In which I look to see how colours other than the six expected by the Boost colour sensor are detected, using the Powered Up app. What I found was…unexpected.

We recently saw the release of the Powered Up 2.1 software, allowing the use of the Boost sensor (colour and distance) with the Powered Up Smart Hub. While experimenting with the use of coloured tiles on the track to control the train, it became apparent that perhaps the dark blueish grey (Dark stone grey) of the track ties were occasionally read as blue or black, and the reddish – brownish-darkish – orange of the table was also detected as ‘Red’ by the system. This has implications for using the sensor in an environment where other colours might be present.

So I thought I’d put the system to the test. Here is the World’s simplest Boost Program: It tells you what colour is currently being detected by the scanner.

I aimed to test as many colours in the modern LEGO Colour palette as I could, to see how they might be read and reported by the sensor. I worked my way through the 2016 colour palette, with the addition of teal – as seen in this year’s brick separators.

Something that became apparent as I progressed was that some colours were unstable in the colour detected – varying a little as they approached the sensor. There was particularly a threshold at approximately 10mm and another around 4. Sometimes, when the brick was placed in contact with the sensor, aberrant behaviour was observed – yellow being measured as red, and so forth.

So here is a summary of what I was able to observe. There are some colours I don’t have easy access to… Light nougat (except as Minifigure’s head) and olive green in particular

The sensor returns a result of Black(Blu), Blue (Blu), Yellow (Y) Red (R) green (G) White(w) and nil (x)

Colour	1.5 cm	5 mm	Others colours seen
Bright Yellow	y	y	red in close opposition
Cool Yellow	y	w	yellow – red in close opposition
Bright Orange	R	R	X in close opposition
Flame Yellowish Orange	R	R
Bright Red	R	R
New Dark Red	R	R	@2cm Unstable between Black and red
Bright Purple	R	R
Light Purple	Y	W	red and yellow as distance approaches 1mm
Bright Red Violet	Red	Red
Medium Lavender	Y	W	Blue at just over 1.5 cm; Y @2mm
Lavender	Y	W	Blu just over 1.5cm. Y @ 2mm
Medium Lilac	Blu	Blu
Bright Blu	blu	Blu
Medium Blue	blu	Y	Also W &Y as approaching sensor
Light royal Blue	Blu	W	Passes Briefly through yellow, then Blu at 1mm
sand blue	Blu	b	Y-W as closer to sensor
Earth blue	Blk	Blu	Blu @ 3mm
Dark Azur	Blu	blu	Blu @ 1mm
Medium Azur	Blu	Blu	Green@1mm
Teal	G	Blu	Green @1.4cm
Spr Yellowish Green	G	W	Yellow @ ~1cm
Aqua	W	W	Yellow and Blue just outside this range (ie>1.5cm)
Bright Green	G	Blu
Sand Green	Blk	Blu	Thin band of green@~1.2cm
Dark Green	G	G
Earth Green	blk	blu
Bright Yellowish Green	G	Y
Olive Green
Reddish Brown	Blk	Blk	Red@2mm
Brick yellow/Tan	Y	W
Sand Yellow/Dark Tan	Y	W
Medium Nougat	Y	Y
Dark Brown	Blk	Blu
Dark Orange	Y	Y
White	W	W	This bands of blue and yellow around 2cm
Medium Stone Grey	y	w	band of blue at 1.7cm
Dark Stone Grey	Blk	Blu
Black	Blk	Blk	Blue @ 3mm

Occasionally, colour detection would be very unstable, and unable to maintain any value. typically it beyond 2cm or <2mm:

I found the variability in the colour detected to be quite interesting: I wonder if, in part, it is due to reflection from of the sensor’s light (which seems to be mainly blue), but might also be due to ambient light. Other things that can’t be completely excluded might be fluorescent pigments in the bricks, with some UV being generated by the LEDs in the sensor. My knowledge of physics is a little limited

So, if we are looking to have a model to be dependent on the colour of tiles triggering a given response, then we need to depend on colour for which there is relatively little crossover at the range considered.

With the train control we looked at last week, perhaps we should be limit the signal tiles to White, Red, Blue and Green. Yellow and black were frequently triggered by other colours, not specifically related, but many might be used in railway ballast or landscaping. Yellow was particularly frequently detected with some of the brighter greens, as well as many of the browns at around 1.5cm.

I appreciate that the accuracy of sensor is good for ‘near vision’ of the colours intended, but the LEGO colour palette is much broader than that, and there is every possibility that people will be looking for reproducible behaviours. I’d be excited to see if people are able to exploit this variability in response to good use.

Otherwise, stick to the basic colours: they are reliable, and don’t tend to cross react in the usual range.

Make of it all what you will. Until Next time:

Play well.

While you are here, have you heard about our building contest, where you can win a set of Disney Series 2 minifigures? You can read about it here. Entries close May 12th 2019