At the moment, the moon-in-lockout checks the floor for both ramps and uses whichever is lower. This makes the behavior consistent regardless of which ramp is active, but it can also make moon too low sometimes. Do people think it should instead use the current ramp floor instead of checking both? This would effectively give it the ability to change the lockout brightness by switching between ramps, perhaps with a day mode and a night mode, but it might also lead to surprises if the user forgets which ramp is active.
I think Iād prefer it use the current rampās floor.
I havenāt reflashed AndĆŗril in a while. Iāve been waiting for Candlelight Mode to be perfected. :partying_face:
Has Muggle Mode been implemented? Does lockout work in Muggle Mode? I like the thought of being able to teach muggles the importance of lockout, both software- and hardware-based.
Ah, been a bit distracted by holiday stuff and fixing FW3A issues, so candle mode hasnāt changed.
Muggle mode is implemented but not published yet. There hasnāt been much consensus about how that should look, so for now itās a simple latching on/off at 1x7135 (~150 lm) with no bells or whistles. I havenāt even made it dim for LVP yet, though I probably should. For now, it just turns off.
How hard would be to do this:
only one mode (100%), it works while you hold down the e-switch, shuts off when you let go the e-switch. Driver is nanjg 105C?
Need this for diving light that had magnetic switch and driver died so I thought to use regular nanjg 105c and replace e-switch with reed magnetic switch!
Siruis9, that should be fairly easy but itāll need to use something smaller as a code base. The FSM library doesnāt fit onto attiny13 like the nanjg driver uses. Iād suggest using STAR_momentary or Ferrero Rocher as a base, and basically just stripping out features until itās a one-mode momentary interface.
Thanks TK, I will give it a try :+1:
TK, do you have a suggestion as to what settings I should use for the BLF GT?
Iāve just gotta have AndĆŗril in my GT when it arrives!
Thatāll make nine lights running my favorite UI. :partying_face:
Oh, the GT will require some additional changes in order to work. Its power circuit functions somewhat differently than the Q8. Iām hoping to get a GT and port FSM to it, but I have no idea when that will happen.
My GT should arrive in a few days.
Let me know if youād like to borrow the driver / head / entire flashlight for āresearch.ā
I demonstrated a few of my lights at a party last night. The GT, Q8, Meteor, and Emisars were all big hits.
It was about 0Ā° F (ā18Ā° C) so thermal stepdowns were not an issue.
About an hour later one of the kids announced, āThereās a flashlight going crazy outside.ā
I had left my D4 outside on a windowsill in AndĆŗrilās āLightning Stormā mode. :partying_face:
Managed to flash Anduril on my D4!
Thatās the first time Iāve ever modded a flashlight, so thanks ToyKeeper both for writing the firmware and introducing me to this exciting world.
Iām keen to tweak the C and have a go at implementing some new modes. Iāve that Iām not an expert at soldering though and wouldnāt be surprised if I damaged the D4 if I tried doing it again.
Did I read somewhere that the Q8 can be flashed without needing to solder anything? That might be a good enough reason to get one if so.
Also, is there a definitive list of lights that work with Anduril? Iām assuming anything with an e-switch and an attiny85 would work but itās often hard to find out what chip a light has. Are there any C8 style lights that might work with it or is the D1s the closest thing?
This is going to get expensive now!
Congratulations! TK and AndĆŗril got me into reflashing too!
My first-run and second-run Q8s both have wires long enough to flip the driver over without desoldering anything. Itās an awesome light, too!
Iām not sure if the current Q8s allow this. Perhaps someone with a recent sample can chime in.
In addition to the Q8s, I have AndĆŗril running on my Emisar D4, D1, and D1S. I also have it on a modded EagleEye X6R with a Texas_Ace driver from Lexel. The X6Rās design takes advantage of AndĆŗrilās dual-switch option, which works great.
Some more random observations and feedback, after playing with it for about a day:
- bike flasher is excellent and means Iāve technically saved money by buying a USB programmer and SOIC clip (compared to a dedicated bike light).
- Momentary moon on lockout is handy and good makes the D4 kid-safe
- Having 2 different lockout moons (from the 2 ramps) would be ideal
- Adjustable party strobe level (eg using using memorized level or bike flasher level) would be a nice to have but realistically, the partyās strobe wonāt get lots of use anyway
- Good night mode being ramp-able, maybe having a config for time to off, would be another nice to have. I havenāt actually tried good night mode yet though.
- A good morning / alarm mode would be very cool. I know thereās issues with power usage though.
Thanks. I guess if longer wires can fit then you could always do a one-off resolder for longer wires if needed. Iād give that a go on my D4 but I doubt thereās space and thereās a good chance Iād damage the driver.
The X6R looks nice. With the cost of getting a new driver for it the D1S looks quite reasonable though.
So far, only a single e-switch is supportedā¦ with the option of also using a power switch to make it do momentary-style use in the regular output mode (no need for a separate momentary mode). Without an e-switch it wonāt do much.
Two different lockout moons is easy to do; it mostly just depends on whether people think itās better.
Adjustable party strobe brightness is a little tricky. The pulses are generally only ~4 PWM cycles long. It turns on for 4*512 clock cycles, then turns off. If the brightness is adjustable, itāll actually pulse 4 times per strobe instead of once. Thatās not ideal. Of course, the method Iām using for short strobe pulses already isnāt ideal. Iāve been meaning to make it control the power chips in more of a raw mode, on for N clock cycles instead of N/512 PWM cycles, but that adds code complications because it bypasses the abstraction used to control output brightness. So it locks brightness to a sweet spot on the ramp where PWM is steady, and sometimes stays lit for 4 cycles or sometimes for 5, depending on random factors.
Anyway, um, kind of a long tangent, but party strobe has some known limitations which arenāt trivial to fix.
Making the sunset mode adjustable, and adding sunrise mode, seems like a better thing to prioritize.
The D1S is an amazing bargain.
I built my X6R before the D1S was available. The smooth, tapered shape is a bit more āpocket friendly,ā and like using it with the tailcap switch for momentary turbo while hiking with a headlamp.
If I had to choose between the X6R and D1S, Iād choose the D1S. :partying_face:
BTW, I posted some beamshot comparisons of the X6R, D1, D1S and a few other lights when my D1S arrived.
What light will you be using on your bike and how will you mount it?
I used the D4 today on the road. Pointed down almost 45Ā° as itās so floody. Mounted on one of these but Iāll probably order a more lightweight clip.
Iām considering getting a D1S or Q8 though, as the D4 is currently my farthest thrower. Just need a good excuse to spend the money!
Both are good, depending on what you need. One of the main things is to match the beam type to what youāre doing.
From floodiest to throwiest:
- D4: ~4 cd / lm
- Q8: ~12 cd / lm (?)
- D1: ~30 cd / lm
- D1S: ~100 cd / lm
For biking I use a very floody light like the D4, and I find the D1 is usually throwy enough for my needs when I need to see farther. The D1S and Q8 are also very good at what they do, but I donāt use them as much because of the size.
Then again, size isnāt always a deal-breaker. I might have a BLF GT on the way (~500 cd / lm), which totally dwarfs all these other lights. It wonāt exactly be an EDC, but Iāll make excuses to use itā¦ because thatās just the kind of item it is.
FWIW, I pushed a bunch of changes to FSM / Anduril. A lot of things should be nicer now, but thereās one specific thing I havenāt tested yet and donāt trust. On a hotrod like a D4-219c, I donāt think itāll regulate down fast enough on full turbo. However, it did pretty well on my FW3A at the highest regulated level, 8x7135. I need to do more testing and tweaking to make sure it wonāt burn people with a quad 219c though.
FSM dev branch, getting closer to stable so maybe I can merge it into trunk soon.
Changes:
- Made eeprom writes somewhat more atomic, less likely to do weird things if power is interrupted mid-write.
- Added a super-simple muggle mode. One level, 150lm, click on/off. Has LVP.
- Added candle mode, next to lightning mode. Is particularly nice in the bathtub.
- Added FW3A support.
- Made lockout moon level use current ramp config instead of whichever ramp is lower.
- Completely reworked button debouncing so itāll be okay even on a really noisy switch.
- Fixed LVP / thermal issues (ADC was triggering 4000X / second instead of the expected 4X, which made things weird).
- Thermal regulation changesā¦
- Instead of ~85 steps, it now uses 256 or 512 (for a 2- or 3-channel driver). This makes adjustments invisible by eye, but they still show up on a lux meter. Try zak.wilsonās ceilingbounce app if you want to measure it.
- Wonāt adjust below 1x7135 level any more.
- Adjustment speed changes with how far it needs to go.
- Added the ability to calibrate the thermal sensor, to reduce issues from the attinyās per-unit variation.
ā¦ and I updated the UI diagram a bit:
The thermal config generally involves a lot of clicksā¦ sorry. Hopefully it only needs to happen one time though, to calibrate the sensor.
Great! Iām gonna update one of my Q8s to give Candle Mode a test.
Did you mean to use this link?
Itās been several months since I reflashed a driver, and Iām still a noob. Iām using my notes to compile Anduril, but am running into 300+ errors.
The first is in Makefile line 79 ārecipe for target āmain.oā failedā.
The next few errors are in main.c in lines 11 to 14 that seem to be caused by some references to folders and .png files in the repository.
The rest are in main.c lines 1500 to 3000 and are āstray ā#ā in programā errors.
It seems I have a fundamental issue with my workflow, but as I mentioned, Iām a total noob.
Are my errors due to some change in the organization of code and header files? I see that Anduril.c has lots of stuff up top that wasnāt there before. :laughing: