hello everyone
such a nice forum… and so big I couldnt read all, Im sorry if I missed something
just received my led flashlight and several mods have been done in the first 5 hours
and more to go…. so here is what I have done and what is still at stage idea…
I bought this from ebay image
look nice but badly made with fake cree LED and poor driver (if I put proper quality cell fully charged it will burn within 10 minutes, full power current was 1.5A(4.2V) and 1A (3.8V))
the LED holder was hollow, the poor guys used a tube with inner diameter 19mm and 20mm star board just catch the edges, with almost to heat exchange
so I made one cylinder piece of ally 19mm diameter and 10mm thick, and pushed it inside very tightly, and then turned the face on lathe to make good flat surface for the LED star board
so far so good… just at this moment I noticed that the led is not that large as XML (it was advertised as XM-L2) but it is not even XML it is the fake one from latticebright ver.2.0…. image
I had one genuine XML so I replaced it… up to 1A I didnt noticed any difference in the light, the only difference found is in current handling, latticebraight was heating too much with higher current
later found that the beam is not used properly in these zoom flashlights, when zoomed to make a spot, only small part of the beam is used… then I found one suitable glass lens in my junk box, 13mm diameter and 7-8mm thick and I stuck it with glue just above the LED, now the beam is much narrow, from 120 degrees now become like 70-80, where much more is used when zooming… and much brighter now
later found about the wavien collar…. I have to try it to see if any better
next step is to purchase XP-L led, is there any brighter? I mean for same current/power 3A/10W
also looking for suitable driver board, but cant find any proper driver
linear - that is fine, I think linear is best option for single battery-single LED flashlight
but I dont understand some ideas… why PWM with linear regulator?!? just no sense… the PWM completely ruin the current regulation… mid and low are just not working as constant current
I have an idea and if someone can comment will be nice
Im reading that the best modes are 100–30–5%
so lets say we use 3A LED that mean 3A-0.9A-0.15A
why we have to switch all regulators with high frequency, when we can turn them off one by one
for example 3 regulators
low - 1pc 7136 chip where the current is set by resistor to 150mA
med - 2pcs 7135 + 1pc 7136 so we got 150mA (low) + 2*350mA + 50mA = 900mA
high - 6pcs 7135 = 6*350 = 2100mA + low + med = 3A
if these calculations are wrong Im sure with the same number of chips it is possible to achieve any combination
total 10 chips… a bit too much but I think they should fit… 6 one side and 4 next to the microcontroller
and also it is pretty flexible, with the number of 7135s and resistors on 7136s you can adjust any other levels for smaller leds
how about this idea?
thanks
Yup, you have a bit of reading to do. The good news is it sounds like you already get most of it so it’s just a matter of coming up to speed. The current popular driver choice is some version of x7135 + 1-7135 or FET + 1-7135 to either dial in a specific high current and use the single 7135 for moon/low modes or an FET to get maximum current on one channel with the same control of lower modes on the other. Lots to choose from either as prebuilt drivers or as Oshpark project boards. Some better buck drivers are around now too as well as the LD-2 true current controlled linear driver.
if we talk about the normal working conditions of up to 3 amps and considering the single cell voltage range and the LED voltage range I think linear regulator will be the most efficient and simple, not worth any switching… therefore I decided to go with linear
LD-2 is really nice
If cant find any similar cheap on ebay probably Im going to build one…
not sure what these are… not much info provided
hello
true, RMM got fantastic store
unfortunately the linear drivers in stock are same as I found on ebay 7135 + PWM
now reading the Oshpark Projects but it look like Im going to build my own driver
but just found another problem - the tailcap… at 3A current only the tailcap itself have 0.6V drop and the spring inside the yellow pin is heating, a bit hard to open because it is not a screw, it was white plastic pushed inside, but finally got it opened. the switch itself have 0.2V drop
it look like I have to make some improvements here….
The moonlight special differs from most 7135 drivers in that it uses a separate channel to control one of the 7135’s for a more reliable and consistent moon mode. Afaik there aren’t any multimode 7135 drivers with no pwm. For that you have to go buck or LD-2. Some members do eagle pcb design if you’d care to do one of those.
the trouble with 7135 is that the PWM is applied on the power supply pin, where for some reason 7135 cannot handle PWM modulation with high frequency and the chip is not current regulator anymore
recently found 7140 chip which have “output enable” pin, it look like very fast as they provide info about the time delay in their datasheet - 3uS
unfortunately 7140 is high voltage and high dropout, not suitable for single cell
I have to purchase some 7136, I wish to check how they handle high frequency PWM applied as additional voltage on their CS input, it could be better than 7135
I couldnt find another low drop low voltage chip with enable input…
just finished the repairs on the tailcap - fixed the switch and doubled the spring with thick wire, now the total drop on the tailcap is about 50mV for 3 Amps
Im going to order from ebay another switch the one which was tested here by another member, 6A switch from ebay… I hope it will fit, their drawing dont provide all measurements required
also did put some wire on the driver spring, it was another 0.3V drop there with 3 Amps … now I see why the current was always low with battery
As always it would be great to see improvements in the drivers we use. Most of us are hacks making do with what we have and abusing the few with know how for upgrades.
just received HP-l HI LED V2 4500K with copper sinkpad
but for some reason the sinkpad dont have the punched rectangle on the back
is it fake?
tried to look under the LED with magnifier, it look like instead of raised copper there is only solder…
anyone noticed these?
it say
luxeon Q
seoul Z5M
Nichia N219
XP I XTE I XPL
1901C
Sinkpad II
where the S is just on the edge, may be slightly chopped
the LED look like genuine but when focused the white square is not same color everywhere, there is some kind of yellowish clouds
and also with same current (1-1.2 amps) it is heating same as the old LED with ally base… I cant notice any difference with the copper base, should the LED be noticeable colder when touched?
thanks
