We've been experimenting with some Velleman PEM10D sensors. (So, something like this.)

The main reason chose, was they have a 10m length (we need a good 6m or so).

I have found that, the response is a bit slow for our needs.

So for example: if you drop a basketball through the beam from a height of two feet, it is travelling slow enough it will register. However if you drop a basketball through from 6 feet height, it is too fast, it will not trigger the device.

I surmise, we need the device to be (let's say) a good ten times faster.

Now, the spec of the device suggests .. "Response time: 5 - 100ms" I don't really know what that means - is that the minimum time of cutting the beam which will register? If so, what does the 10x spread mean? Perhaps someone here will know.

Secondly, in general is this a well-known problem, are there "much faster" photoelectric beam sensors? Or no, or ...?

Thirdly, indeed could the whole thing be a fubar on our end ... say, poor programming on the Arduino, wrong ... power supply or something, or some other mess-up by us?

So, to be perfectly clear, the thing cutting the beam in our setup is a fast-moving ball-like object, which can be moving quite fast .. some 10 or even 20 m/s.

Finally I apologise if this is not in fact the optimum forum. Thx

  • \$\begingroup\$ Are you looking for a product suggestion, or advice on how you can build your own, or improve the existing product? Product suggestions are out of scope for EE stack exchange, but electronics questions definitely aren't. :) \$\endgroup\$ Jul 10, 2015 at 12:20
  • \$\begingroup\$ Commercial sensors very likely have a response speed that is deliberately slow. The typical application is for detecting people, vehicles or maybe objects on a conveyor belt, and they specifically do not want fast-moving objects to trigger them. \$\endgroup\$
    – Dave Tweed
    Jul 10, 2015 at 12:22
  • 1
    \$\begingroup\$ Perhaps if you had read the manual you would have seen the response time adjustment. Product recommendations are OT on this site, sorry. \$\endgroup\$ Jul 10, 2015 at 12:23

3 Answers 3


The manual for your sensor - briefly - mentions "response time adjustment" via a screwdriver, which would explain the spread. Some applications, such as sensing someone coming through a doorway, require longer response times, so as to not trigger separately for, say, each leg.

In general, the response time of this sensor is going to be limited by the fact that it uses an electromechanical relay, which takes time to physically switch on and off. It's likely the internal electronics were designed with low speed operation in mind as well.

Broadly speaking, if you need high speed measurements, you probably want to look for a sensor that has either a transistor output (open collector, open drain, or logic level), or speaks a communication protocol like I2C or SPI. Sensors like these are likely to be designed around higher data rates and response times. Many of these more specialised sensors may be orientated around measuring distance, rather than just detecting beam breaks, but it's certainly possible to construct a beam break sensor with a fast response time; I'm just not sure if any exist commercially.

In passing, Sparkfun have a Lidar sensor that ought to meet your requirements pretty well.

  • \$\begingroup\$ brilliant info. \$\endgroup\$
    – Fattie
    Jul 10, 2015 at 12:30
  • \$\begingroup\$ Sent a token-bounty of my thanks, cheers \$\endgroup\$
    – Fattie
    Feb 16, 2016 at 16:25

Nick's answer is good. If you want a very cheap and probably-fast-enough solution, you could consider using IR remote-control technology like a TSOP38238 receiver and a matching IR LED (950nm) modulated at 38kHz with a 555 or similar. That'd give you about 1ms response times with a CMOS-compatible output that can go directly to your AVR.

You might want to add some simple optics if you're getting too much beam scatter and detections of indirect beams; even just a baffle (black tubes around transmitter and receiver) would be a good start.


The manual is a bit vague as to the actual detector used ("retroreflection" is not a sensing method, it's simply indicating that the beam is being reflected back at the unit, so the emitter and detector are in one housing.)

In general (of the sensors I know)

  • photo-resistors are slowest (10's of milliseconds)
  • photo-transistors are faster (10's of microseconds)
  • and photo-diodes are fastest (100's of nanoseconds, possibly less)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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