How to Beat the $#%! Red Light Sensor (and how they work!)

[dkoerner]

I frequently overhear riders frustrated/confused with red light sensors and figured it would only be fair to share my wisdom and expertise on this subject. Unfortunately, my grandfather is the man who invented the darn things, first at a little company called Canoga Controls, and then for the company I work for now, 3M. Next time you are stuck at a red left turn light, cursing, you have him to thank.

But we've got a light to change!

When you are stuck at a red light, there are only five ways that the light changes:

• A simple timer
• Detection via an optic system of lasers and reflectors or IR
• Weight-activated air filled rubber hoses running below the asphalt
• Triggered remotely by emergency vehicles
• An electromagnetic induction loop.

This last method is by far the most common, and for us, so damn frustrating. Optic systems easily can pick up a motorcycle, and weight-activated systems are sensitive enough to even detect bicyclists, but some induction loops are deliberately calibrated to be able to detect cages and nothing else.They detect when a car arrives at an intersection, when too many cars are stacked up at an intersection and when cars have entered a turn lane... Your KTM? Nothing.

The reason you are left sitting there is due to the nature of the detection method. As in all engineering, the design of a loop detector is a compromise: Set just low enough to detect a datsun making the left turn, but not high enough to sense the metal of a school bus in the lane just adjacent. Combine that with the fact that they are cheap to install (patent on the technology is free, with money made on the hardware, electronics, and inductance meters in the big towers you see set to the side of each intersection), and you have a scourge for two wheelers everywhere.

How can you to tell you are sitting on an induction loop? The familiar square or circular grooves cut into the roadway just behind the hold line that look something like this:



(a older, hexagonal induction loop replaced with a newer, circular one)

Road crews lay the asphalt, then use a tool very similar to a circular saw to make several cuts, producing a groove that is roughly circular or rectangular in shape. For reasons I’ll explain, rectangles are much better for bikes than the hexagons or circles in this picture (circles are becoming more and more popular here in the US, unfortunately). They then lay several loops of thin wire in the groove created and seal with tar (which is horrible for traction, but I’ll save that rant for another day). Sometimes the whole road is resurfaced and there are no grooves at all, in which case the loop is even weaker and you are far worse off.

But enough of that. The light isn’t changing; What are your options?

The first is positioning: The most sensitive spot is directly over the wires in any leg of the buried loop. The farther--higher--the metal of the vehicle is, and the smaller the mass of that metal, the less likely it is to be detected, so on ADV bikes, your best bet is directly over and parallel with, any leg of the loop. Thus the weakness of loops without long straight legs—On a hexagon or octagon you can often line up on a side, but on a modern, round loop, some chord of that circle is as good as you’re going to be able to do.

Also: If and when you are lucky enough to encounter a rectangular loop with three longitudinal legs (called a “Quadrupole”), made up of a long groove cut right down the middle, then, and only then, you should position your bike over that center leg. You’ll find it is very much more sensitive than the outer two, and you should be on your way in no time.

Sometimes there are two or more loops in a lane. This is to detect when more than one car is waiting for a light, and usually, but not always, it is best to come to a stop over the loop farthest back from the hold line (if safe). Should you pull up to a light, position over one of the grooves on the side, and the light doesn’t change try rolling the bike back to a loop behind to signal the light to release the queue that way.

The second method is very slightly revving the engine (and holding it): The loop detector is looking for a sustained decrease in inductance caused by eddy currents set up in the metal of the vehicle. Blipping the throttle and/or restarting the bike are two things that generally won’t work--transient events that do not provide sustained anything. Holding just a tiny amount of throttle usually works on my KLR at the light by my house, though if I modulate it any, I find that it won’t.

As for rare-earth magnets strapped below the bike, putting the kickstand down, leaning the bike over, or electronic IR devices that replicate the signal of emergency vehicles…it’s pretty much the same crapshoot. Induction loops, while working on the same principal, come in several different configurations, and even revving the engine is something I can only recommend because I’ve seen it work.

The last thing that might work is a light sheet of metal or aluminum (mine is 14 inches by 18 inches, and slides behind my left pannier). If I am riding at night and not worried about a car approaching from behind, I lower it on a short length of knotted rope to lay it flat just below my left foot, covering one of the legs of the loop. After the light changes I can raise it quickly and replace it. This is a bit extreme, but between positioning, holding a few revs, and carrying my plate, I’ve been able to take down every induction loop I’ve come across thus far.

Finally, note that with an induction system, the effects are additive. When riding in a group, signal others to pull up alongside over other legs of the loop, or with a circular one, try a diamond, or V formation to cover as much of the circumference as possible. On a bicycle, the same holds true (if they are metal, and not carbon fiber bikes. Ha.)

And that’s it! I’m going to spare you a detailed explanation of how a loop works, but if you have questions about the system or suggestions about how to defeat the darn things, we can ask the inventor (who thinks motorbikes are deathtraps, btw) or just post ‘em if you got ‘em.

Be well,
David Koerner
Manhattan Beach, CA

[viola-tor]

Very interesting, thanks for the write up! There's a light in my 'hood that I usually end up running, I'm gonna try some of these tips...

[Reverse]

I have an 8-mile commute with probably a dozen sets of lights. None of them detect a bike, which is a real PITA when commuting early in the morning when there are no cars to help you trigger the lights.

I've started a google map where I record all of the locations of defective detectors and am going to send it to the city's engineering dept and the chief of police soon.

[tedder]

thanks for the information. Yes, we teach that inductive loops are the most common. I found that my state, for state DOT, *must* reply and fix it within a certain number of days. So they fixed a ramp meter VERY fast after I complained about it, with plenty of status updates.

Keep it in mind if you are frustrated by a specific light.

[buhbeze]

I never had a problem with the loop sensors even with no other cars around. I did get busted though for going through this light in front of this firehouse after waiting a while and getting skipped, must have been one of those other sensors but it was just me at the light with no cars coming from behind, else I wouldn't have gone through of course.

Even the instructor at the MSF who is an ex state trooper told everyone to just go ahead if it doesn't change after a while and it's safe to do so, no use waiting around for nothing. Wish I knew about that revving trick, might have worked!

Another thing, I know cops can make a light turn green but can they also make it stay red or make the other side stay green for a long time? One time after the above incident, I was at a red light on a busy intersection on a highway and it stayed red for a loooong time despite the fact that the turning lane next to me was filled with cars and a row behind me too. I was first in line at the light so I rolled up a little bit so that maybe the car behind me could roll up and hit the sensors, which is what I normally did when lights would take a long time. Rolled up and noticed the car behind didn't budge. Got a better look in my mirror and realized it was a cop. I would hit that light just about every time I would ride and it never took that long, so just wondering if they can control that from the car.

[crofrog]

Some neat tricks but honestly I'll just run them after a reasonable amount of waiting if there's no traffic around.

[CatCow]

Originally Posted by buhbeze

Another thing, I know cops can make a light turn green but can they also make it stay red or make the other side stay green for a long time?

If he left his Opticom(3M brand name, there are some others but they generally operate in the same way - white or infrared light strobes that flash at a certain frequency that triggers a sensor mounted near the traffic signals) flasher on, very likely. Many signals with Opticom controllers are programmed to give a red signal to all other directions, and a green signal to the direction the emergency vehicle is coming from so that the intersection stays clear. I heard a story from a local paramedic that a fire truck had parked at an accident aimed at a traffic signal a block or two away, and someone forgot to switch it off... Traffic was backed up about a half mile on a busy road before they realized what they had done!

[space]

Very nice write up! Thanks!

I'm skeptical that revving makes any difference, though. There's no reason that would affect the mutual inductance of the bike and the loop, which is all that really matters.

[Analog Kid]

In NC there is a new law for motorcycles. If you wait three minutes at a red light and it doesn't change you can run the light if there are no cars in sight on the other streets.

[buhbeze]

Originally Posted by CatCow

If he left his Opticom(3M brand name, there are some others but they generally operate in the same way - white or infrared light strobes that flash at a certain frequency that triggers a sensor mounted near the traffic signals) flasher on, very likely. Many signals with Opticom controllers are programmed to give a red signal to all other directions, and a green signal to the direction the emergency vehicle is coming from so that the intersection stays clear. I heard a story from a local paramedic that a fire truck had parked at an accident aimed at a traffic signal a block or two away, and someone forgot to switch it off... Traffic was backed up about a half mile on a busy road before they realized what they had done!

I see, makes me wonder if they have one of those things in the firehouse since there is a light right in front of it, especially since the officer claimed to be right behind me the whole time(if he was then why didn't the light change, and why would I run a light if there was a car behind me at all, and none coming or I would have just waited for it to hit the sensors) and also might explain why I didn't get pulled over until like 10 minutes later out of nowhere at the beach in the next town over.

[dkoerner]

Originally Posted by space

Very nice write up! Thanks!

I'm skeptical that revving makes any difference, though. There's no reason that would affect the mutual inductance of the bike and the loop, which is all that really matters.

That would be true, but generating the spark on most production bikes is accomplished by the coil, and the electromagnetic field produced is huge. Output voltage to your sparkplug is determined by the rate of field collapse and the windings ratio in your coil (on a car, this is a primary to secondary ration of around 100-1), but basically, because the path to ground is going to involve bridging the spark gap in your engine, the initial resistance is extremely high. The coil must allow voltage to build to a very high value until it gets high enough to ionize the gas between your electrodes and bridge the plug gap.

That ionization process becomes more difficult as cylinder pressure increases, and higher voltages are required, so on a lot of bikes (there are a folks out there much more knowledgeable with ignition systems than I, so feel free to jump in if I am missing something) rely on CDI (capacitive discharge ignition), which basically just uses capacitors to help at high rpm and high cylinder pressures. By putting higher voltages through the primary side of the coil, charge time is reduced.

The upshot of all this of course, is that at high revs, there is a lot of current flowing through your bike to make all this work, and the field change associated when you rev an engine is not insignificant.

Now, whether or not you can sustain a field capable of being detected by an induction loop, which is looking for a very sustained level of resistance--that's another matter entirely. I know for a fact that at least at one light, I can (my grandfather, for what it's worth,

...and great comment about contacting your local city and police department. That kind of advocacy for motorcycles is something that helps everybody. Kudos!

-David

[tedder]

Originally Posted by dkoerner

[COLOR=#cdcdcd]That would be true, but generating the spark on most production bikes is accomplished by the coil, and the electromagnetic field produced is huge.

I'm a little curious about that. The b-field is relatively large, but it drops off very quickly, and coils are a long ways from the ground.

One thing I've heard to generate a nice b-field is to kill the motor and then restart it, as the starter will generate a nice field, and the starter is usually mounted low.

[PacificPT]

I commute daily on my bike, and I have several lights which refuse to recognize the bike. Most of them would recognize my Transalp when I had it, now with the Huskerdoo 610 they don't. I assume it is because the Husky sits much higher. I will try the revving trick on the balky ones Tuesday morning. It will make the commute less exciting if it works.

[GSWayne]

Rather than a sheet of metal, try a loop of wire. I did experiments and could trip an inductive loop with a loop of 14 Ga wire that went around the perimeter of my boot. http://www.advrider.com/forums/showp...27&postcount=6

If you want an even more complete version than the OP's excellent post:
http://www.tfhrc.gov/its/pubs/06139/

another couple of good references: http://www.humantransport.org/bicycl.../detection.htm

http://www.ci.bakersfield.ca.us/city..._loop_info.pdf

Like some of the others, I am skeptical of the reving engine solution. It would seem unlikely to get the spark frequency near the operating frequency of the inductive loop sensors, but perhaps at some RPM a harmonic of the spark frequency could line up with the sensor operating frequency. He needs to go out to a traffic light some quiet night and try different RPMs and do a controlled test and see if it is a real effect or just the human minds amazing ability to see patterns in random data.

[LuciferMutt]

The center wire is the most sensitive?

Dammit. That's the only kind we have here (three three lines down the street) and if you stop on the center one you'll be on your ass because you'll wash out your front tire on the huge grease trap in the center of the lane which is left from all the damn dripping cars.