Connect with us

Relay failures in lighting controller

Discussion in 'Electronic Design' started by Raveninghorde, Feb 1, 2013.

Scroll to continue with content
  1. We have been making a lighting controller for many years. Normal load
    was fluorescent lights.

    We use a 12A relay with a 30A rating for 4secs at 10% duty cycle. The
    contacts are AgNi.

    The customer is reporting that the contacts are failing by welding
    closed.

    The failure is linked to electronic ballasts and LED lamps. One
    reported failure is on a 36W, 230Vac LED lamp. I am asuming, until I
    can get hold of the kit, that the failures are due to inrush current
    into large capacitors in the lamps.

    Will a change in relay contact material to AgSnO make much difference?
    There isn't anything in the data sheets to give me confidence that
    that will solve the problem.
     
  2. Robert Macy

    Robert Macy Guest

    First place I'd look...
    Electronic ballasts require an EMI filter in front of them to prevent
    the conducted emanations. What you have there is a capacitor across
    the AC mains followed by some husky inductance. That makes the load
    look incredibly inductive, not capacitive - with main impact as you
    turn it off. To envision, simply unplug your PC very slowly from the
    AC outlet. Even with it turned off, you'll draw an impressive arc. All
    due to the stored energy in the EMI filter. I suspect it is THAT
    energy that is welding the contacts together.

    Not sure, you can place effective arc suppresssors at your relay
    contacts. Especially, with such an energetic arc occurring just as the
    contacts start to open. HP invented a circuit that is the opposite of
    a zener - dead short until a maximum of current is reached. IC made by
    local foundry in CA, Supertex? You could place THAT component in
    series, then in parallel to the load place a tranzorb to take the
    voltage and the combination should completely protect the contacts.
    From memroy there is some voltage drop across HP's component Don't
    know how much, but less than 1 V, probably Vbe drop like in a 2N3904
    [again form memory, that was the bipolar process for making the die] I
    have a schematic of the IC somewhere.
     
  3. Phil Allison

    Phil Allison Guest

    "Robert Macy"

    First place I'd look...
    Electronic ballasts require an EMI filter in front of them to prevent
    the conducted emanations. What you have there is a capacitor across
    the AC mains followed by some husky inductance. That makes the load
    look incredibly inductive, not capacitive - with main impact as you
    turn it off. To envision, simply unplug your PC very slowly from the
    AC outlet. Even with it turned off, you'll draw an impressive arc. All
    due to the stored energy in the EMI filter. I suspect it is THAT
    energy that is welding the contacts together.


    ** Horse poo.

    Inrush surges with the SMPS associated with LED lights ( and CFLs) are often
    huge and with multiple loads on the same switch quite disastrous.

    With contacts closing and current surging simultaneously, spot welding is
    the outcome.

    Relays have no ability to pull such welds open like a regular switch has.


    .... Phil
     
  4. Jamie

    Jamie Guest

    I think there is something they are not telling you..


    Just a thought

    Jamie
     
  5. Jasen Betts

    Jasen Betts Guest

    contacts can't weld closed while they are open.
    therfore the damage is occurring when they close
    capacitor derived surge current through the bouncing
    contacts heats them up to melting and the weld closed.
     
  6. Jamie

    Jamie Guest

    They most certainly can weld together while they are in
    the process of opening with induction loads. Which is why
    special designs are out there just for handling induction
    loads.

    If you don't use the correct contact design for these types
    of loads, the relay will not last long.

    Jamie
     
  7. Guest

    I have hard time seeing how it can happen when opening, arc
    start when they are already moving apart so unless you close them
    instantly after being melted by the opening arc I don't see how
    they could weld

    at closing it is more like spot welding, melt and press together
    sure arcing at opening will eat the contacts, but that is different


    -Lasse
     
  8. Jamie

    Jamie Guest

    Many relay designs do a lot of surface sliding while they are pulling
    apart or coming together. The pulling apart is worse because you are
    increasing R as you do this. Since the arc has already started before
    the R gets too high, plasma burn is already in the process and heating
    is severe. When closing the contacts, you get only the initial bounce
    and if the load is inductive, the contacts will most likely be closed
    before enough energy is store to cause critical damage.

    Simply put, you get a lot more plasma burn on the surface while
    contacts are opening because the initial plasma has started while the
    contacts where still touching but under a load and arcing, this
    generates a malted metal vapor cloud which is highly conductive and
    perpetuates the process. If the contacts have enough damage to them
    already over time from doing this, the contacts will be pitted and
    carbonized and at some point, the next plasma burn it goes through will
    weld them. Repetitive action on the contacts will heat them
    and they will weld anyway, most likely from incorrect use of the device.

    Induction loads just makes things much worse because they can arc at
    greater distances when opening and cause more damage.

    Relays with spring arms are known for this kind of damage in the wrong
    environment. You need contact designs that pull and push straight and
    fast. Some have rounded heads to help vent off the metallic vapor along
    with using tungsten, etc..

    We have discharge contacts we use for draining off HV in the 20KV or
    more range, from doing HI-POT test on large cables that can hold a large
    charge. These units have 2 round balls that come
    together that form the contact to ground and a air supply that blows
    fresh air across the surface area. These units have been in service for
    years and have never been replace as far as I know.


    Jamie
     
  9. Tim Williams

    Tim Williams Guest

    All the times I've stuck the rod to the work were exclusively when closing
    the circuit. I can safely say I have never, ever had a rod stick to the
    work when pulling it off!

    Tim
     
  10. Guest


    I have done plenty of welding, stick and tig

    have you ever got the rod stuck while pulling it off?

    -Lasse
     
  11. Robert Macy

    Robert Macy Guest

    Makes sense that while being heated the contacts are having pressure
    applied, thereby enhancing the welding process.

    I'll 'retreat' to simply pointing out that the EMI filter is one of
    the worst loads possible for relay contacts.

    So, How about 'trying' to time the closurre so that the contacts hit
    at the zero crossing? Hitting within 1mS, on the 'shy' side should
    help a bit. Whoever suggested, now has me sold on the solid state in
    parallel with the relay.
     
  12. Tim Williams

    Tim Williams Guest

    Gravity, mostly. Ever done vertical or upside-down welding? Lots of blobs.

    Rod never makes a direct ohmic contact, through solid metal or through
    molten blobs. Same as pissing on the 3rd rail, it doesn't make a continuous
    connection.

    Tim
     
  13. Guest

    That's getting close. Something like this shouldn't have a problem:
    http://www.te.com/en/about-te/news/new-products/2012/05/24/relay-rtx-product-news.html
     
  14. josephkk

    josephkk Guest

    There are multiple processes going on. Contacts that are badly pitted are
    easier to weld on closure due to irregular contact area.

    ?-)
     
  15. The answer is yes.

    The relay contcts were welding in the first 10 or so operations with
    AgNi contacts.

    So far AgSnO contacts have survived 10,000 operations on the same
    load.
     
  16. Guest

    Hello there. I think modern HF flourescent light fittings probably use highvoltage mosfets (500-900volts)electronic balasts probably presenting an operating reactive load that is similar to switched mode PSUs (Transmitters,telecoms,server bank PCs) which calls for a substantual derating of your relay in question. 16A resistive(heating element load) but only 2A reactive load is typical of external integral halogen lamp PIR controller or 5-6A maxreactive for seperate PIR module. Probably call for the 15A (600-800mW) or25A (30/20A DPDT Finder) (1-2W) for conventional mechanical relay from Maplin (or rapidonline.com). There are also 25A electronic relay modules (farnell.com) that use triac switching (might NOT be suitable).
     
  17. P E Schoen

    P E Schoen Guest

    "Michael A. Terrell" wrote in message
    It says User Agent G2/1.0 which appears to be Google Groups. They probably
    made another "improvement".
    https://wiki.mozilla.org/Discussion_Forums/Request_For_Comment

    Paul
     
  18. josephkk

    josephkk Guest

    On Fri, 22 Feb 2013 08:48:26 -0500, "Michael A. Terrell"
    It looks like varient412 is using Netscape 4.8 with very messed up
    settings. He seems to be doing an Earthlike - giganews cha-cha as well.

    ?-)
     
  19. Phil Allison

    Phil Allison Guest

    <

    I think modern HF flourescent light fittings probably use high voltage
    mosfets (500-900volts)electronic balasts probably presenting an operating
    reactive load that is similar to switched mode PSUs


    ** SMPS do not present a reactive load.

    Why does this idiotic myth persist for so long ??

    It's harder to kill than the AGW myth ....


    16A resistive(heating element load) but only 2A reactive load is typical of
    external integral halogen lamp PIR controller or 5-6A max reactive for
    seperate PIR module.


    ** **** knows what this total dill is on about - but some folk consider
    transformers to be "reactive loads" and a 2Amp nominal transformer can have
    a big inrush surge particularly when loaded with halogen lamps.

    Not one bit reactive.


    .... Phil
     
  20. josephkk

    josephkk Guest

    On Fri, 22 Feb 2013 22:11:57 -0500, "Michael A. Terrell"
    Oops. You are so right about looking at the wrong post headers. G2 is a
    smartphone/blackberry app though. And he is using gg and i don't get how
    that is done. G2 does have native nntp support.

    Someone should tell him about the other cheap and free news access
    providers.

    ?-)
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-