Twice, when running it through the light box, I saw a brief 11 and a brief 13 on the meter before 25.9 or 26.2 showed. That would have been 3.79 lumens or 4.85 lumens, so there is a moon there but it won’t lock on it.
(11 is 3.79 lumens and 13 is 4.85 lumens, 25.9 is 8.94 lumens, which is the Low mode or mode 2)
And this is with the offtime firmware you say? Hmm.
DBCstm, when you wrote this post it seemed to me that you were saying a PWM level of “1” gave you no light at all. Based on posts #59 and #60 it seems like the mode is “missing / skipped” and you do not have a “black / no light” mode. Hopefully I make myself clear. It sounds a little sketchy to me as I try to describe the two scenarios ;-). Anyway can you clarify? Thanks.
This is the off-time firmware downloaded directly from JonnyC’s site. I added a mode and tweaked the levels of the modes. In this particular firmware, he didn’t list the “0” for a start position, but had the lowest at 13….and looking back at it I see I can’t read
#define VOLTAGE_MON // Comment out to disable
#define MEMORY // Comment out to disable
// Levels should start around 10 with Fast PWM
#define MODE_MOON 13 // Can comment out to remove mode, but should be set through soldering stars
#define MODE_LOW 20 // Can comment out to remove mode
#define MODE_MED 83 // Can comment out to remove mode
#define MODE_HIGH_W_TURBO 180 // MODE_HIGH value when turbo is enabled
#define MODE_HIGH 255 // Can comment out to remove mode
//#define MODE_TURBO 255 // Can comment out to remove mode
#define TURBO_TIMEOUT 180 // How many WTD ticks before before dropping down (.5 sec each)
#define ADC_LOW 128 // When do we start ramping
#define ADC_CRIT 118 // When do we shut the light off
He states levels should start around 10, 1 doesn’t function but 4 does…in this light. This is a twisty, no clicky and no e-switch. Apparently the twisty mechanism is why the moon mode is skipped, with a setting of 1 it’s too low for the twisty type action… it’s not positive enough and allows a mode change.
I changed the levels to 1, 4, 19, 69, 117, 255.
I added a level, called it MEDH for a total of 6 levels, but I only see 5 with the “Low” being the lowest I have access to.
// Levels should start around 10 with Fast PWM
#define MODE_MOON 1 // Can comment out to remove mode, but should be set through soldering stars
#define MODE_LOW 4 // Can comment out to remove mode
#define MODE_MED 19 // Can comment out to remove mode
#define MODE_MEDH 69 // Can comment out to remove mode
#define MODE_HIGH_W_TURBO 117 // MODE_HIGH value when turbo is enabled
#define MODE_HIGH 255 // Can comment out to remove mode
//#define MODE_TURBO 255 // Can comment out to remove mode
#define TURBO_TIMEOUT 360 // How many WTD ticks before before dropping down (.5 sec each)
#define ADC_LOW 128 // When do we start ramping
#define ADC_CRIT 118 // When do we shut the light off
Despite JonnyC’s comments in the code, AFAIK starting level is arbitrary and depends on the hardware. Let us think on it a bit and maybe we can realize why that mode is being skipped.
If you want to blindly try things, you could try playing with:
Right now I’m really not confident that changing these things will help however. As I said, we need to consider where the root of the problem is and why it only affects the extremely low mode.
As any programmer knows… Variables won’t and constants aren’t…
wight, this one’s tiny, one of a kind, been worked on too many times already and the guy that made it is no longer making lights. Gotta be careful!
Beside the comfy modded Convoy S6 triple…
That’s the first BLFTiny10 in the light, with Arctic Alumina Thermal Adhesive potting the 2502 fet to support the copper battery contact. This board is thin, the new ones are much thicker which is what I have in the light now. The newest, V 1.0 with off-time. The thicker board makes it very difficult to fit it into this small twisty.
The fab place we got the thin boards from will do small runs but the cost per board goes up. Partly due to the boards themselves and partly from spreading the shipping cost over fewer pieces. It could still be done though.
I may have a problem, is SOT-23 different from SOT23 (no dash)?
I ordered a new FET that I thought would fit these, it’s spec’d as SOT23, but its about 2x as deep as the old IRLM2502’s thats listed as “SOT-23” and dont really fit the pad’s.
Might be able to get several BLFTiny10’s if paneziled from dirty PCBs (they won’t be OSHPark purple perfection though 
)
10 each 5x5cm PCBs $14 • 10 each 10x10cm PCBs $25
Same thing. Check the dimensions in the respective datasheets?
Yea I could do that, or I could let the digikey customer support lady do it for me lol, they sent the right part the thing is the IRLM2502 is a SOT-23 “micro3”, the AO3400A is a standard SOT-23, thats bad news for my RGBW CC drivers, the larger AO3400’s arnt going to fit, luckily they wouldnt handle over about 3A (per channel) anyway so I dont need to use the AO 3400’s on them, the old 2502’s are more than enough, I can stick with the 2502’s and use the 3400’s on single channel boards with more space, the leg’s have to be bent slightly but they fit the pads still.
Poor lady! It’s impressive to me that Digikey’s CS is knowledgeable enough to check a datasheet and point that out.
When your choosing who to live chat with it gives you three choices, CS, TS or sales, I chose TS. She said had I needed to return something she would of had to transfer me to CS and had I wanted to buy something new she would of sent me to sales so it seems like they actually employ TS staff with (get this) a technical background. Way to go Digikey!
Where is the damn like button 
Are you all using that IRMLM 2502 SOT-23 fet from the OP?
I just perused the datasheet and is shows the RDSon to be 0,035 ohm typical @ 4,2V and the gate charge total Qg is 8,0 typical.
Digikey has a SOT-23 N-ch FET called PMW16UN with a typ RDSon of 0,015 @ 4,2 V with a higher amp rating of 5,8 ohm.
Just thinking that if you buy from them anyways is there any reason not to get that one?? They are like 22 cents a piece if you buy 10.
Did i miss something (again :-D )
I’m using the alpha & omega AO3400A. It’s slightly larger than the 2505 (actually it’s standard SOT-23, the 2502 is smaller than typical) but in a roomie application like the tiny 10’s that’s not a problem.
I’m seeing ~5.5A tail current with them on my 15.17’s (one of my most used drivers).
Maybe some of these options can be located in thus threads op.
So all of the required parts (except for the FET) can be pulled from a 105c correct? Anybody have a link for the correct fet?