Searching \ for '[EE:] Cheap contact sensors?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/io/sensors.htm?key=sensor
Search entire site for: 'Cheap contact sensors?'.

Exact match. Not showing close matches.
PICList Thread
'[EE:] Cheap contact sensors?'
2004\06\12@024344 by Rick Luddy

flavicon
face
I'm working on a project to build a dance pad for video games.  It needs
four sensors, each one detecting whether or not an 11"x11" panel has
someone standing on it.  The end result will look similar to:
www.angelfire.com/d20/ddrhomepad/ddrpad1.html
, although I plan to implement it differently.

Most similar projects (including the example above) work by sandwiching
some non-conductive foam between two conductive panels, detecting when
the two panels touch.  For both durability and comfort, I want there to
be no humanly-noticeable displacement when a panel is triggered.  I'm
having trouble finding a type of sensor that can accomplish this.

Each sensor needs to be durable enough to survive being repeatedly
stomped on while also being able to quickly (within ~1ms) detect contact
or release.  Each needs to be able to detect both sharp (milliseconds)
and gradual (several seconds) contact and release.  Each must detect
contacts  anywhere from a couple pounds of force up to hundreds.  Each
has to be able to work no matter what causes the contact, anything from
an elbow to a heavy rubber boot.  Each also needs to be able to detect
prolonged contact anywhere from milliseconds to [at least] 20 seconds.
All of this with a maximum low-volume cost of $2-5USD per sensor.

Another requirement is that after initial testing, it shouldn't need any
manual calibration.  If it needs to self-calibrate, it should be able to
do so upon power-up and then work properly for at least several hours
without any user-visible calibration (for instance requiring the user to
step off).

The sensors don't need to determine how much force is being used, only
whether contact is present.  Ideally the sensors should be directly
readable using a PIC with analog inputs and minimal external circuitry,
but any reasonable amount of external circuitry would be fine.

I have been looking at cheap piezo transducers, but I'm not convinced
that they will be able to accurately detect gradual or prolonged
contact.  Reading a bit it appears to me that these are unreliable below
10 Hz, and I don't know if software tricks would be enough to compensate.

I've also considered capacitance-based non-contact sensors, but I'm not
sure these could detect contact from heavy rubber boots while avoiding
false positives from bare feet just slightly above contact.

The only sort of sensors I've seen that would seem to work are
heavy-duty load sensors, in the neighborhood of $50-$500 each.  The
budget for the entire electronics portion of the project is $20-50, so
these wouldn't do.

Any advice is greatly appreciated.

-Rick Luddy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@032402 by Russell McMahon

face
flavicon
face
> I'm working on a project to build a dance pad for video games.  It needs
> four sensors, each one detecting whether or not an 11"x11" panel has
> someone standing on it.  The end result will look similar to:
> www.angelfire.com/d20/ddrhomepad/ddrpad1.html
> , although I plan to implement it differently.
>
> Most similar projects (including the example above) work by sandwiching
> some non-conductive foam between two conductive panels, detecting when
> the two panels touch.  For both durability and comfort, I want there to
> be no humanly-noticeable displacement when a panel is triggered.  I'm
> having trouble finding a type of sensor that can accomplish this.

Some variation on common membrane keyboards sounds like it should suit.
Longevity would be the greatest challenge. Basically you have 3 plastic
"membranes". The top one is bottom metallised and the bottom one is top
metallised and the middle sheet has suitably sized and located holes in it.
The holes are arranged such that target force will deflect the top layer
down to touch the bottom, but that the sheet will return without significant
deformation when the force is released. AFAIK the normal material is Mylar
which should suit your application. As you are effectively building a large
1-key keyboard you could probably use a gross conductive material for the
upper and lower surfaces. If you used eg suitably impregnated butyl-rubber
for upper and lower sheets then the middle spacer sheet would have a
mechanically relatively undemanding task. I you effectively have many small
"keylets" the force to operate each can be manipulated by design to suit
your needs.

Another possibility is a sealed "bag" where standing on it causes increased
gas pressure inside.

You could also look at what the security industry does for under-mat
sensors.

Looks like a fairly challenging cost target. Aren't there easier things to
make ? :-)



       RM

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@035137 by hilip Stortz

picon face
strain gauges might work.  basically you'd firmly support the corners
and or edges of each "pad" of the switch and put a strain gauge right in
the middle, using fairly heavy metal (for minimum deflection) or even
wood/plastic for the pad you should still be able to detect the
resistance change.  depending on the application, you could have a
stronger support very slightly separated from the sensing surface for
safety (i.e. so a truck can drive over it or a very violent strike won't
break it).

you should have good frequency response and dc sensing is no problem.
if you don't know what a strain gauge is just try a google search, they
are fairly common and the basis for all electronic scales.  you may need
to buy them surplus too meet those cost constraints and for a non
critical application like this any good epoxy should work rather than
the very good epoxy normally recommended.

another possibility would be an ultrasonic sensor which could tell if
there was something touching the other side of the plate.  you'd want to
put the transponder right against the metal and would have to ignore the
immediate reflections from the surface of the plate, or you could use
something like a piezo bonded to the plate in a resonant oscillator
circuit and detect the frequency shift/damping/oscillation failure as
the plate was loaded and made more stiff.  again the pad would be
supported around the edges and the sensor would be in the middle.  you
might have some problem with mechanical noise but it would also be an
indicator so might not be a problem.

Rick Luddy wrote:
>
> I'm working on a project to build a dance pad for video games.  It needs
> four sensors, each one detecting whether or not an 11"x11" panel has
> someone standing on it.  The end result will look similar to:
> www.angelfire.com/d20/ddrhomepad/ddrpad1.html
> , although I plan to implement it differently.
-------

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@035344 by Robert Rolf

picon face
Interlink makes force sensing resistors whose resistance
changes from megaohms to low hundreds of oohms when loaded with 50#.
About 0.2mm thick and 20 mm diameter. No perceptible motion.
Only problem is creep when under constant load (10's of minutes,
100's of ohms).

For binary sensing as you desire, should not be a problem.
Interface is an op amp to linearize, or comparater/schmidt
if you don't care about linearity. We use them in shoe liners
to control a medical stimulator. Biggest problem we have
is flexure failure due to sole flex (printed ink on mylar
cracks with repeated flexing).

You might also want to look at pain old membrane key switches.
They are available in singles. It's just a matter of devising
a suitable panel geometry to focus the force onto the actuation
area.

And your $50 package will sell for how much?

Robert

Rick Luddy wrote:
{Quote hidden}

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@045931 by Rick Luddy

flavicon
face
Robert Rolf wrote:
> And your $50 package will sell for how much?

It's primarily a hobby project.  Some friends of mine want to run a
tournament with eight pads.  High-quality commercial pads cost $200-500
  each plus ~$40/pad shipping and they can't afford $2000-5000 on what
would just be a for-fun event.

I was a bit imprecise when I mentioned $50; the target budget per pad is
$100.  I've set aside $50 each for electronics/sensors and $50 each for
metal/acrylic/miscellany.  As students, we have a great deal more time
than money so we can afford to invest lots of time in this :).

-Rick Luddy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@051008 by SavanaPics

picon face
Ok hee i an off the wall idea, but sometimes such things work.  I read a long
time ago.  that someone wanted to roll theirown capacitors. The solution was
to use tin foil and wax paper. Perhaps this may work in your  situation in
which each pad is sandwiched with tin foil and wax paper and then measuring the
capacitance or either time in an RC configuration



Just a thought
Eddie Turner, kc4awz




--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@051423 by Russell McMahon

face
flavicon
face
>As students, we have a great deal more time
> than money so we can afford to invest lots of time in this :).

Aha.
Harder but cheaper ...

Main panel sits in a secondary panel - still quite thin. Tact switch set
into hole in base (or possibly switches at each corner - or at several
places across pad - probably 5 is ideal) is/are activated by downwards
pressure on pad BUT completely protected against any amount of force as
system bottoms out before switch reaches allowable max travel. Cheap,
potentially robust.

XXXXXXXXXXXXXXXXXXXX
     |                    |                      |
XXOXXXXXXOXXXXXXXOXX

View with constant width font.
X is pad
O is tact switch set into lower pad.
| is switch actuator

At full compression switches are safe

XXXXXXXXXXXXXXXXXXXX
XXOXXXXXXOXXXXXXXOXX


       RM

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@052916 by Jake Anderson

flavicon
face
if you wanted to get more funky replace switches with optical encoders
tab sticking down
opto sideways, tab moves through opto
all done

you sure you guys need the milisecond timing and stuff about people slowly
stepping on it things like that?


{Original Message removed}

2004\06\12@053330 by hilip Stortz

picon face
OH, i thought you meant $50 total, which would make it hard.  with your
budget, i'd definitely say strain gauges and comparators.  very robust,
decent frequency response and no perceptible movement (they will sense a
very slight deflection of the plates).  plus no problem at all with
dc/continuos sensing and minimal calibration needs (basically one time,
i.e. a pot for the reference level on the comparator), and you can get
some that are temperature compensated as well depending on your needs.
you should also get nearly unlimited lifetime.  for this application you
can just use a resistor in series with the strain gauge to the supply
line rather than the usual current source since you don't need a precise
analog output.

Rick Luddy wrote:
>
> Robert Rolf wrote:
> > And your $50 package will sell for how much?
>
> It's primarily a hobby project.  Some friends of mine want to run a
> tournament with eight pads.  High-quality commercial pads cost $200-500
>    each plus ~$40/pad shipping and they can't afford $2000-5000 on what
> would just be a for-fun event.
----------

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@065542 by Rick Luddy

flavicon
face
Jake Anderson wrote:
> you sure you guys need the milisecond timing and stuff about people slowly
> stepping on it things like that?

On the most extreme possible settings (300bpm, within 1/32 beat), to get
perfect points one pad can need to be hit within +=6.25 ms, giving a
12.5ms window.  On more reasonable settings, it's more like two to ten
times that, but some of the better players actually approach the limits.
 There's also an unavoidable 1-5ms between each time the console polls
the pad.

As far as people gradually stepping on pads, it does happen
occasionally.  Step-and-hold happens all the time, though, and I think
most of the difficulties in detecting step-and-hold are the same as in
detecting gradual pressure.

-Rick Luddy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@074601 by Samuel BOUQUET

flavicon
face
Why not using deformation of an optic fiber?

-----Message d'origine-----
De : pic microcontroller discussion list [spam_OUTPICLISTTakeThisOuTspamMITVMA.MIT.EDU]
De la part de Rick Luddy
Envoyé : samedi 12 juin 2004 12:55
À : .....PICLISTKILLspamspam@spam@MITVMA.MIT.EDU
Objet : Re: [PICLIST] [EE:] Cheap contact sensors?


Jake Anderson wrote:
> you sure you guys need the milisecond timing and stuff about people
> slowly stepping on it things like that?

On the most extreme possible settings (300bpm, within 1/32 beat), to get
perfect points one pad can need to be hit within +=6.25 ms, giving a
12.5ms window.  On more reasonable settings, it's more like two to ten
times that, but some of the better players actually approach the limits.
 There's also an unavoidable 1-5ms between each time the console polls
the pad.

As far as people gradually stepping on pads, it does happen
occasionally.  Step-and-hold happens all the time, though, and I think
most of the difficulties in detecting step-and-hold are the same as in
detecting gradual pressure.

-Rick Luddy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@080259 by Russell McMahon

face
flavicon
face
> There's also an unavoidable 1-5ms between each time the console polls
> the pad.

I doubt it.
If the timing matters to you that much then interrupt driven response would
allow you to refine this timing vastly.
Even polling at 100 uS to 1 mS rate would not place an excessive burden with
a realistic cpu clock speed.
Overhead depends on how many pads are controlled by one processor but if
there is one cpu per pad-set (whatever you call it) you should be able to do
it sub mS with great ease. If you have many pad sets per cpu you may want to
be using a super cheap cpu per pad-set.


       RM

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@152138 by William Chops Westfield

face picon face
On Saturday, Jun 12, 2004, at 02:33 US/Pacific, Philip Stortz wrote:

> OH, i thought you meant $50 total, which would make it hard.  with your
> budget, i'd definitely say strain gauges and comparators.

I think I agree.  Strain gauges are normally expensive, but that's
because you're paying for precision and accuracy.  You ought to be able
to find some bare. uncalibrated, uncompensated, strain sensors for
pretty cheap with a little looking that will work fine for "digital"
applications...

BillW

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2004\06\12@213251 by Rick Luddy

flavicon
face
Russell McMahon wrote:
>>There's also an unavoidable 1-5ms between each time the console polls
>>the pad.
>
>
> I doubt it.
> If the timing matters to you that much then interrupt driven response would
> allow you to refine this timing vastly.

The reason for this constraint is that we're not building the console or
the console software.  It's a commercial Playstation 2 using an existing
game.  It reads the pad using a synchronous protocol clocked by the
console.  I've tested three consoles, getting intervals of 5ms, <0.1ms,
and 2ms (in order of purchase date).  I'm not sure whether the
difference is from manufacturing changes or just really poor tolerances.

-Rick Luddy

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

More... (looser matching)
- Last day of these posts
- In 2004 , 2005 only
- Today
- New search...