Good to know, thanks!
Curious to know your thoughts about the design of Firefly Lite’s “High efficiency 6A constant current buck driver plus FET Turbo,” and what you’d suggest for improvements if any?
Pros :
- It’s an easy way to get very high turbo output (with full cell) while still having efficient low to high output.
- It’s compatible with any number of LEDs/LED chips, a boost driver is limited to 2S and more but not number not divisible by 2,3,4 and 5 (for the MP3431 and most low Vin converter they can’t boost higher) for example it can’t work with the E07x.
Cons :
- It only works with a buck or buck-boost converter for the regulated channel, available buck-boost ICs can’t really do more than 3A and even drop a bit at very low Vin. With buck the current drops at some point, now because I tested the E12R driver with 4xSST-20 it’s not representative of the Vin when it stars to dropout since 12 LEDs would have a lower Vf. The large amount of LEDs (7/12/21) in lights from Fireflies with this driver makes it less of an issue.
- The abrupt step-down from turbo to regulated
Personally I don’t really like unregulated output so I haven’t made any of these types of drivers.
An improvement would be to get smooth dimming from turbo. The problem is that they can’t be ON at the same time, it’s not like linear+FET where the FET channel is in parallel to the linear channel, and we don’t know when to switch from the FET channel to the regulated one since the output current is unknown in direct drive,
It should be possible to read the FET current with the sense resistor if it was wired to an ADC pin and switch to the regulated channel when the average direct driver current reaches the max regulated current. But that’s for the code guys to do as I don’t have any programing skills.
Yes. I’d personally rather have a low resistance high efficiency buck driver over a FET driver at this point.
At this point, I’d rather have 6A buck + 12A linear.
The firmware would be weird, but it would work.
No problem doing buck+linear firmware (Anduril) wise, and that would solve the abrupt stepdown. But it’s an additional regulated channels with all the components that come with it.
Linear drivers can be problematic when there are a lot of LEDs (very low VF), as the power dissipated in the FET can be significant, it needs to be able to withstand it.
The linear channel current could be higher like 20A and it would have direct drive like output.
^ Thanks for the explanation! Given the component sourcing problem, I expect they won’t change revise the E12R in the near term… assuming it’s something on the books to address.
Well they are apparently releasing an updated version but I don’t know the details.
I could really go for high-output Buck drivers with 519A these days
shame about inductor size limitations
For a single 519A a TPS61867 or MP2145 based driver, both with 6Aout max are suitable, they have a fairly high switching frequency of 1-2.4MHz and 1.2MHz respectively which allows the use a small sized inductors with low inductance of 330~470nH, a SRP4020FA 470nH is 6.8mΩ max which is pretty good. In my FWAA TPS62867 driver I used a 560nH one of 8.2mΩ.
The TPS62867 is the better one, actually it’s the best single cell input buck IC on the market, but they pushed back the availability (again) to 2024, which is just extremely infuriating at this point. Anyway the MP2145 is still very good, but a bit bigger, slightly less efficient and more costly.
I have two standard (round) driver designs with the MP2145, a single sided 18mm clearance and a 16mm one with the RPP FET on the back + one resistor, both with 40x0 inductor size, for maximum efficiency an XGL4040 470nH of 3.2mΩ max can be used if the driver cavity is >4mm.
The MP2145 just came back in stock on Digikey/Mouser so I should make those boards available.
At some point I also want to have drivers assembled, I’m still hesitating which model to start with, a single sided (less costly assembly) driver like the 18mm clearance one could be a good candidate.
how about the zebra SC600w Mk IV Plus 18650 XHP50.2 Neutral White Flashlight
I don’t have that one for testing, the difference with the SC700D is the smaller 5030 inductor vs 7030, and the lower drive current (not sure exactly how much). The efficiency should be in between the H600 and SC700.
cool ty i love this chart so much i love details and efficiency is my faverite feature in lights well that and small size
thanks for your driver’s tests.
Are you tested Convoy S21D 8A driver ?
Can you point me to which resistor and which value you used to bump this up to 2.5A? Thanks.
Either change R4 to 255k or the 020 resistor (R10?) To 16mΩ.
Awesome, thanks. What is the SMD resistor size of the current shunt?
0805
Edit : I think I saw a different version with a different sense resistor value though (and so different R4 as well), so better check your driver before ordering resistors.
Yes, my D4K driver is different than the KR4 driver (most obvious is the MCU on the opposite side). That’s why I figured it was safer to swap the current shunt.
Different layout but the reference designators are the same, the R4 value is different though, 124=120k and with R3=4.7k
Vsense=2.8×4.7/(4.7+120)=105.5mV (2.8V being the LDO and PWM voltage)
That means Vsense(R10)=105.5/2=52.75mΩ, that’s not a available value though so I guess R10=50mΩ and the current is actually 105.5/50=2.11A.
Anyway NewRsense=Rsense×Iout/newIout = 105.5/2.5 = 42.2mΩ
Thanks for the calculations. Once I remove the driver I can verify the sense resistor value to see if it is indeed 50mΩ.
I measured the voltage (at max level) at the MCPCB on two Hank boost lights. One is a 519A quad and measured 12.2V (4S MCPCB so 3.05V * 4) so that one can probably be safely pushed to 2.4A (3.1V Vf * 4 * 2.4A = ~30W). It also perfectly matches the recommended max current for a 519A.
The other light I have has an XHP35.2 HI and the Vf is already 14.5V (which seems high compared to an XHP35 HI test) so I’m already at the 30W limit (if indeed it is running at 2A).
Upgraded my D4SV2 boost driver - thanks for the help from @thefreeman
- the information you’ve posted here and on reddit has been very helpful, and especially that last message detailing the current calculation (I’m an EE, but you’ve saved me a lot of time figuring out how this one works :P). Removing the driver was a pain, and I ripped off the button wires, but I got it all back in with enough finangling.
I upgraded to the XAL7030 1uH inductor and changed the sense resister from 120k to 85k (Should be 3A, based on rearranging your equation above). It appears significantly brighter now, gets warmer faster but still seems to have plenty of headroom in the D4S body (4x W2 emitters). I only wonder if I could have gone higher now… XD, but didn’t want to risk inductor saturation (should have simulated/checked things I guess, but I got the parts in today and didn’t think about that till i’d taken the light apart). Moonlight is now a very optimistic name for the lowest brightness, too 
I’ll have to take some efficiency measurements, I guess, but it’s kind of a pain with everything back together now…
2 Thanks