When using an O-scope testing switches?

In the context of switches used for electronic system inputs mostly, rather than switching high power levels...

When a current carrying switch opens the contacts do not break instantly, more like they tear apart like velcro on the microscopic scale, so there can be a noisy change in contact resistance as it opens, depending on factors like the switch design and material, contamination and aging of contacts and current level. This is because the surface of the contacts is not smooth at the microscopic level, and there can be contaminants or oxides present. This period can be brief, so usually well controlled by the filtering necessary for the switch closing.

When the switch closes there is contact bounce, so the contacts physically open and close several times. This lasts for several milliseconds, and has large changes from off to on etc..

Either way the switch does not open or close cleanly or instantly, so some sort of filtering is needed when connecting it to certain circuits like the clock input of a flip flop (multiple counts can occur) or a microprocessor input which could sample the wrong state or multiple states.

Filtering might use a delay circuit (taking into account how the states are being used), or a latch circuit sensitive to one state to remember and hold the change in state cleanly, or the software sampling a microprocessor digital switch input could force a delay long enough to guarantee the bounce has ended and then re-sample, or perhaps compare samples after a delay designed to allow time for a contact to bounce. Note that edge triggered clock lines usually need signals that rise and fall very quickly, so this needs consideration too if a filter is used. Use a level sensitive device like a voltage comparator or schmidt trigger following the filter. With low level analogue circuits like the inputs to a sound system, the switch can cause plops, clicks and worse. This is difficult to eliminate with mechanical switches. Make before break contacts might work.

There needs to be a milliamp or two of DC "wetting" current for most contact materials to work with low level signals and especially low level AC signals. This is to help break down metal oxides forming on the contact surface.

There might also be high frequency oscillations (ringing), due to circuit stray reactances (L and C) resonating at some frequency in the broad range of noise frequencies present.

In the past mercury wetted relay contacts were used to overcome many of the above contact issues. The thin liquid coating of metallic mercury on the contacts breaks and makes cleanly so there is no bounce. They are still available to some extent, mostly the reed relay version. These are more or less discontinued with bans on use of heavy metals like mercury, so ruthenium dry contact relays are an alternative sometimes. Solid state analogue switches are a way to switch low level signals in many situations.

Later edit...

I took it in the context of when the contacts are opened, rather than open, which leads to all the stuff about contact noise. In this case it does depend on the circuit, so that AC signals from the mains etc in a room will be capacitively coupled to the test circuit.

Consider a likely 30pF (or more) from 240V = 680V p-p to a 1 megohm pullup resistor for the contact, assuming the pullup supply is grounded. The Xc is ~100megohms, so about 1% of the mains voltage appears (6.8V p-p). The oscilloscope probe is 10 megohms and pehaps 10pF pulling down, so that will reduce it slightly.

Well, this is a weird question. First I'd ask if the switch is connected to an AC line and plugged in. Because I would tell the interviewer that I'd never check a switch plugged into an AC plug with a scope that is dangerous to the scope and the operator.

Assuming that the switch is on the low voltage side of a circuit. If you were to connect a scope to an open switch, the scope is picking up noise from the components of a circuit. The PC board traces, resistors, and other components are picking up noise from the outside world. The biggest source of noise in a lab are the fluorescent lights. Even unterminated scope probes will pick up noise from fluorescent light. The high impedance of a scope makes it susceptible to noise.

Ok, so your switch is open and you're reading a series of components and circuit traces that are connected to one side of a circuit and your high impedance O-scope lead (1M as a guess) on the other. These components and traces represent capacitance, impedance, and resistance between the scope lead and ground with some non zero length of conductors involved. Basically what you're hooked to when it isn't switched on is one end of an antenna. The scope lead is picking up electromagnetic interference (EMI) from the surroundings such as florescent lamps, radio signals trapped by circuit traces and elements which create tuned circuits at certain frequencies, Also if there are operating transformers or cooling fans nearby their magnetic fields can induce stray voltage. Basically your O-scope lead is reading what amounts to a lousy antenna created by circuit elements downstream of the open switch.

How are you testing them? What is the circuit?

Just connecting a scope across a switch will tell you nothing.

edit, you need to know the circuit. As another answer said, this makes no sense, unless you have the specific circuit. And, if anyone asked this (exactly the way you phrased it) in a job interview, then that interviewer is a total idiot.

.

While the switch is open the noise shows up as a very small voltage on the o-scope. The first picture at this link:

http://www.labbookpages.co.uk/electronics/debounce...

shows that the noise is the part on the flat spots of the trace.

---------------------------------

When a switch opens and closes the voltage across it spikes up and down until the voltage levels off. This is called switch bounce. The switch actually has a springiness to it so it opens and closes in milliseconds.

The following link explains this dirty little secret of switch bounce and they also have o-scope pictures of what the switch bounce looks like:

http://www.maxim-ic.com/app-notes/index.mvp/id/287

Switches do have switch bounce. That's a literal bouncing state and it can last for a few milliseconds. This question is not very specific.

You could be picking up RF, but it's likely not too noticeable.

challenging stuff. lookup in search engines like google. that will could actually help!