I like to set up some QRE1113 Reflective Sensors to count bees and copy this project http://www.instructables.com/id/Honey-Bee-Counter/ The disadvantage of the documented setup is that you need with 44 QRE1113 also 44 input pins. So I decided to use a Mux (or more precisely 3 x 16-Channel) to reduce the number of required input pins. This posting is about the sensor part of the QRE1113 not the IR-LED element.
First I set up the QRE1113 referred to this schematic http://dlnmh9ip6v2uc.cloudfront.net/datasheets/Sensors/Proximity/QRE1113-Analog-Breakout-v11.pdf and did some test. All worked great! You can also use the ~20K internal pullup from the Arduino or a 100K pulldown as described in the original Instructables what leads to different sensitivities, but all worked smooth without the Mux.
In the final version I will use some 16-Channel Analog Multiplexer, CD74HC4067, http://www.sparkfun.com/datasheets/IC/cd74hc4067.pdf. But for testing i grabbed this 8-Channel chip: MC14051B, http://www.produktinfo.conrad.com/datenblaetter/150000-174999/172901-da-01-en-4051_4052_4053.pdf in a local store (the CD74HC4067 was not available).
With the internal pullups enabled I got a enormous sensitivity fare above the spec and not usable (it switches in a range of 70 mm, spec is around some mm, also a halogen bulb in the room "fires" the sensors). Now I tried to replace the internal pullups of the Arduino with external pullups. But I can see only different values on the input with different pullups (1K, 10K, 20K, 100K) but the sensors do not switch by approximation.
I noticed that the MC14051B has a 250 Ohm Typ. “ON” Resistance (I measured a resistance of 100 to 110 Ohm between in and out pin of the Mux when the channel is switched). Could this be the problem with my poor readings? Is this “ON” resistance the reason, or other stuff? How can I fix the problem? Is the CD74HC4067 better (datasheet says 70 Ohm Typ. “ON” Resistance). Or is there a general problem with such sensors as the QRE1113 and multiplexing?