arcing mains switch -> add cap in parallel -> WHY?

[DF96]

No, don't put an MOV across the switch. This is dangerous, as it means that mains spikes will be sent through to the equipment even when it is switched off! It probably won't be as effective in dealing with the problem either. MOVs are for solving a quite different problem.

Use a cap or snubber, ideally across the transformer primary.

 

[AndrewT]

The MOV usually goes from Live to Neutral. It is intended to soak up voltage pulses coming in from the mains.
The snubber to damp/absorb the switch off back emf is across the load, i.e. from Live to Neutral.
Putting the snubber (R+C) across the switch leaves the equipment Live even when the switch is set to off.

 

[tauro0221]

Hi,
Normally I installed my parallel to the transformer and ground. Like AndrewT mentioned. In that way it will damp the kick back from the transformer primary when the amplifier it is turn ON/OFF. Also the voltage rating should be to not too high. I used one with 140volts for 120volt AC working voltage.

 

[tommydoan84]

Thank you DF96, AndrewT and tauro0221

 

[tommydoan84]

Hi,
Normally I installed my parallel to the transformer and ground. Like AndrewT mentioned. In that way it will damp the kick back from the transformer primary when the amplifier it is turn ON/OFF. Also the voltage rating should be to not too high. I used one with 140volts for 120volt AC working voltage.

taruo0221: why we should use the low voltage rating for MOV ? I thought we should use the high voltage rating to damp the high voltage spike caused by the transformer, no ?

 

[AndrewT]

we usually use a 275Vac or 300Vac MOV or Varistor for our 220/240Vac supplies.
Since you have 110/120Vac supplies you can use a 150Vac MOV, or Varistor.

 

[tauro0221]

Hi,
The reason it is that you do not want the voltage of spike goes too high. The 400 volts it is too high.

 

[Pafi]

I agree,

Just for fun..

Does a piece of equipment have to have a switch at all?
When does switching for convenience become isolation?

If we call a switch an isolator it now probably won't have the air gap to be classed as one?

Remember this is just for fun..and regulations will be different in all situations.
Does a lap top PSU have a switch? Does a soldering iron have a switch..(maybe)
Does a flat screen TV have a switch..but is it for convenience IE touch it and it comes on....Strange nothing on the back for isolation..
Could a tube amp be in standby and a touch switch turn it on?
Would it be classed as a portable appliance...?

Does it have isolation (mains plug)..?

Regards
M. Gregg

I like your questions. They tell more than most of the statements.

 

[Pafi]

Ignoring for a moment the safety side of the equation,
all putting a capacitor across the switch (especially something that's only 10nF) does is lower the resonance to something that won't radiate into the rest of the circuitry. with only 10nF and reasonable leakage inductances, there will still be an arc created in the switch, and you might even blow the capacitor after some time.

That said, the impedance of the ac line is not negligible.
20 meters of 12 awg wire spaced 1 cm apart is something like 12 uH
--that will resonate with 10nf at 460Khz and an impedance of 34 ohms. -low enough that when you break 4-12 amps of current, the peak voltage is only 1-3 times the nominal line voltage.

This is perfectly fine of course, until you add another 100uh to 1mH of leakage inductance from the load side. i'll let you do the math on what voltage that will drive a 10nf cap at reasonable turn off currents...

So, just toss a 10-100 ohm resistor and a .1 to 1 uF X rated cap across the load side of the switch, let the emi filter on the line side take care of the line side impedance
(The resistor is there to save the switch from turn on current.)

Im always happy when somebody tries to calculate instead of just telling stereotype values ignoring the actual circumstances.

I used 220nF+150 ohm as long as I used mains transformer and couldn't make a proper GND+signal path. But in the past 15 years I forgot about these snubbers, since I didn't hear popping sound.

 

[DF96]

all putting a capacitor across the switch (especially something that's only 10nF) does is lower the resonance to something that won't radiate into the rest of the circuitry. with only 10nF and reasonable leakage inductances, there will still be an arc created in the switch, and you might even blow the capacitor after some time.

Not true. The capacitor does lower the resonance frequency, but it also lowers the switch voltage - this varies like the inverse of square root of capacitance.

What happens when you break a circuit is that the kinetic energy stored in the inductance (1/2 L I^2) gets dumped as potential energy (1/2 C V^2) into the capacitance (which may be mainly just strays). In many cases the resulting voltage is so high that the air in the gap breaks down. Raising the capacitance reduces the voltage.

 

[AndrewT]

.................the kinetic energy stored in the inductance (1/2 L I^2) gets dumped as potential energy (1/2 C V^2) into the capacitance (which may be mainly just strays). In many cases the resulting voltage is so high that the air in the gap breaks down. Raising the capacitance reduces the voltage.

Have you any guidance on typical values of inductance for loads/transformers that are being switched off?

What if the switch off has been triggered by a fault/short at the output of the amplifier?

 

[DF96]

Sorry, no idea.

The basic aim is to replace 10's or 100's of pF of strays with 1-10nF, so the voltage comes down. The energy can then get dissipated in the wiring/transformer resistance. If in doubt, add a resistor too.

 

[pound]

A click at the point of switch-off may be an arc at the switch. The capacitor will stop this. For a cap bypassing the switch I think I would use something a bit smaller than 10nF, as this will let enough current pass to make you jump (although harmless). 10nF across the transformer primary is fine, and will have a similar effect.

A thump a few seconds later is the amp circuits switching off as the PSU caps lose charge. Similar thump at turn-on.

does that mean a 10nF across the primary is fine will not make one jump also? I hate shocks.

 

[DF96]

Yes.

 

[ctrlx]

Sorry, no idea.

The basic aim is to replace 10's or 100's of pF of strays with 1-10nF, so the voltage comes down. The energy can then get dissipated in the wiring/transformer resistance. If in doubt, add a resistor too.

i have a 100nf x2 type cap - is this value too high to go across the transformers primary?

 

[DF96]

100nF should be OK if that is what you have.