Phototransistor Edge Detection

[slurp]

I'd like to invite comments of the first of a series of application sheets introducing modules for a MiniSumo 'bot.

Constructive comments are welcome on content and format.

Best regards,
Colin


http://www.minisumo.org.uk/EdgeDetection.php

(page update to reflect the move to php)

[slurp]

There are a few alternatives to the OPB608B device we've quoted, many will bring the price of edge detection down. Check the data sheets as these may not be pin compatable...

SG-2BC (Kodenshi Corp.)
QRD1113 / 1114 (Fairchild)
KTIR0811S (Kingbright)
SFH7221 (Siemens) - Surface Mount
GP2S27 (Sharp) - Surhace Mount

regards,
colin

[PeterW]

I am using the SY-CR102 from Maplin at 67p each.
These comprise a surface mount sender receiver in one package which is 3.2x2.7x1.5 mm excluding the leads.

[slurp]

Just a thought or two about these in the run up to the AWS 20 minisumo competition, something I nearly forgot about when putting my servo 'bot back together with a new program...

I use these with an A to D input on the microcontroller, first in mind was the black to white transition for edge detection.

Testing on the sumo ring I got readings of:-

1023 black
970 on black/white edge transistion
160 white

While setting the threshold for action at 970 could be good for a fast edge response might be good it's worth thinking about the Sikiri-Sen (brown starting lines). The bot returned a reading of 850 on this ring, this points to a program threshold somewhere a little lower.

It's worth a thought that the competition dohyo may differ a little from your test surface. Think about building in a margin for error or having some means of adjustment.

regards,
colin

[Gary]

On Origami I used Osram SFH309FA-4 phototransistors (in 3mm LED package) with Vishay TSAL4400 IR LEDs (also 3mm).

The LEDs are spaced 0.2" from the detectors (they are adjacent on a piece of Veroboard). The detector assemblies are about 3-5mm from the floor and are mounted out of direct light under the robot's "skirts".

I used the phototransistors in "switching" configuration fed into a Schmitt-trigger input on the ATmega8. I may have just been lucky, but I have had no problems with edge detection or the brown starting lines. The inputs I used on the micro' also feed into the ADC but I haven't had to use it yet.

I found this solution a lot cheaper than buying the complete modules and also more reliable. I did some experiments with a reflective opto-switch that I had lying around and it seemed very fussy about the sensing distance. My discrete solution works happily over quite a range of distances and I have had no problems with direct pick-up or ambient light. I do not even bother to modulate the LEDs!

[slurp]

From inital finds discrete components where much cheaper but price gap to the some of the packaged components isn't significant pennies.

I was surprised to see that you didn't have anything to stop direct light transfer from emitter to detector but I guess that depends on how hard you're driving the emitter.... the packaged components don't have the problem but do tend to have specified dimensions for surface stand off.

I've not had a problem with these, the devices I've used could stand 2.5 to 5mm. I did note Peter had a problem moving from his darker workshop to the competition venue - didn't take much to fix the problem tho', I've not looked to see what the spec says on the sensors he's used.

regards,
colin

[slurp]

I had an email from Jie Chen who noted our page swapped the Vcc, R and Icc relationships for Active and Switch modes of phototransistor operation - now corrected.

Thanks,
Colin