I bought two Sofirn SF13 lights during the recent AliExpress anniversary sale. I’m not set up for in-depth reviews, but I haven’t seen much comment on this model, so I thought I’d post my first impressions.
Please note that my units are the new model with a 420 lumen turbo level. There is an older version with a 320 lumen turbo level and lower mode levels across the board.
Here’s a photo from Sofirn and a link to their AliExpress page:
Build
The SF13 is a 2×AA light, so the length is about what you’d expect at 155mm. It’s 22.5mm diameter at the head, 21mm at the tail (tapering slightly to about 19.5mm at the rearmost point) and about 18mm at the thinnest point. This allows for a tube thickness of about 1.25mm, so it’s nice and solid.
The light is built in three sections: head, battery tube and tailcap. The threads are square and anodised at both ends, so the light can be mechanically locked out. There are O-ring seals at both head and tail.
There’s enough room in the tube for two white Eneloops to go in easily, with a bit of rattle if you shake the light hard. However, there’s no flicker and no unintended mode changes. The driver board has physical reverse polarity protection, taking advantage of the fact that consumer AA cells always have button tops.
The all-aluminium structure is black anodised. I saw one tiny nick in the anodising on one of the heatsink fins, but no other defects. There is knurling, but it’s not very aggressive. It looks to me as if they made criss-cross vertical cuts into a flat surface to get flat-topped diamond knurls. I’d prefer something grippier, but I should warn you that my favourite knurling is on the Astrolux S41S, which would probably double as a wood file
The light is supplied with 2 spare O-rings and a reasonably sturdy clip, which is pre-attached and appears to be made from stainless steel. The clip only mounts in one place, giving head down carry with about 35mm protruding from a pocket, so no deep carry. The battery tube is not reversible - the threads are a different diameter and have a different number of turns at each end - so there’s no chance of getting head up carry that way, either.
No lanyard is included, but there is a lanyard attachment hole at the tail.
Due to the protruding tail switch boot, the SF13 does not tailstand.
Although the tailcap has raised sections on either side of the switch boot, these are only there to provide a place for the lanyard hole. They don’t guard the switch and it is perfectly possible to turn the light on and off by pressing it onto a flat surface.
Operation
The light has a reverse clicky tail switch and mode memory, so it’s the usual click on, half click to advance to the next mode and click off. Mode changes are strangely slow - the light stays off for about 1 second after you let the switch back up - but I am getting used to it. The mode sequence is Low - Medium - High - Turbo, then back to Low and repeating from there.
Half-pressing again during that 1 second pause gets you into strobe mode, which I estimate at about 4Hz. A half-press in strobe mode takes you back to the mode you were in before activating the strobe.
Modes and Output Levels
Manufacturer mode specs from the manual are as follows:
- Low: 8 lumens for 26h 06m.
- Medium: 55 lumens for 3h 46m.
- High: 196 lumens for 1h 04m.
- Turbo: 420 lumens for 28m.
Turbo steps down to High after 3 minutes, so that 28 minutes is what you’d get if you kept bumping it back up and didn’t worry about overheating the light. I wouldn’t advise that under normal circumstances, because the 3 minute step down is a good design choice here, allowing the light to get disconcertingly warm to the touch - I tested this at 21°C room temperature - but not dangerously hot.
Strangely, strobe is specified at 420 lumens for 28 minutes, just like turbo, when I’d expect longer, given that the light is off and presumably using less power between flashes.
Comparison with Other Lights
I don’t have any optical measurement equipment, but I compared the following lights by eye:
- SF13 (5500K Cree XP-G2).
- Nitecore MT06MD (5000K 90+ CRI Nichia 219B).
- Convoy S2+ (5000K Cree XM-L2).
- Fenix UC35 (6500K Cree XM-L2).
Beam Profile
The SF13’s beam profile is general purpose, leaning slightly to the throwy side, courtesy of its smooth reflector and XP-G2 LED. There’s a clear hotspot; at any range more than a foot or so, it transitions smoothly through the corona into the spill. I couldn’t see any rings or artefacts in the beam.
I wouldn’t be surprised if the SF13 was a bit throwier than its 75m specification; the beam is slightly tighter than my Fenix UC35.
Edit: Sivy has provided a nice beamshot comparison with the BLF A6. You can find it here in Sivy’s post.
Lumen Levels
The four lights I chose for the comparison have many modes with similar lumen levels, so I tried corresponding modes side by side.
I don’t believe the SF13’s specified 8 lumen Low; I’m sure it’s brighter than that. It looks more like 15 or 20 lumens to me. I actually prefer that level, but other people might not.
However, the SF13’s Medium, High and Turbo modes appear to be “about the same” as the corresponding modes on the other lights, so there’s a decent chance that the Sofirn lumen specs bear some resemblance to reality.
Colour Temperature, Tint and CRI
Sofirn specify the SF13 as having a colour temperature between 5350K and 5700K. It looks cooler than the two 5000K lights, but warmer than the 6500K UC35, so I’d say that the SF13 meets its specification there.
I was unable to reliably discern any tint shift away from pure white. The SF13 looks green compared to the MT06MD’s Nichia LED, but pure white compared to the other two lights.
The SF13’s XP-G2 is definitely a low CRI LED, though. That was clear to see when comparing it to the 90+ CRI Nichia in the MT06MD.
Some Comments on the Manual
I noticed some oddities in the manual. For example, it has English and German sections, but the German part is for the Sofirn SP10A, a completely different light.
The manual also specifies “IPX-7” in one place and “IPX-8” in another, so that’s something to bear in mind. IPX7 is submersible, but only to negligible depth, whereas IPX-8 is submersible to at least 2 metres for 30 minutes. Call it the difference between dropping it in a 2-inch puddle versus dropping it in the deep end of a swimming pool.
Finally, the manual says “protected cells are highly recommended” and “900mAh protected AA Alkaline battery used in test.” Your guess is as good as mine, especially given that the SF13’s specified working voltage of 1.6V to 3.2V only supports 2×1.2V NiMH or 2×1.5V alkaline cells. According to the specification, 3.7V Li-Ion cells would fry it, protected or not.
Unusually for a modern light, the manual recommends alkaline cells over NiMH, but I only use Eneloop NiMH, so that’s what I tried it with, and the light works fine on those.
Single-Cell Operation
Although the manual specifies a minimum working voltage of 1.6V, my units do run on one NiMH cell and a short-circuit AA spacer. They flash slowly (at about 0.5Hz) when first turned on, which I presume to be a warning state. Half-pressing stops the flashing, but even then, the lights only do Low and Medium output levels, with a small step up to High and practically no extra output in Turbo.
In addition, they occasionally fail to change mode when half-pressing the switch, so the low voltage is probably at the limit of what the driver can tolerate. This raises the possibility that other samples of the SF13 might misbehave more or just not work at all in single-cell operation.
I ran one of these lights at High for as long as the single NiMH would hold out, and it managed at least 40 minutes at the reduced High level, which is not bad! It then stepped down to Low and continued to run, gradually fading. I stopped the test about 10 minutes later when the output was clearly well below the normal low level, at which point the cell was down to about 0.88V. The light could still be switched on at that level, too.
Incidentally, the fact that the SF13 can run on a single cell makes me wonder whether the low voltage protection specified in the manual is actually present or not. I haven’t seen any evidence of it, put it that way.
Running two cells down to the dregs would definitely risk reverse-charging one of them, and you wouldn’t get much warning, either.
Modding
I haven’t taken these lights completely to pieces yet, but mod potential looks good.
I’ve had the bezel off - it’s immediately removable by hand - so there’s easy access to the lens, reflector and LED. There’s a circle-section O-ring between the bezel and the head, with a flat rectangle-section O-ring between the lens and the bezel.
I haven’t looked to see what the MCPCB is made of, or what it’s sitting on, because I’d have to desolder it, and I’m not ready to do that yet.
The driver and tail switch board have securing rings, so they’re pretty easy to get at. You’ll need something like circlip pliers, though, because the rings have half-circle cutouts on the inner circumference, not holes that you could put tweezers in. Tweezers would just slip out of the cutouts.
Conclusion
I’m quite happy with these lights for the USD $9 each I paid for them (current price about USD $10.50), despite the slightly irritating slow mode changes. They’re perfectly serviceable as general household utility lights.
Questions are welcome, but I’m legally blind, so photography is awkward. I’m not planning to provide beamshots, LED closeups or anything like that.
Edit: Sivy has provided a nice beamshot comparison with the BLF A6. You can find it here in Sivy’s post.