link to my youtube Christmas Lights video..~~~~> http://www.youtube.com/watch?v=eIl05-uguiY

[Einar M]

I know zero about MC, but would assume it uses phase control to vary the lamp brightness. This envolves turning on the SSRs at a different point in each half cycle; 120 times a second. Mech relays cannot handle this, no matter what you try. The only solution is SSRs. Probably you could drive them from your stepped down signal ( through a resistor ). Look for some inexpensive versions; I have seen them in the 5 to 6 dollar range. Anyway, I don't have enough experience to suggest more.

[haklesup]

Can you post a data-sheet link or specs or Part number for the relay you are burning up.

It's important to note that relay contacts have 2 current ratings. One is the static current the contacts can handle once closed and staying closed. The other called switching current is what you can safely switch hot without burning out the contacts. I suspect you are violating that parameter. it is not the rating printed on the relay body in most cases so you need to read the data-sheet.

Also note that even with the twisted pair, the cat 5 cable will add inductance to the output causing any arc to sustain for a moment when the contact is opened. Inductance is only minimized when the force and return current are in the same twisted pair. If you wired it differently, you have not taken advantage of the lower inductance (I say lower, there is still some there). I've not yet located a spec stating impedance per foot (maybe in a spice model)

A series resistance helps calm inrush current but will dim the bulbs, A high value resistor across the relay contacts might work by allowing that arc current an alternate path but the choice of value is not clear. Too big and it might as well not be there and too small and the lamps may draw current when off. A capacitor across the contacts should absorb the voltage spike that occurs during hot switching but would be a short to an AC signal. if it were DC I might also consider a clamping diode.

A useful experiment would be to use an oscilloscope to measure the glitch that occurs when the relay is operated. There should be enough parasitic capacitance and inductance in the wiring to possibly induce a Vx2 spike as the return glitch adds to the main current. Knowing what the spike looks like and where it occurs would give you a better chance of selecting the best corrective action.

You may want to consider avoiding relays altogether, you can use a solid state relay which are far more resistant to this failure mode (contacts opening). These work great with AC loads.

I've been dealing with my own relay nightmare. G-ohm range leakage from a switch contact to the coil. I make a switch matrix and I need 20G or more impedance. Not easy to pinpoint in a 2x54 switch array and at $4 each, not cheap to fix either.

[dyarker]

Hadn't considered that Mr Christmas might be dimming the lamps. If that it the case FORGET the relays and use extension cords.

Cheers,

[Lenp]

Solid state switching is the key here!
Either use SSR's off the shelf or a triac/optocoupler combo.
Look at MOC30210 for ideas and circuits.
Len

[Deac_1]

[quote="Lenp"]Solid state switching is the key here!
Either use SSR's off the shelf or a triac/optocoupler combo.
Look at MOC30210 for ideas and circuits.
Len[/quot

Yup, that's likely what I will end up doing..A couple of interesting things. I had a snubber circuit consisting of a a 100 ohm resistor in series with a .47 cap. On the one channel that drives the 18 led stars, the white stars would go dark during the off times, the blue ones would stay partially lit. Removing that relays snubber circuit cured that.

I was thinking about the suggestion regarding putting a resistor across the contacts. I haven't tried to figure out what threshold of AC would put the lights at "dim" but suspect that the resistor would have be a very high wattage device....