Convoy S21D w/ legendary Nichia 219b :) - Review & comparison w/ Nichia 519a, E21a & other lights with Nichia LEDs incl. Emisar D4V2, Convoy S21F. (Summary & measurements on P. 1)

BTW I am aware some of these posts are for Emisar D4V2, but one, we have “cross” owners lol, and two, what’s interesting for D4V2 should be equally interesting for Convoy connoisseurs IMHO. Anyway, found something interesting while perusing the Carlo Optic web site
https://www.carclo-optics.com/products/optic-10621

It seems you could predict the nature of the optic by this ratio the arrow is pointing at. The higher candela (hotspot brightness representing throw) per lumen (“overall brightness” in simplified terms) means it’s a more throwy optic. By comparing this spec for the same LED by different optics, for example 10621 versus 10622 versus 10623, you could tell which optic has most throwy before testing. For example 7.6 for 10621, 7.2 for 10622, but only 3.2 for the frosted 10623.

This was how I could tell by spec the difference between 10621 and 10622. It wasn’t clear otherwise. It also confirms my observation why frosted 10623 is so dim for me. That is a big jump from 7.2 to 3.2, and I wonder if something in between the 2 would be a more moderate compromise.

The other way to tell visually is that little circle in the middle of each of the 4 small “lenses”. For the same optic series, that circle seems smaller with more throwy optic. But this “method” is more difficult and you have to be a little nutty to be curious about things like this :confounded: :smiley: .

Yet another thing I learned that makes this hobby more fun: AR coatings are not all the same, and they don’t reflect the same color using the light bulb reflection “test” (I just stand under a ceiling light bulb and look for reflection of the light bulb in the flashlight’s lens as if the lens is a mirror. AR coating changes the color of the light bulb reflection.).

I am curious how Duv would differ among the 4 lenses I have:
1. Convoy S21D lens, the latest per Simon with “green tint” - green reflection.
2. UCL Acrylic lens for Wurkkos TS30S - purple reflection
3. UCL Glass lens for Wurkkos TS30S - blue reflection.
4. Wurkkos non AR coated - no change in color.
The question is how to do the amateur level “test” in a reliable manner, not to screw up :slight_smile: as the other lenses are large and obviously don’t fit on S21D.

The UCL lenses are from https://flashlightlens.com/ and highly recommended - about 3-4% improvement in transmission vs non AR coated.

That’s a nice view!

The tested lenses are larger so what I did was taking off the S21D’s OEM lens, then placed the tested lens by holding it, recorded the number, then put back OEM, recorded another number. I did about 4-5 runs or so for each lens, sometimes more. The flashlight was set at 35% brightness and let sit for 15 minutes or so before measuring. I decided to use the middle-of-the-road 30° bead TIR optic this time.

With flashlight mounted on tripod (hence no shaking), I’ve found the Opple Lightmaster Pro (aka Series 3) to be incredibly consistent in its reading of Duv (and brightness & Ra etc.). It reads the same x,y coordinates down to the fourth significant figure between different runs. As such it is a great meter to do comparison work; just remarkable considering how cheap it is. I recommend it extremely highly for anyone into LED collecting.

Duv values for Convoy S21D with Nichia 219b 4500K
The color refers to a reflection of ceiling light bulb, NOT any tint I could see.
Home test for fun, please take result with a grain of salt. There might also be some correction as I double check data.

1. No lens: –0.0127
2. Convoy AR Green-reflection lens : –0.0124 CCT 4471
3. Wurkkos Non-AR: –0.0123
4. UCL AR Blue: –0.0123
5. UCL AR Purple: –0.0102 (Duv change: ~0.0020 more positive) CCT 4365

I find it interesting that the lens with the highest transmission rate, UCL Purple, is also the one that affects Duv the most. Next - beamshots to see if this difference matters in real life scenes.


Great info!

Green reflecting coating is supposed to prevent green tint from getting thru.

As we see from your test, Blue reflecting coating works Just as Well!

The Acrylic w Purple coating lets the most green get through, thereby raising the DUV by 0.0020 compared to the green and blue coated glass. It may not be the fault of the coating, since Acrylic transmits more light than glass, it may also be transmitting more of the Green tint, thereby raising the DUV.

What kind of clamp do you use on your tripod?

The tripod hold I use is this (a legit seller) HERE . It is much better than the small rubber without velcro. I’ve found keeping the light stable is the first step towards measuring reliably, especially for the throw where you have to keep hotspot centered on the sensor.

Thanks Jon. From what I have read (meaning I’m no expert), AR coating targets a wavelength that’s close to the center of the visible spectrum (which is yellow green as explained in article below) and as a result proportionally more green will go through with all AR coating, including Convoy’s. The difference is only how much green, as seen here obviously the Convoy is best, and UVL Purple an outlier worst. To be fair, outside of the UVL Purple, the difference among the lenses is small (it’s in the 4th digit) and within error limit of my amateur testing rig for example, even as I’m quite confident of the result.

A good article on this subject http://rick_oleson.tripod.com/index-166.html

So, if coated lenses are purple, won’t my photos have a purple tint?
No…. actually, they’ll have a slight yellow-green tint – the extra light that didn’t get reflected passes on through to the film, and there’s a little more of the 550nm yellow-green than there is of the other colors. But it’s not very much – glass only reflects about 4% of light from each surface to begin with, and the coating has some effect at all wavelengths – so there’s maybe a couple of percent of extra green light getting to the film. For most normal photography, this isn’t enough to be noticeable in the pictures.

Thanks Cannga!

Btw the Duv test between 1 and 2 (lens vs no lens) should be easy to do for any Convoy owner with Opple 3. I hope to see similar comparison from other owners, especially people who have the infamous lens w/ purple reflection. That way we could compare/correlate our findings.

I think either a. measuring Duv or b. comparison photography (as opposed to visual observation) are important to have some fun with this :-).

I would like to get a hold of Convoy’s lens with purple reflection but can’t find it on AliX. Would very much appreciate anyone who has spare putple lens for any of my Convoy to send it to me . I’ll pay shipping and lens cost of course.

In comparing Duv number, I’ve noted that if the change is in the fourth number after decimal point, for example
“No Lens”: –0.012 7 versus “Convoy Green-reflection”: –0.012 4
it is not easily noticeable on the Duv graph.

What this means is, at least per my test, unless I develop Superman’s vision, I am not sure that I could see a difference between for example No Lens versus Convoy Green Reflection versus UCL Blue. I plan to take beamshots against the wall of truth to confirm if this is true.

OTOH UCL Purple versus Convoy Green Reflection is a bigger difference (it’s on the third number after decimal point) and likely will be seen on Duv graph.

Visually if you compare some 519A lights with different lenses, the difference in DUV is visually noticeable. However if you then add other emitters into the mix such as SSTs and LH351D, then you realise that even the “worst” 519A/lens combination is very neutral and perfectly acceptable.

It's difficult to see a duv difference when both outputs are way into the magenta region--the percent difference between the two is tiny compared to their deviation from the BBL. (Imagine trying to discern between 1 and 10 lumens, and between 1001 and 1010 lumens. The absolute difference is the same, but the percentage difference is huge!)

I also suspect that the green AR lens would make a larger duv difference numerically with a light source that starts off as neutral or slightly green. If the emitter starts out magenta, there is little green to begin with, and not much removal can happen.

It would be lovely if the test could be re-done with emitters that are neutral or slightly green.

Djozz’s quick test

Thank you Haukkeli for linking the test! Looking at -0.003 decrease in duv from swapping purple to green AR. That's a LOT of green reduction if the emitter starts off close to neutral. Also, the +8 increase in R9 is really something.

I would actually expect a more drastic difference for Convoy lenses--their purple reflection, at least in my sample of Convoy lenses, appears much more intense than the lens shown in the D80 photo.

Thanks everyone for your input and @Haukeli great job for finding that post of Djozz’s test - thank you. It’s great to see someone else’s numbers to know I am not entirely lost lol. At any rate after looking at Djozz’s results, I have just rerun my test not just with 219b but also with 519a in my S21D using 60 Bead TIR optic, and could now confirm with near certainty my result :innocent: .

For Convoy S21D with Nichia 219b 4500k:
1. No lens: –0.0127
2. Convoy AR Green-reflection lens : –0.0124 CCT 4471
3. Wurkkos Non-AR: –0.0123
4. UCL AR Blue: –0.0123
5. UCL AR Purple: –0.0102 (the Purple Lens’s Duv is “20 points” (0.0020) more positive) CCT 4365

Note that we know changing lenses affect Duv, but not frequently mentioned is it also shifts CCT about 50-100 points.

Where my result is different from Djozz’s - I found essentially no significant difference in Duv among set-up 1 to 4 above.

Where my result is similar to Djozz’s - The UCL Purple AR turns Duv approximately 0.0020 more positive (note the decimal point) and drops CCT roughly 50-100 K versus the group of 4 above. Not a huge change numerically, but definitely there, and does bring beam into more yellow-green territory.

Thank you @cannga for re-doing the test! Would you share the numerical results for the 519A? (I assume the graphic you shared is for 219B given how low the duv is.)

I swapped out a purple AR coated lens for a non-coated one in my 519A 5700K triple and noticed a huge difference right away. In addition to eliminating much green, the tint shift is a lot less, due to the fact that an AR lens does not alter tint equally in all directions. It alters tint the most near the center of the beam, but barely does anything to tint near the periphery of the beam.

I just ordered some green AR lenses and really hope they will bring the duv down more and eliminate angular tint shift completely.

You’re very welcome. Data from S21D with 519a 4500K and 60 Bead TIR below, very similar behavior to 219b 4500K. Again the UCL Acrylic Lens with Purple reflection is an outlier and the Duv shift is again around 0.002, CCT drops around 50-100. I could provide x,y coordinates if you’re interested.

S21D 519a 4500K with 60 Bead TIR
1. Convoy Green Reflection AR –0.0038 CCT~4041
2. No Lens –0.0037
3. Wurkkos Non-AR –0.0034
4. UCL Blue AR –0.0032
5. UCL Purple AR –0.0018 CCT~3985

Above bolded part in your post very interesting; I didn’t know that. Was about to write that any effect from an AR lens would not affect tint shift since it alters the whole field. Always good to read a second time before responding :person_facepalming: :innocent: . But question: if AR doesn’t alter peripheral field, it wouldn’t help with the peripheral tint shift in this Nichia would it?
I actually found the tint shift in my 519a 4500K not that bad especially w/ wide & beaded TIR and in actual use (vs white wall beamshot). Compared to my beamshot is it worse in your 5700K version?

1 Thank

Thank you so much for the data! I must say that I am very surprised by how little the green AR on Convoy lens changes the tint.

Regarding your question: I think green AR helps with tint shift precisely because it does not affect the periphery much. In a typical TIR setup, the beam tends to be greener at the center and pinkish at the boundary. A good green AR lens should take out much green from the center but not remove green from the boundary, thereby eliminating tint shift.

The 519As actually start out without too much shift (way less than the 219Cs that were in my light before), and however much shift that does occur cannot be accurately captured by a camera, so I couldn't make a reliable comparison to your beamshots, unfortunately. My 5700K triple is under medium spot frosted (Carclo 10508) TIRs, and with a non-coated glass the tint shift is ok enough that you probably won't notice unless you know it's there. With a purple AR, however, the beam turns to fried egg and the tint shift is very obvious on a white wall and in real world use.

It’s numerically small but is noticeable as other users have noted (stephenk above). I am glad that my result is in the range of Djozz’s test. His change between green and purple was about .003, mine is .002.

For those new to this you could very roughly see what the change in beam color is by enlarging the pic below and look at the color field that each of the dot is in. It does become a little more yellow-green with the purple AR lens.

I plan to take A vs B picture of actual scene (wall of truth in my living room :slight_smile: ) for further comparison between green and purple.

another data point:

.