What firmwares have you tried? If you are using programs that use dual-pwm (like most of the newer FW’s), the lower modes won’t light because the nanjg just has a single pwm channel.
I tried a few different ones and then tried to turn off the dual PWM. I tried the Dr Jones Mini too thinking that one would at least get me into the lower range.
Any suggestions for an older firmware that might work for me? Doesn’t have to be exact I can play around with it.
I would try one of the old versions of STAR that never had dual-pwm just to be sure that wasn’t it.
Can you clarify exactly how it’s behaving?
Ok I’ll give that a shot.
What it’s doing is just ignoring the low modes. So when I put current on, there is Nothing, Nothing, Medium, High. So if it’s four modes I have to click past the first two modes because there is no light in the first two modes.
Assuming you don’t have those modes set to a pwm value under ~5, it certainly sounds dual-pwm related.
Yes. If you can run python scripts, battcheck.py can calculate the ADC values for you. It just needs an input file with two measurements, one for a full (ish) battery and one for an empty (ish) battery. It will then calculate other values based on a linear model.
For example:
(~/src/torches/trunk/ToyKeeper/battcheck/)-]> ./battcheck.py readings/tk-s7.volts
Line: 179 - 4.09V
Line: 136 - 3.15V
#define ADC_44 193
#define ADC_43 189
#define ADC_42 184
...
#define ADC_22 93
#define ADC_21 88
#define ADC_20 83
I’m also adding this info to a README file in battcheck/ .
I did go in and set the pwm higher and I still had the same problem. Toykeeper suggested that in a PM.
Whats a good single pwm firmware?
I still can’t get over the idea that I’m just missing something really simple.
What pilotdog68 said.
Also, one project you might find works well on a nanjg driver is STAR-noinit. It gives off-time memory without needing an off-time capacitor. Just make sure it has DUAL_PWM_START commented out.
For a sanity check, you might also want to try flashing NLITE.
Thanks. I have the NLITE someplace and I’ll try the STAR-noinit.
Off to work I go, again! 
Ran the Nlite as it was in my hex file and now I’m getting one with nothing and then medium and high. So still no low.
That’s pretty strange. If NLITE doesn’t work, I suspect a hardware issue.
BTW, should I wait for you to get this into a code repository, or should I go ahead and add it from the zip file?
IIRC, you’re familiar with git… and that would totally work. It’s easy to import from git, and eases the process of merging updates later. I already do that for JonnyC’s code (though he hasn’t updated in quite a while).
I think I still have some of your older code waiting in a half-merged state too. Oops. IIRC, there were a whole bunch of formatting changes and relatively few functional changes, and I was trying to untangle those into two separate commits for clarity… but perhaps I should go ahead and dump them all in at once instead. It’s just hard to diff the original and modified versions when virtually every line has changed in ways which don’t affect the actual code function.
TK, I forgot to thank you for a great laugh I had, when I saw the video you posted of that funny dude, didn’t know that one.
@18sixfifty
About the moon mode issue:
I think NLite and Starnoinit, need a star to be soldered to enable moon mode. Of course you could adjust it so that it’d be the other way around (moon enabled by default without the star soldered)…
Thanks I might actually switch it so the star isn’t soldered because soldering it doesn’t seem to work.
Fresh driver and same problems. I am getting three out of four modes now, so I’m closer. But still no moon. I’m using star 1.1
I’m having problems trying to figure out where to change the pwm settings. That has to be it.
I got it working and I think it was just a combination of me having a case of the dumbspits and possibly bad mcu’s. I ended up using an older driver that I hadn’t used from way back and that along with phase-correct PWM did the trick. I ended up getting the moonlight mode stable at about half of what it was on the original star 1.1 version.
Does anyone have software that will wipe these mcu’s really clean? Maybe I got a batch that was used previously? Can that happen? They came from China instead of Richard, teach me to cheap out.
TK, how do you use the ramp_calc.py?
I’t trying to get some new values as I need it to start at 3 and I’d like 64 steps but I don’t know how to use this file.
I’d suggest using bin/level_calc.py instead… because Ferrero_Rocher/ramp_calc.py is old and less accurate and harder to use, and the only extra thing it offers is pulse frequency modulation to make the low end of the ramp look smoother.
If you aren’t using a single-channel driver with both PWM and PFM, use bin/level_calc.py instead.
For example, a FET+7135 driver with a single XP-L:
(~/src/torches/trunk/)-]> ./bin/level_calc.py
How many total levels do you want? (4) 64
Lowest visible PWM level, for moon mode: (6) 1
How bright is moon mode, in lumens? (0.25) 8
How bright is the highest level, in lumens? (1000) 1500
Use dual PWM? [y/n] (n) y
Second channel, lowest visible PWM level: (6) 3
Second channel, how bright is the lowest mode, in lumens? (0.25) 0.25
Second channel, how bright is maximum, in lumens? (140) 140
1: visually 0.63 (0.25 lm): 0.00/255, 3.00/255
2: visually 0.80 (0.52 lm): 0.00/255, 3.48/255
3: visually 0.97 (0.92 lm): 0.00/255, 4.21/255
...
62: visually 11.10 (1369.01 lm): 208.87/255, 255.00/255
63: visually 11.28 (1433.51 lm): 219.85/255, 255.00/255
64: visually 11.45 (1500.00 lm): 255.00/255, 0.00/255
PWM1/FET values: 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,3,6,9,12,15,18,22,26,30,34,38,43,48,53,58,63,69,75,81,88,95,102,109,117,124,133,141,150,159,168,178,188,198,209,220,255
PWM2/7135 values: 3,3,4,5,7,8,11,14,17,21,26,31,38,45,53,62,72,83,95,109,124,140,157,176,196,218,241,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0
OK how does it work?
I’m only using single channel with 3x 7135
Oh, that should be pretty straightforward then. 
Basically just tell it you want 64 levels, minimum PWM of 3, lowest mode of 0.5 lumens, highest mode of 400 lumens, and no dual PWM. It looks like 64 levels will get you a few duplicates at the beginning though. That’s what pulse frequency modulation helps with, but it’s also more than a little finicky and might not be worth the effort.
Are you putting this into an e-switch light with Ferrero_Rocher, or something else?