Long Story Short

It's a "linear/rotary optical encoder," or, "incremental encoder," like this one, which costs about \$6:


You use it with "optical encoder strips," or "codestrips," like this one, which costs about \$.30:


It uses a single LED with 2 photodetectors. These are exposed as pins labeled Channel A and Channel B. The pins will output either 0 or Vcc V, which is 5V in my case. Since there are 4 distinct states for A and B, you get 4 times the resolution of your codestrip. Since it can handle up to 150 lines per inch on a codestrip, you get 1 600th of an inch resolution. When monitoring the A and B channels, you can determine which direction you are moving by comparing your current state to your last state. For example if A and B are both high, and then A goes low, you've moved up 1 600th of an inch.

It is a great, accurate, affordable system for determining precise linear position.

Original Post

IR emitter / detector pulled off an inkjet printer

The red arrow points to the ir emitter / detector. The blue arrow points to where I got it. The printed circuit board under the blue arrow would slide back and forth under that gray band, which is transparent plastic with lots and lots of very skinny black lines. That was probably how the inkjet printer kept track of the position of the print head.

The 2 ir diode pins were easy enough to figure out. That's being powered by the fat white wire on the far left. I can't figure out the additional 4 pins on the detector part though. Notice, that's a total of 6 pins. I pulled the cover off the part to get a closer look, and those pins look like this, with the skinny white wire attached to the pin on the far right:

Diagram of what the detector part looks like

The black cover on the ir e/d has tiny white letters:

9981 C526

and tiny black letters

Agilent 18

I tried Googling around for datasheets, but I didn't manage to find anything. I tried calling Agilent but just got lost in the phone tree, and they hung up on me at some point, haha.

Does anybody have a datasheet for this or a similar component, or have any idea how this 4 pin detector works, or know where to look for a datasheet or who to call, or know a component I could substitute, or anything really?

This is all part of a bigger thing I'm trying to get a handle on, which is precise 1 dimensional movement, and then ultimately precise 2 dimensional movement.


To be clear, there are 6 pins total. 2 for the diode, and 4 for the detector. I don't understand what all 4 pins on just the detector are for.

Edit 2

I found a very promising blog post about this kind of problem. http://benkrasnow.blogspot.com/2010/02/linear-position-tracking-with.html

Here's a 6 pin linear optical encoder with data sheet which I found from this blog post.


  • 1
    \$\begingroup\$ You might want to read up on graycodes (en.wikipedia.org/wiki/Graycode) -- this encoder gives you a position error of at most 1 bit. So with a 2-bit encoder, you have 1 bit of accuracy - so a 150 lpi strip would give you 1/150-in. resolution. (BTW, this assumes that a black mark and a white space count as two lines for LPI calculation purposes - so in one inch there is 75 black and 75 white.) \$\endgroup\$ – Toybuilder Apr 8 '11 at 7:12

I believe your part is a dual channel IR position sensor. I can't find your exact part, but I did find a datasheet for something very similar although this has 8 pins, this is the OPB822:


However, yours only has 6 pins. I guess that as you suggest, two pins are the infrared LED, so I think that the two detectors share the same LED.

The other four pins are the opto couplers, and you can identify them using an multimeter that can measure the voltage drop of a diode. Try touching two of the probes to two of the pins while switching the transmitter LED on and off, if the reading changes with the LED, then you've found two of the pins you need.

The dual sensors can then be used to sense which direction its travelling along the strip, and how far.

  • \$\begingroup\$ I started working through your suggestions last night. I tested every pair of pins in both directions, and did find a pair of pins that dropped a different amount of voltage depending on if the LED was lit, so hurray! It's a very small amount, about .14 volts. The datasheet seems helpful but I'm struggling to understand it. Any idea what 'Output Phototransister: Collector-Emitter Voltage: 30 V, Emitter-Collector Voltage: 5 V" is all about? \$\endgroup\$ – Dave Aaron Smith Feb 1 '11 at 16:47
  • \$\begingroup\$ This website has a pretty good explanation of Collector-Emitter voltage: johnloomis.org/ece445/topics/egginc/pt_char.html \$\endgroup\$ – BG100 Feb 1 '11 at 16:58

Similar products are simply an IR LED + phototransistor. That makes for the 4 leads.

Look for a diode with a multimeter, polarize it, and you should get the other 2 pins to conduct.

  • \$\begingroup\$ Ensure current limiting.. \$\endgroup\$ – tyblu Jan 31 '11 at 4:44
  • \$\begingroup\$ Oh crap, maybe I wasn't clear enough. The diode is 2 pins, and the detector is 4 pins, making for a total of 6 pins. \$\endgroup\$ – Dave Aaron Smith Jan 31 '11 at 12:51

In case anyone else stumbles upon this IC,

I also just happened to find one of these in an old printer. After some research, i found this datasheet: http://www.datasheetarchive.com/AEDS-9620-datasheet.html, which is pin-compatible and looks pretty much the same as the component in question. I tried hooking it up as described in the datasheet and it turns out to be working. I Used a 220 Ohm resistor at 5V for the emitter, and the two outputs show very nice square-waves when the sensor is slid along the printer's code strip.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.