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

As an estimation to the price range, I would say (BLF Q8 GB price-cost cutting=final price)

Also, DBSAR, add one lantern for me too. Add me to the list.

My eyes are OK, but type C is much more convenient for me as well.

Interested!

Added! : )

type c for me too, you can always get an adapter.

ken

Updates, April,8,2018:

- Ok, I have the new lantern head for the V2 Q8-based prototype completed! I reflowed two 3000K XP-L emitters to the original modified Q8 copper star, (i only had two of these bare LEDs so i could not use all four pads on the star.)
Like the V1 design, i have used a full flat/matte white upper & lower plate light reflector, which works far better at increasing the output through 90 degrees of the lantern globe, while reducing the glare found in most other factory built LED lanterns on the market. Most lights you find on the market commercially make the mistake of using a chrome or shiny “reflector” causing an increase in glare, hard shadows, and does not really work to put it in simple terms. I have tested every possible surface type form gloss, eggshell, matte, chrome polished, and the full matte/flat white is the only one that emits the most lumens, reduces the most glare, and smooths out the hard-shadowing the best.

  • So… right now I am stuck at the driver, and how to safely lower to the maximum amps to the LEDs…

Below, is the new V2 Prototype head ready for driver modification & first testing, (take notice of the center post bolt design, Much stronger & simpler than the V1, and there is a “sleeve tube” over the center bolt that hides the LED star wire leads, ( the tube is also painted flat-white to aid in reflecting light 90 degrees out the sides”

Below showing the solid aluminum head, with the modified Q8 copper star with the two 3000K XP-L 7A LEDs mounted in place using screws & thermal paste, and the surrounding surface painted flat white on the head, along with the base painted the same flat white. This feature seems to create an increase in gathered light, as the light itself is reflected back & forth against each other, increasing the efficiency of the 360 degree output, and smoothing out the hard shadows in the process while still eliminating the eye glare found in most lanterns:

The lower photo is of the driver, whee I am currently bogged-down on how to lower the total amps while preserving all the Q8 modes for the tests, ( to roughly the equivalent of two or three 7135 chips, but not sure of how to do that yet… ( Hinting for the masters of the Q8 driver design & build to chime in here, ( along with ToyKeeper as right now with all other team members MIA ToyKeeper is the only knowledgeable member we have helping with the BLF Lantern build for the driver & firmware part…

Looks nice! Thank you for the update.

You’re doing some very impressive design work here, DBSAR :beer:

Looking at the star with 2 LEDs on it, I just had a random thought: would it be worth having two 3000K LEDs and two 5000K ones?

I have a picture in my head of two Convoy S2+ tailcaps set into the top of the lantern, allowing the user to physically switch the two colour temperatures in and out separately. You could have 3000K for relaxing in the evening, 5000K for a work light, or an extra efficient “4000K” mix getting a few more lumens per watt by spreading the power over 4 LEDs.

I have no idea whether any of this is possible or sensible, mind :slight_smile:

DBSAR…what if you added a white wall around the LED meant to make it invisible from the sides?

Interested!

Add me to the list :).

If you use just two batteries in the battery tube, together with two XP-L’s it should be good to go for testing: too much output on max but he leds should be good.

Or else, you could use a simple resistor in the circuit to tame the driver, try 0.1 Ohm for a start. If you use an old-school 0.1 Ohm 1W capable through-hole resistor you could even use it floating, soldered in the middle of one of the led wires. That will eat up a few 10th’s of a V, drastically reducing the current. It is how many drivers of this type are tamed in stock flashlights.

Electronics experts please correct me if I’m wrong.

I thought the Q8 driver was a FET +1 design. If so, the thing to do is get a two channel driver that uses 7135’s for both channels (or a three channel driver). Have it configured with one bank of just a few 7135’s and the single 7135 “bank” and leave the rest of the spaces unpopulated. Get Narsil installed on it, and you have functionally the same mode sets and UI as the Q8 but with a lower max regulated output. This will allow you to test run-times more accurately than if you just try to hobble the FET on the Q8 driver to make it put out less current. I think Texas_Ace or Lexel could help you out with that driver, if you tell them what you’re trying to accomplish.

FET+1 is not really a problem when adding an inline resistor, if the driver is using the 7135 just some heat-burn-off will be transferred from 7135 to the added resistor, and the ramping curve will be very slightly affected but I doubt that you will notice it.

The V1-Prototype had that design, ( if you go to the OP and have a close look at the photos of it you can see the white band around the top of the lens) but with a higher percentage of frosting for the lens it would not be that noticible.

Interesting idea indeed. Not sure how much it would add to the cost of the lantern overall to add a switch to the top and have two different tint sets of LEDs.

I thought of that resistor method which is what i done to reduce current in most of my driverless lantern mods. But with this thing putting so much amps out i[m unsure of the heat issue with a resistor getting to hot, or of which one to use.

Ok, I just thought ‘regulated’ would be inherently more efficient than a hobbled FET, which is ‘not regulated’ AFAIK. But then, I don’t know much, so the likelihood of my being wrong is great! :stuck_out_tongue:

I’m not sure how physically feasible it is to put a stack of 7135 chips on the FET pad. With the 7135’s middle/wide pin up, IIRC, that means from left to right the pins are LED~~, GND, PWM. Meanwhile, the FET has four pins and a big pad. The three leftmost pins are GND, the right pin is PWM, and the big pad is LED~~.

Adding a Moonlight Special driver as a slave board might be an option, for testing. Or a slave board with only 7135 chips on it, wired to/from the Q8’s FET pad. Or using a FET+N+1 Q8 driver without the FET. Texas_Ace made a nice 46mm one:

https://oshpark.com/shared_projects/tEWsX9HP

Or Lexel sells some pre-made:

He might even be willing to put Anduril on it. The ramping table will be a bit messed up with the second power channel so low… so it’d probably be a good idea for me to make a special build for it.

A floating inline resistor is how lightsaber drivers work. They’re basically direct drive with a resistor on the low-Vf channels. On those designs, the resistor has a tendency to turn into smoke… but the Vf of a red XQ-E is a lot lower than XP-L.

It’s definitely an easy solution for testing purposes, though the output won’t be regulated at all.