Carbon fiber is conductive, but the resin it gets cast in most certainly isn't.

There is one light on my commute near the top of a hill that never seems to sense me. It seems to have a hard time sensing cars as well. I don't worry about it though, it gets enough traffic and a car or two is sure to come behind me soon, and I can usually use the rest (and water break).

ha, finally solved the mystery of the light closest to my house -- only the left turn lane sensor will pick up bikes. Either that, or it's sensitive enough to pick up cars in the other lane, and only one sensor is active. I have seen this at the only other light that I know how to trigger, but on that one I think it's the right turn lane that is active

I wish I knew. Many of the sensors around here are buried under asphalt.

If the cars stop beyond the stop line, they will have driven over the sensor and tripped it. If they stop before the sensor, they won't trip it. A car is a large enough conductive object that they can often trip them by just putting the bumper over...or even near...the sensor line.

The figure 8 loops like you describe are more sensitive in the center because that are 2 wires laid down in the saw cut. This doubles the sensitivity near those wires. This is a fortuitous quirk of the system for those of us riding smaller conductive objects.

You'd do better on a figure 8 or a square to ride directly over and parallel to the wires for their entire length. And you wouldn't have to turn sideways.

Sorry but there is no real good way to detect them after they have been buried. Position over the loops for a small conductive object like a bike is very critical.

Yes, you are correct sir, I misspoke. One key element of being detected is the surface area that is parallel to the road surface - you stand a better chance of being detected if you lean your bike over. As far as CF goes, I recall reading something a few years ago that the conductivity is so low that you are basically un-detectable. We got to talking about this at work and one of the signal techs sent an email to one of the loop manufacturers/pioneers in the field so when we hear back, I will update on what they have to say.

The practice I have seen most of my career is to place the first loop at the stop line or straddling it. There are usually 2 behind the first one, spaced at 10-feet apart. Or, you can ask the road agency to go out and locate the loops for you, we do it all the time for motorcyclists who use the same route each day and have trouble getting a signal call in. It usually takes less than 20 minutes to mark them out and we make someone happy (and stop their pestering phone calls that the "$%^&* signal isn't working").

If they are based on inductance, which does seem to be the case, wouldn't a small magnet on the bottom of a bike increase the inductive disruption, thereby increasing the chance that the signal will detect it? Some of the articles claim a magnet doesn't help, but they might just mean that it isn't magnetically based.

Luckily, most of the lights I have to worry about are camera triggered, and my lights set them off just fine.

Too far away and far too weak. Your wheel does a better job...if it is directly over the loop.

I stop at a triggered traffic signal every morning. I roll up and put my SPD cleat down right over the pavement cut where the loop is, and without fail this starts the crosswalk countdown timer to my left.

So this happened the same way this morning, except that when the counter got to 3...2...1... I picked up my foot and started to roll forward (light goes green at 0), and watched the red hand (STOP) turn back into a white figure (WALK). My light stayed red, and at this point there was a car behind me waiting too. The roads were empty so I ran the light, leaving the car to wait another cycle.

Anyway, try an SPD cleat if you have them, but don't lift until you actually have a green.

Somehow, the sensors here know if you've gone past them. Numerous times I've stopped on the sensor and then started rolling a second or two before my light is green and it will stay red and immediately give the other road a green because it knows I'm no longer on the sensor.

https://www.wikihow.com/Trigger-Green-Traffic-Lights

This article recommends neodymium magnets. I've got a bunch of hard drive magnets laying around, and several stubborn traffic signals on my commute. I'll zip-tie a couple of the magnets to the lowest parts of the chain stays, and see how well it works.

Same problem as above. Even a powerful magnet like a drive magnet is still too weak to have any effect. Position your wheel directly over the sensor wires and, if you have a choice over the middle wires of the figure 8 type sensor.

You could put a couple of the magnets on opposite sides of a wheel. I bet that would work. It gets the magnet right down there near the street level. It would simulate a significantly larger hunk of metal. I have a small (20") front tire on my recumbent, so I could easily roll a little forward or backward to make this work. I'll have to try it.

The wheel works so well by itself that it doesn't need enhancement. Here's a full on geeky version of why. About half way down the page.

Loops can be set to "lock" a call in or not, depending on the lane configuration we won't lock calls because if you have a right turner who can turn on red, you don't want the call staying in when they are gone.

It really is an easy answer. Just be thinking "Good! A red light. I can take a swig of my water, a couple gulps of air, and stretch." As you ride up to the light. It will turn green just to spite you. :-)

Round my way, there's often a button next to the bike zone at traffic lights so the cyclist can press it to change the lights.

Note these two drawings from the posts above:




The right-hand-rule will give you the direction of the electromagnetic fields and the induced current as well as the back-EMF. You can sometimes see the octagonal cuts in the road where a loop was installed. You want your wheel to line up with the edge of one of the loops to be directly over the wire. If you're in the centre of the lane, you're not as close to lining up with the wires as you can be.

The sensitivity of these detection circuits is also varies. When I lived down in Santa Barbara, the lights going after 2005 were much more sensitive and would detect bicycles without any problems. I feel guilty when riding up to the one closest to my house. I'll see a line of cars coming the other way for about 15-20 seconds as it's a long straight. I pull into the left lane and within 3-5 seconds, the light will have changed in my favour. Don't even have to come to complete stop.

I like this one! Murphy's Law, it's pretty much undeniable.

On my commute nowadays I only have one such light. Tried parking all over the sensor area with no effect. Weirdly, sometimes riding down the center cut the "Don't Walk" for cross traffic will start. As if the light is going to change. When it gets to 0, the "Don't Walk" changes back to "Walk" and the light stays green for cross traffic. Is it possible the sensor can pick up a moving bicycle better than a stationary bike?

That always works for me.
I can't remember how many times I've put the bottle back in the cage w/o even getting a drink.

I hate to reply to my own post but I have been noticing something on older paving that covers the pavement cuts. With traffic and time (lots of time, on the order of years), the pavement over the saw cuts will start to crack. I've been noticing the lone linear cracks appearing at several intersections but they are can be difficult to detect because they are subtle. There's no hope for fresh paving jobs.

Do you suppose that a couple of aluminum plates in the soles of your biking shoes would trigger it, if you put your foot down directly on the loop? My instinct is that it would work.

Around here the saw cuts in the pavement are usually just filled with black tar or sealant. Is that what you're seeing? I have NEVER seen stop lights with detector loops that were paved over with no visible cuts.