I noticed today that the light sometimes flicker in Turbo after a few seconds then a few seconds later it stops again
I looked now again close on the driver and its used parts
The driver uses a pretty bad MOSFET compared to new Infineon and Vishay types
and the driver has no Output capacitor to smooth the PWM from the Buck regulator
So I did 3 improvements replacing 2 parts that could cause the flicker investing 2€ of parts
- first remove the old MOSFET and solder one of the best available ones there
- second remove the Diode and put in a 10A SMD type
- third add a 4.7uF 07R capacitor parallel to the main LED
I have also found some inconsistencies in the UI
for example if I switch the light on with an auxiliary button then click the main switch the light turns off
while I turn the light on with main button switch to aux buttons then main switch gets me back to memorized modes
Another odd thing is this
I have a medium mode memorized for main switch
I turn on the light with Turbo on auxiliary button and turn it off with the same auxiliary or the main button,
then the light comes on in moon the next time I switch it on with main switch
after this I turn the light on with a blinky mode on the other auxiliary
now I switch it off with that button or main button
and now switching on the main switch recognizes again the stored medium mode
Next thing
I switch the light on with blink mode on aux1 or switch it on with main button then switch to aux1
then switch to turbo on the 2. aux button
if I click now the main switch I do not get the memorized light level instead I get again the blinky mode of aux1 button
The modded MF02 does roughly 15-20% more Lumens and is driven a bit harder, so the hotspot is overall bigger with more light in it,
but still the Niwalker has a higher peak intensity in the smaller hotspot
The light can be locked out by a full turn, then the carrier does not touch the tail boards ground ring
there are no solder bridges or stuff on the boards of the carrier, the traces are fully covered with solder mask
no risk of a short here
also the Carrier is wired 4S, so reversing a battery wont do a short
Playing more with the light I notice at some points a whining sound from the driver, I am pretty sure the unshielded inductor is causing it
So I switched out the Inductor with a 10uF shielded one
just in case I added another 4.7uF cap on the input in case the used inductor was smaller value
Make a sabot and launch it from a Leopard 2 tank, that should ensure high speed impact testing, and spectacular slow motion footage, cuz i want to see the sabot peel away to reveal the flashlight
added 3 pads for the auxillary switches,
one providing 5V for the Turbo switch
one sets multiples the current by 10 or goes to Turbo
the other divides the current by 10
the throw did almost not increase so I assume they use a CW E2 or very good D4 flux bin
the bezel is not glued, not sure if this will be on production lights as well, as they wrote on my suggestion to thermal epoxy the head and middle piece together that in Production it will be glued on that spot
The reflector is deep
nice square cur threads on bezel
Thermal paste is an uniform thin layer with little air pockets, that pattern comes from removing the board and prooves thats a thin layer
Back of the MCPCB is perfectly flat
Also the heads shelf is pretty flat with a good surface finish looks roughter than it is
reflowing NW LED
using thermal epoxy to bond head and middle for better heat transfer
I see light on the end of the tunnel, of course not so good that actual BL frmwares are single switch, but I can use em to momentary stup up or step down current manipulating the MCU output channels