*BLF LT1 Lantern Project) (updated Nov,17,2020)

No temperature sensing as I understand it. The sensor is on the mcu so temp sensing would involve having a separate sensor on the led mcpcb

Mtn. Don answered this one.

ToyKeeper provided me a trial version of firmware to test that was attempting to fix a problem where short button presses were occasionally being missed. This version had strobing issues that could only be stopped by removing power. I never saw the strobing with the original build of the firmware sent out with the LT1, or any of my other Anduril lights. Wondering if the strobing just stops, or do you need to remove power to get the strobing to stop? For me power needed to be removed. Keep notes if you continue to see the problem and perhaps ToyKeeper will add looking into that to her “list”.

Yes, this is how sunset works. I don’t know what step size the firmware uses, but in reality there are 150 steps from min to max. It seems like there could be smooth steps down for sunset, but I know the Attiny85 is very full, and perhaps implementing such a smoother step down function won’t fit, or has not made it to the top of the priority list yet.

I also would like such modes, but I know the time keeper in the Attiny85 is not terribly accurate, and counting something like 8 hours would likely be pretty variable from unit to unit, at least that’s the way I understand it based on my recollection of hearing ToyKeeper discuss the Attiny85.

Ordered my LT1 + batteries a few days ago with the amazon code. Many thanks to DBSAR, the entire LT1 team and Sofirn for making it happen.

Today I realized that the current AE store price is under 57 USD (with the promo code) which is better than GB price… Oh well. I’ll just stick with amazon.

I don’t follow that…. I paid Amazon 49.69 plus tax of 2.55 for a total of 52.24 US$ (LT1 only, no batteries. ) The actual advertised Amazon price was $69.99 and the coupon code gave a discount of $20.30 The only variance for a no batteries included LT1, shipped to a US address should be a possibly higher or lower sales tax amount.

I guess mine is on the slow boat, this was the last movement:

Thursday, October 31
8:31 AM
Package has left the carrier facility
GuangDongShengGuangZhouShi 51040034,

Says it’s arriving by Dec 19, just in time complete this lantern adventure and sign up for the Mini!

edit: and then it just shows up and blows my mind! Amazon still has it stalled out somewhere in China.

Worth the wait. Darn cool.

Yes, I guess it just depends on where you are - my total remained at $49.69

I’m sorry. To clarify, I paid roughly $63 total for the lantern with batteries as a member of the GB. In Aliexpress, I added the LT1 with batteries to my cart and applied a - $8 promo code that Jackie posted here on BLF. The total was $56.39 with batteries shipped from China to eastern US.

Ok, thanks! Not an issue, just wanted to not have to find out the hard way!

Thanks for the explanation on sunset stepdowns!

Ok, yes, that strobing issue sounds like mine, have to remove power. Seems better after I cleaned the positive contact ring and I’ve taken to using the onboard charging rather than using an external charger.

Thanks for the data. It has been really useful for adjusting the power correction formula.

The prototype hardware and production hardware have different response curves, and the production hardware is much nicer. So the firmware was (over)correcting for an issue which has been greatly reduced by better hardware.

Anyway, the code is updated to better match the new hardware, and it should be included eventually on new lanterns. Or for people with the ability to flash firmware, it’s already available.

1. There is no thermal regulation on the LT1. The sensor is too far from the LEDs, and the MCU doesn’t have enough pins to add an external sensor.

2. That’s a symptom of a crash or a hang. When something goes wrong, the attiny reboots itself… and it will keep rebooting unless it’s specifically prevented from doing so. The strobing on boot means something went wrong, like interrupts interrupting each other until the stack is exhausted, or something like that. This seems to happen when there’s some sort of electrical noise strong enough to trip up the control chip, so it makes sense that cleaning the contacts would help. That would eliminate the source of the noise.

In the past few weeks I’ve rewritten large parts of the kernel to make it more tolerant of electrical noise, by reducing the amount of time each interrupt takes and making code more thread-safe. I also added some code to make it handle crashes differently, by rebooting only once instead of continuously.

So newer firmware should help, but it doesn’t actually fix the cause of the problem — electrical noise. That’s more of a physical problem. The newer versions just makes the control chip respond better when there is a physical problem.

3. Sunset goes down one ramp step at a time, but the ramp’s resolution is limited… especially near the bottom. I have some ideas for how to increase the hardware resolution beyond what is supported natively, but I don’t have it working yet.

Sunrise mode has come up a few times before, but the attiny chip’s timing is really variable and inaccurate, so instead of 8 hours, it would end up being anything from 7 to 9 hours. It’s do-able, but hasn’t been added yet. The main limitations are timing inaccuracy and limited ROM space.

Good to see you hear Julez! Love me some flashlights (and batteries, and chargers :laughing:.

Cheers,
David

Who is this magical Jackie that has Ali discount codes?

Just as an FYI, I’ve been running into the same type of strobing when trying to lock / unlock an EDC18.

On the subject of non-linear battery current draw (and output intensity), ToyKeeper has produced a build that is much flatter across the ranges of tint variation and light output intensity.

http://toykeeper.net/torches/fsm/anduril.2019-11-29b.blf-lantern.hex

This post and a few after it give the final (for now) results of this investigation and updated FW build.

The build also adds the SOS function (after beacon in the blinkies group), and improves a minor bug that sometimes caused the firmware to miss a short button push.

Interesting. That’s the first time I’ve heard of it happening on a light which doesn’t have an active standby mode (ability to blink aux LEDs while asleep).

Then again, I don’t actually have an EDC18 and I don’t know what firmware is used on it… so I have no idea if it was modified. I’m just guessing that it wasn’t.

In any case, it’s generally a symptom of some sort of physical issue which was bad enough to knock over the attiny chip. The ~8 Hz strobey thing is the flashlight equivalent of a Windows Blue Screen Of Death.

However, they could probably fix it by using newer firmware. It’s much harder to cause a reboot now, even with really noisy hardware… and if it happens, it does a normal boot now instead of going into the eye-catching “hey an error happened” mode.

TK - this was a first-run EDC18 pre-order. Odds are high it’s an older version of the firmware. Once I finish moving and get a place to work, I’ll reflash it and see what happens.

I decided to experiment with the “pencil trick” to see if I could enable an additional channel in my lantern. It works. Filling in the space between NW and WW pads for channel 6 increased the current at level 4 from 1.01A to 1.20A. 1.01*6/5=1.21. This would allow for folks to add channel 6 and/or 7 if desired without a soldering iron. It easy to remove with an eraser also. I already removed the connection in about 5 seconds per side.

That will be a welcome trick for many :+1:

How does a person do the pencil trick? I would solder in the long run if I liked the results, but would like to the trick to check it first.

Yes, As amishbill asked… Who is this magical Jackie?

Hello! Why do you need a variable resistor on the driver Board where the contact with the batteries?

Pencil trick - use a pencil to make an electrical connection between two pads on a circuit board, or across a component on a circuit board. In this case, I used a #2 pencil and applied a fair amount of pencil lead (what is this stuff anyway?) between the pads labeled “6”. The pencil lead is a “good enough” conductor to make a connection here. I don’t completely understand how it works, as I measured the resistance between the pads before and after adding the pencil lead, and measured about 10 meg ohms both times. But this connection is in series with the control line of the AMC7135, so there must be just enough improvement in the connection to allow the controller chip drive to reach the AMC7135.

I borrowed the picture below, but the pads circled in blue are 6 and 7 for NW and WW