Tuesday, 8 October 2019

LED vs. Actinic

Three weeks ago, Dave Wilton wrote an interesting post that compared his experience of 125W MV and twin-30W actinic lights in his garden.  You can see it here.  He finished with this: It will be interesting to learn how people get on with LEDs as a light source.  […]  Does anyone here have experience of using LED traps yet?

I have been doing some experiments with LEDs: at the time Dave wrote his post, I was still getting my records in a position to be able to do some analysis.  I have constructed my own LED light which I run on a Skinner trap, alternating with a 15W actinic strip.  To keep this post reasonably short, I won’t give a lengthy description or explanation: if there’s any interest, then perhaps I can make a presentation at a meeting.

My LED light consumes about 13W, though the design could be uprated as far as about 50W if I sacrificed the ability to run it off a battery.  It consists of two flat panels back-to-back (facing opposite directions), and each panel currently contains the same mix of seven LEDs of various wavelengths, including the UV and visible spectrum.  The visible spectrum is not optimal (see the second row of the table below for how poor it is when run without the UV), but I’ve temporarily frozen the configuration in order to make comparisons with the actinic.
My LED light (drivers in separate box).  UV LEDs on centre board.
EDIT.  Dave suggested that I post a picture of the light on the trap.  For the photo, I just threw the sheet I use over a bush, instead of hanging it from a guy rope like I usually do.  Item 1 is the light in position on the light bar - the mounting is temporary until I decide that I'm not going to tweak the LED configuration any more.  Note that I'm not currently using a baffle.  I fix a rain guard to the light bar when I need to, in order to keep water off the LEDs.  Item 2 is the driver box - massively over-size as the three commercial driver boards inside are each only about 42x21x13mm.  I usually place it with the sockets pointing downwards to preserve its otherwise waterproof nature, but to make the connections more visible in the photo, I put the box the wrong way around. My intention is to add a hook so that it will hang on the side of the trap. Three coloured sockets drive the PAR LED (red socket), the UV LEDs (blue socket, rather hidden in the photo), and the other visible LEDs (yellow socket) - see my comment below for more about wavelengths, etc.  The green socket is currently unused.  The two black sockets are the negative rails to each of the two LED panels, so that I can drive just one side if I want to.  Item 3 is a 12V, 16Ah LiFePO4 battery, unplugged in the photo.
The light assembled on the trap, with driver and battery
EDIT: I've taken the opportunity to add a couple of spectra - see my first comment below.  The first is the daylight white LED; the second is the 385nm UV LED.  The visible light is roughly 400-800nm, though definitions (and what different people can see) vary.  The charts show the relative power emitted at different wavelengths; human sensitivity varies across the spectrum, as does the sensitivity of different moth species, but that's a topic I'll refrain from entering here.



Unfortunately, the experimental conditions mean that any conclusions are only tentative at this stage.  I can’t run the two lights on the same night, so I must contend with other variables of weather and moon.  This means I need do to lots of trapping in order to get comparable data, but I only started using the LED on 30th June, and in fact the LED light didn’t reach a relatively mature configuration until 3rd August, so I missed a large chunk of the season.  Moreover, I've had bad luck with asymmetrical weather: moonlit night/cloudy night, dry night/wet night, etc., which has invalidated many of my alternating LED/actinic comparisons.

However, I’m pleased with its performance and I think it is probably doing as well or better than my actinic, with the potential to go further when I modify the visible spectrum.  Here is what I think is a valid subset of comparable trapping sessions not impacted by moon/weather (except as noted below): within a colour band, each record is from trapping sessions in the same spot, a few days apart.  Comparisons between different colour bands may not be valid.
 Date
 Light
 Species Moths  Notes
03/07/2019Actinic40392Position 1 in garden
09/07/2019LED5222.7W of visible light LEDs only, unidirectional
10/07/2019LED335942.7W UV only, unidirectional. 401 C. culmella!
23/08/2019Actinic38218Position 2 in garden
27/08/2019LED56475Full config of LEDs (c. 13W). Record # of species
25/08/2019Actinic45351Position 1
30/08/2019LED38266 
14/09/2019LED1875Position 2.  Full moon, almost no cloud
22/09/2019Actinic18148 

I intend to keep this experiment going, but I'll use the winter period to make some changes to the visible LEDs, and possibly to build a more powerful version to run off 240V AC: only one of the two positions in the garden is neighbour-friendly for powerful lights, but it may attract attention from dog-walkers.

Tim Arnold
Newton Longville, Bucks
Posted by at

5 comments:

  1. Martin Albertini9 October 2019 at 18:26

    Tim, all very interesting and shows LED will certainly be viable for moth trapping.
    Have you any information to comment on a point Dave raised about the effect of the UV LED you used on the human eye?

    ReplyDelete
  2. Tim Arnold9 October 2019 at 20:18

    Hello Martin,
    This was one aspect that I omitted from my original post for brevity, but since you ask: I did look at this before designing my LED light. A full answer would be very long, but I'll give a (slightly) shorter answer in several parts. Firstly, I'll say that I'm only covering UV-A here, which is defined as being in the 315-400nm range. I have nothing that emits in wavelengths shorter than this.

    Next: the effect of UV on the eye. Wikipedia says that the lens of the human eye blocks most radiation in the 300-400nm range and cites a paper by M A Mainster in the British Journal of Opthamology (full reference in Wikpedia under "Ultraviolet", reference 4). While this is true, it's incomplete. I have read some other articles that look at this area. UV light can damage other parts of the eye that aren't behind the lens. People (like myself) who have had cataracts removed - which involves replacing the lens of the eye with an artificial one that is significantly less effective at blocking UV - have an increased risk of getting retinitis and/or age-related macular degeneration if exposed to too much UV.

    Now I'll turn to my light. I'm using three different UV LEDs that emit at the wavelengths of 365, 375 and 385 nanometres respectively. All of them have quite narrow spectra, which was a factor in why I chose them, but it's also why I've had to pick three to cover sufficient spectrum. Unfortunately, I don't seem to be able to include a picture in a comment, so I can't post the spectral output power chart here, but as an example the output of my 385nm LED drops off to about 20% of the peak by 375nm and is effectively zero by 350nm. The visible light LEDs that I'm using are one blue and one green LED with narrow spectra (though slightly broader than the UV LEDs), and a couple of broad-spectrum LEDs ("Daylight White" and "PAR" which is used by plant growers and emits mainly red with some blue). The output of the daylight white LED is zero below 400nm, and the PAR LED has much less than 5% of its output power below 400nm.

    I should mention that there is a paper by Gunnar Brehm, originally published in Nota Lepi. I won't reproduce any part of it here to avoid copyright issues, but the article is available as a download from his website (he sells "LepiLED", typically for around €400!). Amongst other things, the paper compares spectra emitted by a number of sources used for moth-trapping, including MV and blacklights (see p. 96). In principle, I think that well-designed LED lights are similar to other types in the spectrum that they emit: indeed, the challenge can be "missing" parts of the spectrum.

    Finally - and notwithstanding all of the above - precautions. I am a spectacle-wearer, and many lenses are coated to filter out a lot of UV (note that the CE mark for UV applies to wavelengths less than 380nm), but it depends on which coating(s) have been applied to the lens. Regardless of this, if I am going to spend more than a couple of minutes near the trap with the light operating, I put on a pair of safety glasses. These are sold by several of the usual suppliers of mothing equipment, and they are generally ordinary safety specs (as one might use for a DIY job), but checked to ensure that they have the appropriate UV certification: I went through a detailed check of this for the ones that I have. The design of my light also delivers a power connection to the UV LEDs that is separate from the other LEDs: one of the reasons for the many coloured wires in the photograph. I frequently use this feature to disconnect the UV LEDs and to power just the visible light LEDs before sunrise while I'm looking for moths on the sheet behind the trap and in its surroundings, before I close up the trap.

    I think this must be the longest comment ever on this blog: there's probably lots more to say, but I'll stop here!

    ReplyDelete
  3. Tim Arnold10 October 2019 at 00:21

    I have just noticed a paper that I previously missed, which compares 160W MV vs 15W near-UV LED (398nm) vs a 200W incandescent lamp as used by the RIS. The paper is open-access and can be found here: https://www.eje.cz/artkey/eje-201701-0004_Assessing_the_efficiency_of_UV_LEDs_as_light_sources_for_sampling_the_diversity_of_macro-moths_Lepidoptera.php. After a rapid first read, I'm not convinced about the methodology (it really is VERY difficult to compare like with like), but they seem to conclude that the UV LEDs did about as well as the MV, and better than the incandescent.

    ReplyDelete
  4. Dave Wilton10 October 2019 at 10:16

    Thanks for going to all this trouble, Tim - it makes very interesting reading! What would be good sometime is to get an idea of what your Skinner trap actually looks like with the LED arrangement fitted. As Blogger doesn't permit photos within comments, perhaps you could add one in to your original (click on the yellow pencil symbol at the bottom of the post and that will take you back to the composition page).

    ReplyDelete
  5. Tim Arnold10 October 2019 at 13:59

    Good suggestion, Dave, which I've now done. It's not really any trouble for me - I did all the research, etc., before I designed and constructed the light. I'd be more than happy to make a presentation at a BIG or UTB meeting, if anyone would like it.

    ReplyDelete

Note: only a member of this blog may post a comment.

Subscribe to: Post Comments (Atom)