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

That’s exactly what I wanted to hear, thanks very much!

Interesting, not been reported here. I got stuck in Ottawa for two weeks in the ’90s with a big ice-storm. And no power. Even in the city. Helped my counterpart to keep his home running (generator running 24/7 and the sump pump to keep the basement dry)

Oh, he lived three miles up his own road (or drive as we would call it here). A bit extreme, but he had two jeeps and a big thing with a bull-dozer blade on the front. His daughter was a bit mad (one of the jeeps showed the sideways scars from some hits, but usually got lifted up again, usually, no questions asked)

She was delightful, by the way.

Is the countries power supply still so rickety ?

I have two on order already, 688 and 689. I’ll try and read through the last 30ish pages of posts shortly, after being away for a while.

But. If you can switch to USB-C with a sane charging rate, I think it’s an excellent idea. I’m not interested in a frying-my-batteries rate. Just something that will use the future-proof interface for charging. If it costs a few dollars, fine by me.

I didn’t see what the expected cost is anywhere. But odds are, I’ll take at least two more. If it helps in getting an overall better price, put me down for 4 total for sure. Whatever it ends up being is fine.

Great work, take your time. I’d rather have an ultimate lantern than a faster shipping lantern.

Larry

Edit: I noticed you were going to try and open up other battery options. Awesome! I’m an EE/ME though, and I know that complexity truly sucks. 18650s are great. Others can be slightly better. Don’t hose this up trying to have too many power options. The perfect 18650 lantern is a remarkable achievement on its own.

18650 batteries are good, but the new and slightly larger 21700 cells are rapidly gaining popularity and can store up to 40% more energy (probably even more in the future as they seem to become the new standard). Also they are the same length as protected 18650s so that the only thing that we would need to change to be able to use them (at least the unprotected ones) is to make the battery tube a little wider, which isn’t adding much completely in my opinion.

That seems like a very good idea.
The difference could be made up with some type of sleeve.

The difference in diameter is 3MM

Well, that’s only for one cell. According to Engineering Toolbox, in order to increase from the standard 4x 18650 tube to a 4x 21700 tube would take an increase of about 7mm in diameter. Not that it’s a big deal. For a lantern that will pretty-much always be left tail-standing, never meant to be hand-held, it can be a lot wider anyway. But, I’m hoping that’s a consideration for Mark II. This lantern is based on the existing Q8 design, so it will be Q8 sized.

Might be a stupid question as I am new here but why isn’t this driver going to be a buck boost? I think I under stand how the 7135s are being PWM switched directly and additive so you are only switching one on and off at a time but the design of 7135s is inherently inefficient compared to an actual buck-boost driver. With all the machining and materials involved is it such a big cost adder to switch to a buck-boost driver? It seems to me that with a lantern, run time is more important then with a flash light so efficiency, especially at low levels of output is important. I’d be willing to pitch in developing such a driver but its been a few years since my job required designing PCBs and programming micro controllers. Don’t get the wrong impression I just have been exclusively buying only constant current lights for a few years because I was worried about eye strain and like the idea of high efficiency. I’m hoping if this lantern keeps the 7135s there will be a mode without PWM and 4 brightnesses with each step turning on a new chip.

Currenly, the 4x7135+4x7135 driver is well understood and can be easily implemented with tint ramping. Actually, it has already been done.

About PWM, the PWM on Anduril/NarsilM drivers is so high that literally no humans can actually even perceive it.

What you are requesting is quite hard to do. That would require a new driver, which would be easy with a switch mode power supply, but not 7135s. Only 100% full power with both tints, and 50%/CCT can be activated without PWM. Unless I’m wrong.

Finally, it’s because of cost and time to market. We don’t want to rise the final price of the lantern by too much, and we don’t want to push our deadline further. We are predicting to start production hopefully a few weeks before Christmas.

Cpeng, Welcome to BLF and thanks for your input. The driver has been discussed a lot. Even specifically the question of using a buck or boost driver has been. Basically, it comes down to we’ve got what we’ve got because someone was willing to make it, and it will work quite well for our purpose. You’re more than welcome to design a buck-boost driver for us. At this point, it probably won’t make it into this lantern, but it might make it into Mark II, which will make a lot of people happy. If you share your work here, many will even use it to mod their copy of this lantern, and maybe other similar sized lights, like the Q8.

Last update I saw, Den said he was still waiting to hear from Barry (I assumed still concerning design) and that was 3 weeks ago.
First time I have heard any mention of ‘production’…… THAT IS good news……

DavidEF: I’m not trying to change this design, more just trying to understand why seemingly all BLF lights are based around a FET + 7135(s). I’m a mechanical engineer who has dabbled in electrical engineering in school and work so I’m not an expert. I can just search for buck boost circuits and they look easy enough, but I’m sure I’m missing something. I can understand for turbo output its very cost efficient to use a single fet, but this lantern won’t even be using a fet.

BlueSwordM: I’m not sure I understand your point, as I understand a buck-boost circuit can have a ramping brightness by controlling the voltage output which in turn will control the current with a PID loop. As for color shift it seems that a 4 constant current circuit with 7135s would be very similar in tint for all levels except the 0-25% brightness.

I’m not trying to offer criticism just trying to figure out why the current driver is being used. Is it cost, expertise or something else like performance or durability? Like I said I’m no expert and it may be that I would fail trying to make a buck-boost driver. I have other hobbies so I’ve been a consumer of flashlights so far as I believe I have nothing new to offer, but then I found this place…

@Cpeng, we have 2 sets of LEDs, one with a 3000k CCT, one with a 5000k CCT. That is what makes it much more complicated. We are working to power 4 LEDs at 5000k separately, 3000k LEDs separately, and 8x LEDs at the same time, and we can individually control the power level so we get a desired tint mix color temp.

@BlueSwordM, thanks for information. So you would need 3 separate buck boost drivers to control it, seems you are telling me it wouldn’t be cost effective and or too complicated for this project which is a valid reason. I’m sorry if this is somewhere in the thread. I haven’t had a chance to look through the entire thread.

Edit: I’m an idiot, just searched for “7135” in “search within this thread” and I get the answers I was looking for.

@ Cpeng, definitely not an idiot, you’ve not been around for long and lots has happened over the past couple years of this project. Don’t hesitate to ask questions, most of us are a cordial bunch. I don’t think anyone is being rude, but it may come across that way in text sometimes. Folks are just interested in getting this done and into our hands. I know for one that I could really use this in about a month, but don’t expect it to be ready by then.

3000th thread post :+1: :beer:

Awesome. Would be nice if all new people on BLF figured out this function so quickly. :slight_smile: :+1:

Yeah, I agree. Cpeng, sorry if I was coming off as impatient or rude. We don’t mind you asking questions. :+1:

Yeah I don’t think anyone was being rude, I thought there was a simple answer and there wasn’t. Its better to get responses that I wasn’t expecting rather then no response so I thank everyone for that.

@Cpeng

A reason that is easily overlooked for using a simple and well-tested driver design:

We are developing flashlights in cooperation with a chinese manufacturer (Sofirn) who is halfway around the globe for most of us. We can not pay a visit to check what’s going on, everything must be communicated via the internet through cultural and language barriers, and they are not the most technically advanced flashlight manufacturer around. We chose this manufacturer because they are prepared to work with us and they are able to sell the lantern for the budget price we like to see.

Anything what is new to them is received with reluctance, for them to get used to it takes time, and progress must be meticulously checked by us at every stage because they can knowlingly or by naivety change any detail at any moment, possibly ruining the design. New stuff can always be done, but the process can be made more easy and will speed up by keeping it simple and familiar for them.

Sofirn is accustomed with the type of driver we are now using in the lantern, it is very much like the Q8 driver. It is different in some details (the tint ramping requires some changes in the PCB design) but they understand it and have the components in house. Furthermore, here on BLF is so much knowledge on this type of driver that we can predict pretty well beforehand how it will behave in the lantern, with just some finetuning of the software.

Without knowing details about driver designing, from what I pick up, designing a buck driver is already challenging enough for BLF (but can be done by some of the driver guru’s around) but then it must be prototyped and tested and both the hardware and software finetuned in the lantern. And Sofirn is for sure not able to do much tuning, so for several stages in the design process prototypes must be sent around the world for testing. This is a much longer process, the reward in the end is a better more efficient driver, but only achieved with lots more effort.

I think…

All i am waiting for now is any updates from Barry form his engineers, and if Lexel can build a driver test unit to send Toykeeper to test the firmware on.

I’m mostly just waiting. I’ve done all I can for now.

https://bazaar.launchpad.net/~toykeeper/flashlight-firmware/blf-lantern/files/head:/ToyKeeper/spaghetti-monster/anduril/