[Review] SST-20 in FW3A is divine!

Thx for sharing your setups. I’m using my FW for work, nightshift, 8h. Cell is empty at the end of the shift cause i run it only on max and turbo.

Running it like you said, emitters will be fried in few days of constant use. Seen it, burried one triple MCPCB few months ago :(

I was hoping 219C FET will be reduced but guess not. And that horrible tint ....


I’ve seen results from a few people now, and it seems to range from +1.6 to +3.0 mduv on the low modes. It sounds like you got the best tint measured so far.

These things really do look nice on turbo. However, I use lights at 100 lm or less about 99% of the time, so I mostly only care about the tint at low brightness levels… and every SST-20 I’ve tried has been visibly green there. So I usually stick with emitters which are pink-biased at low levels instead, like 219B.

Personal perception varies a lot though, so “divine” is in the eye of the beholder. I find I’m happiest with tints which are roughly –2 to –15 mduv, so even +1.6 is outside of that range. As far as I can tell from research studies, the average preference is about –10 to –15 mduv, and about 5% to 10% of the test subjects disliked a +0.0 mduv light enough to call it unacceptable.

Because of these results, the researchers are proposing a revision to the ANSI standard for white light, to lower the line for recommended tints and hopefully get manufacturers to produce rosier shades in common lighting devices like bulbs.

Interesting. The second thing I did with mine was compare it to afternoon daylight bouncing off the green leaves of a tree… and the SST-20 FW3C was visibly more green than tree-bounced light. Taking the lens off helps a bit though, so I can make it more tolerable by polishing off the AR coating.

Yow, that’s some pretty serious use. I’m not sure 219Bs will last long when used that way, even with the FET levels reduced. Would probably be best to use a relatively low-amp cell like a 35E, in addition to the “219” firmware.

They should be fine when running at only the regulated levels though. That works out to only 1A per emitter, which they’re pretty comfortable with.

Getting the tint rosier is easy enough to do with Lee Zircon filters. Trying to make the emitters more throwy and higher output overall is really hard to do. That’s why I prefer the SST20s over 219Bs.

219B has little to no tint shift in its beam, but it’s not like SST20s have much either with optics lights.

Exactly. You can fix CCT and Green/Rosy shift with some filters. You can’t fix CRI.
If you look at the last page of the report you will see I have it set to recommend Rosco filters. Such as the Rosco R3318 1/8 Minus G if you would like to remove some green.

Would love the standard white to change to DUV between –0.002 to –0.003. Personally I like –0.003 to –0.004 as the optimal range. But I find even SW45K at –0.013 is still acceptable between 4000K and 5000K. Many here love that tint.

Ideally, it wouldn’t be necessary to fix anything with after-purchase modifications… especially if it significantly reduces performance.

I’ve tried FW3A in ~4000K with SST-20 95CRI and XP-L HI 5A. The XP-L HI 5A is by far my favorite of the two. It throws well, the tint is good at all levels, it has good beam consistency, and it doesn’t really get hot.

I could fix the SST-20 by adding a minus-green filter, though it would still make about half as much light and twice as much heat. But I should add a filter anyway, instead of complaining about it. :slight_smile:

There are also some 219B 4000K emitters at Illumn right now, which I should probably buy and put into a FW3A. Those should make a really nice beam.

I’m not sure how the 5A compares to whatever 4000k XHP35 HI emitter is in the GT, but those things look downright yellow next to my green binned SST20s with the minus green filter over them. I’m not referring to just what they look like on white either, but how they render wood and skin tones. SST20 indeed does a good job at high R9 rendering— equally as good or better than R9080 Nichias.

It’s not near that bad for small corrections. The filter I linked has a transmission of 89%. And if you sanded it in half to perfect the strength of correction, it could be even better.
Just throwing that out there for the more ambitious among us. It would be a lot of work for a small gain, and shouldn’t the kind of thing you need to do. But it would work.

For anyone interested, I have just posted a 219b R9080 4000K review in the same format as this review.

A transmission rate of 89% is within the range I was expecting. Looking at Bob_McBob’s review, I see a maximum of 1697 lm with SST-20 and a Carclo 10511 optic, compared to 2967 lm with XP-L HI 3D. 1697 * 89% / 2967 = 50.9% as many lumens, which is why I say it makes half as much light. It also heats up quite a bit faster, though I don’t have actual numbers for that.

At lower levels the difference isn’t quite as severe… SST-20 with minus-green filter makes about two thirds as much light as XP-L HI, instead of only half.

SST-20 has a slightly throwier beam (at about 4.7 cd/lm instead of 4.4 cd/lm), but due to having half as many lumens, the maximum throw is significantly less — about 7100 cd, compared to XP-L HI’s 12500 cd.

So I’m pretty happy with the XP-L HI 3D version as a general-purpose light, though I also use a 219B 4500K FW3A when I care more about accurate colors instead of overall performance. Have been eyeing the 4000K 219B emitters too, which Bob’s measurements indicate would probably perform at about the same level as SST-20… except with a tint I like.

It’s really nice that we have so many to choose from, since the broad range of choices covers a broad range of different tastes. :slight_smile:

Maybe I didn’t explain completely. The meter is recommending I use a minus green filter because I have the target CCT set at 5500K. It is also recommending a CC filter too, and the combination will swing down the BBL to 5500K. It’s a significant green subtraction, way more than the 2 MacAdams it takes to bring the Floor to the BBL.
My intention was to demonstrate a method of adjusting a minus green filter.
Neither me nor the meter recommend putting a filter on this SST.

We don’t need a new line as much as a “Keep Out” zone in the green area.

This stuff is new to me. I know what CRI is but I would love to learn more about some of the units you guys are posting about. Is there an ultimate guide to all the different metrics? CCT, Duv, CRI (Ra), TLCI and even lumens and candelas. Kinda like what they have on baseball sabermetrics on fangraphs haha.

Welcome to BLF!
I can give a quick rundown of the stuff.
CCT is Color Correlated Temperature. It tells you how warm or cold the light appears. Cold colors have more blue, warm colors have more red tones.
Duv is actually how we type Δuv. The greek Delta stands for “change” in math. So Duv is distance from the Black Body Locus (that line we like).
CRI originally evaluated 8 colors. Now it evaluates 15. When we refer to all 15, it’s often called Ra. So “CRI (Ra)” is just a catch-all to make sure everyone understands Ra replaces CRI.
TLCI is like CRI for cameras. Cameras see light a little different than our eyes.

Hope that helps.

Sekonic press release:

Joshk, feel free to try out the new C-800 firmware and do the TM-30-18 report

Correct me if I’m wrong. So if the SST-20 led is rated for 4000K the CCT gives a more accurate measure in Kelvins. It goes from a scale of 1000-10,000, warm to cold. CRI and TLCI is a rating of the ability to produce the full the colour spectrum with 100 being perfect. But what would an extreme example of poor CRI look like? Is it like an old sepia filtered WWI picture?

I’m not sure I understand duv fully. The measurement seems quite sensitive with people complain about the 10000th measure. I found this article.. Does Duv give another dimension to CCT? Is it measured as the perpendicular distance from the desirable locus line? Is rosy, to the bottom right, a negative measure and yellow/green, above the line, positive? Thanks

Kelvin is the unit used for CCT, CCT = 4000K for example. Yes, 100 is perfect for CRI and TLCI. An example of poor CRI would be if you looked at a bundle of wires with a light, and couldn’t tell the difference between pink and red, or blue and purple for example.

Yea people are sensitive to Duv. I think it’s because of our nature to see greenish things as diseased.

That just added some pie slices. What do they tell us?

Hell if I know, but if this is set to replace Ra/R9 we should probably figure it out :slight_smile:

AFAIK you want a big wide circle like this for even, full spectrum. If it jets in or out too much you will have a bias towards that