Straight Dope on Traffic Light Sensors

[DoktoroKiu]

As a formerly ignorant driver of 4-wheeled vehicles, I never put much thought into how to trigger the demand-driven traffic light sensors.  Now that I've taken to two wheels I've heard a lot of rumor surrounding these mystical devices.  You always hear about people attaching expensive magnets to their bikes to trigger them, so I looked into how they actually work before buying into the idea (there's a lot of debate about whether magnets will work).

I found a short article on the subject written by an IEEE member with a PhD (more than my measly BSEE) that explains in (mostly) plain English how the detectors work.  Basically, a magnet will do little to help you, and you can help trigger a light by properly placing your wheels/engine.
Here's the link:
http://www.humantransport.org/bicycledriving/library/signals/detection.htm

Here's my attempt to summarize it (and to describe some things you might not know if you're not well-versed in electricity):

Most traffic light sensors are inductive loop sensors, and you can identify them by the cut-marks in the pavement.  There are a few different configurations, but the simplest has one large "inductive loop" made of several turns of wire that have been buried in the ground.  The buried loop acts as a big inductor and is used with a capacitor to make an oscillator circuit that typically has a frequency in the tens of kilohertz range.

Think of an inductor like a mass that stores energy in its motion, and think of a capacitor like a spring.  Combine the mass and spring, and add some energy by compressing the spring and it's the same basic premise used in an LC oscillator.  (Note: this isn't just an analogy; the exact same equations are used to describe them, only the constants are different)

When current moves through a wire, a magnetic field "flows" in a circle around the wire, and is strongest right next to the wire.  If you make a thumbs-up (using your right hand) with your thumb pointing in the direction of the wire, the magnetic "flux" points in the direction your fingers curl, and is "densest" closer to the wire.  So our big loop in the ground ends up making a large donut-shaped magnetic field that is strongest right over the loop.

When you have alternating current the magnetic field alternates with it, and due to that pesky physics stuff the changes in magnetic flux through a conductor induce changing electric fields (voltage) in anything conductive that the flux passes through.  The most effective way to use this is to have a conductive ring, so the induced voltage makes a current flow through the ring.  Here's a brief demonstration of the effect using rings with and without cuts in them to show the difference:
https://www.youtube.com/watch?v=Pl7KyVIJ1iE

What they make use of in that experiment and in these sensors is that the induced current always makes a magnetic field that opposes the original one, and when this happens it changes the inductance in the oscillator circuit (changes the mass of the weight in the mass-spring oscillator), which causes the frequency to change.  They detect this change in frequency and then trigger the light.  At higher frequencies non-ring-shaped metal will produce stronger eddy currents that have the same effect (this is why motor/generator stators use laminations instead of solid chunks of metal, as it reduces the eddy currents).

So to maximize your chances of setting of the sensor you basically want to set your bike right on top of one of the sides of the loop in the ground so your rims have the maximum flux passing through them and your engine block is as close to the wires as possible.  If you park in the middle or with your front wheel on the front side of the loop (wheel is perpendicular to the loop) you will have no flux in your rims and very little through your engine.  There are other configurations with two loops in a figure eight, or with diagonal loops, but the principle is the same.

Also, the reason magnets wont work is that they produce a constant field and will not decrease the inductance any more than a similarly sized chunk of non-magnetized metal.

Hope that wasn't too confusing for anyone, and I hope you feel like you learned something if you stuck around.  Hopefully there will be less  next time you need to trigger a light.

[RossLH]

Also, the reason magnets wont work is that they produce a constant field and will not decrease the inductance any more than a similarly sized chunk of non-magnetized metal.

I've tried explaining this to a handful of people, and they just can't seem to wrap their heads around it. The inductive loop doesn't care about the magnetic field of the magnet, it just sees the magnet as a tiny bit more metal. You'd get a stronger effect by putting the bike on its side stand.

[DoktoroKiu]

Also, the reason magnets wont work is that they produce a constant field and will not decrease the inductance any more than a similarly sized chunk of non-magnetized metal.

I've tried explaining this to a handful of people, and they just can't seem to wrap their heads around it. The inductive loop doesn't care about the magnetic field of the magnet, it just sees the magnet as a tiny bit more metal. You'd get a stronger effect by putting the bike on its side stand.

The only effect the magnet would have is a tiny bit of induced current in the buried loop, but that would be insignificant since you are already mostly stopped.  And then you've got zero chance of it being fast enough to somehow match a 20kHz signal.

I wonder if there's a case for false advertising claims against companies that sell magnets for triggering lights.  They are ripping people off and wasting rare earth materials for nothing.

[Alan_nc]

I read all the reports.  Basically they say there is nothing you can do except have a mass of metal over top of the wire.

I had to change my route to work because I never could get the left turn arrow at a busy intersection to change for me.  I even asked the local police....they just shrugged.....

If it's not busy I just run the light.

[kensully]

My buddy claims that a1lb magnet will trip the sensor.
I have not attempted it myself

There is no excuse for what I'm about to do

[Electrojake]

Next topic. . . The deer whistle. 

[Big Rich]

If it's not busy I just run the light.

There are some states that have laws regarding this topic. In Pennsylvania, I think the law says that either you have to wait for one full cycle of light changes without you getting a green light (so obviously the sensor wasn't tripped), or you have to just treat your red light as if it was a stop sign for yourself (and everybody else has a green light). I haven't looked too far into it since I rarely ride at night.

[Alan_nc]

Deer Whistles ............

Your right.....maybe they need to be a topic by themselves.

I live in rural NC.  I have deer whistles on my car, truck and all my street bikes.

Do they work..... I don't know....but I do know the deer look up when I drive by.

[yamahonkawazuki]

I read all the reports.  Basically they say there is nothing you can do except have a mass of metal over top of the wire.

I had to change my route to work because I never could get the left turn arrow at a busy intersection to change for me.  I even asked the local police....they just shrugged[\quote]
Here in tn we can run them USING due care and caution

If it's not busy I just run the light.

[DoktoroKiu]

I believe that it was only very recently in Indiana that they added provisions to run the light for motorcycles.  I think you have to wait 2 or more cycles, though.

My buddy claims that a1lb magnet will trip the sensor.
I have not attempted it myself

There is no excuse for what I'm about to do

A 1lb magnet would add a significant amount of metal, but it would also subtract a significant amount from your wallet for no good reason.  You could strap a big chunk of aluminum on the bike and get better results.

[Kiwingenuity]

I used to install these traffic loops.. the main issue is that the inductance loop is sized for cars more than it is for bikes (you can install bicycle and motorbike loop sizes). Same as the head of a hand held metal detector - more turns = higher sensitivity, but you can only get so many turns in the slot.. so you make the loop bigger and hope more of the car covers it.

In Australia and NZ they use a standard pattern - concrete saw to cut a slot which the turns of wire are installed. Usually Tixophalte or some other type of bituminous sealant is used to hold them in.


The biggest issues was resealing over the top would decrease the sensitivity (about 2-3" maximum seal depth), and quite often gravel would get forced into the cut and break the loop. A broken loop should normally force the controller into assuming there is a vehicle always there and force a light change. That can be an issue if you are parked on a good loop and the opposite direction is broken (and how the program is set up to run).

If you can line up your wheels over the cut marks quite often you can get them to trigger the controller sensor circuit. we used to test with a steel toolbox or short length of bar (red item in the drawing). Magnets really have no effect at all since the magnetic field is running at about 20kHz- really you want ferrous materials such as steel belt radials which affect the oscillator circuit sufficiently.

All being said, I can admit to having waited for a few light changes and just made a careful run for it - sometimes there is no other option..

[Watcher]

Just throwing it out there, in most cases in the US a right turn and a u-turn aren't illegal, nor do they take away that much time compared to just blowing the light after waiting a full light cycle.

Over here we call that a box turn. One Right and 2 lefts make a left, and 2 rights and 2 lefts make a straight.


Its only considered dodging the light if you do so through a parking lot, and that will get you in trouble.

[DoktoroKiu]

Just throwing it out there, in most cases in the US a right turn and a u-turn aren't illegal, nor do they take away that much time compared to just blowing the light after waiting a full light cycle.

Over here we call that a box turn. One Right and 2 lefts make a left, and 2 rights and 2 lefts make a straight.


Its only considered dodging the light if you do so through a parking lot, and that will get you in trouble.

On most of the larger roads in my area that make use of inductive sensors it would be impossible to do a U-turn due to a large median, or even full-length medians/separate road surfaces. 

I think the most effective plan of attack is to threaten legal action if they don't keep them in working order.  Nonfunctional sensors encourage riders/drivers to disobey the traffic signals, and I'm sure you could sue for some type of entrapment or discrimination or hot coffee to get them to take the issue seriously.  The law in Indiana is to wait two full minutes before proceeding, and on many lights that's more than one or even two cycles.

I personally don't understand the purpose of roads with left-turn arrows that prevent people from turning when the light is green for straight-through traffic.  Do they really believe us too stupid to make a safe left turn just because the intersection is slightly larger than normal?

[bombsquad83]

If you can line up your wheels over the cut marks quite often you can get them to trigger the controller sensor circuit.

I've tested this on the past few commutes, and it really does work.  It's too early to say if it works 100% of the time, and I'm sure it depends on the condition of the sensor, but I've already gotten a few lights to change for me when there was no other car there to trigger it.  Thanks for the thread!

[DoktoroKiu]

If you can line up your wheels over the cut marks quite often you can get them to trigger the controller sensor circuit.

I've tested this on the past few commutes, and it really does work.  It's too early to say if it works 100% of the time, and I'm sure it depends on the condition of the sensor, but I've already gotten a few lights to change for me when there was no other car there to trigger it.  Thanks for the thread!

Good to hear.  I have yet to ride where there are no cars to trigger the sensors on my routes, but I'm confident that this technique will trigger a properly functioning sensor.

[hInstance]

Years ago when I only had a bicycle I played around with these sensors in the road. An old ballast from a fluorescent light fitting could trigger them. And by carefully placing both steel rims of my bicycle wheels over the tram sensor I could trigger the tram traffic light, which had priority over the car traffic lights and would set them to red. Then I could ride through the intersection without danger from cars 

[Old Mechanic]

If you can line up your wheels over the cut marks quite often you can get them to trigger the controller sensor circuit.

I've tested this on the past few commutes, and it really does work.  It's too early to say if it works 100% of the time, and I'm sure it depends on the condition of the sensor, but I've already gotten a few lights to change for me when there was no other car there to trigger it.  Thanks for the thread!

Recently Va DOT has modified the lights so when the straight through traffic is green the left turn is flashing yellow. When the green light goes to yellow the flashing yellow stops flashing and both trun red for other traffic. Two minute rule here. I generally wait one full cycle then go when it is safe if the signaldoesnot change for me.

regards
mech

[Kijona]

They've recently installed cameras or some type of visual sensor (not sure if it's an actual camera) on all the traffic lights near my house. No more waiting.

[Kiwingenuity]

There are quite a few microwave sensors out there which they can use on roads where the loops get damaged more quickly - try bringing an old school K band radar detector near it and see if it squawks 

(They do look a little like a camera since they have an plastic lens to protect the radar sensor / wave guide which is quite a narrow beam).

That being said, you can use cameras now - just easier to interface the low tech sensors to traffic lights (and far more reliable)

At least no more waiting 

[hInstance]

Deer Whistles ............

Your right.....maybe they need to be a topic by themselves.

I live in rural NC.  I have deer whistles on my car, truck and all my street bikes.

Do they work..... I don't know....but I do know the deer look up when I drive by.

So what exactly is a deer whistle? Never heard of them.