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

Thanks. Makes sense now.

1) Output would be spatially uneven
2) Perceived brightness does not have a linear relationship with power
3) Going off what TheMiller said, there’s higher luminous efficacy when running multiple LEDs at a low current as opposed to one LED at max current

To individually control 7 LEDs, it would need 7 independent power channels controlled by 7 pins on the MCU… or a completely different design. The MCU only has 5 usable pins, which are currently used for:

  • 1x7135: Low modes
  • FET: Med/high modes
  • Indicator LED
  • E-switch

This leaves one, which often gets used for a 7x7135 channel for better medium modes. Or for measuring battery voltage. Or for another indicator LED color.

There are other reasons too:

  • What Miller said about efficiency. Using 1 amp spread evenly between 7 LEDs is more efficient than 1 amp sent to only 1 LED.
  • If it had 7 levels, at 1/2/3/4/5/6/7 amps, it would effectively only have a narrow range of medium/high modes. To achieve levels which appear evenly-spaced, the power use ends up rather uneven, like the graph of x**3 instead of a straight line.
  • Turning on/off individual LEDs in a 7x SRK would make the beam less symmetrical due to the overlapping reflectors.

If we have 1 free pin, why do we not have a 7x7135 channel? Firmware size limitations?

Put me on the list as well please, then I can junk my chinese knockoff SRK.

I’m not actually involved in the project, so I don’t have the details. If I understand correctly though, it’s because:

  • There wasn’t a free pin until recently. (or there may not be a free pin even now?)
  • Three-channel drivers weren’t supported by any firmware when the driver was created.
  • Ramping is generally less smooth and more fiddly with more channels.
  • It’s too late to change the driver.
  • The extra parts could raise the price.
  • Diminishing returns; going from 2 channels to 3 is a smaller upgrade than going from 1 to 2.

Basically, the extra complexity probably isn’t worth it right now.

Hmm well I think Narsil has grown a lot the last year
Now Tom has made a Narsil version for 1*7135/7*7135/FET drivers
Solid triple channel drivers like TA drivers are the result of combining PD68s trple channel drivers and DELs input on stability and parts for the Q8 driver.
NOW at the start of a FET based driver light triple channel is obvious, hence this is planned in the FW3A :slight_smile:
When the Q8 was started people were very happy with the Kronos X5/X6 drivers with FET + 1 so it was the logical choice at that time.

Ehm wait the led in the switch uses a pin too right?

Yup, one pin for the e-switch, one pin for the indicator LED. Unless the indicator is more than one pin, like if it does more than one color. I forget. IIRC it was just one color, and there were no free pins until the recent switch to using VCC for measuring voltage. If I didn’t imagine that. I remember all sorts of things which didn’t actually happen.

early onset prescience, don’t fight it :wink:

The Attiny85 has 8 I/O pins:

  • reset
  • Grnd
  • VCC
  • 5 I/O pins

I had 2 different code baselines: Narsil (2 chans) and Narsil Triple (3 chans). It's now recently has been combined into one code base I'm calling "NarsilM" for multi-channel support. These two versions were diverging and getting difficult to keep both up to date.

Your hardware can be specified at compile time.

For an SRK, yes, we can do 3 channels, a switch and a indicator LED (battery voltage monitoring can be done internally). But for 2S uses (like XHP50's or XHP70's), we need external battery voltage monitoring, so need another pin - an "LDO" must be used to reduce parasitic drain. Right now, the indicator LED is sacrificed.

So NarsilM can support it all, accept for the combo of 3 channels, LDO, and an indicator LED. Bad news is for 3 channels, it's very close to the limit of program space. I haven't released NarsilM yet because of needing further testing of all the configurations, and it would help to do some optimizing to reduce the code size.

Back a year ago, yes - 2 channel was the way to go. Not sure, but think 3 channels was just coming out, maybe not fully supported yet. We committed early to the electronics design though.

Looking at the logs, the first released 3-channel code (that I’m aware of) happened about 6 weeks after the Q8 project started. I didn’t merge it into the stable branch until another 4 months later. So, yeah, it didn’t really exist as an option at the time.

Would future firmware updates be a possibility? Not on the production light, of course, since I guess we are too far into the process to make such a huge change, but as a DIY updgrade?

What is an LDO?

3 channels already existed on the H17F driver, but I don’t think DrJones released the code, did he?

All of this talk proves to me that trying to get into programming a few years ago wasn’t for me. I haven’t got the attitude for it.

It also proves that all of you guys are damn great at programming. I read the documentation for a few of Toykeeper’s firmwares back when I was learning to configure biscotti and bistro, and it’s incredible the amount of features she can fit on the same hardware that is normally limited to “low, mid, high, strobe, sos, off-time memory”.

EDIT: Low dropout regulator. It lowers the voltage from the cells to the attiny, to increase efficiency. Is that right?

Would future firmware updates be a possibility? Not on the production light, of course, since I guess we are too far into the process to make such a huge change, but as a DIY updgrade?

Yes - 2 screws to pop out the driver, need a download cable, 8 pin clip, and USBAVR adapter. Download the HEX file

What is an LDO?

Low Dropout Regulator: we use it to drop incoming ~8V from two cells in series to under 5V to power the MCU. Could be used for 4S input as well (~16V).

3 channels already existed on the H17F driver, but I don't think DrJones released the code, did he?

No - Dr Jones doesn't provide source code to his main line firmware versions he sells, but he was instrumental in providing one of the original sample source code drivers

The Round 3 Prototype arrived today, wiating for me at home. I believe these are true V6 3D's based on the beam tint and output.

Here’s some measurements:

On SONY VTC6 soldered tops at 4.21V, lumens: 6,375 @start, 5,850 @30 secs, throw: 59 kcd (486 meters)

On Samsun 30Q soldered tops at 4.19V, lumens: 6,260 @start, 5,800 @30 secs, throw: 59 kcd (486 meters)

On Samsung 30Q button tops @4.19V, lumens: 6,090 @start, 5,620 @30 secs, throw: 58 kcd (482 meters)

The new designed plastic ring

LED's, not perfectly centered but not bad. Centering pieces fit tight:

Center screw removed, 18 AWG LED wires, 2 mm thick MCPCB (round #2 was ~1.2 mm thick):

Driver views:

Round #3 still has great square threads:

Engravings improved:

Round #2:

Round #3:

Round #2:

Round #3:

4X SRK clone (Securitylng) on left with XM-L2 U3 3D's, BLF Q8 on right. Beamshots appear more bluish than they really are

4X SRK clone (Securitylng):

BLF Q8:

Thanks a lot for taking the time to answer.

I have some experience with the H17F, and I must say, it is the most configurable driver/firmware I have tried so far. Really impressive.

Is there a readme or guide for more current/advanced versions of narsil? I kinda want to see what features to expect on the GT, although the stage of development is a lot earlier in there, so I guess nothing is set on stone.

The driver should be easily reflashable if you have the tools to do so, and the tools cost perhaps half as much as the light. So, yes, you should be able to modify it.

As for drivers with 3 or more channels… Wight released a 4-channel one in mid-2014. As far as I’m aware, no one ever built one and there was never any firmware for it. His FET+1 layout was much more popular. I think Mike C then made one, but only for his personal use. Fast forward a while; DrJones and pilotdog68 both released one within a couple weeks of each other at the end of 2015. A version of blf-a6 could run on it, but the FET was turbo-only with no PWM for dimming. Then the Q8 started in 2016-04. In 2016-05 I got one of the 3-channel drivers and promptly made bistro-tripledown, which IIRC was the first open project with 3-channel PWM. I released it in a dev branch but didn’t merge it into stable until 2016-09. Mike C continued his designs well ahead of most of what BLF was doing, and even released some including code, but they haven’t been part of a big project so they’re not widely used. More recently, several 3-channel designs are available and multiple firmwares support it, and there’s even a new project which should make it possible to buy a light with one.

So… progress is happening, just slowly and it has been a little disorganized.

The LDO is just a way to power the attiny on higher-voltage lights without exceeding its voltage spec, like a zener but with lower parasitic drain.

Right at the beginning of post #2 is a link to the google drive share that has the manual in docx and PDF format - it's 4 pages.

Smooth ramping, quick access to lowest level ,highest level and last level used. 12 sets of modes are featured, plus has timed stepdown or temperature stepdown, all configurable. Lots of options can be configured.

It is gorgeous and looks strong like a tank :heart_eyes:

The FW3A, right? Already on the list for that one too.

That 3rd round prototype looks nothing short of perfect. In fact, there are some improvements that I thought had been given up, like the 18 AWG wire, instead of 20 AWG.