[Update]Wurkkos dual LEDs 21700 Headlamp design Photos/UI update~

Many of the Sofirn lights + Wurkkos FC11 use a CC linear driver AFAIK.

I would also prefer to pay more and have good regulation.

I’ll take CC linear running on a FET + sense resistor or similar over PWM’ed linear chips any day as long as there is a sub-lumen mode.

Once a cells voltage under load drops below the Vf of your chosen output it will drop anyway. Only a sufficiently low Vf emitter and buck driver, or boost/buck-boost will achieve what I see so many in this community call “regulation”.

If a flashlight has any electronics that control the flow of electricity in at least one mode, it is “regulated”.

Behaviour of fc11 driver
There is no stabilisation at all.
As I understand in Hd20 will be the same ugliness.

the factory engineer said the space of Program is too small so have to use a FET driver, i am not sure :person_facepalming:

I don’t really understand the benefit of defined steps to lower modes when cell voltage drops vs this. Either has the same end result of higher modes losing max output as the cell drains, its why I choose lights with boost drivers whenever possible… because that or multi-cell + buck are the only ways to avoid it.

Well that graph is a bit deceptive:

  1. the drain down to 2.65V is excessive. Most of our firmware triggers LVP at around 3.2 - 3.0V. Looks like this light has no LVP
  2. 3.4V down to 3.0V is a significant drop, > 50%. If you look at a normal drain from 4.2V down to 3.4V, the lumens drops in half. That's about right, and matches what many lights use as "Turbo" to "High". 3.4V to 3.0V is a small percentage of the cell range and happens quickly - refer to discharge curves
  3. lower modes are typically in the 1-500 range anyway, still quite relevant. Also lower modes could be using a low channel having better efficiency that the FET shown in the graph

This graph is typical of most of our FET based drivers running on only the full max FET.

When I turn on 100 lm mode I need 100lm during discharge, and then light should give me know that battery low without turn off.

I see no deceptiveness. I know that in real light will be hint of “stabilization” due to discharge curve of li-ion.
But yes it is classic ugly fet.

This chart does not show that the light will do that before the battery is extremely low. You’d need a runtime chart of that lower mode to see if it drops proportionally over a discharge cycle. In Maukka’s test Vf for 100lm from a 90CRI LH351D is only 2.7V.

Again, the only lights that can sustain brightness like that are those with boost drivers, or multiple cells (with combined voltage exceeding emitter VF) + a buck driver. As soon as the voltage at the emitter drops below the Vf required for a given output, that output will start to decrease. This happens with lights that have an extra 7135 channel for low modes as well since they also cannot step-up voltage.

What are some of the “real lights” that don’t have a drop in output like the graph you posted? What type of alternative driver would you propose for this light?

I think the contentious point is using either linear drivers or FET + sense resistor, vs just PWMing a FET.

The former will probably have constant output (smoother with no PWM, and constant current/lumens), vs just having a fraction of max output (FET-DDed LED crowbarred across the ever-decreasing cell voltage).

It’s switching the light on medium for 600lm across most of the cell’s SOC, vs 800…750…700…650… etc. lm as the cell drains.

Did I mansplain that about right? :laughing:

Oh, good point. I was under the impression that most of the single-cell, 3V emitter Sofirn and Wurkkos lights use CC, not PWM’ed FET so that’s what I’m assuming is planned for this light.

Wellp, that constant-downward-slope line AEDE referred to was the latter, ie, the bad one.

To me, that’s like 3×AAA lights that start out nice’n’bright, then start slumping almost in seconds, and just get dimmer and dimmer and dimmer ’til you change the batteries. And you have no idea how bright nor can guesstimate the current that way, either. So runtime is always a surprise, and not necessarily a good one.

Here I was convinced the driver of the FC11 was CC, I’ll be damned. Zak’s review confirms there is PWM on every mode except turbo. :open_mouth:

But still, regardless whether a PWM’ed FET is used or a CC linear, or 7135 chips… without excess voltage, output will decline with cell charge. There is literally no getting around this, and with high Vf emitters it is even more noticeable. The big difference is in efficiency and the presence of PWM.

Graph show driver behaviour only.
No.Voltage drop on 351D is ~2.7V. 18650 under 3V have ~5% of energy. So during 95% of battery energy output can be stable using buck or linear driver. After that 95% I can simply change battery.

Any light with li-ion battery. Just look at discharge curve.

Look at the S2+ Med runtime on this graph, as soon as it drops below required Vf for that mode output declines with voltage. This will be consistent in direct drive and linear vs boost or buck driver.

A Zebralight with boost driver has this result, due to voltage step-up with boost circuit. ^ That is what many reviewers and members on here simply consider “regulation”. There is a a huge difference.

Even the buck driver in the SP31v2 drops output once cell voltage is low. A cell at “X” voltage simply can’t supply more without a boost circuit.

As you said, since the Vf is so low with lower outputs there is little benefit to buck driver in a 3V, single cell light (other than losing voltage as heat earlier in the discharge curve). I would love to see more buck drivers since there’s little disadvantage, but for a budget light I’m never upset with a CC linear driver either.

100% agreed that PWM’ed FET has too many downsides other than a boost in max output with full cells :+1: I was unaware that was what was used in the FC11, and hope this light utilizes something else as well.

Ironically, that’s where linear shines(haha).

Especially with low Vf LEDs, you get stable consistent output for most of the cell’s SOC, and then at the tail end it’ll start going dimmer and allow you plenty of time to react.

And that’s even vs a boost circuit that’ll hold consistent output ’til a few seconds before it craps out completely as the cell winds down to zero.

Behaviour of work at low charge is determined by software in boost driver. It may slowly decrease output, may to step down, may work comined way.

It wasn’t the thermal regulation that was the issue on the C8G, there is a step down that occurs after 30 sec on any brightness above 900 lumens

I apologize for derailing the conversation with a little off topic question.

What would it take to have a multi LED headlamp with a RED LED (XP-E2 deep red) just like in the WK30?