contactor coils and a funny failure mode.

I have just changed a timer for an underfloor heating rig for the second time in less than 12 months. As far as  I can tell it has hardly been used - but like all things MAPJ1 there is a twist.

To save the thermostat (rated at '16A' and a very feeble looking relay) the heating of about 20A is switched by a couple of relays, and they in turn are operated by that thermostat. but the supply they switch  comes from a pair of changeover contactors that switch between Daytime and Economy 7 power, depending on the press of a 'day boost'  over-run timer.


The contactors are Garo 20A units and work seamlessly, and the thermostat and the two relays have been faultless - but the over-run timer has now failed twice, both times when the day time power has been switched off.

My unproven suspiscion is that this occurs when the contactor coils lose supply, but the heating resistance is not in parallel due to a satisifed thermostat.

Now I have heard figures of  an inductance of 20 henry in series with a few hundred ohms for single module contactors, but I have not yet measured these units to see how true this is.

If so the kick from the contactors when supply is removed may be quite noticable, (stopping ~ 40mA each though a pair of  20H inductor I think is 33mili joules. ) but is it enough, really, to break an electronic timer.?

It does not feel quite right.

Has anyone else seen commercial grade  electronics in the vicinity of a contactor fail suddenly  when the supply stops, or analysed the back EMF kick back from contactors to see if that anecdotal figure is way off beam ?

Mike.

(yes I'm going to add a transorb anyway - I'd just like a warm feeling it was perhaps going to do some good.)

Parents
  • It's well known in the small electronics world that the back EMF from a relay turning off can fry a transistor.  The usual solution on DC is a reverse-biased diode to short out the back EMF.

    On AC, a conventional silicon diode won't work, but there are other kinds of snubbers to absorb the energy.

Reply
  • It's well known in the small electronics world that the back EMF from a relay turning off can fry a transistor.  The usual solution on DC is a reverse-biased diode to short out the back EMF.

    On AC, a conventional silicon diode won't work, but there are other kinds of snubbers to absorb the energy.

Children
  • Indeed - which is why I began to wonder about it.

    And it has occurred to me later (the joy of a beer and a bathtub) that the VA rating and the wattage of the contactor coil conspire to give me a 50Hz equivalent circuit,

    8VA implies 8/230 35mA and something like 6500j ohms or 21 Henries.

    and the 1W dissipation implies 28V of resistive drop at that current so about 800 ohms of DC resistance in the series model .

    So at the crest of the 35mA we have 50mA. in 21H so I2L = 25 millijoules. But there are two of them, so 50mJ max  stored energy.

    Now, as we have no idea what that dinky regulator chip can take we could attempt a wild order of magnitude guess and start looking at the Harry Diamond EMP report where we can take a small transistor of comparable die area to that voltage reg, say a 2n2222 (page52 )and read off, 40 watts for 10usec (400uJ) or 135 watts for 1usec (135uJ), or 2.8Kw for 100nsec (28uJ). All of which are massively less than the energy available from the contactor coil by a factor of 100 or more.

    Of course most of the spike gets spread about the wiring,  and does not end up there but even so it seems feeble, especially as it is also  massively less than even the weakest test limit in set  IEC_61000-4-5 .

    Perhaps they have not tested it...

    So, a MOV style  transorb it is. And perhaps an RC trap as well for good measure.

    Mike