Tap Sensors for LED Table Lighting Ideas

[buildernick]

So I'm building this glass table, and I'm going to put a few led's around the edges so i can light up the table if i need to find something in the dark (like while watching a movie). and I want a cool way to turn on the light, so I was thinking about making a circuit that when you tap the table fairly hard it would light up for a few seconds and then slowly fade to off. I'm a mechanical engineer so my knowledge of circuits is limited. the problem with using a microphone is that it will pick up other loud sounds, like a slamming door etc. so if you have any ideas let me know.

 

[berkeman]

What kinds of uses will this table get? If it will routinely get used for putting drinks on, then tapping to turn on the LEDs will have to be in some pattern (like 2 or 3 quick taps as a code to eliminate false turn-ons from placing a drink on the table).

The usual way to do a tap sensor for a surface like this would be to find the resonant frequency of the table (assuming that nothing ever gets put on it), and tuning a bandpass filter (analog or digital signal processing-based) to that frequency. But again, if the table is sometimes loaded with stuff, and other times empty, then it will be harder to detect a tapping to trigger the LEDs.

Alternately, you could put some sort of touch-sensor arrangement around the circumference of the table, just underneath so it's not visible. It could be as simple as a couple of exposed wires, about 1/4" apart. To turn on the LEDs, you just touch the underside of the table near the edge, and your circuit would detect the drop in resistance between the two wires, and turn on (or off) the LEDs.

 

[dlgoff]

"Alternately, you could put some sort of touch-sensor arrangement around the circumference of the table, just underneath so it's not visible. It could be as simple as a couple of exposed wires, about 1/4" apart. To turn on the LEDs, you just touch the underside of the table near the edge, and your circuit would detect the drop in resistance between the two wires, and turn on (or off) the LEDs."

Or you could use an optical sensor for the touch operation.

 

[buildernick]

Thanks for the ideas.

yeah it would usually have varying things on it, but I don't care if it gets turned on inadvertently by putting drinks down. I'm thinking I should make more of a neat switch that's built into the frame around the table instead, like you said with the two wire approach.

 

[berkeman]

Or you could use an optical sensor for the touch operation.

Hey, that's an even better idea! dlgoff is right -- since the glass table is clear, you could just put an emitter/detector sensor under the table aimed up in one (or more) spot. When you touch the table in that spot (or set something on top of that spot), the light is reflected more and that can be used to toggle the lights on and off. Good idea!

 

[berkeman]

Well, whichever way you choose, be sure to post some pictures when you get it done. Sounds like a fun project!

 

[chroot]

The optical sensor idea is getting close to a "frustrated total internal reflection" touch-detection system.

You could fire some infrared LEDs sideways into one (or two) edges of the glass at a slight angle so the energy is totally internally reflected. On the other two edges, you'd put some infrared photodetectors. Whenever an object touches the surface -- any object -- it would spoil the total internal reflection. At least one of the sensors would detect a change in the infrared light intensity, and you could use that to turn on your visible lights.

Of course, the angles and so on might be fairly difficult to get right so this thing is triggered when you want it to be triggered.

Another option is to use an IC accelerometer bonded to the bottom surface of the glass in an unobtrusive spot. You can certainly get one sensitive enough to detect the kind of shocks you'd get from rapping on the glass with your knuckles.

- Warren

 

[buildernick]

yeah I like the accelerometer idea
so i guess i'd need to figure out
-if i epoxy the sensor to the glass, what kind of range would the accelerometer need to detect the vibrations
-does anyone know a cheap source for this kind of thing
-how would i transfer the output of the accelerometer into turning led's on or off

 

[buildernick]

i just found this little project, which is sort of exactly what i need

http://www.dimensionengineering.com/appnotes/alarmclock/alarmclock.htm

 

[chroot]

You should be able to go to Digikey or Newark or any other distributor and look up IC accelerometers and their datasheets. You should be able to find one that has the right sensitivity range.

Getting the data out of the accelerometer, and using it to trigger the lights, is a separate issue. You'll need to settle on an accelerometer first. The easiest way would be to use an analog accelerometer, connected to a small microcontroller with an ADC. The microcontroller would be programmed to trigger the lights whenever the ADC input crosses some threshold.

Another bonus of using a microcontroller is that you'll be able to control the light duration, intensity, fading, and other effects with it.

- Warren

 

[Danger]

You could also cover the rim or underside of the glass with transparent conductive paint and use it as the 'antenna' for a capacitance sensor. You wouldn't have to even physically touch it in that case; just wave your hand under or over it.

edit: You'd have to set it to very low sensitivity so inadvertant approaches such as reaching for a drink or stretching your legs under the table won't trigger it.

 

[chroot]

Do people make transparent, conductive paint?!

- Warren

 

[buildernick]

are there any analog ways for say if i had a switch that would go high momentarily that would cause the led's to turn on, stay on for a few seconds and turn off, or dim, like would having a ciruit that charges up a small capacitor and then running the leds off of that work?

i think I'm going to go the analog accelerometer route. just need to figure out how to make the leds stay on

 

[Danger]

Do people make transparent, conductive paint?!

Yeah. I'm not sure where to get it any more, since it's considered obsolete for the original purpose. We used to paint stripes around the permimeter of windows with it for burglar alarm applications where the client didn't want foil strips showing.

 

[dlgoff]

" i think I'm going to go the analog accelerometer route. just need to figure out how to make the leds stay on"

Should be easy using an op-amp to trigger a 555 timer.

 

[berkeman]

Do people make transparent, conductive paint?!

- Warren

Yeah, Indium Tin Oxide -- it's used for the conductors in liquid crystal displays:

http://en.wikipedia.org/wiki/Indium_Tin_Oxide

 

[xez]

Yeah, Indium Tin Oxide -- it's used for the conductors in liquid crystal displays:

http://en.wikipedia.org/wiki/Indium_Tin_Oxide

True, but I'm not aware of any very simple to use product
form of ITO. Usually depositing it in a film requires
high vacuum sputter deposition equipment, or moderately
complex / difficult chemical deposition processes.

More readily available but not transparent choices
would be:
* some of those flat thin "printed wire"
copper foil over flexible mylar backing kinds of
"invisible speaker wire" type cable.

* using some 30ga. (tiny) wire-wrap wire or similarly
thin magnet wire; if you route it out of the way and adhere
it sensibly it would be very low profile.

There are places that sell conductive polyanaline
(a polymer) paint solutions, but they're rather expensive
and not all that commonly available.

You could use a R/S flip flop to toggle things on/off, or
as others have said, something like a TLC555 (the CMOS
versions will have lower power requirements and will let you
use higher resistances and smaller capacitors) for a
triggerable timer that will go off after a some time.

However since you STILL need to detect / interface the
sensor in the first place, why not use a low power
microcontroller like the Texas Instruments
MSP430F2013 which has a (slow) 16 bit analog to
digital converter and a couple of I/O bits that you could
use through PN2222 transistors to turn LEDs ON/OFF.

You could probably use an infrared LED and
a phototransistor hooked to the I/O pins to detect the
beam being modified by a hand / finger.

Or you could use a piezo disc as an accelerometer to detect
somewhat sharp tapping; I'm not sure how well that'd
work going into the slow sigma delta ADC converter; my
guess is that you'd need to oversample it at something
like 1kHz-4kHz thus losing a lot of bits of resolution, but
all you'd need is about 1-2 bits of useful "is it generating
a spike?" information anyway. Or you could feed the
piezo into a comparator with hysteresis and use that,
but in that case you might as well just use a comparator
and TLC555 + transistor for the LEDs.

Actually if you just had a small electret microphone you
could give it (directly) a flick and that'd work well.

Or for that matter if you're willing to touch the sensor itself,
they have a MSP430F2013 application note showing how
to do a capacitive touch sensor with just a couple of
resistors and the chip itself.

 

[buildernick]

thanks for all the ideas

the reason i wanted a tap sensor is because, when it's dark, i don't want to have to be searching for a switch to turn on the led's, i want to be able to turn it on no matter where I'm sitting at the table.

i don't think i want to use a microcontroller because of cost and i don't feel like learning how to program one.

so i'll probably make it with either some sort of microphone or accelerometer and a 555