Tech / Special Interest → Electronics > how to simulate bouncing contacts?

how to simulate bouncing contacts?

Wonder what would be the best way to simulate bouncing contacts in a test fixture.

Would it be possible to drive FETs fast enough and do the bouncing in SW, or look for some ways to modify relays to be 'bouncy'?

I'm thinking that switching several times the rated current should age contacts very quickly.

TIA
--
Wacky Races 2012!

NM misread your question.

What are you trying to do, simulate a relay contact? I'd use the relay. Why are you going over the contact current or is that just a test?

What are you switching, AC, DC, quenched or not?

The relays should be bouncy by nature. They're a mechanical contact just like a switch (I'm assuming you're talking about an actual mechanical -click click- relay, not an SSR). Honesty, just turning the relay on should give you a bouncy enough input for a test fixture. A latching relay might might be a little more bouncy than a standard relay, as it doesn't rely on a magnetic field to keep the contact closed.

I'm trying to simulate contacts in a consumer device I cannot name. (sorry, very strict NDA )
Based on some cursory testing (scoping), occasionally they can bounce *much* worse than a relay, but they don't always do. There may also be a little bit of delay between some lines depending on the orientation of the mating surfaces.

The lines contain either low level DC (9V, up to 100 mA) or TTL level signaling. Because they can be operated by an Average Joe Blow, pretty much everything imaginable is a fair game.

Relays are normally designed to minimize bouncing, so if I go that route, I'm either looking for some really crappy relays or some that are inherently worse in that respect, or for some way to make them bad, like burning up the contacts before testing. (that was the best I could come up)
--
Wacky Races 2012!

The first crazy thought that came into my head would be to send some short bursts of audio into a Schmitt Trigger. Just adjust the audio level to the amount of bounce you desire.

Hmmm, interesting idea.

Agree with lutful if you are trying to design a manufacturing or engineering test fixture program up a low end micro to drive a MOSFET.

Programming the micro is easier then trying to tweak audio tones and is extremely flexible as to bounce characteristics: cycle time, on time, off time, total bounce period and whatever else you want to throw in. Really depends on what characteristics of the DUT you are trying to test.

/tom

And here I thought bounce was sposed to be random, and well bouncy.

quote:

---

I'm trying to simulate contacts in a consumer device I cannot name. Based on some cursory testing (scoping), occasionally they can bounce *much* worse than a relay, but they don't always do. There may also be a little bit of delay between some lines depending on the orientation of the mating surfaces.

The lines contain either low level DC (9V, up to 100 mA) or TTL level signaling. Because they can be operated by an Average Joe Blow, pretty much everything imaginable is a fair game.

---

Good point - I was thinking about push buttons not micro switches.

To be honest never did a detailed analysis just brute force or if I had the option used a SPDT switch. I've never seen a switch bounce all the way to the other contact.

/tom

Have a solenoid push a doorbell button. Feed the solenoid with random levels of signal (music?). Can't get more realistic than that...

Among other things we have a couple of these: »www.triextech.com/Products/Relay···Pro.php# but since it's USB controlled (plus the app is written in C#) I don't think I can switch the FETs fast enough to realistically simulate real life bouncing.
--
Wacky Races 2012!

here's one you can try. »www.freepatentsonline.com/3925723.pdf

Discussed this a bit more, and it is going to be part of another test fixture where the FETs will be driven by an FPGA.

According to the EEs, some FETs can have switching times measured in nanoseconds, so they should be more than fast enough for bounce simulation. The exact implementation of bouncing is TBD, but we'll definitely be using some ideas from the above patent. There will be a random number generator, be it a linear congruential generator or LSFR that can generate repeatable semi random sequences when started with the same parameters. In addition, there is also a possibility of an audio input and a built in white noise generator (just a zener with an op-amp or two) if we want to use something that is not based on SW.
--
Wacky Races 2012!

Reading through this thread I've been wondering... why is such a test fixture even necessary?

Assuming the DUT consists of momentary push-buttons and a microcontroller which together form the physical user interface, why not just gather contact bounce measurements from whatever switches you're using, and then tweak the programming of the microcontroller such that it will ignore the worst-case measured bounce with an extra margin?