ZozzV6 wrote:

contactcr wrote:

So what can we do about this giant retaining ring? Think i can grind it down at an angle on the inside edge to fit a driver with 7135 on the bottom?

Use fet+1×7135 driver with no parts on spring side.

In such a nice host it’s a shame to not have more regulated output options.

20mm like this? http://www.mtnelectronics.com/index.php?route=product/product&path=67_11...

Mtn also has a 22mm board but then you would have to sand it I guess..

x-ray wrote:

Hi Sofirn

I want to purchase a C8F 21700 from Aliexpress but am now concerned about this defective driver issue.
Can you advise if their stock is being checked/replaced/fixed ?

Thanks

Just a quick check. Does anybody know if it’s safe buying from AE official store at this point or is it likely I will still receive a light with a dodgy FET? Many thanks.

Ask them before ordering just in case.

Perhaps they will just strip the lights of the faulty driver and replace it, or they’ll just sell them as hosts.

BlueSwordM wrote:

Ask them before ordering just in case.

Perhaps they will just strip the lights of the faulty driver and replace it, or they’ll just sell them as hosts.

BTW, does anyone know an EU store that stocks a nice/recommended driver plus a board with LEDs that would fit the host? Cheers.

czajunia wrote:

Thanks. Just sent Sofirn a message to their AE store. Will update the thread as soon as I hear back from them.

czajunia wrote:

czajunia wrote:

Thanks. Just sent Sofirn a message to their AE store. Will update the thread as soon as I hear back from them.

Received a reply from Qidong Yang: “Yes,the MOSFET problem has been solved already, the stock now has already improved”. So it looks like it should be safe to order from AE.

Thanks for checking on this. I wonder when it will be “safe” to order from Amazon to ensure getting the corrected version.

I just got some new FETs from MTN today and swapped it out.

New and old.

I gained 2A more at the tail and 350 more lumens @30s. There is a more instant jump to Turbo as well. I left the stock led wires.

Glowworm wrote:

Thanks for checking on this. I wonder when it will be “safe” to order from Amazon to ensure getting the corrected version.

Also they advised to check directly with Amazon for the latest updates.

Did anything ever come of the flat tail with no switch? I assume it never saw the light of say but worth mentioning cause the C8F 18650 was like this.

contactcr wrote:

Did anything ever come of the flat tail with no switch? I assume it never saw the light of say but worth mentioning cause the C8F 18650 was like this.

You can set your brightness then turn the light off at the tail cap (not the side switch). Then you can use the rear mounted forward clicky in a tactical manner. It’s pretty smart.

If you turn the light off with the side switch, the rear switch does nothing except connect main power. The led stays off. I like it.

JasonWW wrote:

I like it.

So what you are saying is you are going to convince ToyKeeper to write a special C8F firmware to mimic this momentary mode memory deal so we can have the best of both worlds after replacing the driver or building a host? Sounds good thanks for doing that.

I know that. Jason has talked about the idea of replacing his driver so I assume it would be Narsil/Anduril variant and this is the host thread after all. I bought 2 hosts and 0 stock lights hence the nudge from me.

contactcr wrote:

JasonWW wrote:

I like it.

So what you are saying is you are going to convince ToyKeeper to write a special C8F firmware to mimic this momentary mode memory deal so we can have the best of both worlds after replacing the driver or building a host? Sounds good thanks for doing that.

I don’t think NarsilM can remember it’s levels when the main power is removed. RampingIOS/D4 UI seems to remember. I have not tried Anduril yet.

I could see this being a handy feature for a light like the L6, S70 or SP70. I don’t know how difficult the coding would be, though. Maybe I will ask ToyKeeper what she thinks. It may or may not be possible.

With my L6 I had Lexel set up Narsil to go to Turbo when main power was applied. Sometimes this comes in handy, sometimes not. It would be great to be able to switch back and forth.

JasonWW wrote:

I just got some new FETs from MTN today and swapped it out.

New and old.

I gained 2A more at the tail and 350 more lumens @30s. There is a slightly more instant jump to Turbo as well. I left the stock led wires.

I’m not complaining though. I measure about 3,000 lumen with the stock battery and my lumen numbers are a bit lower than others. So it should easily do over 3,000 lumen with the new driver (I assume it will match my current driver with the FET I installed).

Maybe Sofirn could increase the led wire guage to get even more output.

LH351D 4000K 90CRI at turn on:

2,803 at 14.xx amps (Samsung 25S 18650)
2,881 at ?? amps (rested 30T 21700)

This is using very long 20 gauge leads. Like 6” long or something in case I needed to re-flash or test. Numbers may not be exact since this is approximating shoe box.

@JasonWW, do you think it’s possible Sofirn to have changed the LEDs from XP-L W2s to a lower bin, which is perhaps why the average lumen numbers are about the same or down, even with an upgraded FET?

BlueSwordM wrote:

@JasonWW, do you think it’s possible Sofirn to have changed the LEDs from XP-L W2s to a lower bin, which is perhaps why the average lumen numbers are about the same or down, even with an upgraded FET?

Yes, but the forward voltage of the XP-L W2s is lower on average than XP-L V6s, so they can be drawing more current than XP-L V6s. Therefore, the 18650 version, if they have W2s, can be quite a bit brighter as they are drawing quite a bit more current.

Heck, look at @contactr’s numbers. 2800 lumens with 4000k 90CRI LH351D with a 21700 stock C8F is very close to your numbers of the stock C8F with 5000k 70CRI XP-L HDs, which is really strange.

contactcr wrote:

LH351D 4000K 90CRI at turn on:

2,803 at 14.xx amps (Samsung 25S 18650)
2,881 at ?? amps (rested 30T 21700)

This is using very long 20 gauge leads. Like 6” long or something in case I needed to re-flash or test. Numbers may not be exact since this is approximating shoe box.

Does the LH351D they have similar output to the xpl? Meaning similar lumens at the same amperage.

Not even close.

If I remember well, with the new FET, you got 3000 lumens at turn on with a 21700 C8F with 9,5A of current.

He is getting 2800 lumens with 14A of current, and he is using VERY long 20AWG leads, meaning current would be quite a bit higher in a real world situation, or using thicker wire leads.

That means my hypothesis of using lower-binned higher voltage LEDs in the 21700 C8F seems more plausible.

Just to be clear the driver i’m using is a MTN-20DDm (20mm FET+1) and driver spring (small “blue” spring) is through hole bypassed. LEDs are first batch from AEDe.

BlueSwordM wrote:

Yes, but the forward voltage of the XP-L W2s is lower on average than XP-L V6s, so they can be drawing more current than XP-L V6s. Therefore, the 18650 version, if they have W2s, can be quite a bit brighter as they are drawing quite a bit more current.

Heck, look at @contactr’s numbers. 2800 lumens with 4000k 90CRI LH351D with a 21700 stock C8F is very close to your numbers of the stock C8F with 5000k 70CRI XP-L HDs, which is really strange.

My stock driver C8F 21700 was drawing about 6.3A and 2300 lumen from a 30Q battery. With the FET swap, it draws 8.2A and 2600 lumen.

From the 18650 version reviews I found it seems everyone was getting between 8 and 9 amps from a high drain 18650. So it doesn’t seem like there is any big change in amp draws between the 2 drivers when using 18650 cells. This would indicate the xpl emitters are the same, right?

I don’t think I would look at contactr’s numbers as it’s a different driver and different emitters.

Both XPL2 and LH351D have relatively low fwd voltage so it seems more likely something else not allowing it to pull >10A. Seems more likely to me that it would be the driver. The difference between my long 20AWG wires and stock wires that are 22AWG and shorter is only like 0.05V drop isn’t it? Seemingly not enough to make such a big impact.

@contactr, I’m talking about XP-Ls W2, not XP-L2s.

The XP-L2s are a completely different beast.

BlueSwordM wrote:

If I remember well, with the new FET, you got 3000 lumens at turn on with a 21700 C8F with 9,5A of current.

Maybe if they also used some heavier 18ga led wires we would see more than a rough 1A and 300 lumen increase in output?

BlueSwordM wrote:

He is getting 2800 lumens with 14A of current, and he is using VERY long 20AWG leads, meaning real current is probably quite a bit higher.

I don’t want anyone getting confused.

Ah, I’m already having trouble remembering stuff with this light. I think I’ll try some 18 gauge wires and see what happens.

Yes. It would’ve been better to underrate the light most of the time.

Even overrating slightly a light will disappoint customers if they discover it.

BlueSwordM wrote:

@contactr, I’m talking about XP-Ls W2, not XP-L2s.

The XP-L2s are a completely different beast.

You are right I should not comment on stock lights especially since they don’t exist at my house. I got confused because of various mods in this thread.

I think I figured out a small design flaw in the manufacturing that is causing reduced output.

It seems the new aluminum driver retaining ring is not able to fully tighten down on the driver (at least my driver). I thought it felt strange as it only tightened up to one point and then would go no further. I used some marker dots to confirm this. The ring seems to be squeezing the driver lightly instead of firmly.

The 2A difference I saw earlier with my battery tests and again a 1A difference after I swapped to 18 ga led wires. I can move the driver around when the retaining ring is about 1/16 of a turn from its “tightened” position.

So maybe the driver pcb is a bit thinner than it was supposed to be or maybe the threads are not cut far enough. Maybe they should have continued using a brass retaining ring. I’m not sure.

I will relay this information to Sofirn.