Had a look on Amazon and this seems something that I would probably be comfortable gifting to an elderly.
Looking at (and listening to) the single review video on Amazon, its button sounds very heavy.
The fixed 70/180lm output might annoy her (not kidding 
), she is used to lower levels, and has a definite dislike for bright lights.
And it it does not seem to be available around here: I’ve never seen it any electronics stores, cannot find it in any local or EU webshop, and Amazon won’t ship it here either.
And anyway, I already ordered the IF25A, so this ship has sailed…
Especially if grandma accidentally rotates that Energizer pointing upwards shinning into her cataracts could be disastrous
I like the body ergonomics tho…
Hey Guys,
I am back. Sorry for being away for so long. I’ve been busy, both with life in general and with this project.
Plus, of course, the couriers took their sweet time delivering the light from China, to begin with…
But finally, I have some actual news and results. If you are still around and still interested, I am happy to share my experience.
Yeah, definitely share! This topic gets brought up somewhat frequently for a variety of users, so any input helps folks who may ask or search for info in the future. What’d she think…what’d you think?
OK, let’s start with my own impressions on the IF25A:
I ordered it from Sofirn’s online store. It took 15 days to be delivered. Shipping provider: YanWen, tracking available, TN: Ux………YP.
To my fellow Europeans: The webshop is wired into IOSS, i.e. VAT is payed immediately on purchase. No extra hassle with Customs when the package arrives in the destination country. Good job, Sofirn!
The unboxing experience was a bit of a disappointment (in terms of the included accessories).
With my LT1, I’d received an extra USB port cover and button cap (more on that later).
For the IF25, they were not included, only the extra O-rings, battery sleeve, lanyard and cable.
Yes, that was 100% on me. I read the product listing and multiple reviews beforehand, all of which stated the list of bundled accessories accurately.
But apparently I failed to acknowledge that information, because I just assumed all Sofirn lights would come with the same extras. Anyway… no biggie.
Machining quality looked nice, no sharp edges. Lubricant applied to the tube threads at the factory. Battery at ~4.0V.
The beam shape is lovely: floody and homogenous.
Temp measurement off by only 2-3°C. I recalibrated it regardless.
Standby drain 31uA with the switch light turned OFF, 290uA during ADC measurements (i.e. the “spikes” in the current draw every 8s). Measured with an uncalibrated UT61E.
I chose the 4000k variant, and man, do I love those SST20 emitters!
It is difficult to put it into words but their beam looks just very pleasant and natural. When shining it on red(ish) colored objects, they really pop, not washed out at all.
Like some of you had mentioned, the beam indeed turns greenish at super-low levels, but even at a ramp level of 30 it is very nice already.
I prefer them to the LH351Ds in my LT1 (at any tint mix).
About the button characteristics, actuation force etc.:
Well… I was not impressed TBH, but once again, that was partly my fault.
The actuation force out-of-the-box felt passable, but not great. It was noticeably harder than the LT1’s. But that was the less significant issue.
What I was a lot more worried about was how low-profile the button cap was on the IF25A. (Which I am sure also contributed to why the act. force felt higher than it actually was.)
Grandma’s fingers are …well… what is the opposite of “firm”? Squishy? Anyway, the point is she has great difficulty operating any controls which are recessed.
Yes, I did look at multiple photos before making the purchase, but you know… on 2D images, this is not that apparent. And anyway it is part of normal, sensible flashlight design (avoiding accidental presses by avoiding a protruding button), so I guess my brain just filed it under “business as usual, nothing to see here”.
Remember me mentioning the spare button cap I had for the LT1? That gave me an idea: The diameter of the two caps are the same, so in theory a swap could be possible.
I gave it a try and the results were awesome!! The LT1 cap is a lot taller than the other one. So it protrudes pleasantly far from the IF25A body.
But that is only half of the story: Due to the difference between how deeply the switch PCB sits in the LT1 and th IF25A, and between how low-profile the tact-switches themselves are,
the LT1 cap, when inserted in the IF25A, is pressing down on the switch ever-so-slightly all the time (you can see the middle of the rubber cap bulging a little, if you look closely), but not enough to actuate it on its own.
This kind of pre-tensioning combined with the tallness of the cap makes the overall actuation force ridiculously low. I just love it. Even though it was a total accident that I owned a button cap which was “just right for the job”
I hope my description was more or less clear. Unfortunately, I forgot to take pictures of the swapped switch. But I will the next time I visit her.
I normally visit granny every 3-4 weeks, and the flashlight arrived in my mailbox shortly after one such occasion. As a result, I had quite a lot of time on my hands, which I intended to make good use of.
So I put together the toolchain, downloaded the Anduril source code, and implemented a visual low-battery-warning. I did my best to set it up in a way which gives clearly visible indications, while remaining as unobtrusive as possible. The notification is done by the main emitters. The trigger is a combination of time, absolute voltage level, and voltage change.
It also works during strobe modes:
You can read further details and download the source code via my GitHub repo:
https://github.com/madcrow/Anduril_LowBatteryWarning
Like all other parts of Anduril, the new feature is universal: meaning you can compile it for any Anduril-compatible light you want. Feel free to try it.
I included a bunch of compile-time-options, too, so its behavior can be easily customized.
As the IF25A is not very easy to dis- and reassemble, I used the LT1 as my “development board”. All the testing, finetuning, etc. was performed on that one, and only the final release was installed on the IF25A.
Disassembling and flashing the IF25A was a lovely challenge. For the most part, at least…
These are the steps:
- Unscrew the bezel
- Carefully take out the TIR without breaking the standoffs by pulling it straight out. Be careful not to touch the TIR with fingers or anything dirty.
- Unscrew the battery tube and the ring from the switch. Take out the silicone switch cover.
- The LED wires are not long enough to remove the driver while attached.
- As a precaution, attach a layer of crumpled aluminium foil (with Kapton tape) on the MCPCB beforehand to protect the LEDs from accidental contact with the iron tip.
- Desolder the wires from the LED-MCPCB
- As a precaution, attach a layer of crumpled aluminium foil (with Kapton tape) on the MCPCB beforehand to protect the LEDs from accidental contact with the iron tip.
- Remove the switch bezel and pull the switch PCB out
- Now you have to dislodge the glued-down driver:
- Find an area on the top-side of the driver PCB where no components are installed
- Put the assembly down on a flat, stable surface.
- Push down on the driver to pop it out. I used blunt-ended bent tweezers wrapped in duct tape. In my case, the driver was glued very heavily. The amount of force I needed to exert was genuinely frightening.
- Take out the driver PCB stack by gripping the brass knob from the battery side.
- Align the switch PCB so you can push it through its opening (into the housing).
- Remove the driver and switch PCBs
- Grab your favorite AVR programmer and SOIC probes.
- Attach them to the MCU
- Flash FW
- Profit
Here is a gallery of disassembly pics (click on the image below):
As I said a couple of comments before, I was generally pleased with build quality of the light visible from the outside.
Unfortunately I cannot say the same about what I’ve seen inside…
It is really astonishing how little attention was payed on the proper assembly and QC of the MCPCB from the thermal point of view.
Can you see MCPCB fastening screws in this picture? No? Well, I could not either.
But hey, no worries, they are there… installed from the other (i.e. driver and battery) side (doing absolutely nothing for the MCPCB, bc they are too short to reach it anyway.)
But let’s not dwell on this, let’s remove the MCPCB instead!
You see the gold colored parts? Well, those are the areas the MCPCB and the heatsink (the light “head”) did not make contact.
Why could that be, I wonder…
Could it be the burr they left embedded in the contact surface of the heatsink, next to the screw thread?
Probably so, but the accidental solder spill on the bottom of the MCPCB sure as hell did not help either:
Dear Sofirn, may I ask respectfully, WTF have you been thinking??? I mean come on, it’s power electronics 101… Power semiconductors heat up. They need to be heatsinked to avoid damage. When attaching a part to its heatsink, contact area is everything. You make sure it is as large as possible. In part by sheer physical dimensions + accurate alignment, and, equally importantly, by making it as flat and as smooth as possible.
That solder-smeard MCPCB is reject. Junk. It should never have made it off the production line, period. And leaving a burr on the contact surface - making even halfway-decent contact practically impossible - is a no-no.
I understand this is supposed to be a “budget product”. This is why I won’t begin to complain about the unpolished, very roughly machined heatsink surface and similarly rough MCPCB bottom. That is a reasonable cost-cutting measure I can live with. But these ones? Seriously? They are fundamental failures. I was very disappointed.
I removed the burr, then sanded and polished the MCPCB surface before reassembly.
Also bought a pocketful of M2.0x4mm D-head screws and properly fastened the MCPCB, from the top side.
While the light was still disassembled, I non-destructively removed the rubber charging port cover by gently squeezing its “tail” on the inside and nudging it to the other side.
Otherwise it would have been too difficult for her to access the connector. (Not to mention plugging the cover back… that was challenging even for me, LOL)
I know, I know… this won’t exactly help with the waterproofing. But ensuring easy recharging is higher on my list of priorities here.
With the switch mod and reflashing complete, and the time of my next visit imminent, it was time to set Anduril up and pack up the light for the handover.
I configured Anduril like this:
- Simple UI
- Number of ramp steps = 1
- Floor = Ceiling = 30 (matching the output level of her current, incandescent flashlight)
This way it operates as a single-mode light, sort-of.
No matter, what you do, the output will only be 30 (unless you move to advanced UI, but doing that by accident is highly unlikely).
If you accidentally double click from OFF, it will go to 30. If you accidentally hold the button from OFF, it will also go to 30 and won’t ramp and further.
I thought I had it all figured it.
Turns out, I didn’t.
Wonder why?
The moment of double-facepalm (intended at myself, not her) hit me when, after great anticipation, I finally presented her the light, and she tried to turn it OFF…
Back when choosing the “right” flashlight for her, I consciously avoided ones with non-Anduril ramping firmwares for a single reason:
I new from experience that granny does not do short presses / “clicks” very well. She tends to keep his finger on the button for too long. And I wanted to avoid the resulting accidental brightness adjustments.
This is why I went with Anduril, thinking that rendering its ramping feature ineffective by the settings above is going to prevent all that. And it does, of course.
What I failed to realize at the time, was that in that case, 1H from ON won’t just get magically reassigned to the “turn OFF” action, it will remain tied to the now ineffective ramping action.
As a result, each time she tries to turn the light OFF and fails to release the button quick enough, nothing happens: the light stays ON.
She did get better after some practice, but it’s still far from perfect…
Which, unfortunately does not surprise me, because I am quite sure muscle memory is only part of the story.
The other part comes from age and health: extended reaction times, hearing loss, and impaired sensory functions in her fingertips.
Think about it: When you and I click OFF an e-switch light, how do we know the action is complete?
By sound and tactile feedback. As soon as you hear and feel the tact switch clicking, you let go of it.
She, on the other hand, most definitely cannot hear that. And as the majority of nerves in her fingers are shot, I bet she can not even feel the click most of the time.
So the only feedback for her brain to work with in that scenario is the visual kind.
And what does Anduril do on a button press (within the hold and release timeout)? Yep, you got it… It does absolutely nothing. The emitters stay active.
So… this is where we are now.
She insists she is getting better at it but I am not entirely convinced of that. She keeps saying not to worry, she is “not that clumsy”, while I keep reassuring her that it’s ok for someone at 85 not to have the reaction time of a racing driver and the fine motor skills of a watchmaker
So my plan is to bring the necessary tools with me next time I visit her and flash an updated FW on her light, in which I am setting the hold timeout to something like 1.0s instead of the default ~400ms: #define HOLD_TIMEOUT 64 //24. It obviously won’t help with the lack of visual feedback, but considering the average duration of her current button presses, it should work regardless.
Sidenote: In hindsight, I somewhat regret sticking with Anduril. It would have been a lot easier to just build a custom dead-simple 1-mode UI with built-in low battery warning from scratch, on top of FSM (the framework acting as the foundation of Anduril as well). You would not believe the amount of thought that had to go into making sure the battery indicator worked consistently in all edge-cases, accross all the obscure modes and shortcuts of Anduril… In case of a custom 1-mode UI, the switch-off behavior would not have been an issue either. However I did not want to intentionally limit the capabilities of the light. I wanted it to remain a universal, general purpose tool, which I could also use occasionally around the house when fetching stuff from her attic, repairing things, etc.
But let’s not dwell on the negatives:
She likes the general form-factor of the light. She keeps marveling at its hefty weight compared to its tiny size, but it does not bother her, just amuses her. She likes the modded button. And most importantly: she uses the IF25A every day, which, considering the fact how attached she normally is to things she is used to, is a clear indication in itself that she actually considers it an improvement over the old one - even with the switch off difficulty, which I intend to improve as soon as I have the chance.
Maybe they are having the same labor problem the USA is having.
Good stuff madcrow!
It’s the little things that can make a big difference with how a flashlight performs. I’m sure that small little burr from the screw hole really impacted how the well the heat was able to dissipate. A little nudge from a toothpick would’ve resolved that issue.
Almost makes you want to take apart every flashlight just to make sure it was assembled correctly.
Some interesting observations and you really dug into this light right away! That’s a real shame about the QC issues with the shelf and screws…sheesh. I hope that’s just an indicator of new labor or shortage of labor as pennzy suggested. That does seem to be an issue around the world presently, still, and of course there is usually a lot of turnover in factories with CNY every year (moreso in the last two years, so I hear). Are the screws in this one intended just to poke up enough to prevent rotation? That is to say, is the bezel/optic providing the clamping pressure to hold down the mcpcb firmly against the shelf?
Years ago I got in the habit of partially or fully disassembling new lights for a good look at the craftsmanship…fairly common to need to blow away metal dust, smooth burs, pick out stray solder, etc. Even from brands that are generally good-and-clean and from mid-priced brands I still see this stuff from time to time, some more often than others. I would not have caught that bur under the mcpcb, though…that’s a pretty good fail on their part.
I hadn’t thought about the recess area of the switch and “squishy” fingers. lol. Looked again at the WT3M and the cutout is pretty shallow there…trim ring is raised but the button is nearly flush with the ring. With switch pressure and boots, sometimes you can find boots with longer “nipples” on the inside and trim those shorter to suit, or maybe applying a little dot of hot glue or epoxy to the actuator post on the switch itself will give just enough extra height to improve feel. I’ve done the latter with hot glue on some tail switches and have also use tiny bits of plastic rod superglued to the plastic switch actuators (the styrene sprue frames from plastic models are quite handy for this).
Has grandma had the light in hand yet?
I've suggested the Boruit R2 for years as a granny light . Toss the headband and add a big rubberized magnet to the bottom . they were dirt cheap , they ramped ,usb charged and had the easiest button ever made to push . I always found old people had aa much easier time twisting lights vs tail clickies . Another easy solution is the basic bop it,pop it, puck lights . super easy to push on and off. or motion lights that come on were granny needs them to come on.The trick on the cheaper lights is make sure they are at least multiple AA's instead of AAA's if you can avoid them . Old people don't need something impressive and are looking for simplicity .The upside is they don't need massive runtime since they don't use lights very much .
OK guys, I completed the final part of the story. I hope you enjoyed the journey an found it enlightening. I surely did.
There are so many tiny but important design details we never even realize are there. It is truly amazing.
Haha, my thoughts exactly.
Yes, it is a global issue right now, hitting my part of the world as well. Let’s hope the QC mishap was just its temporary side-effect. Not that it is any consolation, as a consumer.
It certainly looked that way.
And yes, that is the million dollar question. TBH, I am a bit doubtful.
What a great idea! I love those clever DIY hacks :+1:
Have you tried that technique with flexible, silicone(-like) button caps? Does the hot glue adhere to those properly?
Sure did Look at post #73
Yes! Couple of years ago I bought her one of those, too. She really likes it, and have used it in addition to her classic flashlight.
I could not agree more. The puck light I mentioned also runs on 2xAAs. It was surprisingly hard to find, while the 3xAAA variants were abundant in stores at the time.
More aspects I did not think of! Slowed motor skills and cognition of cause/effect kind of thing. Had to chuckle a little as I was reading about the “dumbing down” of such a complex and intricate firmware like Anduril to make it do far less than it was created for doing. lol. It’s like all of the design aspects of car exhaust…the things that go into making a system efficient and quiet…there’s a lot to it really…and then kids and shops go and dismantle all that engineering to make them loud again.
I’m glad she’s happy with it. Be sure to remind her that the weight of the light is comparable to the plutonium “demon core” of the first atomic weapons and that she should never operate it during a waxing moon or on Thursdays between the hours of 2pm and 5pm, or place it near raw eggs.
Nice work, madcrow…really. A lot of work and thought in a very loving gesture for your grandmother.
Thanks for the kind words.
And especially for “demon core” comment. Made my day. 