Oshpark Projects

I know y’all are all over this FET driver stuff, and really keen on the big hoss lights. So you should get as much of a kick or exasperated reaction to this as I did. My Dad’s brother is 79, a very practical and successful businessman. Also a perfectionist. He came over today to watch some college basketball with my Dad, 85.

My uncle drove up as the mailman was delivering my new HD2010 from O-L’s group buy. I had built an BLF FET 17mm DD driver for this yesterday, and assembled the copper pill with XM-L2 U2 1D emitter so that it was ready when the light got here. I put it together, dropped in a 26650 and fired it up. Took it in and showed it to my Dad and Uncle. J.D. was impressed, very bright (it was doing about 1200 lumens on that particular cell, capable of 1700)

Then he said this “Who would want a light like that? I mean, why do you make it so bright?”

After I laughed I told him it depends on where you asked, in the right places…Everyone!

Then he asked how long the battery would last. I explained the ratio of amp draw to battery capacity and that it has 4 levels. He very quickly grasped that “you wouldn’t need to use it that high for long at a time, the next mode down would suffice most of the time” (he said this in a sort of muted, thinking about it, tone) So, in a nutshell, even a novice can see the need for modes. lol

My Dad thinks my lights are great, but can’t understand why I keep building them for ME and not selling them, lol, what good are they if you’re not making anything on em? Perhaps, after about 25 lights, he’s got a point…. (I remember growing up out here in the country, with 43 cars around…yes, Dad had 43 cars!) Now if I could figure out why I have over 50 flashlights, I could explain that to my wife and maybe get off that hook…

Nice anecdote. Now get back to work. :slight_smile:

I picked up some of the irlm 2502 mosfets that Werner had mentioned in the development of a driver he is working on and since the larger 70N02 won’t fit on the 15 mm version Mattaus designed the BLFDD15 V1.0-2.0 to use this chip instead. Up until now I don’t think anyone has had a chance to try them out though I did send a few to Dbcstm and Warhawk-AVG. The larger boards have taken precedence in development so far. However, this evening I was finally able to get a bit of time to set up a test board. It is a 105C that some of the drive chips were not working on so I cleared them off and cleaned it up.

I didn’t have to mod the board other than to scrape a bit of the solder mask off the output trace for the drain pin of the FET. It just barely straddles the output trace with the gate and ground pins soldered accordingly.

You might have noticed a short from + to ground in that pic, I didn’t at first so I was a bit disappointed when it didn’t work. Once fixed I tested it with a 10180 and got 33, 235, and 510mA with reliable mode switching. Next up was an Efest 10440. With that cell the levels were 41, 410, and 1450 mA also with reliable mode switching. This is without any extra components other than stock 105C parts. I then tried it with a protected 16340 ultra fire and another cheap, protected 14500 cell neither which were close to the Efest 10440. One topped out at 1A and the other at only 700mA. Lastly was a protected trustfire flame that topped 1.8A but flashed and dimmed in the middle mode but low and high seemed ok.

This was by no means a definitive test and I didn’t track cell voltage other than to verify each one tested was around 4V but it does seem to indicate promise for the 15mm board and possibly even a Tiny10 version where no room is available for a stack of 7135’s. Rather than order the new boards just yet I’d like to try this with the extra R’s on the gate pin. I’ll find out then how much it will handle from an Efest 14650 or unprotected 18650 Mikita pull.

Thanks for sharing Scott!

Someone asked if the 15mm board had been tested yet and I had to say no. Figured it was time to get on this. The rest of you(Comfychair mostly) have been doing a ton of work on this project and I’m just doing what I can when I can.

My BLF17DD boards have been working great so far, both the first version without the ground ring on spring side and the second version. I already ordered some of the "2.0" boards with inline resistor pads. I have been running mine with a 200 ohm resistor in between the gate and MCU, also tried a 50 and 100 ohm which worked fine but had a ton of the 200 ohm already for zener mods. I think that 100 ohms is more in line with what may be ideal although 200 ohms if it works shouldn't hurt anything.

I haven't been using any pulldown resistors yet, are they really needed since we switch the power off on these at the same time we switch the PWM off?

Thanks for mentioning that. I used one of the extras from a Zener kit and it stabilized the modes nicely. The lower modes are too high and star 4 is all ahoo but default, blimkies, and low high work fine. Star 4 does 700, 800, 900, and 1500mA in the 4 modes with the 10440. Obviously the stock firmware is a bit inadequate for this use but the main thing was to confirm the operation of the FET and it certainly looks suitable for small lights.

[Quote=Mattaus]

Just one LED in the center of a 16mm PCB?

[/Quote]


We think it's perfect to put out torches by shooting. 16mm I think goes well!

XB-H

is a package 2.45x2.45 with 1.5Ampere from datasheet (XP-E 3.45x3.45)

Even XQ-E is interesting, maybe twists EDC

is 78% smaller than the XP-E! 1.6x1.6 datasheet from 1Ampere

I'm curious to try them both!

Thanks Mattaus!

CREE XB-H 16mm star

This is what you wanted, correct?

Perfect Helios! thank you very much, I believe that many users can make use of it!

a question, to improve the dissipation of the LED, it is not possible to add copper on top, connected to the base dissipating LED?

I can not figure out where mistake with eagle! work in the "board" but I do not get the parts "solder" as I would like!

Thank you!

I was thinking about adding copper on top but the via holes, in a way, cut off the heat path on top, directing the heat to the bottom instead.

Sorry, I don’t think I understand. If its a problem with exposed / uncovered copper, that is the “tStop” layer on top and “bStop” on bottom. Since eagle is so popular there are some nice tutorials around, videos too.

Now that I know which direction to use that little FET, I aim to attempt an application for my TP. I’d really like to be able to keep everything on the inside of the board if at all possible, this is a twisty after all with very limited space as everything was designed to fit together with little room for error.

That 1.5A on the Efest 10440 sounds VERY promising, exactly what I’d intended from the get go with this light! :slight_smile: (here’s hoping I can do this without screwing anything up!)

CREE XB-H 16mm with copper on top
Shouldn’t really provide much benefit if the bottom is attached to a heatsink / pill.

Should it have the little over flow channels on the actual led pads? Makes the pads look like “T’s”.

:ghost:
You can do 1.5A regulated inside 10mm, non-stacked. You just keep telling them to squeeze in a wee bit more. Promise them sweet copper for dessert if they keep the complaints to a minimum.

Yea, I noticed it was missing. You’ll see the actual oshpark link goes to a slightly different version then that picture was from.

The pulldown resistor is there to drain away any floating charge on the gate pin when the MCU is switching the PWM output on/off/on/off to get the less-than-100% modes. I thought someone had mentioned a while back that the tiny13 has internal puldown resistors to handle that and an external resistor isn't needed, but adding one definitely does have an effect, at least on the boards that have issues with mode switching.

I still don't understand why the 105C components+FET work fine without the resistors when on the Nanjg PCB, but don't work without them on the 17DD PCB. (that's when using the phase correct PWM only, adding resistors to the Nanjg might enable use of fast PWM on that one like it does on the 17DD)

Oh, I didn't post this earlier but I did test an original 105C diode in place of the Digikey ones, and it had no effect. So none of this is caused by incorrect diodes.

Ok, I’ll post it in the op after the triple XML.

Hey Dale, if you can, try it on one of the Oshpark Tiny10’s. If you cut the 7135 ground pad between the tab and the center pin you can solder the gate pin to the now isolated pad and use a 100 ohm resistor to jumper that back to the original pad.

Edit. Until it’s repeated, it didn’t happen or can’t be said to be reliable so let’s see if we can’t make it so.

Comfychair, can you update the shopping cart with the two extra R’s and I’ll swap them into the op?

Which value on the gate resistor? I tested a range of them and it's still not completely reliable even at 78 ohms, which makes me think 100 ohms may be cutting it a little close, I used a 130 on mine. Going too high matters too, more resistance there makes it less sensitive at the very bottom of the PWM values ('1' no longer works with even a 100-ohm at the gate, has to be '2' or above to light up).