4
\$\begingroup\$

I need to electrically "sense" when a brick is thrown through a vertically-oriented pane of breakaway glass, breaking it.

My first thought was to paint a thin strip of clear conductive ink/paint/gel across the back of the glass, then treat it as a normally-closed switch to trigger the rest of the circuitry.

I need this to be reliable and reproducible. Are there any reliability drawbacks to this solution, and if so, can you suggest something better? I'd like the solution to be as passive as possible at the point of the glass (that is, I don't want to do something like mount a tiny digital accelerometer unit on the glass, as this would require much more effort than I believe is necessary).

\$\endgroup\$
4
  • 3
    \$\begingroup\$ There are plenty of 'breakglass detectors/sensors' off the shelf. You should google it. \$\endgroup\$
    – m.Alin
    Oct 26 '11 at 19:20
  • 1
    \$\begingroup\$ If you have the entire glass painted then as long as any part of the glass is left unbroken then you will still have a connection. You would have to perform some sort of grid and do a connectivity check on each part of the grid or find someone that will have the resistance change once part of it gets broken. \$\endgroup\$
    – Kellenjb
    Oct 26 '11 at 20:11
  • 1
    \$\begingroup\$ It is also worth noting that there are very cheap accelerometers that can cause a trigger once they exceed a certain g-force. Might be worth looking into as well. \$\endgroup\$
    – Kellenjb
    Oct 26 '11 at 20:16
  • \$\begingroup\$ TI has a application note about doing this with a cheap MSP430 and some audio detection. ti.com/lit/an/slaa389/slaa389.pdf It detects a thud and then shatter. \$\endgroup\$
    – rfusca
    Oct 27 '11 at 3:06
8
\$\begingroup\$

Various possibilities come off the top of my head:

  1. As you said, a transparent conductive trace on the glass, then check for open. This is certainly easy to detect, but doesn't sound easy to implement on the glass. There are conductive and transparent substances, like titanium-tin oxide, but I don't think these come in anything like a paint you can apply yourself. If you did, the conductive strip would probably be noticable.

  2. IR beam just behind the glass. When the glass is broken by the flying brick something is going to move backwards and interrupt the beam. Even if it's not the brick but a few shards of flying glass it should still be quite detectable. I'm envisioning the beam being only a few mm behind the glass. I think this is the best bet as it's easy to install, should be reliable, and doesn't require additional setup each new pane.

  3. Small piezo microphone stuck to the corner of the glass somewhere. The breaking glass will make strong high frequency sounds that normal activity should not. This might take a little experimenting to get just right, but should be doable. My guess is that rejecting low frequencies will be important. Trigger off of high volume high frequency.

  4. Load cell weighing the window. After the glass breaks, the setup should reliably weigh less. Depeding on your mechanical setup, you can weigh just the glass in the frame, or possibly the whole frame. A load cell can provide a analog signal that is a function of force (weight in this case).

  5. Load cell on the floor behind the glass. The weight of the fallen shards of glass should be detectable easily enough. Possibly just looking for sudden changes is good enough.

  6. Like #2, but refecting at a shallow angle off the glass instead of passing just behind it.

  7. Ultrasound pings accross the glass. The intact glass will conduct ultrasound nicely, with the air gap after it breaks much less. The propagation time will also be longer, but I doubt you'd need to bother detecting that.

  8. Ultrasound sonar far enough behind the glass to be out of the way. The flat intact glass pane will reflect the sound well. A broken pane a lot less.

While it's easy to dream up Rube Goldberg approaches to this problem, many are going to be fiddly to set up. I would go with #2 (interrupted light beam just behind the glass). I think that is easy to set up and should be quite reliable.

\$\endgroup\$
4
\$\begingroup\$

Because you specifically say "vertically oriented" and "brick", I am assuming that this is for something special, like a theatrical performance, and not a security system. If this is for a security system then I'd use stuff specifically for that and not what I'm about to suggest below.

For this, the simpler the better. Use a simple push-button switch mounted at the top edge of the glass. When the glass is in place, the button is pressed. When the glass is broken, the glass falls out and releases the button.

If the glass is very wide, it might not break correctly to allow it to fall out. In this case, get a glass cutter and pre-score the glass so it breaks correctly. This would probably be a couple of horizontal lines across the entire width of the glass, evenly spaced top to bottom.

\$\endgroup\$
2
\$\begingroup\$

For alarm systems, there is a sound-based glass break sensor. They probably wouldn't be too hard to use, although the exact trigger time might not be fast enough for your needs. It could detect breaks from multiple panes, but would not be able to distinguish which one broke.

Try googling "glass break sensors"

Also, you could try placing a microswitch on the top of the glass that would open when the pane shatters.

\$\endgroup\$
0
\$\begingroup\$

If your goal is to detect the event as quickly and consistently as possible, conductive paint will probably be broken by crack propagation even before the pieces of glass had a change to move much.

An even faster approach, depending upon the characteristics of the glass in question, might be to illuminate the pane from one edge (possibly modulating the beam at 1KHz-100KHz or so) and try to sense that illumination from the other edge. Ordinary glass is pretty lossy, but one may be able to get a detectable amount of light through. Any crack between the light and the sensor, though, would destroy the light-transmission properties.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.