Left:
Early morning, October 2002, in Gresford, Wrexham, U.K.Left:
Early morning, October 2002, in Gresford, Wrexham, U.K.
How soon do streetlights switch on/off relative to local sunrise or sunset? Is the order in which streetlights switch on in the evening simply the reverse order of their switching off the following morning?
Since 1987 I have been gathering data on the switching-on times for streetlights after sunset. On this page is a summary of the data for different types of sodium streetlights of various ages and designs, from four different locations in the UK. In particular, I measure a value I call the "sensitivity" for each lamp. Normally I measure this as the average number of minutes after sunset when the lamp first switches on, although I am also (as from 2007) measuring the light level in units of lux (a standard unit of measure) and some results are shown later on.
Unless otherwise stated, all data on this page is for fine weather conditions only (i.e. when the lighting-up time was not affected by cloudy weather). It was taken from 1987 to 2008, and includes 1970s/80s Thorn Beta 5s, Philips MA 90/50s, newer Beta 5s, Philips MI-26s and a handful of others, e.g. two Revo "Daleks", a GEC Z9554, and what were probably a few GR models from the 1970s/80s. Please feel free to make use of my data - all I ask is that I am credited as being the source of it. It's taken a lot of work to put it together!
I
have put together an archive summary file of my evening lighting-up observations,
which is an CSV (comma-separated values) format. To download it, please
click here
(203 kb). This data file spans the years
1987 to 2008 and shows details for 4 'standard' lanterns - it can be used on
its own, or together with my "Lighting-Up" program which is detailed
in the "Free software" section later on this page. More details on
the archive data file are also shown in the "Free software" section
below.
The pie chart on the left shows the sensitivity of all the SOX lamps I observed from 1987 to 2005. It shows for example that 23% of streetlights switch on between 10 and 15 minutes after sunset (in fine weather conditions). It also shows that on average just over 4% switch on before sunset (negative numbers indicate numbers of minutes before sunset).
The two graphs below (with a blue background) show the percentage of streetlights that switch on over time, both for morning and evening, for the streetlights I observed in Gresford (Wrexham, U.K.) in Jan-Mar 2003. Although the sample is fairly small, the shape of the graph is repeated for streetlights in general - a small number switch on/off to begin with (shallow gradient on graph), then the majority follow (sharp gradient), with a few trailing at the end (back to shallow gradient). The two lines (pink and blue) show the difference between fine and more cloudy weather.
The table below is a larger analysis, for 1987-2001 (in all weather conditions):
| Number of minutes after sunset | -30 | -15 | -10 | -5 | -3 | -2 | -1 | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| % of lamps on | 1 | 2 | 3 | 5 | 6 | 7 | 7 | 8 | 8 | 9 | 10 | 11 | 14 | 16 | 20 |
| Number of minutes after sunset | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 |
| % of lamps on | 22 | 27 | 32 | 36 | 40 | 44 | 48 | 51 | 55 | 58 | 62 | 64 | 68 | 70 | 73 |
| Number of minutes after sunset | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 35 | 37 | 40 | 55 |
| % of lamps on | 78 | 82 | 84 | 87 | 89 | 90 | 91 | 92 | 93 | 94 | 95 | 96 | 98 | 99 | 100 |
On the "Data" page (click here) are tables of data that show most of the sensitivity values recorded from 1987 to 2008. Part of this data was used to make the graphs/charts on this page.
Analysis of a single streetlight - B2
Taking
a closer look at a particular streetlight, I recorded the exact time when streetlight
"B2" switched on most evenings from 1991 to 1996. B2 was a Philips
MA90, which was installed in Clitheroe, Lancashire, U.K. in 1989 and removed
in 2007. The pie chart on the right summarises the
time difference between when the streetlight was expected to switch on and when
it actually did switch on. The "expected" value is the sensitivity
(as mentioned above) for fine weather conditions - i.e. the average number of
minutes after sunset when the streetlight switched on during the year in question.
I should point out that the results here will depend to some extent on the local
climate - Lancashire is known to be a fairly wet part of the U.K.! In general,
the darker blue areas of the pie chart represent rainy days, whereas the orange
areas represent clear or fair weather.
It is also possible to see how the weather affects the switching-on time, as per the table below. As you would expect, cloudy/rainy weather generally causes the streetlight to switch on earlier.
| Clear | Fair | Cloudy | Rain | Overall | |
| over 30 minutes earlier than expected | 0% | 0% | 0% | 2% | 1% |
| 25-30 minutes earlier than expected | 0% | 0% | 0% | 2% | 0% |
| 20-25 minutes earlier than expected | 0% | 0% | 1% | 13% | 3% |
| 15-20 minutes earlier than expected | 0% | 0% | 4% | 17% | 5% |
| 10-15 minutes earlier than expected | 0% | 0% | 14% | 26% | 11% |
| 5-10 minutes earlier than expected | 1% | 1% | 28% | 24% | 18% |
| 0-5 minutes earlier than expected | 52% | 61% | 40% | 16% | 42% |
| 0-5 minutes later than expected | 42% | 36% | 11% | 0% | 19% |
| 5 minutes or more later than expected | 5% | 2% | 0% | 0% | 1% |
Finally, the time of year actually makes some difference to how long after sunset it will be before a streetlight switches on. However, this is only noticeable in fine weather conditions, as the effect is lost if the weather is cloudy/rainy. The graph below shows how around December/January and June/July the streetlight will switch on later relative to sunset, whereas around March and September it will be earlier. I also found the same result in analysing streetlamps in Gresford, Wrexham, U.K. for 2002-06.
Streetlight B2 was one of a number of streetlights on the same road, all installed in 1989, and the graph below shows the characteristics of them all. You can see from the graph how the sensitivity of each lamp has changed over time. The abrupt change from 2002 to 2003 is probably the result of a change of photocells. (Some interpolation has been done between data values, but observations done on 1 Jan 2003 show that these streetlights had already been modified).
Lux
(evening): The purpose of the table on the right
is to show how the light level in lux changes in terms
of minutes after sunset. Sensitivity values (i.e. the
time when a streetlight switches on) are shown of streetlights observed
by myself in 2007, both in terms of minutes after sunset and as lux (light
level). The values are shown for days of 'fine' weather
(which I have defined as three-quarters cloud cover or less - when the
lighting-up times should not be affected by the weather). For the 'lux'
calculations I only included days when the streetlight was shining red
(i.e. will have only just switched on) - and because of the small sample
sizes of lux values available, there will be a certain percentage of error
in the data. Order of switching on/switching off: So is one simply the reverse order of the other? From the table below, the answer is that yes, there is a vague correlation, but there are some definite differences in some streetlights: a lamp that switches on early may also switch off early, and vice versa. A "1" in the table means that a streetlight was the first to switch on in the evening, or the last to switch off in the morning (note therefore that the values in this table are simply the ranking for each streetlight from 1 upwards, and not sensitivity values).
|
Lamp | Sensitivity
(mins) |
Sample
size |
Sensitivity
(lux) |
Sample
size |
|
| Q1 | -6.6 |
75 |
1002 |
10 |
||
| Q11 | 6.0 |
87 |
103 |
9 |
||
| U1 | 7.3 |
51 |
70 |
8 |
||
| Q10 | 7.4 |
53 |
83 |
5 |
||
| Q8 | 7.5 |
45 |
67 |
11 |
||
| Q9 | 8.3 |
48 |
55 |
12 |
||
| S2 | 11.0 |
89 |
58 |
15 |
||
| R1 | 11.4 |
6.3 |
48 |
9 |
||
| S3 | 11.5 |
89 |
55 |
11 |
||
| S7 | 12.9 |
88 |
46 |
10 |
||
| S1 | 13.7 |
88 |
39 |
13 |
||
| Q4 | 14.3 |
84 |
46 |
7 |
||
| Q6 | 14.6 |
88 |
37 |
11 |
||
| Q7 | 15.9 |
88 |
25 |
13 |
||
| S4 | 16.8 |
88 |
26 |
10 |
||
| Q5 | 17.8 |
87 |
21 |
15 |
||
| Q3 | 19.4 |
87 |
15 |
12 |
||
| R9 | 20.0 |
88 |
14 |
9 |
||
| T1 | 21.3 |
77 |
11 |
8 |
||
| R6 | 23.3 |
88 |
9 |
12 |
||
| R3 | 24.6 |
87 |
8 |
7 |
||
| R4 | 24.8 |
88 |
7 |
14 |
||
| R5 | 31.0 |
88 |
1 |
9 |
| L7 | L8 | L10 | M1 | M2 | M3 | M4 | M5 | M6 | M7 | M8 | M9 | N1 | N2 | N3 | N4 | N6 | N7 | N10 | O4 | |
| ON eve (2004) | 16 | 4 | 10 | 5 | 2 | 8 | 3 | 6 | 12 | 18 | 17 | 7 | 13 | 9 | 1 |
14 | 11 | 19 | 15 | |
| OFF mrn (2004/5) | 13 | 9 | 12 | 7 | 1 |
2 | 4 | 16 | 6 | 14 | 18 | 10 | 8 | 17 | 3 | 5 | 11 | 15 |
(The data in the above table is based on all observation days, regardless of fine, cloudy or rainy weather, as the weather does not affect the order in which streetlights switch on if they have an electronic photocell.)
Why not download this FREE software?
This
specially-written software (now at Version 4) takes lighting-up data
for locations in the U.K. for 1987-2009. The switching-on times for four selected
SOX streetlights are shown:
Lamp 1 switches on around 8 minutes after sunset on a fine day. Around 20% of all streetlights will be switched on by this time.
Lamp 2 switches on around 12 minutes after sunset on a fine day. Around 40% of all streetlights will be switched on by this time.
Lamp 3 switches on around 16 minutes after sunset on a fine day. Around 60% of all streetlights will be switched on by this time.
Lamp 4 switches on around 20 minutes after sunset on a fine day. Around 80% of all streetlights will be switched on by this time.
The colour of each streetlight is shown at the observation where it was first seen to be switched on. So for example, from the screenshot, on 1st September 2002, lamp '1' was yellow-orange in colour at 8:15, meaning it switched on at 8:11 or 8:12.
As well as streetlight data, the sunset time and local weather conditions (from my own weather observations - another interest of mine since 1987!) are shown. The data can be filtered by date, so for example you can select 28th August and see how the weather conditions have affected lighting-up from one year to the next.
To run the Lighting-Up program you will need to download TWO files:
The program - click here (252 kb)
The data file (which can also be viewed as an ordinary .csv file) - click here (257 kb)
The data file should be stored in the same directory as the main program. However, you can use a alternative path if preferred.
The
software is freeware. Please let me know what you think of it! Contact me
at 
.
I trust that your computer has sufficient security protection against viruses etc., but you have my assurance I am NOT a hacker or any sort of malicious programmer!!
© Matthew Eagles 2006. This page last updated 10th March 2010.