Thread by @sameo416

Some concerns starting with use of far uv "skin safe" lights. I expect the exposure limits will be increased again with research, but right now it's possible to blow through those limits with a light of any power (say 10 or 12 watts and up).

Calculating incident energy and

exposure can be easy, but it can also be complicated.

I can't teach this in a thread, but will try to illuminate the complexities involved.

A simple light source that emits constant energy in all directions it illuminates can be treated as an isotropic source.

That's not exactly isotropic, but works for a bulb in a case that shines in one direction. Might call it 'bounded isotropic'. Anyway, if the incident energy is everywhere similar, you can treat exposure in that area like it's isotropic. Like the Nukit.

Means inverse square law

will work.

How do you know if your light is isotropic within the beam? Manufacturer should tell you, or you can measure it, or look at the physical structure of the light source.

This is where the complicated bit starts. This is not engineering advice

en.wikipedia.org/wiki/Inverse-s…

The physical structure of the source determines the energy pattern produced. Asymmetric structures do not usually produce symmetric energy distributions unless designed that way. These excimer lights are mostly asymmetric.

Means that tube puts out different energy red to blue

Imagine you're starting at the surface of the light box, then arcing in a semi-circle overtop, back to the surface. Do that for the red line, then the blue, and measure incident energy.

If you plot that energy on a polar graph, it will give you the light energy pattern.

This must be measured in both 'cuts' blue and red, otherwise you may miss asymmetries. This light (the UVCAN Delphi Industrial 20W) is actually pretty symmetrical in both the red and blue cuts.

The graph below shows normalized energy as a function of angle in degrees.

Beamwidth is 64.1 degrees. Light is pointing up at 0 degrees.

For this light, if you knew the peak power at 0 degrees for a given distance, you could use inverse square to calculate approximate incident energy anywhere in that about +/- 15 degree conical area.

The reason you can use inverse sq as an ok approx in that +/- 15 degrees, is because you can treat the light energy in that conic area as isotropic. You know that will work because you have the red and blue energy patterns from the manufacturer.

That light is the type I might

do the math to ceiling mount above a dining table to flood the space over the table with enough energy to get a high degree of pathogen inactivation. The relatively narrow beamwidth means people sitting alongside the table will be at lower incident energy. Can calculate that

with trig, and control the width and incident energy by varying the light's height above the table.

That's not a particularly complex design challenge. Couple hours with Excel and a tape measure, could scope things out pretty well and be confident your dinner guests won't burn

Now it gets trickier. That Delphi has a semi-circular reflector in behind. That produces a nice pattern of energy, nice in that it's mostly symmetrical.

If you take a light that uses two flat reflectors, like the UVCAN Gerani, isotropic goes away.

The same polar plot of the two 'cuts' of incident energy is now very different. In the blue cut, those flat reflectors now produce a pattern with two distinct lobes. The graph is normalized energy, normalized against the peak measured.

What's going on here? What does it mean?

The reflectors create a unique pattern that concentrates energy in the two blue lobes, but only in the cut across the middle. The other pattern is a broader kindof isotropic. If you can imagine that in 3D, this would form a very interesting shape.

Now recall the Delphi and my dining room application. If you're sitting 30 degrees off centre, the power in the pattern is quite low - under half the peak.

If I used the Gerani in that application, now the peak power is pointed exactly along the 30 degree lines...right at guests

Performing an exposure energy calculation in this case is much more challenging, and past the ends of the filter window I'll have to interpolate between the two patterns in 3D. I likely can't use this in that ceiling mount because of the lobes, and I'm still working on the model.

You need to know the beam pattern and beam width of any uv source that is of any larger power. This is because of the potential of exceeding recommended exposure limits for eyes and skin.

A low power light (Nukit 3W) that comes with a distance recommendation (1 m) is no problem.

Isotropic approximations can be very useful, but only in the case that a light does not have a lobed pattern. This is the Lily, 3W, but can see it's reflector design is also producing lobes.

On the Nukit, kind public domain measurement by @EwanEadie is available in tabular form. Since the Nukit has a semi-circular reflector, my guess is it has a nice pattern in both cuts like the Delphi Industrial.

The data are here for the Nukit. These are super lights imo, very flexible. I've used these lots. Thanks to @NukitToBeSure for your commitment to quality and cost-effective design! (just received a note for a US brand of uv @ $2,000 'on sale').
researchgate.net/publication/37…

This is plotted cartesian not polar, but the same data. Can use that isotropic approx within the main beam, which is about +/- 33 degrees beamwidth to the half power points (50%).

I've changed the scale a bit.

If the Nukit had two flat reflectors, would expect to see a strongly lobed pattern.

Designs with flat reflectors may be looking to concentrate energy. More likely it's chosen for it's lower cost.

For most applications a symmetrical energy pattern is preferred.

So, tldr, be aware of what the light design is putting out in terms of an energy pattern, to determine if you can use simplified exposure math, or if you need to break out a spherical coordinate system (or ask an expert).

This is not engineering advice. If you're going to use higher power uv at any frequency, consult appropriate knowledgeable people concerning safe exposure limits. This is particularly true with very high power (100 watt+) fixtures.

Not a light expert, but know ++ about antennas

I have a 20W Gerani fixture and the Nukit set.

Use the 20W in an upper air application where those two lobes are pointed along the ceiling. Not a bad location for that pattern.

Use Nukit on table top regularly. Being able to use 4x 3W resolves much of the dose concern. /fin

PS. If you’re trying to do irradiance measurements, you need to have some idea if the light is pushing out a symmetrical pattern (balloon) or not (lobes). If it’s lobes, you can’t do a simple measurement on centreline with distance.

For the Gerani you’d measure half peak.