[Review] Convoy S21F EDC with tint shift

Hello everyone!

One more review finished. Convoy S21F high CRI EDC floodlight.
Hope you will enjoy watching (if you liked video, please subscribe, more reviews are coming… :slight_smile: ). As always suggestions and questions are welcome :slight_smile:


3 Thanks

Hi thanks for sharing. Please allow me to add as I had used the same wrong term myself:
“Tint shift”: this refers to undesirable change in color as one goes from hotspot outwards. For example Nichia E21a has yellow tint shift at periphery of hotspot when used with 10622 optic, etc.
“Tint ramping”: This is used by Hank in his variable CCT lights. Technically not entirely correct because although tint (Duv) does change, it’s the CCT that is ramping. I’ve seen the same “tint ramping” in Anduril 2 instruction sheet and maybe that’s where it came from.
“Variable CCT”: This is used by Simon/Convoy. But it’s not a verb to describe the action so I use “CCT ramping.”

To see either end of CCT ramping :innocent:, I turn down brightness and look at the LEDs. It’s easy to see 2 turning off at each end. But agreed, a double flash would be nice.

My first S21F (the very first batch) had upper end CCT at 5000k, like yours. I since have received a second S21F, the second batch, and the upper end is now 5700k.

I love the light and CCT mixing changes CCT AND lowers the Duv to level never seen by 519a without dedoming. Very highly recommended for Nichia hobbyists.

1 Thank

Nice review thanks man !

1 Thank

Very nice review L.R.O of a pretty cool flashlight! :+1:

Hey, thanks for clarification. I also thought that this term is not the right one, but there are a lot of people who do not know what is CCT. That is why I used that term, I think the best would be color shift, simple for beginners.

Sure and thanks for the nice review :+1:.

Thanks for posting runtimes as images, I’m on a remote site a lot of the time (satellite internet) don’t have the patience to load YouTube videos.

You could peel the domes off the LEDs for c.1800-4000K range and more negative duv (rosy pink tint) than you’d know what to do with :stuck_out_tongue:

Nah I don’t think it’s wise to dedome this light. The mixed color is already in the very negative range. Why do you need a pink light?

Can be said for majority of posts on here…

I did say it would have ungodly negative duv… I’m tempted…

A dedomed 2700K 519A is around 2200K, not 1800K. I believe an 1800K LED is only achievable with Nichia E17A LEDs which is essentially candlelight.

I’d probably dedome the LEDs since they look so nice with the negative DUVs. I really don’t notice an issue with my dedomed 2700K and 5700K being too rosy. My dedomed 3500K (=2700K) is certainly noticeably rosy, but I still like it more than my 2700K 219B.

I think the extra rosiness with the lower CCTs helps make things appear less dull and yellow.

But perhaps the mix of the dedomed LEDs would be overwhelmingly rosy in the middle CCTs? So it may come down to where you want rosy tint to come from, either at the extreme ends or in the middle.

And just for comparison, here’s the 1800K E17A vs the dedomed 519A 2700K:

1 Thank

thanks for review of 2nd version of S21F.
O, 2nd mode now has ~180 lumen’s instead of ~60 in 1st version, with ~9 hours runtime from fully charged battery. For my taste, it’s better, and closer to “10%” mode.

Can you measure PWM in mixed color mode in 1st-2-3rd modes ?

You guys talked me into trying a s21F out and it is very nice.
I was thinking it would be as bright as my s21D 4500k (8a cc) but it’s not quite.
My main complaint is convoy’s ramping is painfully slow at the top of ramp and feels unresponsive.

1 Thank

My first batch S21F’s outputs are not different from second batch. I posted the numbers in other thread. As best I could tell both versions measure 140-170 lm in 2nd mode, not 60.

No PWM at 2700k, 5700k, and Turbo 3800k. The rest has PWM.

Yep outputs are U shaped. Highest at either end 2700k and 5700k, and lowest in between - hope this helps you to decide where to set it. I set mine at 5000-5700k.

To compensate Turbo mode @3800k is 2000 lumen and hence a little higher than S21D 519a 4500k and 219b 4500k.

The CCT shifting is actually pretty fast when I measure with Opple while shifting. It helps if you are trying to set to a certain CCT, but I guess doesn’t help if you want to switch CCT quickly.

Tint ramping speed is fine, the brightness isn’t. should be quicker like anduril

1 Thank

Convoy S21F tint ramping with 8A CC driver uses PWM to split the power between two channels

Just thought you might be interested to know. The micro-switch is acting up in my unit, so I took it apart and tried to clean the switch using alcohol. I had a good look at the driver board, sorry I didn’t take any picture because I forgot to do so.

I always wondered how they managed to achieve dual channel and 8A CC together. On other brands a dual channel light would be a FET + linear drivers? In this S21F driver it uses the 8A CC driver and added two MOSFETs downstream to split the power to 2 x 5700k in parallel and 2 x 2700k in parallel.

Essentially in top or bottom of the ramp a single channel receives 100% of 8A power, in the middle each channel receives 50% of the power via PWM. I guess it’s an okay design as PWM gives them high channel splitting efficiency while the single buck driver provides the source of power which is also very efficient. The frequency is high enough to be completely imperceptible.

So here is that. If you want tint ramp with good efficiency, you can have it all. Convoy drivers are well designed, except the UI.

1 Thank

Thank you. I’ve given this much thought also and have also arrived at the conclusion that PWM is used to reduce output during CCT mixing.

One more setting you haven’t mentioned: Turbo, also in the “middle” @ 3800k, but at this point 4 LEDs split the 9A total, and no PWM?

This creates the curious situation wherein 3800k achieved by CCT ramping is very low output, whereas 3800k by double clicking to Turbo has the highest output for this light (2000 lm).

Turbo mode on this light is a separate mode that can’t be achieved by brightness ramping. I suspect what it does is it opens both channel FET’s so it’s essentially 4 emitters in parallel splitting 8A of power, or 2A each. This should result in slight brightness increase. According to 519A’s power-lumen curve, 2A gives 600lm and 4A gives 1000lm. Multiplexing two groups of 2 emitters at 4A each gives you 2000lm combined output (since only one channel is active at any given time, perceived brightness equals to one channel, not two) but a single group of 4 emitters at 2A each should give you 2400lm. I don’t have the equipment to actually verify this theory so if anyone can do it that would be great. A 20% lumen increase is not much, but it’s still an increase nonetheless. This didn’t take into account that 2700K emitters have less lumen output than 5700K ones so it’s purely in theory.

It was my review that had the low measurements on 2/4 brightness.

Mode 2 can’t be 60 lumens. At least not my first batch sample.