Sofirn C8F 21700 Teardown and Review

Dont know about the skyray king bit my c8f 18650 I measured 3,155 lumens running on fresh vtc5d batteries and cleaned contacts. My 700 lumen Skylumen Delta throws several times farther than my C8F. Your customer is confused between lumens and throw. You shoukd tell him to ask on BLF.

Is lux the intensity of the hotspot/place of the beam being measured?

Okay, I always thought lux = lumens when calculated correctly.

Lux and lumens are completely different.

Lumens is the total amount of light regardless of beam shape.

Lux is the highest intensity at any one point in the beam and is only useful to determine how far the beam will travel.

You could have a 200 lumen light that was floody and only measure 80 meters or you could focus it really tightly and make it go 800 meters.

if it was XML it’s just as likely he perceived it as more floody and looks more impressive at short distances or indoors. Either way it’s perceived brightness. You wont be able to really notice much difference between 2000 and 2700 except for the throw/beam shape.

On the “bright” side now he has a light suited for more flood and more throw, whichever they may be!

I just sent him the Sofirn Q8 so I’m sure that’ll impress him at 5000 lumens with spring bypasses.

Thank you for both of your input answering my questions

I just tested my new batteries in the C8F (has new MOSFET and stock led wires). I measured both lumens (TA Lumen Tube calibrated with Maukka’s light) and tail amps (using UNI-T UT210E clamp meter and heavy wire) with the Samsung 30T, 50E and Sofirn 4000mah 21700.

Something funny happened as all my first round results were down in power by about 500 lumen and two amps (left side of chart). I checked the driver components and saw no issues, but around the driver edge and driver shelf there seemed to be some dirt or maybe flux build up. I cleaned it all up with a q-tip and alcohol then reassembled. This fixed it. So new readings are on the right hand side. I did not recharge the batteries between the first two 30 second pulls and the second two 30 second pulls. I took Turn On measurements as well as 20 seconds and 30 seconds to give an idea of how output was dropping.

All 3 cells did very well. I think the 30T did not have as much voltage sag which is why it had slightly higher lumens while drawing the same amps as the Sofirn battery. It goes to show how good the Sofirn battery is.

Since I like this light so much, I may swap in a Lexel driver in order to get that instant jump to Turbo (no more ramping up in output) and to know the battery voltage better. The stock driver does not start showing red until 3.0 volts.

Oh yeah, I charged each cell on my Miboxer C4-12 and tried to get a rough idea of their Internal Resistance. This charger is known for reading IR a bit high, but the results seemed fine.

Did you buy another one or did you swap in the revised driver they sent you? Thanks for the very useful data!

I only have the one light with the original driver. A few posts back I showed that I swapped the FET out for one with lower resistance.

I read that Sofirn had confirmed it was the FET causing the reduced lumens, this can’t be the only cause looking at your results

I am not ragging on this new version but I just have 2 questions;
So only after these mods of replacing the FET and beefing up connections is this flashlight just now getting to the output I already have out of the box with the 18650 version when used with a good cell?

Just a newfangled 21700 battery capability.
Where’s the love?

Will wait for the factory BLF approved/vetted Hotrodded version.

The FET they used in the early drivers did have a little more resistance in it which reduced output. The details are in post #17

I actually don’t know what a triple xpl can do amperage wise. Right now each emitter is doing about 3.2A and about 1000 lumen.

Should we be expecting 15A and 4000 lumen?

Is there too much resistance in the led wires?

A completely new driver might tell us if there is something in the Sofirn driver that is creating a bottle neck.

Frankly, I was okay with the stock driver that only drew 7.25A and 2600 lumen. I tend to set it at about 500 lumen for walking around and do the occasional blast of turbo. It’s a big jump in output and it doesn’t get too hot or run the battery down too fast. It is a good general purpose light.

Using the 30T and 4000mah battery gets the light hot so fast it’s not that practical anymore. If I put a Lexel driver in it and got even more output, I’d probably remove the spring bypasses to try and reduce output on turbo to make the light practical again.

I know, it’s not the BLF way to reduce output, but I use this light at work where it needs to be a bit more practical.

The latest versions already have the updated FET so you don’t need to do anything.

What do you mean by beefing up connections? Both springs are factory bypassed. There is nothing to do there either.

Do you mean replacing the led wires with bigger ones? How big are the wires in the 18650 version? Bigger or the same? I don’t know.

How many amps does the 18650 version pull? I don’t have one so I don’t know.

I looked up reviews for the 18650 xpl version to find amp draws. (This is usually more accurate than lumens)

Bilakos10 measured 8.33A with a Samsung 25R cell, but that is not a very good battery.

Robo819 measured 8.4A with the Sofirn 18650 and 8.9 with an LG HE4 cell.

ImA4Wheelr measured 8.8A with the Sofirn 18650.

Barkuti measured 9.7A with a Samsung 30Q. He used a shunt as apposed to a clamp meter.

DB Custom drew 13.57 amps with a very high drain VTC5A 2600mah and with slightly lesser batteries, like a 30Q, he got between 10A and 11A. This was with 18ga led wires and an upgraded MOSFET though. So it doesn’t really count as stock.

Since the 18650 version is supposed to be about 2500 lumen and the 21700 version is supposed to be 3500 lumen I’m guessing there is some kind of bottleneck with this 21700 driver. The 21700 light should probably be drawing at least 12A to 13A stock to get 3500 lumen. Sofirn may have over rated this 21700 version. It might be a true 3000 lumen light.

Referencing this post;

Dale shows the following;

Sofirn ……………………………30Q…………………………Sony VTC4

  1. 0.552 Lumens…………………0.552 Lumens………………0.587 Lumens
  2. 37.50 Lumens…………………22.29 Lumens………………22.81 Lumens
  3. 83.15 Lumens……………………82.8 Lumens………………84.53 Lumens
  4. 890 Lumens………………….872.85 Lumens………………931.5 Lumens
  5. 1338.6 Lumens……………1314.45 Lumens…………….1400.7 Lumens
  6. 3015.3 Lumens……………3018.75 Lumens…………….3125.7 Lumens
    ……2756.66 at 30 sec…………….2691 at 30 sec

This is stock from what is described.
Amp draw do not know.

After 18 gauge wires and new mosfet went well over 4000

Bottom line, If I can’t see the difference I do not need to be buying a new battery form just to be “like the new kids”
Like the New Version just can not justify another version at this time.
Thanks!

Keith

His lumen tube reads way higher than mine. This is why I don’t use his lumen numbers. The amp draw is more accurate.

The best thing about the 21700 version is the new ramping mode UI with one click on/off. This is why I never bought the 18650 version.

OK, that’s cool
Good info.

I just spent some time to dig up my post with my own lumen numbers and seems like Dale’s measurements are spot on

I grabbed a few 40T to try them out in my light.

I’ve got the SIRa800 and 18 gauge from the driver to mcpcb installed.

Turn on with sofirn battery gave me 3,904 and 3,504 at 30s
Turn on with 40T gave me 4,024 and 3,764 at 30s

3% difference, negligible IMHO

Both tests, batteries fresh off the charger. Measured in a maukka light calibrated sphere.

Mtn. D4-v2 driver in mine, I got 3890lm on the Sofirn battery 4430-4400lm on 2 of my 30T batteries.

This does not invalidate my statement. It still holds true.