Sensing Impact

[RyanG]

I am trying to make something to detect an impact and close a circuit momentarily following the impact. In the picture you can see the idea. The impact will come from the right hand side and the sensor will be on the back of the piece that takes the impact. I have thought about a few ways of doing it, like a tilt switch or something along those lines, but I am not sure it would work. Does anyone have a suggestion? Preferably, as simple as possible with as few pieces as possible.

[Rerouter]

Analog accelerometer chip and a comparitor for the electrobic approach,

For the mecahnical approach, a microswitch connected to a mass, if the impact is enough to impart monentum to the mass the microswitch toggles.

You never clarified is this single activation or  more, and the likely impact force.

[Nusa]

Look for "vibration sensors". There's stuff out there from $1 and up.

You need to nail down your requirements more. How much of an impact? How solid is the object and does it move when impacted? Will there be impacts from other directions and do they need to be ignored? Will the object reset to a stable state after impact? Etc. Some answers may require experimentation.

[frozenfrogz]

There are some solutions on the market you could implement. The cheapest and simplest "sensors" consist of a spring, that closes a circuit when inducing vibration, or a sideways force exceeding the projected value. Search for 'impact switch' or 'acceleration switch'.
The more sophisticated way would be to use an active electronic solution like a MEMS accelerometer (ADXL, or similar). But there are also pieco-electric sensors that might be suitable.

It all depends on what exactly you need to detect (momentum, acceleration, duration, ...), how precise you need to measure the event and how tolerant you can be in terms of false, or no detection.

Edit: typo

[calexanian]

The poor mans version of this is a piezo element and a filter. Could even do it with a DSP micro.

[RyanG]

There is a wide range of impact force values that it would have range from about 100 Newtons up to around 2500 Newtons. The impact side of the object being hit is rubber and the whole thing does move somewhat when hit. It does reset to its starting position very quickly after being hit. I would probably only need to detect impacts in 1 direction, with all the others ignored. I would also want a minimum amount of acceleration to cause it to trigger so that just moving the object by hand wouldn't set it off.

[frozenfrogz]

Force on its own is somewhat irrelevant, as what you have discribed so far is about acceleration, inertia and impulse, where time is of concern. You need to determine a range of g-force the switch should operate at.

Switches from Select Controls
Switches from Comus
Switches from Inertia Switch

[DTJ]

+1 for the cheap piezo sensor.

I've seen a few different domestic solar inverters that use this method as a switch for changing the displayed data.

[MagicSmoker]

And an even poorer man's version is a largish, High-K MLCC capacitor! The Y5V and Z5U dielectrics are hugely microphonic. For example:

https://e2e.ti.com/blogs_/b/precisionhub/archive/2014/12/19/stress-induced-outbursts-microphonics-in-ceramic-capacitors-part-1

[Len]

And an even poorer man's version is a largish, High-K MLCC capacitor! The Y5V and Z5U dielectrics are hugely microphonic.

Dave shows that in a recent video:

https://youtu.be/WE9pYUVvr00

[MagicSmoker]

Dave shows that in a recent video:

Wow... I never would have thought to test my scopes by whacking them with a plastic rod! 

Okay, to be fair to Dave, that first R&S is surprisingly sensitive. That could be really annoying if just tapping on the touch screen to save a waveform caught in single shot mode caused the scope to trigger again.

That said, it's hard to avoid the problem these days - High-K MLCCs are used everywhere, and damn near unavoidable if you want your product to be compact.

[RyanG]

If I do the piezo, how do I adjust the sensitivity? The signal voltage that I have to switch is 32vdc

[CatalinaWOW]

If you want something relatively inexpensive, very robust and widely available the "ping" sensors used in newer automobiles may be of interest.  Given that you have a rubber bumper they may not respond to low enough frequencies, but that comes back to the homework you need to do first.  You have to define the impacts you wish to respond to in terms of frequency content, amplitude and repetition rate (the latter will affect several things including how long you have to perform signal processing to discriminate between desired shocks and "noise").

[RyanG]

Thinking about it again, I think that the piezo wouldn't work because I would like to detect impact from 1 direction only. I think the accelerometer and MCU is probably the best option. Would an ATtiny work with a suitable accelerometer for the impacts I need to record?