Page 1 of 1

Phototransistor sensitivity question

Posted: Tue Feb 17, 2009 12:13 pm
by steve547
I seem to be running into a problem with my finish line sensors. I'm using a set of Osram SFH314 phototransistors (Digi-Key 475-1081-ND) set into 1.5" deep black tubes under the track, with ultra-bright LEDs overhead (about 6" over the track).

I'm having a tough time getting the sensors to see the lights reliably, though. The transistors just don't seem to be pulling down far enough (not saturating) to bring the sensor inputs to a "0" logic level. They're pulling it from 5V to about 3.5V or 3V. The phototransistors are supposed to be visible-light sensitive, but respond more to the red end, so swapping the white LEDs for red ones (10,000 mcd) helped but still not quite enough.

Is there something in how I'm hooking them up which might be an issue, or any way to make them more sensitive? What's working for others? Maybe shorten the tubes they're down in? Any suggestions are welcome.

I have the transistors hooked up like this:

Code: Select all

        +5V
         |
         /
         \ 10K
         /
   ______|_______>to pic input pin
|/
|\
  V
  |
 ___
  _
When this didn't work, I enhanced it to this:

Code: Select all

        +5V
         |
         /
     --->\ 10K
     |___/
         |
         /
         \ 10K
         /   |\
   ______|___| \o_____>to pic input pin
|/           | /
|\           |/
  V         
  |
 ___
  _
Not sure if continuing to fiddle with the resistor to +5V on the collector is going to help or if i just need a different transistor or simply brighter lights.

Re: Phototransistor sensitivity question

Posted: Tue Feb 17, 2009 8:29 pm
by craig_newbold
Your phototransistors are probably designed for infrared light. You might try changing over to IR LEDS for your source or just use an incandescent bulb which will have enough light, including some IR in it.

- Craig

Re: Phototransistor sensitivity question

Posted: Tue Feb 17, 2009 9:56 pm
by steve547
They're actually designed for visible light, although they peak in sensitivity around the red end of the spectrum. I have some IR emitters too, but was shying away from IR, though, because it's a lot harder to aim a light you can't see.

Re: Phototransistor sensitivity question

Posted: Tue Feb 17, 2009 11:16 pm
by Randy and Son
I don't think you need individual sources of light for each individual sensor. Simply a source of light bright enough to drive all of the transistors. If that can be done with LEDs then that's great. But it might also be accomplished with a street light too. OK, maybe that's a little overkill. But there doesn't need to be a one-to-one relationship between the emitter and the receivers. I would take Craig's advice and try a really bright incandescent bulb.

Randy

Re: Phototransistor sensitivity question

Posted: Tue Feb 17, 2009 11:28 pm
by steve547
Ok, I'll give that a try. I liked the idea of the individual lights mostly because I thought it looked better, but that's ultimately not nearly as important as the finish line sensors actually functioning :)

Re: Phototransistor sensitivity question

Posted: Wed Feb 18, 2009 5:42 am
by Randy and Son
steve547 wrote:I liked the idea of the individual lights mostly because I thought it looked better
I totally agree and you don't have to use big ol' bulbs from out of a lamp. You could still use individual bulbs, just get something with a small form factor and yet high intensity and then you could still have something with better looks. In fact, I would build some sort of enclosure around the bulbs to eliminate glare so people can see the finish easier. That would help clean up its appearance too. Just leave the bottom of it open to illuminate the track and sensor area. Perhaps a ventilation hole or two in the top. Might even line it with foil to make it more reflective. The enclosure could be decorated with a racing theme.

Randy

Re: Phototransistor sensitivity question

Posted: Wed Feb 18, 2009 7:07 am
by gpraceman
steve547 wrote:Ok, I'll give that a try. I liked the idea of the individual lights mostly because I thought it looked better
A friend has a 4 lane RaceMasterIV timer, which requires him to supply illumination for the sensors. He was using a desklamp but it was not working well with all lanes on his BestTrack.

So, I built him a LED light bar. I took a piece of leftover oak, added a push button switch, 4 ultra bright white LED's from Radio Shack, appropriate resistors, and a 9V battery connector and holder. It provides really good illumination for each lane and doesn't have the heat buildup issue of incandescent bulbs.

Sorry, no pics.

Re: Phototransistor sensitivity question

Posted: Wed Feb 18, 2009 9:01 am
by John Shreffler
Your photosensor has peak response at 870 nm, so it helps if the emitter is also putting out that light wavelength. For example using Digikey's HSDL-4250-ND should help immensely. Also, only that portion of the light cone that actually hits the sensor does any good. So go for narrow beam focus. The 4250 has 15 degree half power points, which is fairly narrow.

The higher the resistor value, the higher the sensitivity. The downside is that high sensitivity gives slower response time. But if all lanes have identical circuits, that is not very important. I use 22K, which is a good trade off. It gives reliable operation with 7 inches separation.

Putting the sensors at the bottom of a deep tube does not do any good, and has drawbacks. The light bounces around inside the tube and makes for a mushy, rather than sharp shadow. The extra distance needs to be overcome. Put the sensor right up at the track surface. As long as you are operating under flourescent room lights, there is no problem.

Re: Phototransistor sensitivity question

Posted: Thu Feb 19, 2009 2:55 am
by steve547
IR emitters? Ok, actually I have a set of those too. Am I reading the datasheet on that right? 100mA? That's a bit of a jump from the 20mA LEDs, but not a problem for this circuit.

Maybe I should just get over my reluctance to use IR sensors. I actually have IR emitters and IR-only photodiodes (with visible-light-blocking filters). Would those be a better choice?

My reasoning for going visible light was that it would be easier to see if the emitters and receivers were out of alignment, and the lights over the track make a nice visual effect (each is separately driver from the timer circuit, and as the cars cross the line, it turns off the overhead light for that lane, so you get a visual signal of the car registering... but that's more for show than for timing.

Would it hurt anything to use both? (the visible LEDs for show and the IR LEDs for actual detection?)

And now you've got me wondering about the long tubes the sensors are hiding down in. I'm not sure how much shorter they can get, though... this way they're mounted to an angle bracket bolted to the bottom of my BestTrack track, with the tubes spanning the vertical distance up to the actual holes. To get closer I'd have to actually mount the sensors to the track itself, which is certainly possible but makes it a little harder to work on the timer without lugging the track around and makes me a little more nervous about transporting the track... how much of a real effect would that have? If it's significant I'll certainly consider moving them up there.

Re: Phototransistor sensitivity question

Posted: Thu Feb 19, 2009 8:55 am
by SlartyBartFast
I've seen circuit schematics for detectors with frequency comparison to detect a specific emitter signal and eliminate ambient light issues (can't find the schematic online right now, it operated at around 48kHz) and sensitivity adjustment.

Can be implemented with standard components and op-amps or with microcontrollers.

Seems to me the best approach would be to use an adjustable sensitivity circuit so that you can properly calibrate the lanes. Alignment issues lane to lane could be overcome by placing a flat edge across the finishline and adjusting the sensitivty with a custom flate edge type go-nogo guage that can use the finish line tower as it's reference point.

On the go side, slide the guage up against the tower and the portion over the finish line can be used to adust the sensitivity for when each lane sensor "just" detects the emitters. On the nogo side, the flatedge is just a hair further over the finish sensors and the sensitivity can be verified that all lanes trip.

For the specific signal frequency circuits, I've seen circuits that use standard IR detectors and emitters and components and some that use transmitter detector units from remote controls.

Here's one:
http://home.cogeco.ca/~rpaisley4/IrProximity.html
Although it only uses the receiver. Using a transmitter and receiver pair the emitter circuit would be as simple as the receiver side.

Re: Phototransistor sensitivity question

Posted: Fri Feb 20, 2009 3:33 pm
by John Shreffler
Am I reading the datasheet on that right? 100mA?
That does not sound right, but there is a lot of errors in data specs That could be the max current at a low duty cycle pulse. Remember catalog writers are not engineers. They don't have an idea of what specs mean.

I use 10 ma of current into my emitters. You must consider that the sensitivity of phototransistors vary all over the map. At a distance of say, 7 inches separation, some will not register a thing, others will go all the way to saturation. I am not familiar with your component selection.

Send me a SASE and $1, I will return my detector pair from the Judge. They will give you a benchmark of what you should be seeking.

Re: Phototransistor sensitivity question

Posted: Sun Aug 02, 2009 8:38 pm
by altontoth
I found that using IR sensors in our Groups older track, we were just using a desk lamp from Ikea as our light source (after screwing around with IR emitters and alignment). That being said, our new timer system I built up has an overhead bar of plexiglass maybe 3/4" high, that the leds are mounted into (1/4" hold drilled into the plexi as deep as the leds). We found that this plexi provides quite a nice amount of diffusion to illuminate the sensors evenly (don't know how scientific, but it works well). Also, our resistor values for the photo transistors are only 10k, and we haven't had any issues. Again, as long as they're all the same (I use only 1% tolerance resistors in my projects), then you shouldn't end up with much variation.

Re: Phototransistor sensitivity question

Posted: Fri Jun 21, 2013 2:35 pm
by sirma
Are you sure you can use the pull down source as a sink? If yes, then please check the rating of the maximum current that it can sink. Phototransistors are normally used as switches instead of a pull up or pull down configuration. If the problem persists and you have to use the configuration, just use a comparator which converts any voltage below 3.5V to 0V using an OP470 or some other comparator.