Sofirn C8F 21700 Teardown and Review

Mine is extra tight. Even tried a jar gripper. I was thinking it would just spin off, but since it wasn’t I was thinking maybe I am missing something and/or have gotten in over my head.

I am looking at replacing it with the yellow one in this bunch. https://www.ebay.com/itm/Set-of-4-42mm-diameter-Lens-for-C8-Flashlight-Blue-Green-Red-Yellow-/183098112801 - the link is not spam. It is an eBay seller.

I get a bit of dense fog where I live and driving snow think star field effect - like driving along in your car in heavy snow at night) - in either case, a white led is not helpful. I tried a 2700k from Sofirn (IF25) and it was mildly better. So I thought I would swap out the clear lens for the yellow one and see if it improves visibility under either condition and if it does post pics here - to the extent it is something people would be interested in. So the goal is just to simply swap the lens - nothing else at this point. Thanks!

Ah OK, Yeah there isn’t much heat at the lens, so may want to try something like they use for car fog light film, that will provide a better yellow filter. In your case you will just need to unscrew the bezel. Nothing else. Not sure why yours is so tight. I’d think a high CRI version of C8F 2700 along with a yellow lamin-X film would help?

Shows how the bezel comes off it is pretty tight. Took me a bit to get it off to take this photo.

Thank you for the information and especially the picture. I am going to try the lens and see how that works out. I have a C8F with the three XPL LEDs 5500k and then I also ordered a C8F with lh351d-2700k. I was going to compare the two with the lens swap. I didn’t know if I could use the fog light film safely - so I will track some down to try that next. Thanks again for the help and picture!!

The best thing to have for unscrewing bezels is a set of rubber strap wrenches like these.

They are $5 a set at Harbor Freight. Your local hardware store my carry a cheap set as well. You set them up in opposite directions and you get a lot more leverage.

It it confirmed later batches of this light gives more lumen? i have one from banggood bought in May last year and its no way close to the stated 3500 lumens… now also not long ago the spring bypassed wire fell off from the tailcap… really bad job on that one… my l6 bypassed still havent come off….

but for a budget price this light is ok robust and not much to complain about really…

No, based on reports from users they still seem to vary a bit.

I think the only people complaining about it not meeting specs are the few who can actually measure it. Most users only care if it’s really bright or not.

I wonder if they would sell less units if it were labeled 2500lm or 2750lm or 3000lm?

Yeap possibly may sell better if they just call it a 2500lm and all will be fine. Still a great flashlight IMO and worth the $27 I paid for it. I am in the process of a complete tear down and rebuild with new wires spring bypass (maybe even on the tail cap with double spring), better thermal paste, etc… to see if I can get 3000lm or higher. :slight_smile:

I think they just like most other chinese budget brands just do an estimated guess on the lumens… and if its a bit more then what u actual get it helps sales hehe…

No, they have some kind of expensive machine that their company bought that is supposed to measure lumens and lux. I don’t know the brand or details of it, but it doesn’t always match what we measure. On most of their lights their lumens match what we measure and sometimes it measures higher than we do, usually not this much higher. Some lights tend to measure lower candela than we do (like the Sofirn Q8), but usually their candela is about right. Sofirn is one of the more accurate companies for specs.

I saw a picture of the outside of the machine (big rectangular box), but the model name didn’t turn up much when I searched for it. It was a Chinese device not sold outside of China.

Well……My new batch of C8F are back down to 2710 lumens at 30 seconds. It must be the quality of components that are used. I checked Jason’s assumption being poor contact of the retaining ring and driver by putting a ring of solder around the driver but made zero difference.

Can you start your light on moon and check all emitters are equally lit? Just received a batch and one light has two leds that are dimmer on moon which tests at 2400 lumens at start up and drops to 2100 after 30sec.

I’ll get a replacement for this one.

but this isnt bad really ?

No, it’s fine.

Correction - was using an old 30Q, used a 40T which makes it 3050 lm at 30 sec. This is still lower than my previous batch at 3270. It’s just strange behavior.

The TK70 & SP70 behave the same between different batches.

It’s not too bad. The choice of battery used does make a difference and I’m not sure what battery Funtastic used. The thing with FET drivers is they are notoriously inconsistent.

Sofirns 3500 lumen rating is not realistic, so 2700 can seem disappointing, but it’s quite bright to the eye. Most people would not be disappointed by the output, but BLF’ers are not most people. :wink:

Hey, since you have a TK70, what driver did they go with? Regulated or FET? What output did you get with it?

Pretty certain it’s a fet since it jumps to 7000 at 30 sec with a bypass. Low level is almost 800 lumens which isn’t ideal. Stock is around 5800. Advertised at 5500 so it’s one of the few that test higher.

I will note that the Sofirn HD cells maxed at 6500, the PLB 5000mAh get 7000. The PLB won’t connect in series without a magnet or solder.

I won’t be stocking more since springs are too short or tube too long making the light flicker tail standing.

I notified Sofirn ages ago but since I’m the only one who noticed, it’ll remain unchanged.

ah okay… ja that would explain it then.

The Sofirn C8F comes in XPL-HD and LH351D (5000k, 4000k, 2700k). Although their green and brown body color C8F seems to only come in XPL-HD LED type.
I wonder if Sofirn will accept request for having the C8F with LH351D LED in green/brown body color though…

Sorry didn’t see this until now, put my sample did not have this problem. I bought it direct through their site. I have 2 samples - one with the XPL and the newer one with 90CRI LH351D and neither one exhibits uneven LED’s at moonlight level.
Thanks,
Al