I was trying to measure the bounce time of a tactile switch so that I
could design a debounce circuit based on a timer. But when I look at
the switch output on a scope, I don't see *any*! The output is as
clean as it gets and does not show bouncing down to 100 ns/div.

I am very used to scopes and I understand what can be hard to see and
what can be easy to see. The sort of bouncing that a switch normally
has should be very easy to see on an old fashioned triggered scope.

This switch has no electronics inside, it is just a mechanical
switch. From the feel, it is a dome type with a dome switch with a
snap action. But I would still expect to see bouncing.

One of the five on the board actually seems defective in that the
bouncing is to a point that it is intermittent. But I really do thing
it is defective. The other four are just fine.

Anyone seen this before?

On Jun 13, 9:09 pm, rickman <gnu...@gmail.com> wrote:
Seen it.

The ones I'm familiar with are domes that are soldered over a pcb
contact.

Probably something about the shape of the contacts and the pressure
involved; I never saw any bounce.

G.

rickman wrote:
Did you look at both edges ?

It can also depend on the load - For example, a switch with some
capacitance, driving a 80C51 port, has a large current ratio between
activate, and any bounce. So Circuit capacitance can slow the rise
enough, that there is no apparent bounce in the voltage domain.
(even tho there was in the physical-contact domain)

You can also get a 'bouncless' switch action, with a a SPCO Microswitch
into a CMOS input. With this, the first contact make, charges the
capacitance, and a later bounce-open, has no threshold-sized effect.
When the Switch is releases, the other contact pulls the IP to the
other supply, and again a bounce-open is held-over by the circuit
capacitance.

-jg

rickman wrote:

rickman <gnuarm@gmail.com> wrote:
Yes, but only on some brand new switches of the small-piece-of-steel-taped-
on-the-PCB-type. Bouncing seems to increase with age, probably due to
corrosion and piling up of dirt. I wouldn't count on the switches to
stay bounce-free in the future, though.

--
:wq
^X^Cy^K^X^C^C^C^C

On Jun 14, 5:09�am, rickman <gnu...@gmail.com> wrote:
You should look at the problem from the other direction. How fast can
a finger operate the switch? I use 20mS for each edge.

On 14 Jun 2008 12:30:25 GMT, the renowned Ico <usenet@zeev.nl> wrote:

Yes, I've noticed this too- even with certain designs of mechanical
relays. It follows that one should not assume that debounce routines
are working properly unless they actually see some bouncing. I
generally use switch datasheet values for maximum bounce time.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com