High Efficiency BLF Switching Driver? - Request for Desired Features

gchart · 2019-10-21T16:31:15.000Z

Subscribing! My 2 cents…

Output: I’m probably in minority here, but I fully support the 3-5A limit. I much prefer a constant current output instead of the FET “wow” factor that gets the flashlight blazing hot and then needs to step down right away. Sure, 3000 lumens in a tube light is neat for 25 seconds, but give me 1000 lumens for several minutes any day. For this, our traditional 7135’s aren’t bad, but regulation is limited by the voltage drop and their linear nature (no buck/boost). I’ve also used the QX7138 lately with an external FET - so far, so good, but I haven’t fully tested them yet.

Programming: I really like the simplicity that 3-pin UPDI programming brings, but that’s limited to the newer AVR chips (0 and 1 series). While I (and Mike C) really like those new chips, we’ve otherwise seen almost no traction in their adoption. It’s too bad… they’re great to work with.

WTF:

It would be interesting to hear if Gchart has one series firmware working for both clicky and e-switch lights.

I do have firmware rocking for e-switch lights using the 1-series. I’ve got some custom stuff written, but I’ve also ported TK and Tom E’s excellent RampingIOS over and it’s working flawlessly - I’m using it in my Convoy H1 (stock driver with new MCU), Emisar D1S (custom driver), Sofirn D25 headlamp (stock driver with new MCU) and others. I also hope to have Anduril ported over soon-ish, but I just need to get re-engaged on that effort. As far as clickies go… I just designed a new clicky driver using the 1-series, but found that I couldn’t use our tried-and-true noinit trick to detect fast presses. The SRAM was taking around 45 minutes to decay instead of the ~1/2 second that we’re used to. So in order to use the 1-series with a clicky, you either need to use on-time memory (yuck), OTSM (as Mike C is doing), or an OTC. To be continued…

Joshk · 2019-10-21T16:50:33.000Z

So I have an idea/suggestion for simulating the benefits of having like 8x 7135 with just the FET. But it’s so simple I don’t know why I haven’t seen talk of it.

Currently if you want to maintain a brightness that’s 60% (of maximum), you stack 8x 7135 and send them a 60% PWM signal. As the battery voltage falls, the 7135 maintain a constant brightness down to 60% battery.

Well space is limited, so why not do it like this:

Let the code monitor battery voltage. If the battery is fully charged, PWM = 60. If the battery is 60 PWM = 100%. Now for smooth transitions you want to use a multiplier instead of just if/then statements. And the perception of brightness comes into play. But with the extra code space of the new drivers, this seems quite feasible and should free up a LOT of board space on the drivers. And lower cost.

mattlward · 2019-10-21T17:02:19.000Z

After reviewing this thread again…. I only have a couple of relevant thoughts.

First off, I know you are going for sensible… I think that would be about 3 amps in a single and I really think more like 5 amps in any triple. 1 amp per led will not get most of the crappy emitters even close to out of the green!

Second thought… I think a 1S design would be ideal. Now, if possible a 1S/2S could be handy if one could dump in lithium primaries or for example a pair of 18350’s if need be.

Have been watching this discussion from the start, really hope it goes somewhere. I for one would use it in my EDC FW1A and FW3A.

iamlucky13 · 2019-10-21T17:35:55.000Z

Joshk:

So I have an idea/suggestion for simulating the benefits of having like 8x 7135 with just the FET. But it’s so simple I don’t know why I haven’t seen talk of it.

Currently if you want to maintain a brightness that’s 60% (of maximum), you stack 8x 7135 and send them a 60% PWM signal. As the battery voltage falls, the 7135 maintain a constant brightness down to 60% battery.

Well space is limited, so why not do it like this:

Let the code monitor battery voltage. If the battery is fully charged, PWM = 60. If the battery is 60 PWM = 100%. Now for smooth transitions you want to use a multiplier instead of just if/then statements. And the perception of brightness comes into play. But with the extra code space of the new drivers, this seems quite feasible and should free up a LOT of board space on the drivers. And lower cost.

That seems viable, but is a distinct concept from what loneoceans is proposing.

Another concept is the LED4Power drivers. I’m not sure how exactly they manage it, but they achieve a non-PWM output, and some of their versions have fairly high current ratings.

Joshk · 2019-10-21T19:00:19.000Z

Yea, it’s different, but I thought of it while reading this thread, so I shared.

I don’t know anything about LED4Power drivers, but from what you stated, I imagine they use a very high PWM frequency, then maybe even feed that into a capacitor to go analog.

Barkuti · 2019-10-21T19:50:38.000Z

Both Led4Power and the latest drivers available at the Convoy store use MOSFETs as linear regulators. They use their microcontroller to monitor the voltage drop at a sense resistor, and according to this they precisely tune the MOSFET gate voltage for it to exactly drop the required voltage delta. This means the dissipated power at the MOSFET is voltage differential times current, being this the main driver hurdle (dissipated power in MOSFET and driver, nothing to do with amps rating here). (^̮^)

iamlucky13 · 2019-10-22T01:30:09.000Z

Barkuti:

Both Led4Power and the latest drivers available at the Convoy store use MOSFETs as linear regulators. They use their microcontroller to monitor the voltage drop at a sense resistor, and according to this they precisely tune the MOSFET gate voltage for it to exactly drop the required voltage delta. This means the dissipated power at the MOSFET is voltage differential times current, being this the main driver hurdle (dissipated power in MOSFET and driver, nothing to do with amps rating here). (^̮)

Ah, that explains why he talks about being able to dissipate 15W if you use a seperate FET board.

My deepest dive into circuit design, not being an electrical engineer, involved accidentally dissipating too much power in a FET, due to not having a pull-down resistor to drain the gate voltage while PWM’ing it - the FET was spending some time every cycle with its gate voltage in an intermediate, partial on state. I burned up several FETs this way before a tech helped me figure it out.

My experience makes suspicious of that approach, but I could see it work if there is an appropriate amount of cooling.

JasonWW · 2019-10-22T02:05:36.000Z

mattlward:

First off, I know you are going for sensible… I think that would be about 3 amps in a single and I really think more like 5 amps in any triple. 1 amp per led will not get most of the crappy emitters even close to out of the green!

(Generally speaking) This is only true if you are talking about a linear 1 amp. If you pulse a FET to create a 1A load your tint will be closer to what you see if the FET was on 100%, like 6 amp or whatever. It’s because you are literally getting a fraction of that 6A. So it does matter how you get that 1A.

Joshk · 2019-10-22T01:48:36.000Z

JasonWW, I would think so too, but look at the 219b plot in my signature, the tint is wildly rosy with the FET at 100%, but not similar when it PWMs.

JasonWW · 2019-10-22T02:04:58.000Z

Joshk:

JasonWW, I would think so too, but look at the 219b plot in my signature, the tint is wildly rosy with the FET at 100%, but not similar when it PWMs.

My statement above was pretty general, like for most leds. Maybe the 219b is an outlier. The A6 driver is FET+1, so Med level is definitely using the FET. If you are a person who wants to build a light that keeps the green away you would probably want it to have a FET only (or FET+1 if you want very low moonlight levels) and then match it up with the right led, probably not a 219b. I wish I knew more about the subject, but I only know a few things. Lol

Texas_Ace · 2019-10-22T13:17:14.000Z

Joshk:

JasonWW, I would think so too, but look at the 219b plot in my signature, the tint is wildly rosy with the FET at 100%, but not similar when it PWMs.

That is really strange, what driver was being used? Sure it didn’t have a second channel? My testing and commercial lighting practices says the tint should remain very close when PWM’ing and everything else being equal.

Except for super low PWM levels where the FET might not be turning on all the way or in setups that use very slow FET’s or purposely slow the FET down. The tint should be quite close to turbo tint when PWM’ing the FET and all else being equal.

The PWM pulse will actually be feeding the LED the same current as turbo, but only for a fraction of a millisecond.

contactcr · 2019-10-22T13:37:44.000Z

Texas_Ace:

Joshk:

JasonWW, I would think so too, but look at the 219b plot in my signature, the tint is wildly rosy with the FET at 100%, but not similar when it PWMs.

That is really strange, what driver was being used? Sure it didn’t have a second channel? My testing and commercial lighting practices says the tint should remain very close when PWM’ing and everything else being equal.

Except for super low PWM levels where the FET might not be turning on all the way or in setups that use very slow FET’s or purposely slow the FET down. The tint should be quite close to turbo tint when PWM’ing the FET and all else being equal.

The PWM pulse will actually be feeding the LED the same current as turbo, but only for a fraction of a millisecond.

He was using BLF A6 FET+1 so, as you say, multi channel. Also driving a 219B as a single LED w/ FET is hardly a good case study since cooling is inadequate.

Joshk · 2019-10-22T15:45:27.000Z

I looked at the reports I posted, and yea I see the 7135 is still running for Med. but not for turbo. Probably not a relevant example then. As far as cooling, there was no heat build up during the 1 second it took to collect the data.

KaraPelosi · 2021-02-05T15:37:23.000Z

Hi….as far as I might be concerned, 3A in a light that size is somewhat needless excess, yet I’m unquestionably not the electric lamp nerd that would get the most happiness out of this driver.

The Anduril firmware would likewise be needless excess for me (in the wake of watching a survey of the FW3C), however I do like the sloping capacity. I fabricated a dimmer for my undercabinet lights that works that way (tap for on/off, hold for incline) when I assembled my kitchen numerous years prior.

order pcb

loneoceans · 2021-02-05T18:34:53.000Z

Thanks to the comments, suggestions and feedback from all in this thread! This discussion has help shaped the Lume1 and Lume X1 switching drivers running on Anduril.

Lume 1 - https://budgetlightforum.com/t/-/60947

Lume X1 - https://budgetlightforum.com/t/-/65319

gchart · 2021-02-05T18:37:01.000Z

Hey LO, is the new Fireflies driver an exclusive to them, or will you posting details about it here?

loneoceans · 2021-02-05T23:01:33.000Z

gchart:

Hey LO, is the new Fireflies driver an exclusive to them, or will you posting details about it here?

The Fireflies driver was designed to their specific requests / functional requirements, so it doesn't quite represent what I would design for myself or my own flashlight. I believe it is up to Fireflies' prerogative to post details about it so please feel free to reach out to them if you have questions. Since I designed it, and because the driver is part of a production flashlight, I think it's best for me to refrain from commenting about its performance or express my personal opinions to avoid biased. I see that some people have started writing some reviews about it and comparison with other drivers in terms of performance and efficiency, such as here: https://budgetlightforum.com/t/-/64022/349, and I'll be interested to hear feedback and comments. I should be able to answer general questions about it too. :)

gchart · 2021-02-05T19:27:37.000Z

Understood! While I like the Lume1 with it’s 3A + FET setup, I’m with you in what you’ve mentioned before that you prefer more control than just an unregulated FET channel. If I were to come up with my “ideal” driver, it would be an adjustable current buck (likely topping out around 5 or 6 amps) with good output resolution. None of this FET+1, multi-channel, depends on the LEDs and their quantity non-sense. I like the Lume X1, but there are plenty of nice 3V emitters that I wouldn’t mind running on their own, no boost required. I’m probably in the minority around here. But it seems like with all these FET drivers, we’ve got like 15 seconds of blazing brightness, followed by 2 hours of 20% output because the host is thermally soaked and can’t cool down. I’m a bit tired of that. [/rant]

thefreeman · 2021-02-05T19:34:37.000Z

I fully agree, and the E12R for example would have been better with a boost driver.

Regarding 3V single LED flashlights it’s too bad there aren’t more powerful buck-boost converter (integrated switches) than the TPS63020

Scallywag · 2021-02-05T21:42:23.000Z

gchart:

Understood! While I like the Lume1 with it’s 3A + FET setup, I’m with you in what you’ve mentioned before that you prefer more control than just an unregulated FET channel. If I were to come up with my “ideal” driver, it would be an adjustable current buck (likely topping out around 5 or 6 amps) with good output resolution. None of this FET+1, multi-channel, depends on the LEDs and their quantity non-sense. I like the Lume X1, but there are plenty of nice 3V emitters that I wouldn’t mind running on their own, no boost required. I’m probably in the minority around here. But it seems like with all these FET drivers, we’ve got like 15 seconds of blazing brightness, followed by 2 hours of 20% output because the host is thermally soaked and can’t cool down. I’m a bit tired of that. [/rant]

My ideal driver would probably be a 6A buck/boost with anduril and good moonlight modes. Share the oshpark builds and component lists (a digikey or mouser cart or list of links is even better) in a few key sizes and I’d be set for years.

Problem is, I imagine it’s tough to fit all that, plus flashing vias, in 17mm…

At this point I’d settle for some open-sourced linear drivers that do moonlight well, yet depart from 7135s. Bonus points if it can be easily adapted to other voltages (looks longingly at my SP70).

Note: I still like the boost capability because I still want turbo if my battery is low, and I’d also like to be able to pump some amps through LH351D.