Warrants and Justifications for Traffic Signals

October 4, 2016 at 10:26 pm | Posted in Traffic Signals | Leave a comment

Other articles of mine cover just about every facet of traffic signals, including how they work, how timing and phasing are set, and even how to identify who made a particular signal. But what about how it’s decided whether to add or remove a signal in the first place? That is done through “warrants” and “justification.”

The MUTCD (Manual of Uniform Traffic Control Devices) is the engineer’s bible for traffic control, so of course it is the basis for whether a signal is needed or not. Minnesota has warrants-01it’s own local version of the MUTCD, but it’s mostly copied from the national MUTCD, and the section on signal warrants is identical.

The following is a primer on the language used in the MUTCD:

Shall means something is an absolute requirement. For example, the colors of traffic signals cannot vary, evenif a city wants to honor Prince and put purple lenses in their traffic signals (although purple was a very early pedestrian indication).

Should is a very strong suggestion for items less acutely dangerous than things covered by shall, but supported by years of engineering experience and studies.

May is an option. You don’t have to put up a sign saying there’s a McDonald’s at the next freeway exits, but standards exist if you want to.

The MUTCD Speaks:

So lets see what possible warrants there are for traffic signals (in all of these, emphasis mine).

An engineering study of traffic conditions, pedestrian characteristics, and physical characteristics of the location shall be performed to determine whether installation of a traffic control signal is justified at a particular location. The investigation of the need for a traffic control signal shall include an analysis of factors related to the existing operation and safety at the study location and the potential to improve these conditions, and the applicable factors contained in the following traffic signal warrants:

Warrant 1, Eight-Hour Vehicular Volume.

Warrant 2, Four-Hour Vehicular Volume.

Warrant 3, Peak Hour.

Warrant 4, Pedestrian Volume.

Warrant 5, School Crossing.

Warrant 6, Coordinated Signal System.

Warrant 7, Crash Experience.

Warrant 8, Roadway Network.

Warrant 9, Intersection Near a Grade Crossing

The satisfaction of a traffic signal warrant or warrants shall not in itself require the installation of a traffic control signal

The last sentence is extremely important. Just because the warrants are met does not mean a traffic signal will be installed, or even that it’s necessarily a good idea. If the intersection meets objective warrants, a subjective, common sense call based on engineering judgement is made to see if it is “justified.” Every justified signal is warranted, but not all warranted signals are justified.

Although these aren’t too hard to understand in English as opposed to “engineer-speak,” I’ll explain a few nonetheless. The first three are vehicle volumes at given times of day. To meet the warrant it has to meet the threshold for at least eight, four and one hours in a 24 hour period. It doesn’t matter which ones, or even if they are consecutive. Warrant 1 is the normal one used to warrant a signal, and the MUTCD is clear Warrant 3 is only to be used for unusual situations:

This signal warrant shall be applied only in unusual cases, such as office complexes, manufacturing plants, industrial complexes, or high-occupancy vehicle facilities that attract or discharge large numbers of vehicles over a short time.

Warrant 8 is if you want to encourage traffic to use certain streets to try and establish a hierarchy where none exists (say encourage traffic to use a collector as opposed to a local street to exit a neighborhood), and the signal doesn’t already meet other warrants, and Warrant 9 is if a stop or yield sign on an intersection near railroad tracks is causing queued traffic to back up over the tracks.

Analyzing Some Signals

It’s been suggested that Minneapolis has too many signals, some of them perhaps unwarranted. So I thought it would be illustrative to do my own “warrant analyses”, starting with vehicle volume, the normal one that is used. In many cases only six hours of data is available from the city Traffic Management Center web site, which means Warrant 2.

Here is the chart for Warrant 2. If the plot is at or above the red line for any four hours of the day, the warrant is met.


Here are the charts for Warrant 4, pedestrian volume. The first chart is any four hours in a day; the second is any one hour, if either threshold is met the signal is warranted.


Bicycles get counted as vehicles or pedestrians, depending on if they’re on the road or on the sidewalk.

For the first signal, let’s look at Bloomington Ave and 46th St in Minneapolis


46th Street and Bloomington Ave, Minneapolis

Our “X” data points are 313,300,443,341 and our “Y” data points are 529,518,719,662.


Since all the data points fall above the red line, this signal is warranted. But is it justified? Maybe not, since it doesn’t exceed it by much (it’s still on the chart), and since the original article reported it functioned fine with a temporary stop sign. Maybe this was a conversation that should have happened before money was spent rebuilding it.

For the second signal, lets look at Hiawatha Ave and 46th St.


Hiawatha Ave and 46th Street

Commons sense would tell us that this one is warranted, but lets be sure. Our “X” data points are 2843, 2836, 3187, 2910; our “Y” data points are 629, 663, 720, 741.


This one exceeds warrants so far we had to blow up the graph.

For our third signal, let’s look at Grand Ave S, and W 34th St.

I saw this one browsing around on the TMC web site, and thought “how in the world could this possibly be warranted? Our “X” Values are 70, 150, 152, 176, and our “Y” values are 23, 36, 29, 22.


This one is so far from meeting warrants it’s off the plot area to the bottom left, even if it still fits on the image. But shouldn’t we have a signal here because of the school? Well, let’s check Warrant 4 to see if it’s warranted for pedestrians. Our X values are the same and our Y values are 8, 11, 8, and 5. This is so low I’m not going to even try graphing this one.

But what about Warrant 5, because it’s near a school? Well, it still has to meet a certain, albeit lower, pedestrian count:

01 The School Crossing signal warrant is intended for application where the fact that schoolchildren cross the major street is the principal reason to consider installing a traffic control signal. For the purposes of this warrant, the word “schoolchildren” includes elementary through high school students.

02 The need for a traffic control signal shall be considered when an engineering study of the frequency and adequacy of gaps in the vehicular traffic stream as related to the number and size of groups of schoolchildren at an established school crossing across the major street shows that the number of adequate gaps in the traffic stream during the period when the schoolchildren are using the crossing is less than the number of minutes in the same period…and there are aminimum of 20 schoolchildren during the highest crossing hour.

03 Before a decision is made to install a traffic control signal, consideration shall be given to the implementation of other remedial measures, such as warning signs and flashers, school speed zones, school crossing guards, or a grade-separated crossing.

 So it doesn’t appear that this signal meets Warrant 5 either, and there’s no reason this signal shouldn’t be removed.  I’d even make the city an all cash offer for the old Eaglelux signal heads like they have on this intersection, the kind  desirable on the collectors market that they currently throw into the scrap bin.

Why so few kids is an interesting question; the data was collected on from what I could tell was a warm, dry fall school day. (Wed Oct 12, 2011). Are kids living a half-block away put on a bus to go to some far-flung school to achieve racial balance or attend a magnet school? Are parents giving kids a ride in their car? Are they riding bicycles on the street and thus being included in the vehicle count? Is the data somehow misleading? I don’t know, but the point is that the official data does not warrant a signal here.

I’ll also comment on Lyndale Ave and 25th St.

After a car vs pedestrian crash that was caught on a security camera and made the local news, there were calls that a signal was needed. As it turns out, data for this intersection exists. The city response to calls for a signal were “warrants aren’t met,” and I confirmed that to be true. Traffic on 25th St would need to double to meet vehicular warrants, and with five pedestrians crossing Lyndale Ave in an eight hour period, it’s even farther from meeting pedestrian warrants. If there is a particular, unusual crash problem at an intersection that could be corrected by a signal, that’s another story (Warrant 7), but skimming the crash data in the neighborhood suggests this doesn’t seem to be the case.


Now, to get a sense, let’s analyze all the signals in a corridor. I picked South Lyndale Ave (with the exception of the signal at the ramp to MN 62, which is Mn/DOT operated. I’m going to cheat and just say that one is warranted rather than try to get data). This corridor has many busy intersections, some not so busy. Some signals have been there forever, and some have been recently reconstructed.


So the  Four Hour Vehicle Warrant, the signals at the I-94 Ramp, Franklin Ave, 22nd St, 24th St, 26th St, 28th St, Lake St, 31st St, 35th St, 36th St, 38th St, 40th St, 46th St, 50th St, 54th St, and the west ramp of MN 62 meet vehicle warrants.  33rd St, 34th St, 43rd St, 48th St, Minnehaha Pkwy, 53rd St, 56th St, 58th St, and 61st St do not.

A few random notes:

1) I didn’t note what times I took data for because it’s not important beyond being the highest time available, it was generally 7:00-9:00 AM and 4:30-6:30 PM, provided data existed for those hours. If the PM rush hour wasn’t available, the noon “mini-rush hour” was. Asterisks are where Lyndale Ave is the side street.

2) The volume required to meet warrants I got from eyeballing the graph. Real engineers use complicated spreadsheets with macros. Undoubtedly if you’re a programmer (which I most certainly am not) you’d be able to scrape data from the TMC web site and plug it in to the warrant spreadsheets.

3) 22nd was the only intersection that also met pedestrian warrants; I didn’t include the data for that here but I looked at pedestrian warrants if it looked like there was a chance of meeting it.

4) I was a bit surprised by how few pedestrians were crossing Lyndale Ave at Minnehaha Parkway, nowhere close to meeting pedestrian warrants,  only 34 in the four hour-long periods. But when I’m there it’s always on my bicycle on a nice summer weekend, not a typical rush hour when traffic is counted.

5) The 4-3 conversion south of Lake St. magically made some signals warranted that were not before.


Here’s a map where I’ve indicate warranted signals in green, and those that do not meet the warrant  in  red.

Why no Warrants?

So why are there warrant-less signals? Warrants have existed since the beginning of the MUTCD in 1935. So “warrants didn’t exist then” isn’t an excuse. Sometimes a signal was warranted at one time, but traffic patterns have changed. This is particularly startling in Detroit, the archetype for urban evisceration, where a lot of traffic signals aren’t needed simply because no one is around anymore. But patterns change in other cities too. Usually, whenever local residents hear a traffic signal may be removed, local residents are strongly opposed, predicting all kinds of death and mayhem.  So cities are reluctant to be too motivated to remove them; normally it comes up when continuing to operate the signal would involve spending money because the signal needs replacing due to deterioration or road reconstruction.


Eaglelux “Tall Fin”, Minnehaha Ave and 46th

Sometimes too, cities just do whatever they feel like. Just about every signal on Michigan Ave. in Chicago has a configuration that is a direct violation of the national MUTCD, prohibiting left turns with a sign instead of a red arrow. And Washington state for years used a flashing yellow ball to indicate permitted left turns.

A Notorious Intersection

Going back to 25th St and Lyndale Ave, where a widely publicized car vs pedestrian crash was caught on security cam, it’s hard to make a judgement about who was immediately at fault for the crash. State statute 169.21 says basically that motorists must stop for pedestrians in unmarked crosswalks, but pedestrians must not cross if a motorist cannot reasonably stop. That the pedestrian said it was her fault is not relevant, and all the public has is grainy, zoomed in CCTV footage.  So rather than point fingers, lets look at potential ways to fix the underlying problem. Engineers are very pragmatic; if there’s a reasonable engineering way to fix a problem that won’t create other equal or worse problems, they’ll try to fix it.

Lyndale Ave is above the threshold where a 4-3 road diet is workable so we’re stuck with the 4-Lane Death Road. Generally speaking, anything over 15,000 to 20,000 vehicles a day isn’t workable because congestion overwhelms the road, and if there’s a lack of a hierarchical road network, motorists start using local streets instead (this part of Lyndale is at 22,000). Motorists become angry and impatient, thus making poor and even reckless decisions zipping down local streets that weren’t designed for them and where residents may not expect. This seems to be a recipe for disaster.

If you look only at the arterial, the costs of congestion might be worth the benefits of safety on the arterial. But dealing with motorists diverting onto local streets adds a lot of costs and complexity. As we’ll find out later, stop signs (to say nothing about “Slow, Children” signs and similar) are completely inappropriate, ineffective, and even dangerous. Thus they’re very strongly discouraged in the MUTCD for traffic calming or speed control. And when you start stationing police there on a semi-permanent basis or making physical changes like traffic circles, diverters, and chicanes to each and every local street, soon you’re talking about a lot more money than some new paint on the arterial.

I’d also suggest that despite the tendency toward the reaction to blame suburban commuters for all the cars on the streets, Lyndale Ave is probably almost entirely Minneapolis residents. Despite the horrific congestion on our freeways, it’s simply not attractive for suburbanites heading home from downtown to try to get off and navigate city streets, especially with the MnPass lanes available for people that are unlucky enough to have to drive in rush hour regularly. Traffic counts would seem to back me up. There’s over 30,000 vehicles a day north of Franklin. Half the traffic is gone by the time you get past Lake St, and on the MN 121 entrance to I-35W and MN 62, there’s only 7000, and only a subset of that would have traveled the entire length.


Lyndale Ave at 25th St.

So what else can be done?  Legally it’s a crosswalk (despite the media calling it jaywalking) and there’s not a ‘No Pedestrians’ sign. But the lack of a painted crosswalk, and the lack of even curb cuts, sends a message to both motorists and pedestrians that this isn’t a place for pedestrians to cross. Adding marked crosswalks and curb cuts would be a good place to start. Maybe there’d even be space to build a pedestrian refuge island. But what if we could do even better and make motorized traffic just disappear for a short time to allow pedestrians to cross? We’ve talked about a lot of the warrants, but here is the final one: Warrant 6: Coordinated Signal System:

Progressive movement in a coordinated signal system sometimes necessitates installing traffic control signals at intersections where they would not otherwise be needed in order to maintain proper platooning of vehicles.

The need for a traffic control signal shall be considered if an engineering study finds that one of the following criteria is met:

  1. [not relevant]
  2. On a two-way street, adjacent traffic control signals do not provide the necessary degree of platooning and the proposed and adjacent traffic control signals will collectively provide a progressive operation.

The Platoon

When you have a number of signals, you can time them to create a “Green Wave” with platooning. The idea is that you want to keep the cars together and try to give them green lights as they arrive at each intersection. Spaces between platoons give pedestrians a chance to cross and motorists a chance to enter from side streets and driveways. Doing this in a stream of traffic would at best slow the traffic down, and at worst, it results in crashes when pedestrians and motorists get impatient and start to take chances.

In the name of research, I did something I normally try to avoid at all costs: drive surface streets at rush hour. Driving southbound at 6:00 PM, I tried to drive a steady 30 mph. North of Lake St I was unable to do that with the streets being swamped with traffic and turning cars and stopping buses. South of Lake St I was able to do it for the most part. I was stopped for seven out of the 21 signals between Franklin and 58th, in some cases arriving a few seconds too late or soon.

There’s obviously been some attempt at progression but it needs tweaking.  Minneapolis is way, way behind in terms of traffic signal technology and ideology; in the suburbs every one of the signals would have sensors for the main street (as well as flashing yellow arrows) and with advances in traffic control, the signals could figure out the queue length of the next signal and traffic speeds and adjust themselves accordingly. There was obviously no attempt to ensure platooning in both directions, leading to excessive delays when people want to cross at unsignalized intersections. There’s plenty of time when traffic has ceased in one direction, but it’s almost always flowing in the other. (It should be noted that this is where refuge islands would help immensely.)

So how do we set up a situation where motorists on Lyndale Ave can travel without hitting too many lights and motorists and pedestrians at non-signalized intersections get a chance to enter the roadway and cross it? An engineer I talked to, in typical engineer fashion, didn’t want to give an answer without having done a traffic study, but when pressed, the engineer commented that generally on a 30 mph urban street, a 1/2 mile spacing is enough; a source online suggested 1/4 to 3/8th mile. In order to enable a bi-directional green wave, they they should be as evenly spaced as possible, or failing that, in even multiples. (In no case can signals be less than 1000 feet; that’s not a problem on Lyndale Ave, but in downtown it’s best to just forget about it and turn all the signals green at once.)

Obviously some unwarranted signals would have to be removed, some warranted only by Warrant 6 kept, moved, or built. Maybe even some warranted signals might have to be removed–remember, just because you have a warrant does not obligate you to build or maintain a signal. But in the end, you make things better for all people, whether in a car driving down Lyndale Ave, in  a car trying to enter or cross, or on foot trying to cross. Even in an urban environment, transportation doesn’t have to be a zero sum game.

Back to North Star Highways Home


The Myth of the Dangerous Right Turn on Red

September 17, 2016 at 1:30 pm | Posted in Traffic Signals | Leave a comment
Tags: , , , ,

The Right Turn on Red

We all know right turn on red (RTOR) is extremely dangerous, leading to motorists mowing down workers carrying plate-glass and old ladies pushing baby buggies. Or is it? Well, it seems to lead to a lot of aggressive behavior from both motorists and pedestrians, which has been well documented here in official studies, but actual crash data shows it’s very safe.

California, with its wide streets and auto-centric culture, has had RTOR since 1939. The first study was in 1956 by James C. Ray in San Francisco, Berkeley, and Richmond which found just 0.3% of intersection crashes involved a RTOR. Also of note: RTOR maneuvers involved 11% of the right turning crashes but 18% of the total right turns (And the same motorists that can’t make a RTOR would make a right turn on green during the next phase, which has the issue of motorists not yielding to pedestrians, along with much higher speeds).

Eastbound Exposition Boulevard approaches and crossses at Farmdale Avenue intersection traffic signal green lights but No Right Turn on Red, No-U-Turn signs and No-Right-Turn light sign illuminated when the Expo Light Rail Transit Train is crossing

Later studies have similar results. For vehicle vs. vehicle crashes, a National Highway Traffic Safety Administration (NHTSA)  1989-1992 study of Illinois, Indiana, Maryland, and Missouri (selected because crash reports from those states included data on whether a RTOR maneuver was being performed) found they amounted to 0.05%, of all injury crashes, and 0.06% of all fatal crashes.  A 1994-1996 study limited to San Francisco found they amounted to 0.45% of all intersection crashes.

Nor do things change much for car vs. pedestrian crashes. A 1994-1998 San Francisco study (to determine if RTOR should be banned citywide) by Jack L. Fleck and Bond M. Yeeshowed just 0.8% of all car vs pedestrian crashes involved a RTOR maneuver. To try to see if there may have been some problems with reporting, they picked 100 random car vs pedestrian crashes to analyze in detail. They found that of the 25 that occurred at signalized intersections, zero involved right turn on red and 12  involved right turn on green. Quebec, long the lone holdout against RTOR, finally allowed them in 2003 after reviewing the studies.

There are some conflicting data: one study of several states (by Paul Zador, Jack Moshman, and Leo Marcus) showed that when RTOR was enacted crashes at intersections increased 20.7% from what they would have been but there was no significant change in severe or incapacitating crashes.  A similar before/after study by Claude Dussault had similar results. One possible explanation for the discrepancy is a learning curve from both motorists and pedestrians, and that when the law was changed there were specific intersections that should have had restrictions that hadn’t been identified and so marked yet.

I acknowledge the possibility that a lot of very minor crashes between cars and pedestrian are not reported. This is not something I’ve personally observed (I’m rarely outside a car unless I’m on recreational trails which tend to have few signalized intersections and I try to avoid driving in pedestrian heavy areas if at all possible).  However there have been no studies on this, just anecdotes, and thus I can’t support or refute them since there’s no data on how common they are in objective terms. No one chimes in and says “I was not knocked over by a car today”. Maybe that’s a good area for a future study?

The MUTCD Speaks

Here are the MUTCD guidelines for NTOR, section 2B.54 03:

A No Turn on Red sign should be considered when an engineering study finds that one or more of the following conditions exists:

  1. Inadequate sight distance to vehicles approaching from the left (or right, if applicable);
  2. Geometrics or operational characteristics of the intersection that might result in unexpected conflicts;
  3. An exclusive pedestrian phase;
  4. An unacceptable number of pedestrian conflicts with right-turn-on-red maneuvers, especially involving children, older pedestrians, or persons with disabilities;
  5. More than three right-turn-on-red accidents reported in a 12-month period for the particular approach; or
  6. The skew angle of the intersecting roadways creates difficulty for drivers to see traffic approaching from their left.

For a long time there was a no RTOR at Lyndale Avenue and 66th Street, but this was removed because there wasn’t engineering justification.  Of course this specific example is soon going to be a roundabout…

Lyndale Avenue at 66th Street

“Unacceptable number of pedestrian conflicts” is kind of subjective, so if they really wanted to keep the ban it they might have been able to justify it here, and I agree this might be a good place for one.  And this is only a recommended practice (The MUTCD has three levels of guidelines: “May”= Optional, “Should”=Recommended, and “Shall”= Required). San Francisco, in fact, bans RTOR in some places just to prevent motorists from inching into the crosswalk. .

But take another kind of intersection: If you ban right turns from a shopping mall to a wide suburban style road, traffic on the main road has to screech to a halt every time a motorist wants to turn out of the mall, and there’s probably not a pedestrian there and hasn’t been for the past hour. When you count all the suburban intersections around, there are probably more places they should allow RTOR than shouldn’t.


MN 252 at 85th Ave N, no reason to ban RTOR here. Pedestrians aren’t even alloweed at street level because of the pedestrian bridge

My Own Thoughts

1) As should be obvious by now, due to the improved LOS for motorists and the preponderance of studies indicating it is actually extremely safe, I fully support allowing RTOR where appropriate. 

2) I don’t like city-specific bans like NYC, or Montreal where you have to be aware of different rules of the road. This diverges from everything we’re trying to accomplish now with standardization, and I think NYC can afford signs at this point.

3) As for whether the problem of motorists making illegal turns where banned is an engineering problem, enforcement problem, or both, I don’t know. Lighted blankout signs grab your attention, but have typically been used only at specific times of day, or when there’s a train on adjacent tracks. We use lighted indications for everything but a NTOR prohibition, so while not excusing bad driving, it’s understandable why it’s sometimes hard to see.

An Idea: Flashing Yellow Arrows for Right Turns

Directly related to the problems with right turning traffic at intersections, I propose increased use and modification of four section arrow heads on the right to try to make intersections safer.

The states that allow a right turn on a red arrow would have to change, and the MUTCD would have to allow a flashing red arrow, but this is how I see it operating:

  • Green Arrow: Go, no conflicting vehicle or pedestrian phases
  • Flashing Yellow Arrow: Yield to pedestrians or vehicles
  • Red Arrow: Stop and do not turn / Flashing Red Arrow: Stop and turn if safe

A four section head with a red ball on top is already legal if the intent is to allow RTOR, but I would prefer that we narrow the meanings of red balls , just as we are for green balls.

Three section flashing yellow arrow heads are now permissible, but installing a standard 4-section head may be more problematic when physically mounted on the post for right turns. The lack of positional change from a flashing yellow to a steady yellow would be mitigated because the through signals would always change at the same time. It seems motorists have a problem with the “yield to pedestrians” on a green ball, just like they do for “yield to oncoming traffic”on a green ball.

Back to North Star Highways Home

Traffic Signals Abroad

September 16, 2016 at 1:01 pm | Posted in Traffic Signals | Leave a comment
Tags: ,

Minnesota vs the Rest of the Country

1) As I’ve mentioned before, protected only turns (where left turning traffic has a red arrow and must wait a whole cycle if they arrive one second after the typical 10 second green), are much less common. Only in California have I seen more than in the Twin Cities suburbs. These are becoming less common with the debut of flashing yellow arrows, the feeling being that with the extra warning to yield, left turning drivers can now be trusted to make good decisions.

3M Model 131 Traffic Signal Left Turn Arrow

The Capacity Killer

2) The usual arrangement for 5-light signals is the “doghouse”, rather than the 5-light vertical. Vertical signals are mainly used in Minnesota, Wisconsin, Illinois, South Dakota, and Arkansas; other areas use the doghouse configuration. With the adoption of the 5-light bimodal flashing yellow arrow, probably at the urging of the FHWA, Minnesota and Wisconsin are adopting the doghouse. (And much more common is a 5-light vertical  to be replaced with a 4-section flashing yellow arrow, although agencies that refuse to use flashing yellow arrows, like the city of Minneapolis, are still erecting new vertical installations.)


5 lens vertical signal

2) Signals hung from overhead span wires are not used in Minnesota, except in temporary installations (although some, like MN 62 at Hiawatha, persisted for decades for various reasons). MN 36 in Lake Elmo will likely be the same story. These are more common in permanent installations out east, which are usually metal poles, and usually a lot neater than the wood poles that Minnesota slops together. There are also quite a few surviving 4-way signals, some converted to LED. With the requirement of more than one signal head in each direction, these were augmented with two single faces tied together on either side, for a total of two facing each direction. Looking at old photos, Minnesota always seems to have preferred putting them on posts on the side, I’ve never seen a 4-way here.


4-Way Signal Cleveland, TN

3)  Minnesota always puts signals on the poles on each side of the mast (and these are useful if there’s a truck blocking your overhead view. Other states do not always do this. It might be carried over from masts supplementing side signals in Minnesota, and replacing overhead span wires in other states.

3) Of course, some signals are mounted horizontally elsewhere, as we’ve all seen in Wisconsin (although their new standards are vertical only). These are also common in Texas and Florida.

4) Eagle brand signals are a lot less common elsewhere. Possibly because of their Quad Cities location (stuff cost a lot more to ship in days gone by), Eagle in all its various incarnations, from the 1940’s Eagleluxes up to the 2000’s Bubblebacks, had a virtual monopoly on the signal market. In other areas of the country, Crouse-Hinds (of New York), and Econolite (of southern California) are a lot more common. McCain, a new player making a big splash with boring but functional equipment, is becoming more common both in Minnesota and nationwide.

Spotters 11

Type SIG “Bubblebacks, later version, 98th and Dupont Ave / I-35W, Bloomington. Eagle visors tend to point downwards more than other brands.

5) Of course other agencies have their own quirks. Obviously there are too many to list, but Chicago DOT does whatever they feel like, nationals standards be damned.


Noncompliant Chicago signal

The US vs Canada

1) The flashing green ball in BC is used at intersections where a pedestrian can stop traffic to cross. The light will flash green until a button is pushed, then solid green, yellow, and red.

In Ontario it has the same meaning as a left turn arrow in the US. In some places the flashing is much faster than in the US. I called them “seizure lights” and my sister called them “disco lights”.

2) Some provinces use square reds and yellow diamonds as assistance to the colorblind. Except for Quebec, they seem to be on their way out.

Prince Edward Island Traffic Light with Shapes

Charlottowne Traffic Light

3) Pedestrian symbols are red and green, and are always outlines (as opposed to the US where they’re now required to be all solid). There’s a subtle difference to the hands, but a large difference to the mans.  This one is Canadian standard but is actually in Osseo. A Canadian company called Ecolux was one of the first with LED signals, so some Canadian stuff snuck in down here. Canada also generally uses circular lenses.

Canadian Standard pedestrian signal in the US

Canadian Standard pedestrian signal in the US

Not really related, but because I think they’re cool, a moose warning sign in New Brunswick, and a stained glass representation of the Confederation Bridge at the visitors center. Tolls on PEI are one way leaving the island, leading to the slogan “You only have to pay if you want to leave”, but still was a bit of a shock paying a $45 bridge toll.


Moose Warning Sign in New Brunswick

Colored Glass Art of the Confederation Bridge

Colored Glass Art of the Confederation Bridge

The Rest of the World

I’ve not been out of the US and Canada since 1985, when I was a child, but here are some of my understandings based on conversations with other road and traffic signal enthusiasts.

1) Fixed time signals are much more common than actuated signals.

2) There’s a much wider variety of configurations, and some vehicle signals have countdown displays until the light turns green light, or until the light turns red. These are specifically banned by the MUTCD on the grounds that studies have shown no benefit, and there’s been suggestions to try to hide pedestrian countdowns from motorists because motorists were looking at them.  I’m guilty of that myself. After driving around San Francisco for several days where the vehicle light turns yellow at zero, I angered a truck driver behind me in Grant’s Pass by slowing down where the countdown went to zero with the light still green.

To get vehicle countdowns to work in the US requires technology that hasn’t been implemented here yet. The pedestrian countdowns were designed to “drop in” to existing installations and memorize how long the change interval is, which works because unlike a vehicle signal the pedestrian change interval never varies; there’s no additional wiring beyond the mains wiring for the two indications.

The beautiful “Spiderweb” lenses were only used in the US in the 1940s; they stuck around in Europe


Traffic Signal with GE Spiderweb lens, Cleveland, TN

2) They have stuck with LED look indications, and in turn have developed thinner signal heads that can only accommodate them rather than switching to “incandescent look” modules that require as much depth as an incandescent reflector.

Lenses 24

Comparison of module depths

Early Leotek LED module

Early Leotek LED module

The famous Ampelmännchen of East Germany.


Animations are common in Asian countries. Notice that the animated pedestrian speeds up and starts “running” as time expires

Uniquely, some German pedestrian push-buttons allow you to play Pong with the person across the street while waiting for a “Walk” signal.

This was actually an ad campaign: people could walk into a booth where their dance moves were copied on the Don’t Walk light, but 81% more pedestrians waited for the light.

Fundamentally too, other countries, including Canada, use red and green for pedestrian signals, instead of orange and white. I covered the development of pedestrian signals earlier, but I’ll repeat, expand,and speculate. In the US the word “Walk” was originally in thin art-deco letters on  8″ round lenses, usually mounted on the same head as vehicle signals.

A green lens would have made it even harder to read across an intersection and possibly be confused with the adjacent green vehicle light. When the orange “Wait” indication was added and moved to a separate head, an orange stripe was added to the top and bottom to make it easy to see and differentiate. The neons, which were a large separate head and one color, were as efficient as another and no way to be mistaken for a vehicle signal, used red and green until they were forced to change. (Red is the “default” color in neon because it is cheaper and lasts longer than other colors). I believe Europe used icons (easier to see) on a separate head from the beginning.

Early pedestrian accommodation, first with the green lens designed to throw some light downward, then with the separate “Walk” lens

Early pedestrian accommodation, first with the green lens designed to throw some light downward, then with the separate “Walk” lens

Later evolution to two indication Wait / Walk and then square Walk / Don’t Walks.

Later evolution to two indication Wait / Walk and then square Walk / Don’t Walks.

Crouse-Hinds Neon Don't Walk Signal

Crouse-Hinds Neon Don’t Walk Signal

4) Some areas have a yellow before the green. This reduces reaction time and gives motorists with manual transmissions (the norm in much of the rest of the world) time to drop the clutch, but can also encourage people to jump the light. Interestingly some early American signals by their design also used a yellow before the green, like the el-cheapo 3-lamp Darley

5) Installations are usually less elaborate than in the US. 

6) Right turn on red is prohibited unless specifically allowed. Universal RTOR is one of the legacies of the 1970’s oil crises,  the other being the 55 mph speed limit. The first is definitely pro-motorist, the second anti-motorist, so it’s not hard to understand why the first hung around in North America but not the second. Various countries use various means to indicate that a RTOR is allowed. Some use signs and some use signal indications, for example Germany uses a sign and France uses a flashing yellow arrow.

Of course RTOR is controversial in the cities, with some feeling they should be banned and some noting a problem of drivers doing them illegally anyway in places they are banned. And that’s a subject for another article, along with my ideas to make right turns safer and improve LOS for pedestrians.

Back to North Star Highways Home

Common Questions About Traffic Signals

September 15, 2016 at 12:36 am | Posted in Traffic Signals | Leave a comment
Tags: ,

Here are some random questions and answers about traffic signals that didn’t deserve their own article.

1) What can pedestrian sensors do?

Although pedestrian sensors are uncommon for all the usual reasons (cost, engineers don’t see need, old equipment, etc), there are some neat things they can do. Besides being able to call a pedestrian phase if a pedestrian is present, a second sensor can be mounted pointed toward the crosswalk. This can extend the pedestrian phase if the pedestrian is especially slow. Or conversely terminate it early if it’s a runner or bicyclist. Perhaps by mounting one farther back on a multi-use path they could call a phase if a bicycle or pedestrian is approaching, hopefully giving a walk signal by the time they reach the intersection. The idea is not new; some early ones experimented with pressure sensitive plates, but modern ones use microwave detection.

2) Why do we still have push-buttons if the “Walk” sign always goes on (ped recall)?

This is covered pretty thoroughly in previous articles and my comments to them, but basically ADA (at least as implemented in Minnesota) absolutely requires a station with  audible and tactile feedback (in the form of an arrow that vibrates) next to the crosswalk.


Shiny new audible-tactile push-button on and old signal

First, a station that includes feedback, with all the sophisticated electronics and simply lacking one mechanical button doesn’t make sense, either to produce as a product or for agencies to stock spares for. You’re probably talking about a part that wholesales for a couple of cents for a $400 station that’s part of a $6000 system (and can cost well into five figures to retrofit an existing intersection).

Second, if there was a station without a button, it would still look like it had a button, people would push at it and be confused when it doesn’t move or do anything (if ped recall is enabled or the button has already been pushed the speaker says “Wait”, but it doesn’t affect the cycle.) There’s already a myth that pedestrian buttons don’t “do anything” (that apparently started in New York when they converted from actuated controllers to fixed time and disconnected the buttons without removing them).

The Econolite push-buttons, as used with the famous California neons, would turn a small red light to green to acknowledge that the button had been pushed. This was probably not more widespread due to the maintenance of the incandescent lamps, and has now finally been coming back with the new electronic buttons.


Agencies vary widely in how rapidly they are bringing existing infrastructure into compliance with the ADA. Not just traffic signals, but removing utility poles from the middle of sidewalks.

Guidance in the past has been vague, for example the MUTCD does not require accessible signals, and disability rights groups have in the past conflicted with each other, but finally the FHWA made the statement:

“Implementing regulations for Title II of the ADA, which covers State and local governments, also address “communications and information access,” requiring ‘effective communications’ with persons with disabilities. In the sidewalk/street crossing environment, this would include accessible pedestrian signals, markings and signage.” (FHWA, 2004).

And with the “Public Right-of-Way Accessibility Guidelines” (PROWAG) under development, it is expected accessible pedestrian signals will be mandatory.

But the ADA only requires modifications to existing infrastructure when it can be done without “undue financial or administrative burden.” So far, modifications have been up to the whims of individual agencies and how far disability rights groups want to push things. Minneapolis and Hennepin County, for example, aren’t always using them in new construction. By contrast, MnDOT, which was pushed by disability rights groups, feels they have to include them every time they go smooth out some bumps in the road, and will actually remove crosswalks rather than leave them noncompliant or upgrade them.

3) Why aren’t there more leading pedestrian intervals?

There generally is an impact to motorists, as you more or less double the time they cannot use the intersection (the exception being if the pedestrian phase is long enough and motorists few enough that motorists don’t fully utilize the green time if a pedestrian phase is present), but they are such an advantage to pedestrians that they are beginning to be implemented in areas with heavy pedestrian traffic anyway (one study showed a 60% decrease in vehicle + pedestrian accidents).  The MUTCD only specifically allowed them in 2009, so there are the usual issues with older equipment not having the capability, and of resources required to implement them.

One area where I feel a leading pedestrian interval would be appropriate is Berry/Calhoun Parkway and Richfield Road. This has a lead left and in practice pedestrians and bicyclists on the very heavily used trail often jaywalk as soon as the cross street turns red, or at least get confused when they don’t get an immediate walk (It doesn’t help matters that during non-peak period it is permissive only). It would be better as a lag left with recall enabled and a leading pedestrian interval.


Calhoun Parkway and Berry Road

4) Why won’t the light turn to “Walk” if the same direction has a green left turn arrow?

Sometimes it’s just the engineers didn’t think of it and implement it, but this requires the “pedestrian overlap” outputs on the controller be used, not the standard outputs, with associated reprogramming as well as additional load switches. The controller may be old and simply not have that capability. There may be no room in the cabinet for more load switches. Implementing this certainly requires wiring changes in the cabinets. Always much more complicated than setting a menu item on a controller


Simple Traffic Controller Cabinet

Here’s a typical cabinet, you can see in #4 that load switches plug directly into the backplane, #3 If there are no empty sockets you basically  have to replace the cabinet, probably with a bigger one, and thus probably requiring concrete work… You can’t simply shove an extra load switch in some nook and cranny.

5) Why, when I pull into the turn lane, does the side street turn green first with no one there?

First, it could simply be a fixed timed light where there is no sensing of vehicles. Second, skipping phases where there is no demand is a fairly recent concept. It simply didn’t exist in electro-mechanical or early electronic controllers. So it might be that no one has worked on and upgraded the intersection recently.

Or, the loop detector could be broken, as often happens. If they break they are supposed to default to the “sense” position so calls are always generated.

6) If the cars have the green light why don’t I get a “Walk” signal when I push a button?

“Pedestrian Reservice” does exist and is an option on most modern controllers, but is subject to the minimum pedestrian phase time. You can’t promise the controller “If you give me a Walk, I’ll hurry across.” Since walk intervals tend to be short compared to clearance time, there’s usually not time to give a walk if a pedestrian is not there before or at the beginning of the vehicle phase.  Simply extending the vehicle phase until the pedestrian phase can get done would lead to all kinds of issues with coordination. The “Home Depot” intersection previously discussed does have reservice even if it doesn’t have recall to cross the driveway and side street.

To me this is somewhat redundant, since if a phase is long enough to allow ped reservice, it’s long enough to allow ped recall, but the concept does exist.

7) Why don’t signals flash at night or during times of low traffic?

Nothing’s more irritating than coming across a red light late at night when no car is in site. In Minneapolis where there are many fixed time intersections, it can even be a while before it changes. Basically night flash was removed in Minnesota because there have been a number of studies that it’s safer to have full operations 24/7, as well as actuated lights becoming increasingly common. Various studies show a reduction in night-time accidents from 27% to 95% when signals run 24/7. In theory if a light is red, it should turn green as soon as a motorist arrives at 3:00 AM.

As I’ve mentioned repeatedly, engineering fundamentally balances efficiency and safety; Minnesota engineers try to tilt the balance towards safety as opposed to efficiency in a lot of ways, most visibly with how prevalent protected-only turns are (Only in California have I seen more); this balance explains why night flash is extinct in Minnesota. With the increasing number of actuated lights and the introduction of right turn on red, motorists at least don’t have to wait as long as they would have in days past.

8) Can I change a traffic light to green with my TV remote or flashing my headlights?


Longer answer:-That Youtube video is a myth/prank/joke.


Preemption is triggered by a strobe light mounted on emergency vehicles that flashes at a rate different than what a commercial strobe light will do. After electronic geniuses started building their own triggers, Minnesota specifically outlawed having one in your vehicle (although I have one for my home setup). And in case anyone still has ideas, they’re not easy to build; they require more precision than can be obtained with simple “Electronics 101” type resistor/capacitor timer circuits.  Also, they won’t work on the street anyway anymore, because the flashing is now encrypted.

9) What are those blue lights on some traffic signals?

To tell the cops waiting in the next driveway you just ran a red light. The first installation was at the now replaced signal at County 5 and MN 13 in Burnsville and was improvised with a standard industrial fixture; now there are specialized LED lights available. Also, those cameras on top of the signal mast are not “red light cameras,” they’re generally video cameras or microwave sensors for detecting cars, and go no farther than the traffic cabinet. Some MnDOT intersections do have traffic cameras, but they are mounted much higher up.



10) What about driver-less cars?

I really am excited about driver-less cars.  Imagine being able to start out to Chicago at night, fall asleep, and then arrive in the city with my car in the morning. Imagine editing a post for streets.mn while stuck in traffic. Imagine ordering fast food, or even slow food, on a touch screen and it would be ready when you pull up (and a lot easier and safer to eat in your car). I’d probably eat at Applebee’s instead of Arby’s if it were just as convenient.

But getting back to the topic, think of how fundamentally unchanged the interface between traffic signals and motorists is despite nearly 100 years. A traffic signal controller can only tell if a car is there or not, not how many are there, or how many will be coming it’s way in five minutes. A driver can only see and react to a red light, not know there’s a string of red lights five miles away. So it’s easy to see how revolutionary cars communicating with traffic control equipment could be: No wasted reaction time, signals which know all about traffic and can react accordingly; cars which can reroute if there are a bunch of red lights ahead. Despite frustration with the present, I see great optimism for the future.

Back to North Star Highways Home

Traffic Signal Timing and Phasing

September 13, 2016 at 12:36 pm | Posted in Traffic Signals | Leave a comment

This is an overview of traffic signal timing and phasing, and the implications of various sequences. As of note about terminology, I’ve generally used the term “vehicle” and “vehicle operations” to match engineering terminology, despite that fact that it is ultimately the human motorist whose time is wasted with poor vehicle operations, and that a bicycle is a vehicle, but it can be on the sidewalk or sidewalk / trail rather than the street, and the rider is affected differently than motorized vehicles. This part focuses on timing and phasing issues.

The Case for Long Cycle Times- Intersection Efficiency 

During each cycle you have some time where no one is utilizing the intersection. For simplicity let’s say the yellow and all red times are when no one can use the intersection.  (Although it’s well known motorists will “run” the yellow, and they can’t start immediately on green, in the real world these considerations usually exactly cancel each other out.)

Let’s take two extremes. 20 second Main Street green, and 7 second Side Street green, three second yellows, and  two second all reds. That means for that cycle only 73% of the time is green.


Now how about an 8 minute Main Street green (as actually implemented for US 10 in Royalton on Friday summer evenings) and say a 30 second Side Street green. Here 98% of the capacity is green.


Obviously there’s disadvantages to longer cycle times. Pedestrians have to wait longer in the cold if they want to cross the street that’s green. Motorists start to fume and may turn around and cut through neighborhood streets. The engineers’ phones start to ring with complaints. Like so much else, setting cycle times is a compromise based on judgement and trial and error.

Protected-only turn phases (you can only turn on a green arrow) have a pretty negative impact to vehicle operations. With protected phases both in directions, you’ve just doubled the yellow and red times per cycle, plus some of the turning traffic that time is devoted to may have been able to utilize gaps in the main phase had they been allowed to.

3M Model 131 Traffic Signal Left Turn Arrow

The Capacity Killer

The Case for Short Cycle Times: Progression 

But why can’t lights be timed so you hit a series of greens? They try to (sometimes). And have been since almost the beginning of traffic signals. And computer modeling is making it a lot easier. But the suburbs, with their irregular traffic, irregular spacing between traffic signals, and frequent use of turn phases make it extremely challenging. One of the reasons you’re seeing many more where the turn arrow is at the end of the phase in one direction is that it was determined that is the most efficient way to coordinate (and the flashing yellow arrow makes it safe to do so with protected/permissive by eliminating the yellow trap).

Another problem is that although long phases are more efficient at moving traffic through an isolated intersection; short phases are better for progression. For progression on a two-way street you want to have a cycle take twice the time it takes a motorist to travel from one signal to another. Here’s a video of a NYC driver catching a green wave for 125 blocks. This is the ideal situation: late at night with little traffic on a one way.

Fixed time lights can actually work quite well in dense areas. A while ago NYC disconnected many sensors and reverted to fixed time (and left the pedestrian buttons abandoned, leading to the myth that they “don’t do anything”.


Portland and San Francisco, with their noted anti-car, pro-bicycle/pedestrian culture, have timed signals to create a green wave for bicycles.

In downtown areas with signals every direction every block, it may be better just to have a string of lights turn green at the same time. This is easier to implement and if traffic is gridlocked it all starts moving at the same time.

Pedestrian Times 

Moving into pedestrian times first we need to define some terms. We have the walk interval (the time the “Walk” sign is illuminated), the pedestrian change interval, (when “Don’t Walk is flashing), and the buffer interval, (when the “Don’t Walk” sign is illuminated steady). The clearance time is the combination of the change interval + buffer interval. Here’s a graphic illustration, Minnesota uses the first option, where the pedestrian buffer interval starts with the vehicle change interval (the yellow light). Overlapping the buffer with the red clearance interval, like the 2nd, 3rd, and 4th options can reduce overall cycle times or increase the walk interval, but is problematic to program and reduces the time for pedestrians to scramble out of the intersection after the sign goes steady and before vehicles are released.


Pedestrian times are set with the following parameters:

1) The walk interval must be at least 7 seconds, unless it’s determined that 4 seconds is adequate in a specific location.

2) The clearance time is the distance from curb to curb in feet divided by 3.5 feet / second walking speed.

3) As mentioned above, in Minnesota, there is the two second red clearance interval after the clearance time.

4) There is another calculation that is made: the time of the entire phase must equal the distance from the near push-button to the far curb divided by 3.0 feet / second. In practice this doesn’t become the limiting factor until road widths become extremely large, in the order of 100 feet or more, so we will ignore it for a while

5) 4.0 feet / second walking speed was formerly the national standard, but is now only allowed if there is a provision for a pedestrian requesting extra time at the button, or pedestrian sensors in use. I am not aware of these features being implemented in Minnesota.


Refuge islands  count as far as setting time to the far curb; in practice the time is usually more generous than the minimum, since the goal isn’t to make every pedestrian wait in the middle, just reduce the times so the slowest possible pedestrian doesn’t impact vehicle operations too much.

Pedestrian Recall- Why or Why not.

Sometimes the “Walk” sign goes on every cycle (ped recall), and sometimes not. Generally speaking:

1) In a fixed time intersection, there is no provision for vehicle sensors or pedestrian push-buttons, so ped recall is enabled all the time in all directions.

2) In a semi-actuated intersection, some directions have sensors and some don’t, so ped recall may be enabled some direction but not others. In the past the presence or lack of pedestrian buttons was a good indication of whether an intersection was actuated or not, but with the advent of accessible pedestrian signals they all have buttons.

3) For fully actuated intersections, ped recall may or may not be enabled if it can be without degrading vehicle operations. Sometimes the limiting factor in how short an intersection phase can be is the pedestrian phase, sometimes the vehicle phase. Lets look at two scenarios.

Scenario 1: 56th Street at Penn Ave , 56th is 35 feet wide, Penn is 40 feet wide. 40 second minimum vehicle phase on Penn, 15 seconds on 56th. The minimum pedestrian phase is  21 seconds (7+40/3.5 + 2) in theory across Penn, 19 seconds on 56th (but is 25/22  seconds as programmed- they made the change interval longer than what is required) Vehicle phase is the limiting factor across 56th, Ped phase is across Penn. So ped recall is programmed in the former but not the later.

Penn Ave S and W 56th Street

Penn Ave S and W 56th Street

The only impact to vehicles is if a vehicle arrives on 56th and there is no traffic on Penn after the minimum green time, they may have to wait through Penn’s change interval before getting their green. This is mitigated somewhat by having the ped signals repeatedly count down and then go back to walk, at the cost of making a pedestrian wait if they don’t arrive during one of the walk intervals.

Scenario 2: American Blvd at Grand Ave (better known as  Home Depot Driveway) Six lane wide suburban-style road about 90 feet wide,  15 second vehicle phase for Grand Ave.  If a pedestrian phase is present, the minimum time is 35 seconds (7 + 90/3.5 + 2) both in theory and practice, so a pedestrian phase impacts vehicle operations pretty severely. It’s such a problem that engineers try to deal with it by either by adding a refuge island in the center (as has been discussed as a solution for various places on American) so a slow pedestrian has to use two cycles to cross. Or by fudging on the timing so the 3.5 feet / second rule only gets to the center of the last lane (the theory no doubt being the driver in the last lane is going to look before starting…).

American Blvd at Home Depot

American Blvd at Home Depot

There is no ped recall on any crossing at Grand and American, but with the 70 second phase for American Blvd. an engineer could enable it to cross Grand with essentially zero impact to vehicle operations. Why don’t they? A blanket policy not too? Didn’t think of it? The way it’s always been done? I don’t know. (The city of Bloomington has been less than friendly and responsive with my dealings with them so I didn’t even attempt to get an answer). Engineers (and from what I perceive of  the general public) just don’t see the act of pushing a button either as especially difficult nor the supreme ideological insult.

 Exclusive Pedestrian Phases (X-Ped)

Ped Recall may or may not have an impact on vehicle operations, but exclusive pedestrian phases almost always have an extremely severe impact.  Rather than sharing the intersection, vehicles and pedestrians use the intersection sequentially, the most inefficient way possible, and a ban on right turn on red is essentially required.  The problem is compounded by the recent reduction of assumed pedestrian walking speed from 4.0 to 3.5 feet a second. The vehicle delays are so extreme that Denver, not exactly a bastion of conservatism, removed the last of theirs fearing gridlock with the reduction of walking speed and the coming of light rail.

So let’s chart our intersections “as is”, with ped recall on all phases, and with an exclusive pedestrian phase. (To keep things simple I’ve used the theoretical 3.5 ft/s walking time even though they are more generous than that on Penn.) Wasted Green Time is when the vehicle signal has to be green because of the pedestrian phase, but no vehicles are there. And we’ve already calculated concurrent phasing, so let’s calculate exclusive phasing. Penn Ave S. at 56th is 65 feet diagonally, 7 + 65/3.5 = 26 seconds. American Blvd at Grand Ave is 11o feet diagonally from curb to curb, 120 feet from button to far curb. This is getting pretty wide, so we better check both calculations. 7+ 110/3.5 = 40 seconds, 120/3 = 40 seconds. So interestingly this is exactly where the two calculations meet.







Obviously this has a big impact to vehicle operations. But what if we decide that motorists can sit and stew because all we care about is pedestrians? Well, it’s not so clear then either. There are drawbacks to X-ped to pedestrians too. It increases their average wait time. (One intersection that was modeled increased it from 34.7 seconds to 49.5 seconds- LOS D to LOS E.) The tendency is to get impatient and jaywalk against the “Don’t Walk” sign, where motorists see they have the right-of-way and thus are not looking for pedestrians. An Israeli study showed  that X-Ped is safer as far as vehicle vs pedestrian crashes at heavy vehicle volumes (18%), but more dangerous at lower vehicle volumes (-8%). Like most everything else in engineering it’s a trade-off between safety and efficiency.

Back to North Star Highways Home

How Traffic Signal Controllers Work, Part 2: Programming a Modern Controller and a Look at Their Limitations

September 11, 2016 at 1:34 pm | Posted in Traffic Signals | Leave a comment
Tags: , , ,

This is Part Two in a series about Traffic Signal Controllers. Part One showed various types of controllers and cabinets, here we continue with a closer look at a 1980s-2000s vintage controller: the Eagle EPAC 300, a 16 phase NEMA (National Electrical Manufacturers Association) TS-1 controller.

Eagle EPAC Controller

Eagle EPAC Traffic Signal Controller

There’s not much to it on the outside: a membrane keypad and LCD screen for the interface and the “ABC” cables at the bottom for the inputs and outputs. The “A” cable feeds AC power to the controller, DC 24 volts from the controller to run the load switches, and all the inputs and outputs for a simple 2 phase intersection. The “B” cable adds phases three and four, and the “C” cable add phase five through eight. Each wire has a single fuction; there is no serial communication with the cabinet.

To the lower right of the controller are serial connectors for communications equipment, generally to a master controller or a traffic control center.  Most newer controllers it’s possible to hook up to a computer with serial connections, and even newer ones may have integrated USB and Ethernet ports, but this is generally only done for initial programming. EPAC stands for Eight Phase Actuated Controller, despite some early units available with only two or four phases, and newer units having 16.

NEMA, conceived in simpler times, only had physical inputs and outputs (I/O) for eight phases. To use more than these, it’s necessary to electronically remap some of the I/O used for other marginal to useless features. NEMA has many features that either were never or are no longer useful; the I/O can be remapped and reused for other things on newer controllers, for instance there is a provision to have a “yellow” light for pedestrian clearance phases. Since instead the “Don’t Walk” is flashed, and these are in the cabinet wired to a the spare center section of the pedestrian load switches, these are normally used to drive things like pre-emption lights and blankout signs. Some controllers also had a proprietary “D” connector for such things.

NEMA controllers can operate by themselves as free-running controllers, or as a master controller that coordinates many of them. The EPAC 300 series is more or less still in production as the M50 and M60, though the Eagle name has been removed and is still owned by Siemens after the Eagle signal head business was sold off to Brown Traffic.


Some of the Screens

Although there are many screens, these can be broken down into (1) “Status” screens, that show what the controller is doing or logs of what it’s done, and (2) “Programming” screens where you enter what you want it to do. I’m omitting anything having to do with: Density and Time of Day Programming (where cycles change based on traffic volume and time), Coordination, Logging, Communication, Start up Sequence, Preemption, and Flash Mode. These are probably of marginal interest to non-engineers, some of these I don’t understand myself, and skipping them eliminates probably 95% of the complexity. Also omitted is anything having to do with actuation. Although some collectors will add video detection and pedestrian push-buttons to their setup (and there’s an amazing variety of buttons to collect), I have chosen not to. I have four phases with two associated pedestrian phases that run in sequence without waiting for calls.

Here is the main screen:


Let’s see what it’s doing first: go to 1-ACTIVE STATUS and then 7-INTERSECTION. The second line shows the phases and then “V-SIG” says they’re all red except phase 4 is green, below that P-SIG pedestrian outputs are “D” for “Don’t Walk” except Phase 4, which “d” means it’s flashing the “Don’t Walk” for the clearance interval. All the phases with physical outputs (1-8) default to red and “Don’t Walk” all the time unless programmed otherwise. In this case 4-8 are not programmed. If there were any Calls,  these would be indicated under V-CALL and/or  P-CALL. The Overlaps, with letters instead of numbers are additional outputs for driving such things as a protected/permissive display where you have a green ball and green arrow in the same direction. Not all of the vehicle overlaps have designated physical outputs. There are also pedestrian overlaps on a different screen (where you might give a “Walk” signal on one side only if there’s a green arrow the same direction; none of these have physical outputs).



Now suppose we want to play with things and see where values can be entered and changed. Go back to the main screen, select 3-PHASE DATA, then 1-VEHICLE TIMES. The MIN GRN  is the green  time in seconds for each vehicle phase. Phases 5-8 (and 9-16 on the second screen if you page down) have default values except for the zeros for MIN GRN, which disable them. PASS/10, MAX # 1, and MAX # 2 would extend the green if needed vehicle detection was used, since they’re not here they have no effect and  the controllers default values are left in place here. YEL/10 and RED/10 are the yellow and red times in tenths of seconds.



Going back a screen and then to 3-PEDEST. TIMES. Phases 2 and 4 have pedestrian signals hooked up, with WALK time at 30 seconds and PED CLR. at 20 (5-8 will not activate since there are no associated vehicle phases and the numbers you see are default values). The FL WK will flash the walk light, now a MUTCD no-no, but previously used in DC to indicate a crosswalk where there vehicles might attempt to turn across. EXT PCL will continue flashing the “Don’t Walk” through the red and yellow vehicle phases. ACT RIW will hold the phase at “Walk” until a conflicting call come in.


Going back a screen and selecting 5- V&P RECALLS, we see Recall is selected for all four vehicle phases and both associated pedestrian phases (indicated by a “2”) So it will go through the cycle and service all phases without any demand from pedestrian push-buttons or pedestrian sensors. Entering a “1” will generate a call to test the controller in case physical test buttons aren’t provided in the cabinet. DELAY allows you to set a time before Recall is activated.


Why Can’t Signals Do What We Want?

Quite often the question gets asked  “why can’t a traffic signal do this, or why doesn’t it do that”.  Now with a bit of understanding of controllers and  cabinets,  lets go over a couple of  broad reasons why things are or are not done a certain way.

mnmutcd-cover1) It’s required or banned. The Minnesota Manual on Uniform Traffic Control Devices (MUTCD) based on the federal one with some very minor changes, has options , strong suggestions, and absolute requirements. Engineers are not likely to deviate from suggestions, and may not deviate from requirements. Banning creativity might preclude the best response to a specific situation, but on the flip side road users can expect uniformity nationwide. When traveling, they’re not going to encounter a flashing purple arrow in Peoria or  a pink strobe-light in Paducah and then have to try to figure out what those mean. If a city really wants to try something new out, say a red arrow on top of a standard three light signal to emphasize no turns on red (not currently a legal configuration), they can apply for permission to experiment, do a study, and then either remove it if a study shows no benefit, or it may be adopted into the MUTCD if it does.

2) It’s often a zero or negative sum game with intersection capacity a fixed resource,  since  pedestrians and motorists compete for the same resource, and the needs are completely opposite. Right on red or longer overall cycle times: pro-motorist, anti-pedestrian. Exclusive walk phase or leading pedestrian intervals: anti-motorist, pro pedestrian. Other people may not see it this way, and are surprised when I describe “pedestrian improvements” as “anti-motorist”, but ultimately that’s the effect even though there may have been intent to harm vehicle operations (although sometimes there is). Additionally since pedestrians move a lot slower, sometimes a very modest benefit for pedestrians can produce an extremely severe impact for motorists, like ped recall across a wide suburban-style road. More details on some of these scenarios will be in the next part.

3) A lot of equipment on the field is very old. Even computerized controllers tend to use  1980s vintage microprocessors like the Motorola 86040, and the need for standardization  dampens innovation. This is more true in the cities were demand for pedestrian amenities is greater. Until very recently Minneapolis even had electro-mechanical controllers.  Old controllers may not have the capabilities that people want. But it still works and cities are not made of money. In some cases even what seems like a minor change could require a complete cabinet replacement, or even a bigger cabinet requiring a bigger concrete pad and redoing a lot of the wiring, and the cities are not likely to do that just because someone wants a leading pedestrian interval or something.

Eagle Electromechanical Traffic Signal Controller Cabinet

Eagle Electromechanical Traffic Signal Controller Cabinet

Although we’ve come a long ways from a free-running E/M controller in every light, even the newest controllers are nowhere near as powerful as a PC. (You may have noticed the line “16 MHz CPU” and a 2000 firmware date on the main screen of mine.) Even if they were, in most situations the input and output are still limited, with the only input available being  vehicle sensors and pedestrian buttons and only output being lights. They can’t look at a a bunch of cameras focused on the intersection and down the block and make decisions based on it, or sound a fog horn if they see a motorist inching into the crosswalk.  Things are getting better with multiple vehicle sensors that can thus measure vehicle speed (one controller even has a “feature” that will turn the light red if an approaching motorist is speeding) and the new Linux based controllers that are more flexible in certain ways.

Sometime upgrading the firmware on existing equipment is possible, but besides the expense and need to go out and swap out chips or update the flash memory, engineers like to have known, stable, and consistent firmware. At least once Mn/DOT has had to back out and revert to an earlier firmware version because of bugs. Mine is running 2000 firmware and would require an upgrade to do leading pedestrian intervals.

There is also the “SMART” Signal system, which tries to harvest and aggregate data gathered by standard controllers together using industrial PCs and use it to dynamically optimize timing in a way that’s not possible with traditional manual coordination. The controller is an Econolite ASC series, the Mn/DOT standard and most widely used in Minnesota.


4) Old-Fashioned engineers. To end with the elephant in the room, sometimes it really is engineers that are just set in their ways and not taking the needs of pedestrians seriously or being open to modern ways of doing things. And it’s not just signals; they admitted the trails associated with the St. Croix Crossing are only worked on “as we have time”.  And closed down the sidewalk on the Hastings Bridge for the entire duration of the project even when there was no work being done anywhere close to it. However the point I’m trying to make is there usually more too it than some “dumb engineer sitting in a chair in Medina” not wanting to put down his donut in order to click an icon on the computer in front of him.

Overall, the idea isn’t to pass value judgments or state my opinions on what is or is not a worthwhile change, even though perhaps the fact that 95% of the time I’m  a motorist shows through. Rather, what I’m trying to do is lay out the reason things are the way they are.

Back to North Star Highways Home

How Traffic Signal Controllers Work, Part 1: An Overview of Controllers

September 10, 2016 at 2:37 pm | Posted in Traffic Signals | Leave a comment
Tags: , , ,

Electro-mechanical Controllers.

Crouse-HInds PCE-3000 E/M traffic signal controller.

Crouse-HInds PCE-3000 E/M controller.

Electro-mechanical controllers date from the beginning of traffic signals up to new installations in New York City into the 2000s. I’ve placed a few short videos below, and there really is no substitution for viewing them to get a sense of how them parts move, but here’s a description and some still pictures first.

1) A dial with keys on it rotates, very much like a mechanical timer. Some used multiple dials for different timings on different times of day. The speed of the rotation is controlled by different gears that can be installed.

Crouse-Hinds Electromechanical Traffic signal controller dial

Dial and Keys

2) As the keys pass a micro-switch they temporarily close it to activates a solenoid or motor. The motor rotates a camshaft with cams. Each cam has lobes that can be moved or broken off around the perimeter. The knobs hold an array of piano switches open. Removing a lobe causes an individual switch to close when the cam is in that position. Each signal indication on each street is wired to one of these switches.

This is the simplest possible installation, a two phase, no red clearance operation. The first key changes Main Street green to Main Street yellow, the second changes Main Street yellow to red and Side Street red to green, and so on. The key in the middle prevents the cams from getting out of sync by not allowing it to advance unless it hits a switch.

Crouse-Hinds Electromechanical Traffic Signal Cotnroller Cams

Camshaft, Cams, and Switches. Note some switches are in the closed position due to the metal lobes being removed from the cam, and some of the cams are color coded. Here the Main Street Green and Side Street red switches are closed and thus the lights would be lit.

Coordination is actually possible with E/M controllers. A mains wire runs from one controller to another, and a relay removes AC power to the controller downstream to prevent it from getting out of synch.

Video 1: Shows the outside of the controller

Video 2: Shows the inside and the dial rotating:

Video 3: Shows the dial, then the cams

Video 4: Closeup of the cams

This area tended to use Eagle E/M controllers. Crouse-Hinds controllers like mine (from probably the early 1970s in Moberly, MO), used motors instead of solenoids creating a whir instead of a clunk when moving the cams. Also, the cams have metal lobes that are replaceable, so if you make a mistake arranging cams or want to change things you can. With Eagle and other brands if you break a lobe off that’s it, you need to replace the entire cam to undo it. Cams are still available from one supplier, but there’s talk from signal collectors at having them 3D printed due to the expense and hassle of working with distributors.

In some ways E/M controllers can be even more flexible than early electronic ones. There’s no reason you couldn’t “program” leading pedestrian intervals or even flashing yellow arrows, something early electronic controllers are not capable of.

Hybrid Controllers

E/M Controllers may work well for fixed time applications, but they did not really handle actuation (where they respond to serve a vehicle or pedestrian phase).  This was more important with the rise of the suburbs, where traffic demand is less predictable and fewer traffic signals exist, rendering fixed time inefficient. Basically the timing was generated by vacuum tube electronics charging an absolutely huge oil filled capacitor, which would eventually discharge through the solenoid, moving the cams. The rate of charging, and thus timing was controlled by knobs on the front which are variable resistors.

Other items in an E/M or hybrid cabinet might include E/M timers to switch dials or put the cabinet in night flash, an E/M or electronic flasher, and transfer relays.

Electronic Controllers.

The first electronic controllers were designed as drop-in replacements for E/M controllers and used a lot of the same terminology and philosophy. The times were set by knobs and dip switches and such and each brand was proprietary. Gradually a standard (NEMA TS-1) was implemented to the connector from brand A was the same and each wire would do the same as brand B, and certain functionality was required, so you wouldn’t have to essentially replace the cabinet to switch controller brands. Later controllers went to a keypad with a few numeric LEDs, to program them you’d look up on the manual something like “Function 132 Main Street green time”. Later LCD displays were much more user friendly. Philisophically there was a switch from “percentage of dial times’ to programming in phases.


There are a few traffic signal collectors that tinker with these early controllers. Very few. Some of the flashing LED displays are fun to watch, but the connectors and documentation are hard to find, electronics do not age well, especially when unplugged for extended periods of time, and newer, much more user friendly, reliable and versatile models are available at attractive prices. So collectors usually use newer electronic models, or else old cool E/M ones.

Electronic Controller Cabinets 

Here is a simple demonstration cabinet with the components annotated.


Simple Traffic Controller Cabinet

1) The controller itself: The actual controller. Inputs and outputs are 24 volts DC,

2) The conflict monitor: makes sure that the controller doesn’t mess up and do something like turn both directions green. If it detects something like this of if the controller fails, it will put the signals into flash

3) The back panel- contains the terminal blocks (towards the left), flash transfer relays (towards the right), and sockets for the load switches and flasher (towards the bottom)

4) The load switches: These take the 24 volt outputs of the controller and control the mains electricity to the signals themselves, essentially solid state relays. Each load switch can control up to three indications, generally the red, green, and yellow of a single direction.

NEMA Traffic Light Flasher Module

NEMA Flasher Module. the load switch is similar but has a third solid state switch module instead of components on the main board.

5) The flasher: Generates a flashing signal independent of the controller, so the signals can still flash if the controller itself goes down.

6) Circuit breakers and power filters

7) Manual control and test switches

In the real world it would likely be jammed full of equipment. The second shelf would have fiber optic communications equipment and pre-emption equipment, the manual controls would be mounted to the door and the top shelf filled with loop amplifiers or video detection equipment.

A newer standard is TS-2, which replaces some of the hard wiring with serial connections. Also available are 2070 cabinets, where the hardware is purchased separately from the software, and the controller is more modular, accepting various function cards. Minneapolis uses 2070 cabinets, the other agencies use NEMA.  The newest controllers are ATC, which combine some of the elements of NEMA and 2070 controllers and run on standard Linux, enabling more flexibility with software programs.

My Own Collection:

Here is my home setup. It’s a four phase sequential setup.  It lacks a conflict monitor (they are difficult to wire and in a home setup serve no real purpose unless you own a real cabinet, in which case it’s a pain to rewire the whole thing to not need one). Beyond the bare minimum to run some lights (controller, cable, load switches) it has a flasher and flash transfer relay, and some manual switches to put it into flash mode. The controller is an Eagle EPAC 300. Despite using Eagle signal heads, their controllers are not really used in this area. The EPAC series is a favorite of collectors since it’s (relatively) simple, new enough to be reliable but not too new to be complicated and expensive, and other collectors use them so help programming is available. I know there’s exposed wiring, but  it’s semi-protected and is in the basement where the pets don’t go.


Home Setup

The controller I got for $75, and the load switches I got eight for $50 on eBay. Surplus dealers will buy a bunch of stuff at municipal auctions and try to flip them, not knowing that they’re essentially worthless unless an occasional collector wants one. No city is going to buy a 20 year old controller off eBay. Unfortunately the sellers can’t test the equipment and sell it “as-is”; they can’t even plug it in to see if it powers up because there’s no standard AC power cable, and it may be password protected which would require new firmware to clear.

The cables are by far the hardest to find because they tend to stay with the cabinet and get discarded when the cabinet is recycled, new bare connectors are close to $100 from electronics suppliers. I finally got some cables from another collector who is also a real signal tech. Obviously some knowledge of electronics and ability to work with mains current safely is required. But of the 175 pins on the three connectors, I’m only using less than three dozen, and a simple setup with two phases can get by with as little as 14. (10 outputs, one for each signal light, AC power, neutral and ground inputs, and 24 volt output to the load switches.)

I’ve used a bunch of terminal blocks from a local surplus store for much of the wiring. The load switches are designed to plug into Jones sockets in cabinets, but with a little coaxing they will accept standard female spade lugs. Load switches aren’t especially expensive or hard to find, but a bunch of 24 volt relays can also be used.  Wires to the signals are 4, 5, and 7 conductor “SOOW” cable, usually used for wiring industrial equipment.

Home  Controllers

It should be mentioned that all this is way overkill if you just want a traffic light in your garage or man cave. Single board  controllers are available that will drive a single signal and are easy to use. All that is involved is wiring a mains cord to the controller board, then each of the three lights to the board.  The 150 watt rating is good enough for any signal, even a 3M. Some serious signal collectors will use these too; some of them have no interest in real controllers, and may have signals scattered through their property instead of a single “signal saloon”. The main problem with these is there’s no protection on the output circuitry like a real load switch; a bulb burning out can occasionally damage them.

Finally, a video of my setup in operation. I  have fairly broad tastes (as broad as signal collecting can be) with stuff ranging from 1940s art deco signals to modern but unusual stuff to examples of signals common to this area with an interest in both lenses and LEDs and a particular interest in pedestrian signals.

Since I know people are wondering: Most of the signals I got from eBay or other collectors. The cheapest one was $25 (the St. Paul Durasig pedestrian signal); I have over $1000 invested in obtaining and restoring the neon pedestrian signal. Some collectors have bought new from distributors, but understandably they tend not to be friendly to collectors. They’re intrigued by the interest private collectors show, but at the same time they have a job to do and are more interested in a city that might buy 500 as opposed to some guy that wants one for his basement, and then decides it’s too expensive when he finds out the cost.

Back to North Star Highways Home

Some Sign and Signal Fails

September 9, 2016 at 3:33 am | Posted in Traffic Signals | Leave a comment
Tags: , ,

Here’s some Sign and Signal Fails (as well as some that are simply unusual) from around the state.

Upside Down Traffic Signal Warning Signs

Irish Traffic Signals?

I can understand getting one sign upside down (well, no I can’t, but whatever, but three of them in the same area? These were when US 212 was being built around Prairie Center Drive.


Is this Hamline Ave or Hamline Ave?

No question what the name of this street is.


Walk or Wait

These signals are sending a mixed message. Likely the one on the right got twisted into the wrong direction by a snowplow


Surplus 61 Shields

When US 61 became MN 61 they missed a couple in out of the way locations; these were years after the change.  I believe these have finally been replaced.


A little late

This sign was made decades after US 61 had been removed from this road, I guess the locals still call it “Old 61”


In Wisconsin…

These have all been replaced with real MN route markers.


Sloppy Contractor Work


Some Interesting shields


Signs mounted funny to alter their intended meaning


Another contractor innovation. Photo by Andrew Filer


Wisconsin and US markers are so similar that screw-ups are almost non-notable

Is it a Ball or an Arrow?


Elm Creek Blvd and County 81

Not a fail, but interesting. This intersection of County 81 and Elm Creek Boulevard doesn’t seem to be particularly interesting, but look closer when the light turns green. Then move closer.

The green ball magically changes to an up arrow! 3M, in addition to common balls and arrows and much less common pedestrian lenses, had all kinds of special contraptions for use in their Model 131 programmable visibility signals. One of these was a bimodal indication that would show a ball from a distance and and an arrow up close.

The reason for this is that balls are much easier to see from a distance. Once drivers approach the intersection, however, it changes to an arrow to direct the driver where to go. At County 81 and Elm Creek Blvd., the eastbound and westbound lanes are widely separated by MN 610. What you have is a one way with 70 mph traffic in a place you wouldn’t expect one. Using only left and up arrows reinforces the point that you are absolutely not supposed to turn right.

Like most of 3M’s odd stuff, these bimodal indications are extremely rare. Only one is known to exist in a private collection and only a couple are known to exist on the street. Here’s a video of what you see as you approach, in a car the transition would be much faster.

A Danger to Pedestrians

At the same signal, facing the other way is a problem.  Ten points to whoever can instantly see what is wrong with this:

Traffic Signal at Elm Creek Blvd and County 81

Elm Creek Blvd and County 81

Although this configuration is allowed by the Minnesota MUTCD 4D.4 #2 and #4, (which states that pedestrians may enter against a green arrow when indicated and drivers following an arrow must yield to pedestrians lawfully in the intersection) that doesn’t mean this is a good idea. Most drivers and pedestrians aren’t in the habit of reading obscure sections of the MUTCD or state statutes for exceptions to normal rules, and on the surface this signal is telling both motorists and pedestrians simultaneously they have the right of way.

As for fixing it, one option is to just replace it with a standard green ball, and lose the extra feedback the signals provide to keep drivers from turning in the wrong direction. Probably a better option is a flashing yellow arrow right turn signal, that would give a green arrow under normal conditions, and a flashing yellow arrow during a pedestrian phase (this light was installed before this option existed). Although a red ball mated to a flashing yellow arrow is an illegal configuration, state law allows a placement of a sign allowing turns on red arrows after a stop, so that’s what I’d do here.

Since this intersection operates in split phase, there is nothing wrong with the 3M arrows facing north since the west crosswalk will never have a walk while northbound traffic has a green.



For want of an unrusted bolt, this this signal failed.

Collapsed Traffic Signal

Signal Fail

Signal Fail

More subtly, do you see anything wrong with the black signal on the right?


Silver Bell Road and Cedar Grove Blvd, Eagan

How about now…

Yes, this picture is right-side up. The

Yes, this picture is right-side up. The “dive-bombing” Eagle logo

This was a mistake during an entire production run. The companies official line is “China sent us a bad mold but the signals themselves are as usual, the highest quality”. Now that Siemens has spun off its traffic signal head business to Brown Traffic, they plan to replace the Eagle “thunderbird” logo (which somehow never got removed from these during the entire time Siemens owned it), with their new Eagle logo.

Looks like someone is going to lose a game of “3M Jenga”.

Signal Fail at 46th St. and 46th Ave, Minneapolis

Signal Fail at 46th St. and 46th Ave, Minneapolis

Old brittle poly signals + wind


Signal Fail at W Lake Street and Thomas Ave, Minneapolis

Signal Fail at W Lake Street and Thomas Ave, Minneapolis




The space between the jersey barriers is a temporary pedestrian crossing. Note the “No Crossing” signs that weren’t removed.


This one’s kind of subtle. Remember a bicycle is a vehicle


Back to North Star Highways Home

A History of Pedestrian Signals

September 5, 2016 at 3:03 pm | Posted in Traffic Signals | Leave a comment
Tags: , , , ,

Early Days.

Since almost the beginning it was recognized that additional accommodations needed to be made to pedestrians. As early as 1930 federal standards encouraged signals at the corners of the intersections, where they could easily be seen by pedestrians and cars alike, however in practice agencies would go the cheap route and hang a single stoplight over the center of the intersection. (Earlier I mentioned the Darley, the ultimate in cheap, which used only three bulbs and had a simple, integrated controller). Such a signal would be hard to see by pedestrians and traffic near the intersection alike (in fact there were gizmos that consisted of a prism to be mounted to your rear-view mirror to make it easier to see overhead signals). As a rudimentary accommodation, the Crouse-Hinds “Smiley Face” lenses came out, that were especially designed to throw light down.

Later, encouraged by traffic signal companies that wanted to sell more segments, a separate white on black “Walk” indication was introduced. Typically they would be added to the bottom of an existing signal.

Early pedestrian accommodation, first with the green lens designed to throw some light downward, then with the separate “Walk” lens

Early pedestrian accommodation, first with the green lens designed to throw some light downward, then with the separate “Walk” lens

Orange and White vs Red and Green

From the early days orange and white were standard for pedestrian indications, but red and green were also used. Also, later on some of the whites were modified to “lunar white”, by using a light blue filter over the yellowish incandescent lamps. The theory is that this made the white color more distinctive as opposed to all the other incandescent lamps around, but it decreased the overall brightness, and was never universally adopted until the LED era, with LEDs being naturally lunar white. There was also experiments with aqua blue “Walk” lenses that never caught on, nor were officially sanctioned.

Further improvements.

8″ circular lenses were hard to read across increasingly wide streets, so 9″ square lenses were introduced, mounted on adapter sections that attached to the bodies of standard vehicular sections. With the advent of 12″ vehicle signals the pedestrian signals grew also, although for many years it was common to have 9″ pedestrian signals with 12″ vehicle signals.

In 1961 “Wait” was replaced by “Don’t Walk”, and flashing the clearance interval was introduced.  Telling pedestrians what to do rather than what not to do was seen as more direct (and audible pedestrian buttons still say “Wait”), but “Walk” and “Wait” looked similar enough to be confused on a small lens across the intersection. People being like the are, pedestrians tended to notice how long the clearance interval and buffer were, and would start walking on the “Wait” indication, sometimes with very little time actually left. Flashing the clearance and thus separating it from the buffer gave differentiation to what in reality  meant “start walking if you can move fast” and “get out of the intersection now”.  The later introductions of countdowns gave further feedback as to exactly how much time they had.

As a side note, there used to exist flashing “Walk”s, and most controllers have a setting for selecting it. Notably Washington DC would use to to indicate when permissive vehicle turns were allowed to cross the crosswalk, and thus for pedestrians to be on the lookout for cars failing to yield.

Later evolution to two indication Wait / Walk and then square Walk / Don’t Walks.

Later evolution to two indication Wait / Walk and then square Walk / Don’t Walks.

The California Neons

Although used in other warm climates and New York, neon pedestrian signals are most associtated with California. The original versions had an red “Wait” and green “Walk” spelled out in neon in very heavy casting. Later models were lighter weight, spelled Don’t Walk and “Walk”, and were bimodal- the clear red tubing was overlaid on top of the opaque green tubing resulting in the “Walk” being able to light up two colors. This was the only model of pedestrian signal ever to have the correct apostrophe in “Don’t”. In the early 1970s new California standards required standard lettering rather than spelled out in neon, so plastic lenses were backlit with neon, and the colors changed to orange and lunar white.

Below is a classic bimodal “Walk / Don’t Walk in operation

Newer Backlit E8 Neon Pedestrian Signal

Newer Backlit E8 Neon Pedestrian Signal

Newer Backlit E8 Neon Pedestrian Signal Internals

Internal view of the Econolite “E8” neons. The black boxes to the right are the high voltage (5,000 and 9,500 volt) transformers for the tubing, which is contained in the metal trays mounted to the door.

The hot dry climate of California was ideal for neon, and many existed from the 1980s up to the last one being converted to LED in 2012. By contrast in New York the cold affected the “Walk” tubing . The “Don’t Walk”s were filled with neon, but the “Walks were filled with mercury/argon, similar to a fluorescent lamp, and with the same dimness problem in cold weather. And the high voltage tended to short out in humid climates like Florida.

As a competitor to Econolite, a company called Indicator Control Corporation began by making their own neon signals in the 1970s. Originally these still had heavy transformers, but the neon tubing was in a sealed plastic “modular” unit. Later the magnetic transformers were replaced by switching power supplies and also incorporated within the modules, so maintaining neon became as easy as a modern LED; just toss the entire defective module and put a new one in. Notably, ICC introduced the standard 16″ by 18″ form factor that is still with us today.

ICC Neon Pedestrian Module

ICC Neon module removed from external housing. These are about the same size and can be used like LED modules.

ICC Neon Pedstrian Module Internals

ICC module with cover removed. The power supplies have been reduced to small circuit boards behind not visable behind the tubing rather than huge, heavy transformers.

Incandescent Bimodals

The neon signals had the advantage of being large and easy to see, however they were expensive and did not perform well in cold weather. As an alternative some companies made bimodal signals that were back-lit with incandescent bulbs; the “Don’t Walk” would light up red or orange, and then only the “Walk” would light up green or white. The Marbelite LPS-20 was one example. The lamps for the “Don’t Walk” indications were 65 volts and wired in series so if one burned out the signal would not light up with an orange “Walk”.  (These bulbs are now completely unobtainable, and collectors that don’t have working original lamps need to rewire their signals to accept regular voltage bulbs).  Unfortunately the plastic reflectors would get brittle and break, and even catch on fire. (Wink-O-Matic signals also had problematic plastic reflectors). Later Marbelite released the MPS-20 using the same casing that had separate sections for the “Walk” and “Don’t Walk” using metal reflectors.  Other companies like ASI and E-Man had casings of cheap stamped aluminum instead of cast, but at least had metal reflectors.

LPS-20 Pedestrian Signal Internals

nternal view of the LPS-20. The metal plates inside were done to repair the reflectors, which had broken under the weight of the sockets while being shipped

As a side note, a fellow collector came up with a LPS-20 still in it’s original box that indicated it was to be shipped to the city of St. Paul. As I’ve never seen anything like it there, this is a mystery. Maybe they ordered one as an evaluation or something.

The Barnes Dance (Exclusive Pedestrian Phases)

Almost from the beginning it was recognized that exclusive pedestrian phases (all vehicle traffic is stopped and pedestrians can cross in any direction, including diagonally) were desirable in certain locations that had both heavy turning traffic and heavy pedestrian traffic. In most cases this is done with standard signalling just by having the vehicle phases red and the pedestrian phases white in all directions, but some notable variations. The Harriss signal was used experimentally in the 1930s in New York, and was an effort to be more emphatic than a standard 8″ light, with a huge red hand (representing a policeman’s upright hand).

Peds 08

Harris Signal.

The bottom green light indicated vehicle traffic should go. A red hand surrounded by violet lights indicated that vehicle traffic should stop. When the orange light at the top was also lit pedestrians should go in all directions, the light went out as a blind clearance interval before the  hands changed to the green light, but this 5 second clearance interval meant that they better run. Later this became more widespread in New York, and became known as the “Barnes Dance” after a 1960s era transportation commissioner.

The experiment with these rather elaborate signals were not continued. After this another thing tried was a backlit sign panel that could flip between “Stop” and “Walk”. This was not widely adopted either.

In Boston, which had a lot of very old equipment that was never updated and many intersections without pedestrian indications, a simultaneous red and yellow light would indicate that all cars should stop and pedestrians should go.  By 2012 when this video was shot such installations were very rare, and there may be none left.

Exclusive phases were never the norm, and recently they have been falling out of favor even more. Except for the very specific scenario of heavy pedestrian traffic and heavy turns it’s one of the most inefficient ways possible to run an intersection, and the new federal mandate of 3.5 ft/second of pedestrian clearance time instead of 4 compounded the problem. In Denver they removed the last of them to prevent gridlock with the coming of light rail.

Recent Changes

Although the “hand” and “man” had been experimented with since the 1970s (they were adapted from signal designs in Europe), in 2000 is when they became mandatory for new installations. In 2003 came the requirement that they be filled in rather than outlines, and in 2009 countdown displays became mandatory and leading pedestrian intervals (to give pedestrians a head start before releasing traffic) were formally allowed. Here’s a comparison of five different pedestrian signals from my collection- neon, incandescent, and three styles of LEDs.

 An Idea: Ditch the Pedestrian Change Interval

The idea for a pedestrian signal got started when streets got wider, traffic got faster, and things got more dangerous for pedestrians stranded in the intersection when the light changed. So the “Walk” light was added to tell pedestrians when it was safe to leave the curb. Since then, we’ve added the “Don’t Walk” and countdowns, but we’re still telling pedestrians that we’re the better judge than they are if they can make it across. And the long clearance interval, designed to accommodate the slowest of pedestrians, encourages jaywalking and disrespect for traffic control devices, and destroys the LOS for law-abiding pedestrians.

Now that we have countdowns, what about eliminating change interval entirely and let pedestrians use their own judgement as to whether they can make it across the intersection? An 80-year old has a different crossing speed than a bicycle on the sidewalk or a multi-use path, but they are treated all the same. The countdown would start at the beginning of the Walk Interval, then count down to the buffer. We could extend the buffer a bit if we wanted to (but making it too long would again encourage jaywalking and reduce LOS). An eight second buffer, rather than the standard three, would allow 10 seconds of the red hand before the light turns green for cross-traffic.  Even at a narrow intersection, you could probably triple the walk interval. If a visually impaired person pushed a button, it could say “20 seconds left, Penn Avenue is 45 feet wide”. Would this increase crashes? I don’t know.

Back to North Star Highways Home

A Gallery of Traffic Signal Lenses and LEDs

September 4, 2016 at 1:56 pm | Posted in Traffic Signals | Leave a comment

Early Lenses

Traffic signal companies have always been more in the business and sales rather than production, and thus most of the manufacturing was contracted out. This is especially true with lenses. Originally there were four major glass companies making lenses: Kopp, Corning, Macbeth-Evans, and Holophane.  Macbeth-Evans, notable for it’s Depression Glass, was merged into Corning in 1936. Holophane eventually chose to concentrate on it’s much larger line of industrial light fixtures. Today  Kopp Glass, founded in 1926 from the reorganization of the Pittsburgh Lamp, Brass and Glass Company is the primary manufacturer. Although there’s not much of a market for glass traffic signal lenses nowadays, Kopp still has the capability and will accept orders if someone wants a thousand of them.

Crouse Hinds “Command” lenses.

Early lenses sometimes had “Commands” printed on them. In the early days traffic signals were new and the position of the lights not uniform, so instructions as to what to do were seen as beneficial. These are rare nowadays and a set usually sells for over $300 on the collectors market.

Lenses 01

Crouse-Hinds “Command Lens”

Adler “Bar” Lenses

These were an attempt to accommodate the colorblind (in the days before the placement of colors was universal). There was a bar running across the lens- vertical for “go”, diagonal for “caution” and horizontal for “stop”.  Although these didn’t catch on, the design can still be seen in today’s light rail signals. These were never common and are extremely rare today, a set of these usually fetches $1000 on the collectors market.

Lenses 02

Adler Bar Lens

Kopp “Diamond Pattern” lenses (#27 and #44)

A common early pattern used by a number of manufacturers.

Lenses 03

Kopp Diamond Pattern Lens

Kopp 27 Lens

Holophane “Spiderweb” Lenses

In my personal opinion the most beautiful of lenses, collectors and enthusiasts refer to these as “Spiderwebs” due to the obvious resemblance. These were used exclusively on 1940s era GE signals. Siemens adapted these for Europe where they are very common, but in the US these, not lasting long and used by only one manufacturer, are somewhat rare. A set of these usually fetches over $300 on the collectors market.

Lenses 04

GE Holophane “Spiderweb” Lens

Corning “Smiley Face” lenses (Type T and Type T-1)

Known as “Smiley Face” due to the distinctive pattern designed to throw light downward for the benefit of vehicles and pedestrians near the intersections when the signal was mounted on overhead wires.  There was an earlier version with larger beads, but it was revised once ITE standards came out which regulated the light output of lenses. Although distinctive, the later version isn’t especially rare, the CH “art deco” signals were extremely popular back in the day and as are desirable to collectors and thus many have been preserved when they’ve come down.

Lenses 05

Crouse-Hinds “Smiley Face” lens

Later Lenses (1940s to present)

Over time lenses evolved to be less unique. Most of these are still rather common and a set goes for less than $50 a set.

Corning T-3

The T-3 was used mainly in later Crouse-Hinds signals, some of the later ones are (apparently) mis-labled as 1-3. There was also the T2, the “Faint Smiley” that maintained the ribbed outline of the original “Smileys” but not the beads.

Lenses 06

Corning T-3

Kopp “Brick” lenses (TL-4655 and TL-4955)

Following the “Spiderwebs”, GE again commissioned their own design with their own logo, commonly called the “Brick” pattern. This was later adopted by Econolite when they bought out GE’s signal business. All but the oldest GEs and all but the oldest and newest Econolites will have this pattern, including all in the time frame that they were used in the Twin Cities.

Lenses 07

GE “Brick Pattern” lens

Kopp “Sawtooth” lenses. (TL-4677 and TL-4777)

Smaller manufactures would typically use “Sawtooth” lenses so they are extremely common in other areas of the country, but Minnesota has always shown a preference for Eagle and to a lesser extent Econolite, both of which used their own patterns.

Eagle Lenses (Kopp 88, 88.1, and 88A)

Very early Eagles would be equipped with #27 lenses, but Eagles own design, the Kopp 88 series, came in the 1940s and the same basic design was used by them until the end of the glass lens era. Although the pattern was not that distinctive, they featured Eagle’s logo and are by far the most common 8″ glass lens still in the area (often in the yellow sections of older Minneapolis signals).

Eagle 88A Traffic Signal Lens

Eagle 88A Lens

Polycarbonate Lenses

With the coming of polycabonate signals came polycarbonate lenses (although aluminum signals could also be ordered with them, and polycarbonate signals could also be ordered with glass. These tended not to last, they’d oxidize in the weather, and get scorched when used with higher wattage bulbs that were on a lot, like main street greens or side street reds.

Eagle Polycarbonate Traffic Signal Lens

Eagle Polycarbonate Lens

12″ Lenses

Glass and Polycarbonate Lenses

Like most collectors and enthusiasts I have little interest in 12″ lenses. There isn’t anywhere close to the same variety of the earlier, more interesting 8″ lenses, and 12″ signals are awkwardly large for indoor display. But here are a few of them that I have laying around, a new 12″ Lexalite red, a Eagle branded polycarbonate yellow, and a generic glass green from Poland.

3M Lenses

Lenses for the 3M Model 131 signals were square plastic with a Fresnel pattern. Balls or arrows were stenciled out.

3M 131 Lens

3M 131 Lens

Earlier LED Modules

The first LED modules were actually designed to utilize existing sockets, reflectors, and lenses, and were initially available in red only. the primary companies making them were Electro-Techs, Ecolux (later bought by GE), Cooper (which was popular in New York but exited the business, Dialight, Leotek, EOI, and Swarco.

Early screw-in LED. These are still the only way to use LEDs with an 8″ 3M adapter

Early screw-in LED. These are still the only way to use LEDs with an 8″ 3M adapter

The earlier modules all had the individual LEDs exposed, this is from Leotek. Dialight and Cooper were arranged in circles rather than rows.

Early Leotek LED module

Early Leotek LED module

GE RX-11s.

These were an attempt to compete with the later emerging “Incandescent Look” or “Uniform Look” designs while using the older technology of arrays of many low-power LEDs by putting a diffuser lens on top to spread the light. There’s a lot of these in the Twin Cities, but they have not held up well, are now reaching the end of their life with many partial failures, and  tend not to be liked by collectors.



GE RX-11 with the outer lenses removed

GE RX-11 with the outer lenses removed

Ecolux / GE “Honeycomb”

These were a variation of the exposed LED modules with a small magnifying lens over each LED. There are a number of these still in Minneapolis and St. Paul and they tend to outlast even the newer RX-11s.

Lenses 18

GE “Honeycomb” LED

Later LED Modules

“Diamond Pattern”

The first of the “Incandescent Look” LED modules, Dialight and Leotek both produced models that were a throwback to the early diamond pattern lenses. These are impossible to distinguish from each other in the field.

Dialight “Diamond Pattern” LED

Dialight “Diamond Pattern” LED

Later Incandescent Look

The later GE, Dialight, Leotek, and EOI modules are very similar and are difficult for even collectors and enthusiasts to differentiate in the field. Swarco LEDs, used some by Minneapolis and to a lesser extend by other agencies, are different with a noticable brighter spot in the center.

12″ Incandescent Look LED at Valley View Road and Flying Cloud Drive in Eden Prairie

12″ Incandescent Look LED at Valley View Road and Flying Cloud Drive in Eden Prairie

Spotters 02

Swarco LED Lower Left, the red is an RX-11

New GE Incandescent look module with the lenses removed- a fresnal lens, a beaded lens, and a clear outer lens (that’s tinted if so ordered)

New GE Incandescent look module with the lenses removed- a fresnal lens, a beaded lens, and a clear outer lens (that’s tinted if so ordered)

New GE Incandescent look module with the lenses removed

New GE Incandescent look module with the lenses removed

Also interesting is that the US has switched to incandescent look modules almost exclusively, while LED look remain common in other countries. Incandescent look modules are much deeper than LED look, almost as much as the old incandescent reflectors. Other countries have adapted traffic signals that are much shallower, and can only accommodate LED look modules.

Lenses 24

Comparison of module depths


Several companies make LEDs to retrofit 3M and McCain Programmable Visibility signals. Like all traffic signal LEDs these are not dimmable, so a retrofitted signal can be rather bright at night, and if a tech forgets to remove or disable the dimmer the LED module will flicker horrible and eventually burn out.

PV LED Modules

PV LED Modules

Back to North Star Highways Home

Next Page »

Powered by WordPress.com.
Entries and comments feeds.