Many thanks for the in-depth science. I’m definitely one of those OCD fools drooling over perfectly flat run-time data. And I certainly don’t care for FET drives for that reason. But we also know that the excellence of the M43 is definitely not a result of an accident or luck; everything is well thought off. It is what it is, is often said here on this forum, and it certainly is applicable here as well.
I just have so much confidence in Hank that I’m convinced the Emisar models are going to please me a lot. Of course time will tell. :partying_face:
(Of course Hank needs to sell the lights to me first instead of giving everything to Richard, hehe )
That’s true.
The generated heat on medium modes (or rather high Amperes combined with big voltage difference between battery and LED, which can also occur on high mode) being the risky thing…
You’re muddling cause and effect here.
Ramping isn’t something particularly desirable by itself, it’s only a crutch to make unregulated FET drivers at least somewhat comfortable for ordinary use when you need stable output for a duration of time, not a quick wow-flash.
Constant current drivers don’t need ramping because they already can maintain their output level without any user intervention whatsoever. While direct-drive FET drivers are practically useless without it.
Same for me.
Especially when you’re forced to un-optimize the electrical path to prevent too high currents with a freshly charged cell, then i don’t like the idea at all.
Maybe in the future there will be high capacity LiFePo cells and lower Vf LEDs to make perfect combinations, since LiFePo have a near flat discharge curve.
So, you wanna say, that linear driver has better heat performance and efficiency on high power rather than boost? FET+1 is less addicted to overheat, so why? It’s much more important how the driver thermal interface developed is. Everything depends on developer. How about stabilization?
Well M43 driver is big because it’s calculated this way for required output power but there are other compact boost drivers. You can smoothly adjust the mode in M43. Have no idea what you meant about M43 ramping, really.
@Reyden
Do you mean a boost driver for 2S2p LED configuration (for this D4 light)?
It will need a big inductor and diode, i think that will use a lot of space…
Actually a quick WOW-flash is what I want from this light. I used to have zebralights, armytek and eagtac lights all with constant current, but for my EDC I chose to go with a modded jetbeam mk II with a FET+1 and XP-L HI, because of the wow flash. For any other “serious” situations I bring a larger light that can handle the heat at a regulated output. Can’t speak for the others though.
I thought ramping is more of a desirable feature rather than a needed one to maintain output? It’s not like the user will have to ramp up manually to regulate the output.
…
Me too consider a 4S boost driver would be ideal for this light, but it’s a bit unfair for the FET+1 to be receiving all the bashing.
Ramping with an e-switch…
Hmm… Always ramping too far up or down, or not far enough, or you need slow ramping maybe…
We’ll see, i think i’ll buy a grey 18650 D4
No, I suppose the best way is 4S for this flashlight. Yeah inductor takes a lot of space but compromise can be found. But seems it’s too late 2S2P also accepteble. Other question what is desirable output power…
Who said that ramping is “must have” function? Reasonable modes that’s what important.
FET+1 for marketing, profit and wow effect, nothing to add. 18-20W is more than enough for such flashlight .
Potential buyer it’s not about me so have no idea what im doing here
It seems “ramping” is not clear enough to describe the concept. It is unrelated to heat or power regulation. It is an interface style.
The Meteor has 9 brightness levels (plus short transitions between each):
4mA (UI1 1A, UI2 1A, UI3)
60 (UI1 1B, UI2 1B, UI3)
500 (UI2 2A, UI3)
750 (UI1 2A, UI3)
1000 (UI2 2B, UI3)
2000 (UI2 3A, UI3)
3000 (UI1 2B, UI3)
4000 (UI2 3B, UI3)
8000 (Turbo, UI3?)
It doesn’t offer 20mA or 200mA, for example. Those are between the stair steps.
Narsil (as used in the BLF Q8) has 150 brightness levels, a smooth visually-linear progression from about 2mA all the way up to ~20,000mA. If I understand correctly, the D4 uses the same smooth ramp as Narsil. It is similar to how the SWM V11R works, or the Jetbeam RRT01, except it uses an e-switch instead of a magnetic ring.
This makes modes and mode groups irrelevant. There are no stair steps to arrange into groups. Instead, simply hold the button until it is bright enough. It can be set to the exact level desired.
Mode groups can be implemented for those who prefer it, and Narsil includes a UI which uses mode groups. But the smooth ramping UI is simpler and people seem to prefer it so far.
This is how the Meteor’s levels compare to the Q8 and D4, in graph form:
No. You or anyone else still not able to answer my question.
Your graphs are shown as lines. For q8 and d4 they should be extended and drawn as big area (and you are not able to say in which point will it work actualy). Thats why q8 will be hard to sell over $40 price - if seller will claim some output and runtime numbers, lots buyers wont see same numbers in own tests and some of them will open cases and ask refunds. Like led lamps from ali - you never now how much power they will consump.
The graph above could be viewed as a slice of a larger 3-dimensional surface. If a time dimension was added depth-wise through the screen, it could be used to show runtime graphs.
For example, take one X value on the levels graph above. Let’s call it 60% max brightness. If it extended back through the screen, it might look something like this graph selfbuilt made: (green=Meteor, red=Q8, purple=D4)
This isn’t the best representation since it’s for a completely different set of lights, but I hope the idea is still clear: The Meteor has better output regulation on medium and high modes, so it looks almost flat. The Q8 and D4 have only partial regulation, so the output follows a gradual curve over time.
This depends on the output level though:
On low modes, there isn’t much difference because both methods are regulated on low.
On turbo, there isn’t much difference since they’re all pretty much direct-drive.
On medium and high modes, the Meteor’s flat regulation is significantly different than the FET+1 driver’s partial regulation. However, on a small light like the D4 these modes are generally limited by temperature, so the output depends on thermal regulation rather than power regulation.
This regulation difference is something people can’t easily see during use. If output slowly drops 20% over a period of 2 hours, it’s virtually impossible for a human to notice without a lux meter. However, the interface is something people notice right away, so it might be a good idea to prioritize interface over regulation if a tradeoff is required. The SWM V11R was a hugely popular light because it had such a nice smooth ramping interface, even though it was notoriously bad at regulation.
Anyway, nobody knows the details yet, so instead of saying rude things about it based on speculation, let’s wait until reviewers have a chance to try it.
I don’t know all the details and it’s not my story to tell but, from what I’ve heard, this quote might be more true and more unfortunate than Serp intended. A big part of why I make everything open-source is to make sure nothing has to depend on one developer.
One thing is certain though — Intl-Outdoor is not Astrolux. Hank doesn’t sell badly-designed junk like the S42.
2 hours? 20% drop? Are you sure?
Several messaged ago you was not happy with steps in m43 ui. Now you understand that smaller steps are not reasonable by eye.
And no, you cant show all factors in 3-dimensional graph because cell internal R and discharge curve are not same. Ramping fet is smth very strange. You are trying to manage some parameter that you havent calculated and dont know exactly its value right now.
P.S. Im not against cheap custom drivers, Im not going to share links to foreign forums and make adv. for developers that wants $30 for one driver. I just dont like when programmers making hardware developments (and vice versa).
P.P.S. Cell voltage and R can be tested each time you turn switch on and some standart discharge curved can be used to make lower modes (50% and less) more stable (compensate cell discharge).
While I definitely prefer boost drivers with stable output to FET drivers, I have never seen the combination of:
Good efficiency
High maximum current
Ability to fit in a pocketable light (let’s call that about 20mm driver diameter)
Low price
It’s usually 2/4. Zebralights, for example are small and efficient, but they’re not hot-rods by BLF standards and they’re pretty expensive. The Haikelite MT01 using the off the shelf LD-35 boost driver is cheap and efficient but doesn’t come anywhere near fitting in my pocket, nor is it capable of a lot of output for a 3x18650 light. The Meteor might arguably hit three points: it’s a good value for what you get, but it’s not cheap in absolute terms.
A compact triple with a high-current boost driver would be awesome. Wait… that exists, but you’ll pay through the nose for anything from Lux-RC. The D4 costs $35 and I’m fine with a FET driver at that price point. If anybody would like to bring the Indigo driver to market outside Russia and can hit a competitive price point, I’m sure BLFers would be interested. I, for one need more Nichia 144As in my life.
I personally don´t care about the driver type in the new D4 - I want the most lumen out of it in turbo mode (hopefully nearly 4400 lumen in the XP-L HI version :smiling_imp: ); batteries can be recharged!
)
