Ultra Cheap flexible optical linear sensor


Phew!! - Yes it is a mouthful.
Ok - was looking at a cheap way to make a sensor. It needed to be…

a) Sensitive enough so I could connect it to a lip or eyebrow using makeup glue,
b) Fast enough to follow the twitches of the muscle,
c) Small enough so that it could be fixed to a face mask
d) Cheap… cheap… nearest equiv was $400 Ea…
e) Flexible to fit over curves.

Yeah - I was asking a lot. I started to look at LVDT’s but they were expensive $$$, Then stretchable carbon elastics but the strength to move them was not linear. and many many others sighs

Finally I had the idea of using a easily printable gradient fill scale using GIMP. Once I had that I could laminate it for reliability.
Next I had to read from it - Using a QRE1113 in Analogue mode I hooked it up to my trusty Arduino Magpie (cause I love the blinking lights :slight_smile: ) and scanned the QRE1113 module over the paper.
It worked - I got some lovely readings that were very precise AND repeatable at any spot on the lamination.

So using some flexibile polythene I made a little enclosure with a channel inside, stuck the QRE1113 on to it and fed in a slip of the laminated paper. This is the prototype and it can be slimmed down a lot. It can measure factions of a mm movement and depending on how good your laser printer is will depend on how fine a scale you can make. My target was to measure 30mm max but it will work with a lot less. Also the orientation of the QRE1113 makes a difference too in sensitivity.

Here are some pics of the item :slight_smile:


That’s cool.

I see the QRE1113 is available in different mountings. Looks like you used the Sparkfun breakout.

A suggestion; you can use overhead transparency film in many laser printers. That might avoid the lamination step. Though make sure the upper surface doesn’t slide against anything with a hardness greater than the baked toner. I’ve just checked, and transparency film is now about $3 per A4 sheet retail in .au; $76 for a 25 sheet pack.

Linearity is probably enough for your needs. You might make it even more linear by measuring the reflectance at several points, drawing a curve that relates image pixel brightness against measured, fitting a formula, then adjusting the gradient by coding the formula into a python plugin for The Gimp.


Thanks for the suggestions. I’ve had to put this on the back burner for a short while.