Switch lifetime at a capacitive load

[nbucska]

The lifetime of a switch is usually 10,000 cycles at AC
line voltage. How does the lifetime changes if I use it to
switch a 12V circuit which has 1000 uF input filter
and much lower steady state current ?

<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[itsmoked]

If you are saying there is a 1,000uF cap across the contacts I would guess that the switch would probably last its mechanical life as compared to its contact life. I would expect no arcing or heat.

 

[sdmays]

nbucska, are you saying across the input of your circuit you have a 1,000uF cap and you are wondering if this capacitive load would affect the contact life with the surge current associated with DC input voltage source suddenly applied to an otherwise discharged cap? I suppose there is a chance the contacts could slightly weld shut on occasion if the contacts are rather 'light'.

 

[nbucska]

The load is 2W at 12V i.e. 72 ohm . The steady state current is 166 mA.

The circuit has 1000 uF parallel to the input
i.e. at turn-on the current is limited only by
the wire and the battery's internal resistance.



<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[nbucska]

How do you select a switch? Do you use a swich
specified for 60 amps if the parasitic resistances are
.2 ohm ? What is a reasonable safety margin ?
361 seems to be slightly too much !

<nbucska@pcperipherals DOT com> subj: eng-tips
read FAQ240-1032

 

[jraef]

One rule-of-thumb that I have used for years is to look at the rated current capacity of the swith, and derate that by 125% for a purely capacitive load. So if you have contacts rated for 10A resistive, they will be good for only 8A of purely capacitive current.

The theory behind that was that the cap switching causes a slight amount of additional heating effect, so by reducing the current capacity rating of the contacts, you minimize the net effect that would have on their lifespan. If your contacts are already rated to switch far more than you are switching, then as itsmoked said, it probably doesn't matter.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[VEBill]

The problem is quite real. I've got some equipment at home that turns on with a mightly snap (at the switch contacts) due to this issue.

One option is to add some series resistance (a few ohms) to the circuit so that the capacitor isn't charged 'instantly' by battery when the switch is closed. Also, re-examine the decision to use 1000uF - perhaps it should be less ?

The switch manufacturer should (in theory) be able to provide some technical data regarding the capabilities of the switch when faced with such capacitive loads.

Another thing to keep in mind is that the peak current would depend on the type of batteries used. If the user switches to NiCd battries (for example), then the peak current would be much more than with regular alkaline batteries.

 

[OperaHouse]

A lot of AC line rated switches fail from oxidation at low voltage and current. That cap might even increase the emperical life of the switch by keeping it clean. I don't think 1,000uf is that much at 12V. I always like isolating caps from the power with something. A 1-2 ohm resistor wouldn't hurt anything and reduce the current. Might consider a small inductor. If you are asking how many cycles, I sense you are using this cap as an "off delay" and you will be pulsing it quite often.

 

[jraef]

Thisis a little more conservative than the rule-of-thumb I was using, but it addresses the same issue.

http://myronzucker.com/docs/CAPACITALK 105.pdf

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more."
Nikola Tesla

 

[BrianR]

The usual de-rating factor for a lamp/capacitice load is to de-rate it to 20% - 30% of the resistive load current.

Eg 115VAC 10A = 28VDC 10A resistive or 2A lamp load.