Not bad:
We might be able to use the internal voltage reading for the thermal regulation. Higher voltage -> more heat -> higher phase shift -> faster thermal regulation response.
Indeed. :+1:
[quote=Unheard]
I can only strongly advise to not use anything newer than r453. The thermal regulation is almost completely broken in my eyes. It’s pure luck if it works.
I just wanted to throw Anduril into my H03, which has been waiting for it for two months. I think I’ll wait a little longer.
The default VOLTAGE_FUDGE_FACTOR seems to be a reasonable choice.
End result of the new thermal regulation code (link to the code: fsm : Code : Flashlight Firmware Repository) :
It’s… impressive
Yeah, the regulation looks pretty good for this light (and even the D4Sv2). But it wasn’t tested on other flashlights and is based on a older version of Android Anduril. I think we need ToyKeeper now for review, testing on other flashlights and possibly cleanup of the code (although it’s already clean, but maybe not what ToyKeeper likes).
edit: Damn autocorrection…
Ok, I made FW3A build from code and it seems its working. Will need to test it outside, its about zero degrees celsius now.
Anyone else tested it on ohter models? I am waiting for TK comment on this new code too…
As a follow-up to my earlier post, I just received the new aux board from Hank and soldered it in. Worked for a few minutes and I had my hopes up, and then went dead just like the old one. So there’s something else going on… must be in the driver.
I initially suspected the old aux board had a bad resistor for the blue, but I measured it while running and it was dead on with how it was labeled (0.9k). So then I assumed it must be the driver but tried the new aux board anyway. It was worth a shot I guess.
Time to contact Hank again…
Updating firmware daily on FW3A is not fun because of soldering 
Do I need to make some adjustments in FW3A cfg file before I test?
In that case I would just wait for the next stable release.
No.
| D4V2 XPL-HI V2 5D | Sony US18650VTC6 | Samsung INR18650-30Q | Sanyo NCR18650GA | Keeppower IMR18500 UH1850 | Panasonic NCR18500A | Keeppower IMR18350 UH1835 |
| Age / Remaining capacity | new / 98 % | 2 years / 90 % | 3 years / 90 % | new / 100% | new / 100% | new 100 % |
| Lumen | ca. 4300 | ca. 3700 | ca. 3400 | ca. 3650 | ca. 2840 | ca. 3350 |
Interesting. Any data using cells of the same age?
No, but I got this (the levels are the ones from the stepped ramping mode):
Yeah, I’ve already started but it’ll take a while. This part of the code is tricky to modify and the tests take a long time to run.
Over the last weeks I was in contact with 0-8-15-user and reviewed the code several times. Not the latest version yet, but there were only small changes added. IMHO overall the code looks good. It is much simpler than the previous implementation and seems to work with lights of different properties (D4Sv2 and D4v2). Nothing I would send to a manufacturer without further testing, but it should perform much better than the current version.
Ramping level vs. Lumen:
|
(the battery that was used is roughly 2 years old and has a remaining capacity of around 2700 mAh)