A driver rated at 20A means it can deliver up to 20A under optimal conditions; in practice the actual current may be significantly less due to factors such as Vf rise, battery voltage sag, voltage drop from contact/resistive losses, etc. This is how old FET drivers manage to drive a high Vf LED without frying it, despite no current-limiting element built into the driver.
6 Thanks
This is the key point regarding the SFT90 @20A discussion. Most batteries will not actually deliver 20A to the SFT90 due to the high Vf and diver, spring resistance matching the battery voltage sag before 20A. Only the most high performance 21700 cells will drive it at the full 20A - and only at the top 10-15% of their charge level.
That is why he does not need to worry about the SFT90, whereas low Vf emitters, like the LHP73B, need to be limited by the driver to be safely operated.
6 Thanks
Convoy pairs SFT-90 with a buck driverā¦
A buck driver only bucks when the input voltage is greater than the target output voltage. Otherwise it simply falls out of regulation and enters what is essentially direct drive phase.
With high Vf emitters like SFT90, turbo on a buck is probably indistinguishable from direct drive. What benefits from the buck design is lower modes, which widen the gap between emitter Vf and battery voltage under load.
7 Thanks
Buck differs from fet at turbo in 2 main ways. The first, convoy isnāt using the best buck ICs so the duty cycle is only like 92% so not full 1:1 voltage direct drive. You get dropout earlier and lower output.
Second, a buck has significantly more resistance. At turbo, a fet has less components and less resistance, which is less voltage sag and less dropout.
1 Thank
Thatās true for the buck IC used in the 8-10A drivers, But Iām not sure thatās the case with the 20A ones, it uses an asynchronous controller and itās pretty common for them to drive a P-channel MOSFET, thus 100% duty cycle capable, but itās not impossible for it to use an N-channel one so Iām just guessing, do you know the part used ?
3 Thanks
I seriously doubt Convoy is using the best FETs out there. Sofirn along with Haikelite used good Chinese versions of the low resistsnce Infneons. Mtn doesnāt use the greatest ones either.
I remember looking up the datasheet for the FET on the 3x21C driver and thinking it has the appropriate ratings for the purpose. On the other hand, the Alpha & Omega FETs in the Sofirn Q8+ have crazy high resistance in the low-voltage regime where the LEDs operate, and there have been reports of the FET burning up with bypassed springs and high-current cells. It is either a low-quality component or one unfit for the job.
I did burn up 2 consecutive Mtn FET drivers back in 2018, and the FET was the point of failure. Quite surprising as I used them to drive the old XPL HI, which has very high Vf and cannot draw too much current.
They must have changed them since it seems the Q8+ uses different FETs than the other Sofirn FET+1 drivers with Anduril like from early 2018 up to like 2021. The MF01 Mini, EC and EA series, and the FT series all used decent mosfets and could take loads of amps.
I would almost say that this problem with inferior MOSFETs is common to almost all manufacturers. The Emisar D18 also uses a Chinese MOSFET, which can break down relatively quickly like in my caseā¦
Itās a shame, because name-brand MOSFETs are only slightly more expensive, but in the end they can make the difference between longevity and quickly produced electronic waste because of defective driver.
5 Thanks
I heard from an electrical engineer that the 3x21c/e fet inside the battery ring is not the fet turbo. Thatās actually a reverse polarity protection fet. And the fet turbo is just running the buck in direct drive mode.
I do not. Stephen mentioned it
You have mentioned this before. To be clear: which of your 3 claims came from the engineer you heard from, and which ones are your guesses? Could someone confirm or refute, explaining their reasoning?
I find them hard to believe because
- The design of a contact ring already provides physical RPP, since the - contact of a battery cannot extend past its wrapper to touch the flat brass ring,
- The FETās specs are quite appropriate for driving this particular LED configuration, and
- Simonās listing states ā54A FETā, which seems to imply that the turbo is indeed driven by FET.
Datasheet for the component in question. Even as low as 2.5V, the resistance stays well below 4mOhms up to at least 80A.
Sofirn Q8+ MOSFET in comparison. Even with current as low as 20A and voltage as high as 3.4V, the resistance shoots past 5mOhms, rising āoff the chartsāāin the literal senseāfor voltages lower than 3.4V.
1 Thank
Changes have definitely been made between batches, and there are probably also mixups between the Q8+ and EC06. Iāve seen earlier generations use a ChipSourceTek FET that is supposedly much better than the latest Alpha&Omega stuff, per this teardown.
1 Thank
Check discord
1 Thank
Do you guys think this would this be worth trying in an old m21a host i have with the 10amp buck? Too under driven?
It might achieve decent sustained lumens perhaps.
Trying to find something to do with it, but limited to 20mm or 17mm dirvers. The 10mm opening reflector with the 7070 gasket could work maybe.
Tis a matter of how the beam might end up i guess. Its ok if it turns out floody. Might even be great for that.
Thoughts?
Yes
Yeah, 10mm OP Reflector should be perfect for the LHP73B.
2 Thanks
Is there a difference in output between the 5000K and 6500K versions?
You can probably infer from the LHP531 test that features all CCTs; there are good reasons to expect the same proportional differences.
2 Thanks
At what output/current are these?