OP updated with

OK time for a little update

We cannot show the battery tube and tail since that is Thorfire design.

Cross section

At last, the outside of the head is almost good. Check the minor problem with the button not being integrated looking from the side.

Some details showing a driver retainer ring, the LEDboard.

NOTE: the outer edges on for example the fins (left) now nice and smoothly rounded.

Yesterday the head of engineering department was very responsive.
They are going to correct the button.
Also the bezel was corrected closer to the result of the poll (though we will ask for a flatter one)

Then, DO NOT BE STARTLED BY THIS PIC, read the text below it first :wink:

It startled me, that is for sure.
But Daniel was responsive so we could email directly and explain why copper is needed and finally got this reaction:

PCB design in the combination with dielectric quality gives MUCH MORE influence than changing base plate from alu to copper.

Nice update. Looks like this is getting ever closer. Maybe by Christmas? :smiley:

The strong idea is that the board has direct thermal paths for the middle pad of the led. We know the costs for such a large quad DTP board and it seems feasable for this light within the price we are aiming at. It is something we do not have an affirmation from Thorfire about yet (every detail goes with repeated requests from our side, followed by absent or incomplete design suggestions by Thorfire, but in the end things are happening).
If in the end a DTP board is not happening (we consider that as a fail from our side, and we will make sure that Thorfire knows that adding a DTP board will sell way more flashlights), we make the board design such that the heat transfer is maximised for that sub-optimal situation. A copper board does help somewhat compared to aluminium. Of course the performance must be good enough to go forth with that option.

I suppose that here is no person that will be available to see any difference in ALUvsCU in this light. Or measure any temperature difference.
PCB shelf is thick enough, surface is much bigger than cutter quad boards for example.
Cu will gives almost nothing, and such board is useless for any other future projects. May be it is better to try ordering custom DTP pcbs for other, more suitable situation?

What makes the real difference in heat shedding is the direct thermal path (which btw makes it a copper board, the only aluminium DTP boards I have seen sofar are Sinkpads), but a test that I have done on 6 different 20mm XM-boards two years ago suggest (not proves, too many variables for that) that also for non-DTP boards it helps if the core is copper.

Looking good!

Is there a need for all that space in the driver cavity?

Seems to me that a much thicker shelf is easily doable. Looks like the shelf could almost be doubled in thickness without interfering with the e-switch. The more meat there is in the shelf the easier it is for the heat to move through it. Plus from experience the extra weight helps balance out the weight of 4x 18650’s.

My only other real question is the deep bezel Thorfires idea or ours? Seems that space would be better utilized for deeper reflectors instead of empty space?

TA, it took a while before we got something back with an integrated shelf, the one you see now is about 4 times thicker. I think changing this to a thick ness where it aligns with the hole for the switch (I guess ~80 of the current thickness added) should be fone and would not add costs.

Bezel should also be a tad flatter

This will all be send to Barry and Daniel in our next mail IF no tech is brought up by them.
Since these are relative easy adaptations I expect them to be done.

LOL, it is only since two weeks that the integrated shelf is at all present in their design, and it has already become 3 times as thick since then. But agree, the design allows it to be even thicker. Not sure if you will see any heat path improvement from that btw.

The deep bezel is a direct copy from the traditional SRK, but personally I agree with it, first because you need some decent threading because the bezel has the added function of pusing the ledboard against the shelf, and second the glass lens is pretty big, the deeper it is sitting, the less vulnerable it is for cracking.

+1 As flat as possible for bezel. Don't want it to cause any delays though, but the shallower the better, though an SRK style light has a bezel that threads inside, not outside, so you can't get it that close, so there's always some depth. Threads could be reduced, but that's a potential issue.

For the shelf, the cavity looks pretty shallow to me, equal to the total width of the button. Maybe a mm, possibly 2 mm, could be squeezed out, but it's pretty limited already. I agree though, if anything can be done to reduce, better to expand the shelf. I know we've been pushing for a thick a shelf as possible all along with ThorFire.

Ooops, cross posted with djozz - but say'n same things basically .

I've been wanting a little space for mounting a light pipe next to the switch, so I think it's do-able. The ones I've been using with good results is 4-5 mm in size. I'd prefer it mounted a little more in front of the switch, but the design has the flat surface a little behind the switch, but it would work.

For the record I know I am nitpicking. I have this thing about wringing every possible improvement out of things. I firmly believe that everything can be improved, that includes everything I do myself, there is always something that could be improved. So don’t take anything I say personally, I will find something that can be improved on just about everything this early in the development cycle. Most of the time the improvements will be small, sometimes a bit larger.

I know dealing with china is fun to say the least, the shelf will work fine as it sits, but I figure leaving it thicker would actually reduce machining time/cost slightly and save money (be it a few cents), so I saw no reason to not do it.

Also I know some have talked about getting reliable temperature readings with the internal temp sensor, I have experimented with using a silicone thermal cube on top of the MCU directly in contact with the shelf overhead, temperature reaction time was as good as an external sensor from limited testing FWIW.

Really enjoying watching this shape up!!

The main reason that early design stages are posted is to have people like you nitpicking and bring up ideas we have not thought about :slight_smile:

That is a key aspect of the genuine BLF special. All that input and thinking along makes it a project of us all!

Nice drawings!

What is the head diameter?

Where opens the housing to change batteries? On the side of the red PCB? Or the green PCB?

Tail magnets:
If magnets go in right between the green PCB and tail cab, it seems the magnets are 2mm away from the end where it should contact with an iron surface. And there are groves at the end / outside.
I gues 2mm thickness is hard for sticking a lamp this size magnetically on the fridge door. 1mm would be better, 0.5mm deliver more force but could be a bit thin.
And it seems there is only space for 1mm thick magnets. A bit more would be nice. Room for a 3mm magnet and Insulation

I was hoping to mod a Qi-Charger in the tail cab.
But I think its hard, for that I need pos. and neg. at the end. Anny ideas for that?

The red is the driver
That is where it opens for cell removal.
Though I have seen nice mods tobtge tail PCB so that could be used

We have no exact sizes yet, but a great difference with a SRK is not to be expected, the Q8 being able to hold SRK reflectors.

Tail is Thorfire design so what er know we cannot share at the moment (but we fo not know a great deal :wink: )

But it seems to be again similar to a lot of soda can 4*18650 lights.
How big are those wireless charging components? You could maybe sacrifice the space of 1 cell and install the charging stuff there.

Sorry to be so late, but I’m in…

Here is an example of, probably well designed, cooling fins.

Triangular threads seem to be preferred for lower cost, so square fins are probably not the cheapest thing to make. It is clear from the physics that tapering the fins improves cooling both by convection and by radiation.
There seems to be room in the design to also make some of the fins deeper, but that would increase machining somewhat.

8/5 I found a better version of the image with more accessible discussion.

I can’t see the image and the site tells me I am not allowed to enter via image hotlinking :frowning:
Yes in a situation where the fins are to create convection other then square lines are used.
In our case it would mean, removing material (thus less thermal mass) and ironically less contact with the hand (and well as soon as parts of the Q8 head get warmer then the hand that holds it the blood will take away warmth. In a way flashlights are blood cooled :smiley:
EDIT: i might miss a lot, not being able to see your pic, so if I do, I hope others will chime in here :wink:

How is the picture now?

The hand does complicate things, but even thin ends on the fins will conduct much faster than the skin, so I don’t think that is a big point. The hands are very important for a one or two amp. 10440 light, but I don’t think they can absorb 50 watts or more. As far as the mas, one can taper by making the top of the fin narrower, which makes it lighter, or by making the bottom of the groove narrower which makes it heavier. The airplane engine has the additional constraint of light weight so it can get off the ground, but my intuition is that thin fins cool better. I think closer spaced fins cool better with forced air, but maybe not with thermal convection.

The fins here aren’t deep enough to make a lot of difference, but since people are talking about being picky and heat transfer is the main design goal, this seems a good time to voice my view of square fins. I think square fins are an example of a more general problem, which is that things made on numerically controlled lathes imitate things easily made on manual lathes, when numerical control can just as easily make much more general shapes. I think the reason that triangular or trapezoidal threads and fins are quicker to machine must be that there is more support and cooling for the tool, so it can cut faster.