SOX lamps: Lighting-up data

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?

From 1987 to 2011 I gathered 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 measured a value I call the "sensitivity" for each lamp. Normally I measured this as the average number of minutes after sunset when the lamp first switched on, although I also (as from 2007) measured 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 visit the Software page. This data file spans the years 1987 to 2011 and shows details for 4 'standard' lanterns - it can be used on its own, or together with my "Lighting-Up" program which is also downloadable from the Software page. More details on the archive data file are also shown there.

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.

(Data such as that shown on this page is covered in more depth in my "SLO statistical report on the Data page)

The table below is a larger analysis, for 1987-2001 (in all weather conditions):

On the Data page (click here) are tables of data that show most of the sensitivity values recorded from 1987 to 2011. 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 1997. B2 was a Philips MA90, which was installed in Clitheroe, Lancashire, U.K. in 1989 and removed in 2007. The data that follows is from the period 1991-96; additional data from 1997 was found later and this is included in the SLO statistical report which is downloadable from the Data page. 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.

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).

(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.)

Back to title page

Back to top

Next section

© Matthew Eagles 2006. This page last updated 18th January 2012.