Tuesday, 30 June 2020

2020 so far; LED vs. actinic part 3

When I started mothing almost exactly two years ago, I decided not to use MV lights: I didn't want to buy equipment for which bulbs were no longer made or imported, and I also wanted something that I could run off a battery. I initially bought a 15W actinic strip light (with both mains and 12V DC drivers), but I was interested in LEDs. A year later, having read a lot about LED lights for moth traps, I concluded that those on the market at the time were more-or-less divided into two groups: those that were effective and very expensive, and those that were cheap but not terribly effective. I thought that there was a lot of potential in this area and I wanted to experiment, so starting in late June 2019 I started to design and construct my own LED light. You can read my original post on this here, describing what I built.

At the end of 2019 I made another post (here) with a progress report. As we are now half-way through 2020, I thought I would give a further update. In 2019 I was running a single trap, alternating between LED and actinic lights on different nights. Variations in weather and moon between "actinic" nights and "LED" nights made comparison rather difficult. At the end of the year, I worked out how (i.e. where!) I could run two traps in the garden simultaneously, so I bought another identical trap, spent the first three weeks of January on maintenance and for all of 2020 I have run them in parallel. I have tried to operate the traps twice per week in order to get enough data for comparison, and as there is a difference between the two trapping sites, I alternate the lights between the two locations. During 2020 I have made some other changes, which I'll mention shortly.

The results for these last six months as a whole are below:
2020 H1 Results Actinic LED Overall
Nights run 40 40 40
Species caught 154 166 210
Macro species 101 109 133
Micro species 53 57 77
Best night for species 23-Jun 23-Jun 23-Jun
Species on best night 47 46 73
Best night for moths 23-Jun 15-Jun 23-Jun
Individuals on best night 233 211 392
Total number of moths caught 1136 1449 2585
This makes the lights look similarly effective, with the LED slightly ahead of the actinic. However, this masks the effect of changes I made in April and May, prior to which the LED was slightly behind the actinic. With lockdown, I decided to increase the power of the LEDs, exploiting the availability of mains power in the garden. The LEDs had originally been a total of 14 one-watt components, but in April I replaced them with eighteen 3W LEDs - adding more ultraviolet components. However, the LEDs were not driven at their full 3W rating: I limited the total power to about 18W, compared to the 13W of the original version. The reason for limiting the power is that it is difficult to dissipate the heat; even though I have mounted them on an aluminium heatsink and used thermal paste, they are not tightly bonded and so there is a risk of overheating.

I then found that the UV LEDs were extremely inefficient: this is a common problem at UV wavelengths, but hard to identify because most datasheets do not show the radiation flux (ϕe) that they emit. Only about 2% of the electrical energy was being converted to light and the rest was coming out as heat. After more research, I identified a source of UV LEDs that are about fourteen times as efficient, so in May I replaced six of the LEDs with these better versions. Heat dissipation is still a problem and I am limiting the current so as to keep chip temperatures below a conservative 70°C, meaning that the total power to the LEDs is just under 20W (the drivers consume another 2W).

These changes have made quite a striking impact on the relative performance, best illustrated by the cumulative number of moths caught. Since making the changes, the LED light catches about 35% more moths and 13% more species than the actinic.

7 May 2020 onwards Actinic LED Overall
Nights run 16 16 16
Total species caught… 127 144 182
… of which macros 81 93 112
… of which micros 46 51 70
Total moths caught… 934 1275 2209
… of which macros 605 909 1514
… of which micros 329 366 695

I will continue this experiment for the rest of the year. I may make more changes to the LEDs, but when I think I have finalised the configuration, I will then look to improve the bonding so as to be able to run at a higher power: better bonding generally means it's harder to swap LEDs. I will probably also produce a new set of drivers that will power a subset of the LEDs at a lower current, in order to restore the option for compatibility with running off a battery. Finally, when I am no longer taking the light to bits to change components, I will use a less Heath-Robinson (pun acknowledged) way of mounting the light on the trap, and add baffles - but that might be a job for 2021.

Tim Arnold
Newton Longville, Bucks


  1. All very useful stuff, Tim, and it will be interesting to see the results produced by a full year. I hadn't realised there was a problem with heat when using LEDs as a light source.

  2. Hi Dave, yes heat has to be taken into account. LEDs are (usually) far more efficient at converting electricity to light instead of heat than MV bulbs for example, but LEDs have their small size working against them in respect of being able to make the heat flow away.

    For example, the datasheet for my new ones gives Φe as 900mW with an electrical power input of a fraction below 2W. So that means about 1W of heat to dissipate, but the physical size of the substrate through which all that heat has to flow is only 3.5 mm square. Normally, that's not much of an issue because it can be dissipated by good contact with a heatsink: in fact, most low-power (domestic) LED lamps don't need a heatsink as such: they just use their metal fittings to conduct the heat away. My problem is that the best ways of making a good thermal contact usually involve permanent adhesive - and at the moment, I want to be able to take off one LED and substitute another as I tweak the design. That makes the thermal contact less effective, so the LED can't dissipate the heat as easily. Consequently, my heatsink is only at about 30°C when the LED itself is at 70°C.


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