[Review] Convoy S21D || 4 x Nichia 519A w/ TIR, 12A FET Driver, 1 x 21700 || Output, Thermal Regulation and Beamshots

LED Flashlights – General Info 21700 Flashlights
1

The flashlight was sent by Convoy review.
Here's the product's link: Convoy S21D

The Convoy S21D comes in black matte anodization.
As with all Convoys, the quality of the product's finish is excellent.



The body of the flashlight includes horizontal grooving for better grip.



The head makes use of some heatsinking grooves.
This come especially handy whenever the 100% output mode of the FET driver is invoked



A polished metal bezel sits in front of the glass lens.



Looking at front part of the head, we can see a TIR lens along with 4 x Nichia 519A high CRI emitters.
I've went with a 60° TIR and 4500K 519A emitters, though more TIR (10°, 30°, 60°), emitter (219B, 519A, XPL HD) and tint options are available.



Taking apart the head is super easy, since the bezel is not glued into place.
As we can see, the TIR lens is protected from scratches using a glass lens, while an o-ring is also in place.



Of course, the TIR lens can also be removed, giving us access on the quad LED copper board.
Plenty of thermal paste is in place.



The Convoy S21D comes with two different driver options:

  • 12A FET driver - providers the maximum possible output, though doesn't provide regulation.
  • 8A CC driver - less output, but it's constant current so its output is fully regulated.


The driver of the flashlight uses a thick bronze spring as the positive terminal.
A retainer ring is what's holding the driver in place.



The Convoy S21D is operated via a mechanical, reverse-clickie tail switch.
The switch rubber boot sits flush with the outer metal flaps of the tail, which allows the flashlight to tailstand.



Looking into the tail, we can see a thick bronze spring being used.
A brass retainer ring is what's holding everything into place.



Here's a tail disassembled.



And a closer look at the switch board.



The tail side threads are anodized, which allows for mechanically locking out the light.
Both threads are very cleanly cut and arrived nicely lubricated.



The bundle I've chosen, also comes with a Liitokala 5000mAh 21700 cell.



The cell is flat top, and it's more than capable of providing the required amperage for the flashlight's FET or CC driver.



User Interface

The Convoy S21D comes with the 12-group firmware.
Here's all the supported actions:

Turn of/off: Single press
Mode cycle: Half press
Configuration mode: 20 clicks

The user can choose from 12 different mode group. Feel free to check the product's page for info about each mode group.
I've settled on mode 2, which supports the following output levels: 0.1%, 1%, 10%, 35%, 100%
Mode memory can also be configured - I've chosen to keep it off.
Of course the driver itself also supports Low Voltage Protection as well as Thermal Protection.

Output

Here's my output measurements along with the current draw per each output level using the include 21700 Liitokala battery.



As we can see, the output of the S21D is very impressive, considering that it's using a high CRI emitter.
On Turbo the flashlight will push more than 2800 lumen at turn-on.
Of course, since the variant I chose uses the FET driver, this output isn't regulated and thus will depend on the input voltage.

Thermal Regulation

Here's a thermal regulation I've recorded using the included Liitokala 21700 cell over a 20 min period.

What we can see in the graph:

  • Turbo's (100%) full output is sustained for 45 seconds. After that, the flashlight will slowly step down to ~350 lumen over a 85 second period.
  • High's (35%) output is sustained for 2 minutes. After that, the flashlight will step down to 350 lumen
  • Medium's (10%) output is sustained for the whole span of the test.
  • The flashlight regulated the temperature pretty well, since the body never got too hot to handle.


Outdoor Beamshots

Here's a couple of outdoor beamshots with the S21D.


Thoughts

The Convoy S21D is a very interesting flashlight
It's capable of producing a wall of high CRI light (I've gone with TIR 60° and Nichia 519A 4500k).
As always, the build quality is on point, while the flashlight itself is quite compact provided that we're talking for an 21700 quad-emitter setup.

Note: I'm in the process of preparing some Thermal Regulation graphs, which should hopefully be ready in the upcoming days.

2

I got the 12A version first, because it was the only one available. A couple of months later, I got the 8A version.

At 35% output (the mode I use every night doing farm tasks) I am unable to observe any difference in brightness between the 8A and 12A lights. If you have a choice, the 8A buck driver is the way to go IMO. Longer runtimes are always nice.

3

I agree. I’d say the FET driver is only recommended if you really know what you want (a hotrod flashlight :D). From a practicality’s standpoint the 8A CC driver is the better option most probably.

4

Thanks for the nice review. Sounds like Turbo with FET is quite a bit brighter than 8A.

I am very interested in your thermal regulation graph. I have not done formal testing but my Convoy lights when started at Turbo would seem to settle down to around 40% of max brightness then stay there forever, comes hell or high water lol. External temp would hit 60 even though it’s supposed to be 55 regulated. Let’s see if this one with FET does the same.

Just curious what did you do to make the pictures flash like above? And is there some reason why they are flashing - am I missing something? :slight_smile:

5

In regards to the flashing beamshots, I’ve basically stitched two shots together. First one is a control shot - a photo without having the flashlight turned on.
Since I was taking the pictures at pitch black, the control shots turned out completely black.
The second shot is with the flashlight at Turbo.
Stitching those two shots together can be done with many different software. I happen to use the imagemagick/convert CLI tool to generate the GIF.

6

Thermal Regulation

Here's a thermal regulation I've recorded using the included Liitokala 21700 cell over a 20 min period.

What we can see in the graph:

  • Turbo's (100%) full output is sustained for 45 seconds. After that, the flashlight will slowly step down to ~350 lumen over a 85 second period.
  • High's (35%) output is sustained for 2 minutes. After that, the flashlight will step down to 350 lumen
  • Medium's (10%) output is sustained for the whole span of the test.
  • The flashlight regulated the temperature pretty well, since the body never got too hot to handle.
7

Thank you. Did I read that right ~2800 lm at Turbo start, then reduced to 350 lm? Down to 12% of max after just 2 minutes? This seems to be so low. If this is the FET driver’s behavior than I am not sure that I would want it at all.

I only have numbers from ceiling bounce test, of both 219b and 519a in my two S21D’s. Both with 8A Buck driver. I superimpose its runtime on your graph, making assumption that my 100% is 2000 lumens. Even if the 8A Buck is not as bright at start, it seems to sustain the stepdown brightness so much better.

8

My results with the 8A driver are similar to those plotted by cannga, in fact I’m getting slightly higher sustained brightness. The 8A driver thus appears to have considerably better sustained performance than the 12A.

9

Thanks Stephen. Always feel good for my result to be confirmed by a vet.

BTW I’ll find the post later, but I remember you called this (the behavior of S21D Buck vs FET), on the nose. :+1: :slight_smile:

10

Nice review bilakos10!

I’m surprised by how messy the quad MCPCB is. Solder blobs, flux, wires roughly soldered. Doesn’t seem typical for Convoy.