What device / adapter to update fw? 0.04 lm wow
It’s Atmel’s UDPI programming. More info about it in my thread here. Programmers for it aren’t too common. You essentially need a common USB to Serial TTL adapter and then a pogo-pin adapter to mate up to the flashing pads. I make (and sell) some but shipping from the US is expensive. Check with thefreeman as he’s in Europe and might be able to help out.
1 Thank
Thx a lot !!!
thanks for the Brass TS10 prototype photo… Maybe they will also make a Red Copper version.
The Chinese word for Copper (Tóng) includes Brass
White Copper: Bái Tóng, German Silver, Alpaca, Copper Zinc Nickel
Yellow Copper: Huáng Tóng, Brass, Copper Zinc
Red Copper: Hóng Tóng, Pure Copper
Purple Copper: Zǐ Tóng, Bornite, Copper Iron Sulfide
Green Copper: Qīng Tóng, Bronze, Copper Nickel
Great modding tip Lightbringer!
Hopefully I’ll never need to fine-tune an SMD resistor this way, but good to know that it’s possible. :+1:
Yep, I’m wonderful. 
Terry, thanks for going to the effort to address concerns! Do you know if the light was tested with H10 while using turbo repeatedly? I think the previous events/concerns had more to do with excessive heat caused by continuously forcing the light into turbo mode after it had stepped down, so the light was running hotter than it usually would, for longer periods than what is shown in the video. I’ve been using mine with H10 and trying to keep it hot, but not excessively, and so far so good…if it fails I’ll just consider it a donation to science. 
Well, just what I had expected. 
I used this ramp:
level_calc.py 7.01 2 150 7135 1 0.5 130 FET 1 10 1200 —pwm dyn:64:2048:255
First, the lowest level can be really low. I mean, ZebraLight low. Problem: It’s all in the undefined behavior region of the 7135 because the pulse length is just too short.
Here is an example ramping through levels 1 to ~45:
Level 2 is darker than level 1 because the frequency is doubled (dynamic underclocking, quarter speed below 2). There is another jump from level 4 to 5, again dynamic underclocking (half speed below 5). Then it starts increasing, but in an undefined ramp shape. At level 45 the 7135 starts turning fully on (maybe a bit earlier because my light sensor is too slow).
What’s next? I might try to start at level 16 where the slope changes. That might be the point where the error of my light sensor is most influencing. Or it’s just the non-linearity of the 7135. From there I might add one or two “special” levels with ultra low moonlight and from level three upwards it will be lot brighter, but more reliable. Not sure what to do with dynamic underclocking…
Anyone with an idea how this could be set up in a nice way? ToyKeeper, gchart, thefreeman?
So, now that Wurkkos has found out the emitters actually shouldn't die with a high drain cell, is a linear driver still on the table or nah? Cause I just want that linear driver for slightly higher efficiency and that smooth, smooth ramp. 7135 also seems to be somewhat unstable and it varies from one part to the next... I haven't bought any TS10 yet because I was waiting for the linear driver upgrade :/
Nice test! I’m a bit torn about be dynamic underclocking. We same a few mA of current draw from the reduced clock speed, but at the cost of slower PWM and some minor inconsistency in the ramp output. It’s all in the trade-offs. So what do we value more?
- Lower lows with lower power consumption but slower PWM?
- Lower lows with higher power consumption but faster PWM?
- Higher lows with lower power consumption and faster PWM?
yes…
fwiw, the floor on TS10 and SC21 Pro and FWAA is 0.2 lumens
the floor on the SP10 Pro is 0.01 lumens.
Does that have higher power consumption but faster PWM?
Just curious. I know that much of this discussion is about just seeing what can be done with the software, but of what use is having an ultra low output.
I personally find that the 0.2 lumen output is pretty useless. Even at night from sleep with fully dark adapted vision. I have set the floor on my TS10 to be level 5 of 150 which is about as low as I can go and still get enough light to see anything. Just a sincere question…
Maybe this question would be a good one for a dedicated thread ??
For me the 0.2 lumen output of the SC21 Pro is too much in the middle of the night with dark adapted vision. It takes very little light to completely disrupt my sleep and I have arranged to sleep in near total darkness. The SP10 Pro at 0.01 lumens is actually more light than I need to get to the bathroom.
OK, I think I made some progress. Problem is that we should keep the pulse time above ~10 µs if we want to use the 7135 in the proper region. This can be achieved only if we increase the start PWM value or underclock the MCU long enough.
This is my configuration from my previous test:
As you can see, many low levels have too short pulses and will cause non-linearity.
With some ramp tuning, starting with PWM 10 and increasing frequency divider to about 12/22 it looks better, but I haven’t tested it on the flashlight yet. Sadly I don’t have more time today (and maybe until next weekend).
Btw. the stock configuration is deep in the non-linear region as well. This affects most if not all 7135 lights using Anduril.
I wish to order a silver 4000k but it is OOS everywhere. Anyone know when it will be restocked?
Wurkkos Terry wrote:
A great job!!!
For me: Lower lows with higher power consumption but faster PWM.
I’ve just received mine today, must have been close to the end of the second batch. The very pale silver and cyan aux lights look great. I’m just wondering if a cooler white LED might have complemented the icy look of the light better?
I just got my red 5.8kK TS10 yesterday. This is my first tail e-switch light, and I have a question.
What, exactly, does loosening the head do in this type of light? My aux lights slowly ramp off, and the tail switch activates these weird, random momentary ramping (sometimes flashing) modes. I can’t find them on any Andúril ui chart.
Loosening the tailcap disconnects the inner tube and disables main power.
However, the outer-body is still connected to the head which powers the switch channel. Anduril 2 does weird stuff when only the switch channel is connected to power. I think the stuff its doing is intentional, but I’m not sure what it’s for.