Led4power.com : LD-4 CC linear drivers, ILC-0/1 illuminated tailcaps, optics, MOSX, copper DTP MCPCBs...

One needs to consider all the contact resistances in the flashlight though. I think even the Samsung 20S won’t give you 30A for all that long (even if it is capable).

SubZero should be compatible with any standard 4.2V clicky driver since it mimics half press with 0.1sec power interrupt.

For now I will have only 1S (4.2V max) version of subZero.

Max current depends on LEDs and battery most, max. current I got with 6x LH351D was about 24A due to relatively high Vf of LEDs, 30Amps will be "useful" only for 6x LuxeonV and probably 4x MZ.

SubZero is not just for 30Amps, I'm sure a lot of people don't want such high powered burst lights with total runtime of 3 minutes on high, 12-20Amps would be still impressive in most hosts, and this is area where omten clicky switches starts to be very unreliable.

This is especially true for small run commercial custom lights (which are clicky lights 90% of the time due to simpler machining compared to side switch lights), nobody want's 100+$ lights to fail because of switch, and that's why you don't see >12Amps triples and quads in such lights. To limit current most of lights use outdated XP-G2,XM-L2,XP-L/HI emitters because their high Vf limits max. current.

What about the LED’s in the sub zero switch? Would they still work with another type of driver like the MTN boost driver?

Switch doesn't know what type of driver is on the other side, as long as it have bleeder (which is mandatory for subZero switch) and battery voltage is <4.2V, it should work.

Thanks Neven!

What resistance value is needed for the bleeder? This means that the battery will be continously discharged even when the light is turned off (I didn’t use lighted tailcaps until now)?

300-400Ohm should work (LD-x4 drivers have different type of bleeding, so no resistor is needed).

Battery is discharged via bleeder to GND when light is turned ON, bleeding current is Vbatt/Rb, so it can be from few mA to up to 15mA. This parasitic current drain can reduce runtime on moonlight significantly.

That's why LD-x4 drivers use different type of bleeding which consume 1mA max.

When flashlight is OFF, tailcap gets supply voltage over bleeder, current consumption from battery depends only on tailcap electronics(if there is such thing) and illumination leds.

For example ILC-0 illuminated tailcap consumes ~3uA for electronics and 20-600uA for LEDs. But when voltage drops below 3V, electronics shuts down everything and current consumption is only ~1uA.

Do you recommend using LD-A4 for relatively small hosts?

I would like to make Jaxman E2L builds with this driver, and triples of different white emitters (Luminus SST-20 4000K 95CRI, Nichia 219C 90CRI, Samsung LH351D 90CRI).

How safe is it to use 9A or 6A as maximum current for this host and the above-mentioned emitters?

Thanks for the comprehensive answer. I suppose the super cap is also charged this way?

Yes.

I have thought about this some more. Initially I wanted to use just three Luxeon MZs to get maximum throw at 30A. Getting a constant 30A does seem a bit unlikely though.

At 10A the Luxeon MZ has a Vf of around 3.24V (see here ). Since I will be using the C8TT head, I will need to use an 18650. The best 18650 at 30A is the Samsung 20S. At this current the average voltage under load is around 3.35V (see here ).

So the resistance of all the electrical connections between these two points needs to be smaller than 0.0037Ohm = 3.7mOhm for a constant 30A in the turbo bursts. Your switch has 1-2mOhm, so that leaves 1.7mOhm for the rest. This seems very unrealistic.

Using 4 LEDs instead (which should be possible because the Cute4 optic has the same dimensions as the Cute3) will help a bit. Each LED will get around 7.5A and thus the Vf will be around 3.135V. The resistance of all connections between the battery and the LEDs now needs to be lower than 12.2mOhm. Thats better, but it still seems unrealistic.

Edit:
Here you state that a bypassed sping has a resistance of 1-2mOhm. So for two springs we can assume 4mOhm. This leaves us with 6mOhm for the wires and the flashlight (and the solder connections).

I estimate the wires will be about 4cm long in the C8TT head.
Using 0.5mm^2 would mean the resistance of each wire piece is 0.00156Ohm = 1.56mOhm. So we need to add 3.12mOhm to the total resistance.

This leaves us with 2.88mOhm for the flashlight, the pcb traces, the FET (another 0.7mOhm?) and and the solder connections.

Edit:

Of course I can use larger wires. The hole in the pcb has a diameter of 5.5mm and the one in my C8TT head was enlarged to 6mm. I can fit two 1.5mm^2 wire pieces (outer diameter: 2,5±0,15mm) through there or even eight 0.5mm^2 wires. The wires I’m talking about here are special, they have a very thin silicone insulation.

Using 1.5mm^2 would mean the resistance of each wire piece is 0.000532Ohm = 0.53mOhm. So we need to add 1.1mOhm to the total resistance. Using 8 0.5mm^2 wires would be even better. This would leave us with ~4mOhm for the flashlight and the pcb traces.

You are using average battery voltage for calculations? Average voltage is not really useful in this case.

Of course you won't get regulated 30Amps for long time("long" is relative here, 2000mAh battery would theoretically have 4mins run time at 30A), after few 30Amps bursts current will drop to less than that.

As I said, subZero switch can handle 30Amps, you don't have to actually build every light with 30Amp driver,for ex. 12-18Amps triple/quad MZ, Luxeon V or LH351D in EEX6 host would be still very bright and more practical.

30Amps single cell light? Yes - for fun, practical - hardly.

Yes, I want the full 30A. I realize it’s only for very short bursts. Still, I find this fascinating. I was not really possible until now.

I use the average voltage at 30A because I need to use something realistic. The battery wont always be fully charged. If I want to get constant 30A during normal use, I need to do it this way.

This way I get at least 2min of turbo bursts.

The low modes of the driver will still make the light practical.

If you want regulation >50% of time with 18650 battery at 30Amps, maybe the best option would be 6x MZ in Convoy C8 with medium ANNA-40-6 optics( I have only spot optics in stock).

Vf of cold LEDs would be ~3.07V according to maukka's test, hot LEDs would have Vf lower by ~0.1-0.15V, so below 3.0V (pretty crazy).

I already have the C8TT head, got it for this project. It’s better because the fins allow it to cool off much faster. This way I can use the turbo mode more often.

BTW: I just noticed that at 30A the tailcap could get warm from the FET? At 3.5-4W loss in the FET-switch it should warm up noticeably after using Turbo for a few times.

I also still need to account for the heat dependancy of Vf of the LEDs.

EDIT:
If it lowers by 0.1V I get an extra 3.3mOhm of possible resistance.

I’m hoping to hit 7000lm (otf) at turn-on.

At 24Amps I didn't notice tailcap heating,but whole host get hot,so it's hard to tell. Battery tube act as a heatsink, so tailcap overheating is not really possible.

Turn-on OTF lumens for 6x Luxeon Vs in C8 host at 30Amps should be more like 9000lm.

Waiting for those x6 spacers!

I did some reading back to tread page 15, and found this detailed description.

Now I am sure, that 9A would be clearly over both host and driver limits. I just would like to ask, that would I be safe with 6A as max. current in Jaxman E2L host?

I don't know what do you mean by "safe", it could be safe to use that host at any current, only difference is lumen output and time until host reach temperature that will trigger OTP protection.