Pre production sample [Review] Astrolux MF-01 4x18650 flashlight supplied by Banggood

Thanks for the review Lexel. Looks like a nice light once the kinks are worked out.

Quick learners. Definitely a good sign for Astrolux/Banggood. :+1:

@Lexel
Please tell Banggood their lamps would gain a lot if they use atiny chips or tell how to flash each light.
Some Astrolux lights could be brilliant lights with the right software. They should enable the flashaholics to fix their errors easy.

2S battery config with buck driver feeding 2S emitters? 2S li-ion operative voltage window (8.x down to 5.4V in this case) overlaps with 2S emitters driving voltage, which means you lose regulation early at some point. Stupid decision imho.

Dealing with certain limitations can be understandable with cheap flashlights in the name of cost. This is not the case.

Good for ANSI lumen rating, and good for ANSI runtime rating. So yeah, I want to know what output you get at 4.2V, 3.8V and let’s say 3.5V.

P.S. It’s not just 2S emitters, but those 2S are in 9P, so you have a lot more voltage sag.

hIKARInoob, I clearly see the torch uses a 2S2P high drain battery feeding 3 buck engines, each one driving a 2S3P emitter stack. The limitation I am referring to is the same you can find on any other 2S torch driving 2S emitters (like a Convoy L6): once the battery voltage drops below Vf + driver overhead, you lose regulation. This means turbo won't fully work once the first low battery warning has appeared, or thereabouts (a 30Q loaded with 5A outputs below 3.4V once you're in the 2 to 2.8Ah of its discharge capacity curve; also bear in mind carrier and other contact losses).

To me, regulation matters. ¿2S2P battery? Go fully regulated buck with a beefier engine and 1S18P emitter stack. Also, some efficient boost engines available now: 1S4P battery feeding powerful boost engine driving 2S9P emitter stack. For examples, of course.

^

I think we’re on the same page here. I think I fully understand the limitation you are referring to, and I’m definitely not a fan of “you only get that output with fresh 4.2V cells” either.

UI is awful.
Modes :
100, 600, 5100 ,12000.
Power consumption:
0.125W , 6.4W , 57W, 120W

LPW:
800, <100,<100, <100
First obviously is mistake. But if the other are right that is shame. Newest led and constant current driver and lpw < 100! WTF?!
To get 5100lm from 18 xpg3 s5 it needs about 30W not almost 60.

This is a prototype with Nichia 219, so who knows if they drive the XPG3 with a little less current or this is not the final driver

it is possible the light does more than 12k with XPG3, this Nichia one is really bright

Its likely they do not use highest flux bin to keep the costs low

and the TIR plus AR glasss will have like 15% loss
so your 5100 lumens need 36W @80°C Tj

driver efficiency 85%
42.4W

S3 flux bin
47.85W

Price differense CW S5/ CW S3 is very small. They can not much reduce cost this way.
So looking forward test in sphere. Maybe they change UI meanwhile.

I’m thinking about using this for wintertime mountain biking, the Nichia’s should look nice in snow. It might even work as a hand warmer. With four cells it should be able to run for a couple hours at a high enough level to keep the batteries warm. Do you think the light will handle the vibration or will parts of the battery carrier or driver start shaking loose?

Biking on snow is usually pretty slow but I might need more throw. How are the Nichia’s for dedoming?

The mode spacing could be better, probably want 1400 to 2200 lumens depending on beam width and reflective glare from the snow. How do buck drivers respond to doubling the resistance of the sense resistor? Is it a linear change, mode spacing get wacky or does the driver just quit working?

do you have any output lumens VS time chart ?
I can’t see the picture of the post.

Hi,

I cant measure the real output, but I estimate about 10000 lumens

The light steps down after 2 minutes when cold, when really hot its about 1 minute.

But when the light steps down first time from being cold its on the edge to be held comfortably on the tube and way too hot to touch the head,
you would not do another turbo unless wearing gloves.

I tested the light in ice water and it seems it the timed stepdown is always there just shorter when the light is hot.

So if I understand this right, the cells are 2S2P and the LED's are 3 groups of 2S3P. Therefore, you can run the 3 groups in parallel and have a fully 2S LED setup with 2S cells, therefore a FET based driver can be used. I think that's right??

yeah you can run it DD, but it gets hot quick as it is now

with DD you may get like 25A

Yea, awesome. Just wonder if the std SRK size boards will fit. Might not have clearance for piggybacking in there? From your pics, looks pretty tight. Not sure if a SRK style driver would fit or be compatible as-is.

Beam shots are awesome! That thing is a search light. I’ll definitely be getting one!

I don’t know if they are comparable but would you describe this as Q8 killer?

For $20 it would be a Q8 killer . I wouldn't bury the Q8... yet. Without knowing lumens, throw and how the UI works, hard to compare. Only 4 output levels compared to smooth ramping plus multi mode sets, etc... I can also still pocket a Q8 but with the wide head on this, I don't think I could.

Plus the Q8 is a fully open source design.

Hi

I measured the driver with 47.5mm like the Q8 one

give me some time and I make one for you