Q8, PMS SEND TO THOSE WITH ISSUES BLF soda can light

Perhaps by “indexing” the battery tube, meaning carefully sanding down the end, bit by bit, until it lines up nicely. But only try this if there is enough thread left for the tube to be screwed on further. A quick check would be to take out the driver and see how many turns are available before it bottoms out on something. If this is say one more turn than needed to fit it with the driver in place, then it should work.

The above is very valid in theory but my experience with lighted tailcaps, that use the same type of leds and single li-ion cell, is that green leds light up fine during most of the drainage of the cell, while red leds of course light up always, and blue leds do dim significantly when the cell is half-drained.
So these green leds should do fine, and indeed they do in my Q8 prototype and so it seems in many production lights.
Or it has to be that Thorfire used a green led with particularly high Vf (-variation), but I always thought that the Vf of 532nm leds is determined by the technology and does not really vary that much.

I can have the proto work with the flat aligned, but it is less bright. It is such a short twist in unscrew direction it seems hard to fill it.

No don’t think so, er don’t know the extend and so far it is 4 in the delivered and no way all 500 are in. At best it is 1% and well we don’t want 20 with the issue, each one is one too much.

I measured 2.10V with fresh 4.2V batteries.

yep, been done already. Measured 2.247V across each LED on one light that has 2 different brightness's, on the prototype measured 2.167V. Think the higher voltage is why the production Q8's switch LED's are brighter than the prototypes. Ideally design wise, would have been better using 1 resistor per LED. We've relayed all the details to ThorFire.

If the LED is being driven from a straight MCU output, rather than a low side drive, then it is not the same as a simple lighted tailcap, there would be additional voltage drop across the reverse protection diode, and any other series resistance in e.g. a transient snubbing circuit, as well as within the MCU output driver itself to consider.

I’m used to doing worst-case analysis using max and min datasheet parameters, which is why I said I l probably could not sign off such a design, but in the real world thing usually work better than that. Suck it and see is also a valid approach, but can sometimes lead to surprises once production starts, that were not evident in small quantities of prototypes etc.

There is a difference between e.g. a modder hand-tuning a few lighted tailcaps to their personal taste, and the small batch of LEDs that they have on hand, compared with committing to production run of 2000, where the knock-on consequences of having to repair or replace even a tiny percentage of failed torches after they have been delivered, are significant.

Anyway, if a few people could measure their LED Vfs we might soon get an idea of the mean value and actual variations, whether the batch of LEDs has significant variability, and whether the design is solid, or a bit marginal.

Any reason not to get a protected battery like the panasonic ncr18650b?

You are bound to measure the same voltage across both LEDs (due to single resistor supplying both.)

But any small difference in Vfs between the LEDs would mean the lower one would hog the current, hence one bright, one dim.

Agreed, 1 resistor > 2 LEDs isn’t the best way.

Guess my OCD is worse than yours… I actually fixed mine. You can see that the SW leds are lit so it is contacting. I have swapped my SW LEDs to dimmer white.

The high power consumption of this light may shut off the protection circuit if you try to use the highest levels.

Sony VTC6, Samsung 30Q, button tops. The Sony VTC5A is also a top performer but difficult to find with button tops applied. I got 8 of the 30Q with button tops for $5 each at LiionWholesale.com, they had 400+ then, but I don’t know if they still have them.

Okay, so I have done some measurements on my switch board LED and resistor. The resistor is measuring at 15,130 OHMS and I am passing 1.453VDC on to the two LED’s. The switchboard LED’s stopped flickering after some more use of the lights. The voltage on my batteries is around 3.8V. Applying a full regulated 4.1VDC makes the LED’s light up like they are supposed to with 2.4VDC going to LED’s. Now here is the odd thing… After putting my partially drained batteries back into the light, they two switchboard LED’s lit up again for a short time, although one a little brighter than the other. I have measured the resistance across the LED’s themselves and found one to be at 61k OHMS and the other to be at 68.5k OHMS. Not sure what to do with this information, but I will probably be replacing the LED’s in mine as well as the resistor to hopefully higher quality components. Disappointing, but I guess that is to be expected from a first run. Is there any possibility of Thorfire offering replacement switchboards for those who wish to repair their lights?

Is there any way to modify flat tops? LiionWholesale doesn’t have any more button tops :frowning:

Aside from that battery not being any more “protected” that the 30Q?

Not really, but before buying one’s first “protected li-ion battery”, one should be aware of who manufactures what. “Protected” Panasonic NCR Bs are comparable to, say, Trustfire, Keeppower or Nitecore branded batteries. Not comparable to normal Panasonic, LG, Samsung etc. bare cells.

Company A (typically someone big) made a li-ion cell. Company B (when buying “protected panasonics” an unknown sweatshop) rewrapped them, sometimes in clear plastic and fit a protection PCB and nipple. Company C made the PCB. From saying “a protected battery like the panasonic ncr18650b” one will have a long way to making an educated guess who made what.

TL;DR: It’s not Panasonic who markets those.

You can’t really learn much by trying to measure the resistance of an LED probably nothing, because a typical multimeter won’t apply anywhere near the Vf when measuring. Particularly when still connected to the de-powered MCU on the driver (whose internal protection circuits might be expected to provide another current path),

If your meter has a “diode test” setting that could provide useful information, but most such meters are set to measure switching diodes, typically less than 1V Vf. I do have one meter (a very cheap one) that can measure Vf on it’s diode test setting, up to about 3 V

The diode test setting usually applies a constant current of 1mA through the diode, and measures the resulting Vf. It’s actually very handy to check parts before or after assembly, because 1mA is enough to light up even big LEDs sufficiently to show that they are working, or with these little SMD leds light them up very brightly.

Oh ok. Well thank you… Is there any problem buying the 30qs from banggood then? Since no US retailer has them in stock?

got my 1st light today. had the charged up LG cells ready while the samsungs were charging so i tossed them in.

man is it bright. cant wait for night. :slight_smile:

my switch LED’s are working fine so far.

thanks to those who made this light possible.

Brian

Do you mean the LED stays on too long?

I’m not sure how people use beacon modes. I’ve seen them with an on-time anywhere between 1ms and 1000ms, and an off-time of anywhere from 1s to 10s. Brightness has been anything from moon to turbo. What kind of beacon is ideal?

On my most recent beacon I took a guess and made the on-time 500ms, with brightness and off-time both configurable. But I don’t know if that’s what people consider good.

Probably not. I’m assuming you’re looking to buy 30Q with aftermarket installed button tops(but no protection PCB added). They’re probably OK. However, it’s always an unknown factor. They may or may not have remarkable added resistance.

30Q unmodified could be available too. No problem buying those from BG either, if you’re willing to solder those to make kind of button tops.

> production stoppage

Don’t get bent out of shape.
Consider the odds are that they did stop production for good reason.

Like, they need to understand why the problem is happening and, importantly, figure out whether it would appear over time with more flashlights.
You don’t want them to rush and ship you something that is going to make you unhappy later on.

They are smart to be concerned not about just a one percent chance of disappointment but about the possibility of shipping a known failure or problem that will appear after a while.

Could be there’s a problem with construction technique. Find that guy and educate him.

Could be there’s a problem with a bad parts lot.
Could be all the good source parts have already shipped and the next batch to be built would all or mostly have (or develop) the problem.
Are there bad parts in the bin they’d be drawing from if they kept building lights? They need to find that out.

And remember — it’s several companies supplying parts, several people putting them together.
All you need is just one person on the critical path with the chabuduo — “hey close enough just ship it”— attitude, covering up a problem and passing it on.

Finding that person can be a real difficult task, as that person’s manager could have the same “close enough” problem and be helping hide it.

Could be they really care about their reputation for doing quality work. Signs are that’s what’s happening.

If you really want protected, I would go for better protected GAs (i.e. protection circuits around 10A) or one of the new breed of protected 30Q or VTC6. Examples include Imalent 15A protected 3000mAh and Acebeam 3100mAh protected.