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

Switch to pin 2

Switch LED is driven from pin 3, active high.

WW 7135 bank pin 5

NW bank pin 6

Remote thermal sensor appears to be provisioned on pin 7, which is pulled up to +V via a fixed resistor. Pads are provided to wire in a remote temperature sensor, maybe a PTC. Perhaps could instead be used to power aux-leds etc. if e.g. the fixed resistor was removed. Not too familiar with what pin 7 can and can’t do.

That’s what I think I’ve deduced from a first glance.

PS: if looking to evaluate remote temperature sensor PTCs, I’s suggest looking at the KTY range.

I’ve been using KTYs in automotive applications for decades, starting with the KTY13, they work well, and if you calibrate them individually (probably not necessary), can be extremely precise. This would be easy in a torch, a 2-point cal. say after soaking at say zero C (plastic bag wrapped, in bucket of melting ice), then the second cal. point after soaking at ambient, or a known higher temperature.

Agreed. Unnecessary for electrical, or thermal reasons, possibly damaged if the tube contact surface rides over them, and maybe structurally weakening. Imagine a heavy drop, upside down, and perhaps the cells smashing the centre out of the driver, or just cracking it.

Every via on a PCB adds to the manufacturing time, i.e. production cost. They are drilled one at a time using very expensive machinery, and drills that wear out. It is not good practice to use them unnecessarily. Simplistically put, if you can halve the number of vias in your design you can also halve the run time in the drill shop, get twice the throughput, or only need half as many machines.

The Q8 approach is tried and tested, and I think would be a better design to follow, for this important detail. Even then, early Q8s had problems with contact between the tube and the driver, mainly down to mechanical issues with PCB flatness, screws, finishing of the tube etc. which were later satisfactorily resolved. I’m simply pointing out that getting this seemingly minor detail right, turned out not to be quite as easy as it seemed.

There were many lessons learned during Q8 development and early production, hopefully there won’t be too many new ones (or old ones re-discovered) this time around.

Again, if that is a HASL finish (it certainly looks like it), and it is also intended to be used in production, then it would be incorrect for this application. It would wear through rapidly at the contact surface with the tube. It would also have to be RoHS compliant lead-free HASL.

The finish specified for the Q8 tube contact face should be used. It was carefully considered. Even then Thorfire did not follow the specified design, and what was delivered is not as good (too thin), but works well enough.

I also suggest double-checking the clearance of the components at the very edge, in the USB area. Are they definitely going to clear the metalwork inside the head ?

Otherwise it’s very interesting, I see lots of good ideas on this layout, and some experimental stuff too. Nice work.

Currently I’m on the list for one but please add me for 2 more lanterns in total of 3.

Updated OP, Thanks!

Wow! Has another stock driver been designed with more consideration for mods after buying? I may or may not get a BLF Lantern, but I certainly hope this trend in driver design continues!

No, I think that the switch LED, called indic is driven from pin 3 of the MCU, via a fixed R_indic.

The pot is associated with something different called loc. LED, and can either be permanently powered by R_loc, or instead via POT_loc and thence supplied either permanently or by a connection to the powerbank inductor, solder-splash jumper selectable.

The Loc LED circuit is shared with the remote thermal sensor pads, so with suitable configuration could drive some thing else. I see a lot of experimental ideas going on here.

It would be a lot easier to understand if Lexel would just publish a schematic diagram, and maybe a brief technical description.

the loc. Potentiometer and the resistor for it is simply connected to batt+, not sure if it will be fitted or wired to the internal switch at this point,
but having it there for user mods does not hurt

same with the NTC, it can be used without the high resistor as an addiotional LED output whatever in the firmware gets supplied
MTN uses NTCs on his firmware builds a lot why no add it to the firmware later
I voted to fit an NTC pads to the PCB it cant hurt there, even if not equipped

right now there is it on the driver it cant hurt even if its possibly not used in the final production light, but may find modders to get it working

Thank you for the clarifications, and it is very ingenious. Thank you for putting in so many innovative ideas and possibilities for further development. The large area of the driver allows so many more things to be used, or trialled, thank you for filling it up with yours.

Just one point, for accurate temperature sensing the last thing I would use is an NTC. I referred to “PTCs” because that is something easily understood, but what I really meant is silicon diodes using base-spreading-resistance to measure temperature.

Which have PTC characteristics, but are quite different. I provided a steer to a good source, many of the world’s best car manufacturers use these particular components, for good reason.

Here is another steer:

Please add me to the interest list for one.

Ok, -a thanks go out to member SKV89 for sending me four of Osram’s High-CRI emitter to test in the V1 BLF lantern prototype. (I use the V1 as the test-mule for measuring outputs, tints, etc of LEDs for lanterns)

The four LEDs sent to me for testing are the Oslon Square LCW CQAR.CC MQ-6U-L2 , MQ flux bin, 6U tint bin, and L2 forward voltage. MQ = 194 lm – 210 lm 6U = 2700K L2 = 2.90 – 3.00 V ) emitters.
They are a smaller substrate foot print than the 3535 XP sizes, ( more on the 3030 size, but they will flow onto a 3535 Cree XP board with no problems as i measured the center thermal-pad to be the same width, just the + & - pads are narrower on the Olsons.

- First positive i noticed is that the Osrams as nearly no angular tint shift, almost similar to the Nichia 119 or 219 series, but more efficient per amp input. (As almost all Cree emitters have very bad angular-tint shift, except for the MT-G2) the Samsung 351 series has almost as bad tint-shift, while Nichia and these Osrams has almost none, which is a great thing for a lantern to have full natural light tint consistency from all viewing angles.

For those wondering what angular tint-shift is, if your familiar with the horrible green tint caused by the original shallow-angled reflector Olight S20 then you know, or see this topic to understand: The CREE "Angular-Tint-Shift" flaw syndrome..It's causing brain damage.

The Lux output at 1 foot is comparable, (Samsung 351B > 42 on Medium , 167 on High) ( Osram > 43 on Medium, 165 on High)

as for tint to the eye, the Osram is much more “natural” . The Olson Square 2700K appears to my eyes as the closest light tint to what of a 25 ~ 40 watt Incandescent light bulb, probably due to the very slightly warmer tint, but much better tint emitted from all angles of the lantern, (less angular tint shift) than the Samsung 351B 3000K.
If i had a choice to have the production BLF lantern to have for the warn tint channel, it would be to use these Olson LEDS for the warm tints. I don’t know if they have a larger 5000K version, but the best 5000K tint LEDs to use for the cooler-channel is still the Samsung 351D 5000K, but definitely these Olson Square 2700K, as i really like the linear tint form the light angles.

As seen below, the greens & reds and ever other colors seem to just stand out much more natural from these Olson 2700K than it does from the Samsung 351B 3000K. (photos take with lantern on medium mode)

Yeah Den, keep teasing us while we patiently wait for the lantern :laughing:

Those osrams are special for sure!

I’m surprised that Oslon Square 96CRI compares so well to the LH351B in efficiency because the color rendering is so much superior to the LH351B. The R9 is a whopping 96. This has the best color rendering of any emitter I’ve seen besides the Optisolis. I really hope we will see this emitter in the production lantern.

Here is the test result someone posted for the Oslon Square 96CRI 2700k

Here is the test result for the LH351B 90CRI 3000k by Maukka. The R9 is only 70 and it has an above the BBL DUV of 0.0023, which puts it on the green side.

Depending on price, i would like 5 of them….if that’s ok

What is the Duv of the Osram? Generally it’s typical for warm LEDs to have a lower Duv compared to neutral ones.

I would love to know too but comparing with my other WW emitters, it’s definitely not above the BBL. Its one of the best WW tints I have seen and I have seen about a dozen or more

I am very impressed with the Osram emitter. From my testing of the ones you sent me, the output for the same amp-load is roughly the same as the 351B, but i do like the more natural rosier tint better of the Osram, because it just looks more “natural” and a better replication of an incandescent light bulb than most others i have tested. Barry of the factory also told me that those Osrams are easier to get in large numbers than other High CRI warm white LEDs, so it may work out. ( to have the Osrams as the warm-channel at 2700K and Samsung 351D 5000K (or Cree XP-L 5000K) for the cool channel as a reserve LED if the Samsungs are hard to get. (with nearly 1400 BLF lanterns awaiting to be bought, Barry’s factory has to order 5000+ of each for the warm channel and 5000+ for the cool channel.

Here is a close up comparison to what the Osram Olson Hi-CRI 2700K to the Samsung 3000K 351B ( the 351B as a smaller die than the 351D, so it compares better to the Olson die size, or somewhat to the very in-efficient Nichia 119 smaller-die series which can’t match either the Osram or Samsung variants in the 2700 to 3200K color range for lux output versus amps-load.

Below the samsung 351B on left, and Osram Olson Square on the right: ( lantern was on medium-mode, both pulling 0.74 amps from the four 3600mah Panasonic cells, (roughly 0.18.7 amps per cell, so tested to give a run time of roughly 27+ hours of light output equivalent to a 25-watt incandescent light bulb on one charge.

Update to the interest list, the next 10 entries:

1132 Plotthound
1133 niktak11
1134 ZozzV6
1135 ZozzV6
1136 jaharkes
1137 kntx
1138 kntx
1139 kntx
1140 kntx
1141 kntx

interest list sorted by entry number

interest list sorted by user names

Thanks :slight_smile: OP list updated!

DBSAR,

Here’s the source of the measurements.

The reason these are so good are the tint bins. These are analogous to a 219B SW45k. Tint binned on and below the BBL.

If Barry can get these awesome bins I’d be one happy camper. I :smiley: :innocent: