I’ve had a low-severity light addiction for many years. I never get into the > $100 lights that some of you love, but I can’t resist cool $30-$40 lights.
That’s why I’m loving the S2+ (that I only recently discovered). I know it’s old news to most of you, but to me, this is a nearly perfect light. It’s compact, uses 18650 s which I have hundreds of, and it can be both a modest practical light, or an “OMG how is that much light coming from that tiny flashlight” show stopper, with the SST-40 + programmable 12-group 8*7135 driver.
But here’s what I don’t get… why the heck is it using a linear driver and not a switching driver. I haven’t really paid attention to driver tech in probably a decade, but even 10 years ago, 7135s were already considered outdated. Yes, they are dirt simple to parallel on a board, but they’re so inefficient.
Looking at the voltage curve for an SST40, at 3A, you only hit 3V. That’s means with a linear driver and a fresh 4.2V 18650, you’re turning 25% of your power into driver heat. That’s bad for runtime both in terms of battery capacity and overall heat.
7135 chips are small, but 8 of them still take up space, and you should be able to fit an SMD inductor plus switching controller, mosfet, current sense IC and diode, in the same space.
So why is it so hard to find an S2+ with a switching buck driver?
I think that’s actually what I have. I thought it was 8*7135 for the normal modes, and then direct drive (or through a FET) for the 5A. But I misunderstood.
Even still, it’s a linear driver, so I’m still wondering why there’s no switching driver option.
Are the schematics for that 12 group driver known / open source? I’m tempted to go find my decade old copy of EagleCAD and try to make a switching version.
That is a pretty efficient layout. But I think you could get those components onto a 17mm board with a few tricks like vertically stacking those two sets of two caps, and moving some of the ICs to the underside around the spring.
Is the microcontroller on Convoy’s 17mm 12 group driver a known IC? Is it just an ATTINY or something?
I’m really tempted to take a buck driver design I already have and try to make it fit on a 17mm board. I just don’t have the equipment (or desire) to fiddle with the microcontroller part of it. It would be great if I could just buy a Convoy linear S2+ driver and reflow the chip off it and onto my own board. Of course I’d need to know what the chip is first, and which pin it uses to drive PWM.
Or, alternatively, if Convoy’s 22mm schematic is known, I’d love to try and remake it at 17mm, and then just buy the 22mm board from them and reflow + move the major components to a 17mm board ordered Oshpark (I’d probably just re-buy they small resistors and caps).
If Convoy’s driver schematic and microcontroller IC aren’t know, are there any decent open source projects that have the microcontroller group/mode stuff already figured out?
That doesn’t sound correct to me. It shouldn’t be that hard to fit a switching circuit on a 17mm diameter board.
I’ve been looking back through old buck driver designs and I have several high current buck drivers that I made two sided at 9x9 and 12x8 mm. Even one side of a 17mm round board has more real estate than that. I don’t have any designs based on ICs that run from a single cell. But I just spent a bit of time looking at ICs on Digikey, and they’ve come a long way since I last looked .You can get tiny ICs in the 3mm x 3mm range that have integrated 5A+ switches. The LTC3309 and RT8073 both caught my attention.
The RT8073 comes in a 3mm x 3mm package, and at 2MHZ, it could use a pretty small inductor. A 6A switch is integrated so it doesn’t need a mosfet. It would really just need the RT8073, current sense IC, microcontroller for modes, caps, resistors, a diode, and a decent inductor.
Is anyone doing an open source driver project these days?
Just the board itself would be a small fortune to be able to accommodate a 3mm×3mm package. You’re already into fine-pitch traces, probably multilayer, etc. Those chips are probably made that small to pack into 1kbuk cellphones and such, not 10-15buk flashlights.
You already can get an 800mA 1-/2-AA driver on a 17mm board, but the inductors are small, there’s no µC, and it’s open-loop output, ie, no regulation to speak of. But if you want those extras, yeah, that’s more real-estate.
And if I had my druthers, I wouldn’t want any components on the spring side.
Convoy 8A buck driver uses the MPQ8612-12
While it’s a high current converter with fairly low Rdson FETs I don’t think it’s a good choice for flashlights because it’s not 100% duty cycle capable. The MCU is from Sonix, the firmware is proprietary.
17mm buck driver is possible of course, especially with high frequency buck converters. This is a 18mm, 14mm clearance one I made for the FWAA , it’s double sided but it has a RPP PFET and 2 sense resistors for high dynamic range and Aux LEDs, without this it could be one sided.
I used the TPS62867 (11mΩ,10.5mΩ, 2.4MHz, 6A), it is AFAIK the best buck converter for 1S to 1S as of today, unfortunately all TI DC-DC offering is out of stock until 2023. Another good option is the MP2145 (20mΩ,12mΩ, 1.2MHz, 6A), I designed a 22mm, 18mm clearance one and tested it, also one for the FWAA, with 16mm clearance :
Again getting rid of HDR, RPP FET and Aux LED pads it would fit in 15mm clearance (17mm total diameter) on one side, maybe even while keeping RPP.
Here is the schematic :
(Q2 and R7 can be removed if moonlight modes aren’t needed).
Note that it uses the DAC of the T1616, removing the need for a LDO, battlevel resistor divider, LPF filter for PWM, Convoy drivers all use filtered PWM to set the current sensing voltage and so have these additional components.
Of course for the S2+ you need to add either OTC or OTSM for the clicky switch as well as use an appropriate firmware, since my drivers are for e-switches. Gchart metionned in the Attiny 1 series thread that he tested a clicky firmware IIRC, using PWM, I don’t think work as been done on a DAC controlled T1616 clicky firmware.
That RT8073 looks pretty bad, no way that it can do 6A with 50 and 35mΩ FETs.
Not at all, plenty of cheap services are fine for this, Oshpark is 6mil (0.1524mm) spacing and trace width and cost 2.25$ for 3x 17mm boards, it’s virtually free, JLCPCB is 5mil IIRC, Aisler, 0.1mm, it’s largely enough. In my drivers the smallest pitch is 0.4mm (0.2mm trace and spacing), the ICs cited by rhd have a 0.5 and 0.45mm pitch. Basic 2 layers boards are sufficient.
It can not be a buck driver, it has to be a boost buck for 1 cells, and they are more complex than linear. What practical advantage would such driver provide with 1 cell for power?
Yes, it does. however, will go into boost mode from a li ion cell? i had a similar driver a decade ago, shining beam used to sell so caller “perfect driver. 0.9-4.2v, it would work fine from a alkaline cell, and it worked great with li ion, until it dropped to about 3v, after that is would start blinking (a warning for cr xxx) cells, and it would turn dimer and dimmer, it would not go into boost mode on a li ion cell, looked like it went straight to DD.
@thefreeman: This is where me not being in tune with the state of flashlight tech comes back to bite me. I was thinking very simplistically about just a basic buck driver with modes added by a microcontroller like an ATTINY85. I wasn’t as up to speed on all the other features (like aux leds, battery level monitoring, etc) that are now apparently common in drivers. Even for mode memory, I wasn’t thinking about OTC / OTSM. I thought it would be simple to just read the saved mode from memory on startup, then increment it to the next setting immediately and store that to memory, and then reset it back to the original mode after x seconds. Admittedly this creates 2 writes to the EEPROM every time you turn on your flashlight, which isn’t great. I see that the driver world has better solutions for this.
I also wasn’t thinking about something as complex (better) as your OP AMP solution for FB. I was just thinking of using something like ZXCT1009 to set a constant current, and then PWM the EN pin of the switching regulator from a pin on the ATTINY. Yes… probably not the best way to do it, but simple.
I guess, bottom line, my idea is too simplistic and hacky for the times.
I think it can probably be a buck driver for one lithium cell. Looking at the IV curve for the SST40, at 2.8A (the amperage of an 8*7135 config), forward voltage is under 3V. I don’t typically drain my 18650s under 3V, so I think a switching regulator could buck for most of the discharge curve of an 18650, if at moderate currents (this obviously stops being true if you’re trying to run at 5A, but I think that runs into thermal limitations of the tiny host pretty quickly).
But, no doubt, a boost/buck would be better. The practical advantage is increased run time and less heat. This is hacky math, but with a 3.5 AH cell with a nominal voltage of 3.7V, an emitter current that produces a drop of 3V, and a linear driver, you’re burning 0.7V * 3.5AH = 2.45 WH of energy as heat through the driver. That’s a 12.95 WH hypothetical cell, so you’re losing about 19. With a well designed switching regulator, you might get that down to 10.
I’ll admit though, I was dissapointed when I did that ^ math just now. I was expecting to be able to give you a more dramatic comparison.
No it doesn’t have to, here is the SC64LE tested by Zak, cooled so that it doesn’t thermally step-down, it has a 3A buck driver :
Yet it still outputs 70% lumens just before LVP step-down and the decrease only starts at ~80% of the runtime before step-down. Consider that the LH351D doesn’t have the lowest Vf, 219C and 519A are lower for example and would be even less affected.
Another thing to consider is that the drivers I showcased have lower dropout voltages, the TPS62867 one have a resistance less than 40mΩ, I haven’t measured the SC64LE but It could be double this amount (TLV62085, 31mΩ HS FET, 15mΩ RPP PFET, 15mΩ inductor, and they use fairly high sense resistor usually).
Buck-boost converters have much higher switch resistance and low switch current limit, in the end they don’t do much better at low Vin while being less efficient.
While a boost driver is the best choice for highest sustained output over the longest period of a time, a good buck driver is a close second at lower currents.
