Ok but the mtn driver can give 25 amps no? And the blf…?
The mtn fet drivers use a >6mOhm resistance fet (IRLR8726TRLPBF) which is… not ideal for such a high amperage application. Several watts of heat will be dissipated in the fet. There are inexpensive fets in the same package size which have 1/10 that resistance.
1 Thank
Sorry I make a mistake
Hello everyone
Sorry for my English.
I wanted to show off my modification led h4 from the car, which I converted into a bicycle lamp.
Specyfication:
-driver 3/5 mode 2.8-4.5v
-2x xpl hi
4 Thanks
Any idea on why Richard would choose a FET like that for his drivers? Seems he could save some money and get better performance with the other ones.
It was one of the options in ~2016 and it was inexpensive. As of late last year anyway, it was still the FET being used. It is still less expensive than better options, we’re talking $0.25 vs ~$1.00 at 1k qty, but it has significant impedance at all voltages. Also worth mentioning: Mtn fet drivers do not feature transient suppression, which matters for high amp builds if MCU’s other than T13A are to be used. Also the OTC cap used is likely not thermally stable enough to support firmware’s which feature timing based mode switching, so when the light gets hot, the mode switching behavior will change or otherwise malfunction.
If you build your own driver you will achieve a superior result. Use a TA driver board or one of Del’s driver boards, they are shared on OSH park. You could order the boards with 2oz copper from OSH PARK if you are building a high amp driver. The 2oz boards are only .8mm thick though, so you may need a GND ring spacer, or solder a solid core wire around the GND rim, etc.
I have a superior design to any other Fet+1, but I have yet to build and test it.
1 Thank
I’ve experienced this with the MTN-DDm driver with Bistro firmware. When it gets very hot, the medium press timing changes and is very hard to use.
Usually fets with low impedance have low swishing speed and vice versa. So for DD- low impedance and for low output- fast switchers.
BSC009NE2LS5ATMA1 is very fast, and <1.5mOhm.
A fet which is ~10 times slower at switching, but has 1/2 the resistance, is SIR178DP.
Flashlight PWM is relatively slow, ~20kHz. Until a head to head comparison of switching losses between BSC009 and SIR178DP is performed, I don’t know if the switching speed even matters, or to what extent.
What will matter in some applications though, is the Vdrop and heat loss at turbo current:
At 40A, 3.5V, no PWM:
Sir178DP would burn ~1W, .025V drop.
BSC009 would burn ~2W, .05V drop.
Sir404DP would burn ~2.4W, .06V drop.
IRLR8726TRLPBF would burn ~12W and drop ~0.3V (desolder itself)
Do the “switching losses” matter at the next lowest mode? 30% output or whatever? I don’t think so.
I feel like most of the FET driver theory crafting won’t matter as long as you aren’t able to buy a finished product.
I can go to MTN electronics and order all FET drivers in basically all sizes.
There just isn’t an alternative for buying ready FET drivers
2 Thanks
Put anduril on a Convoy H1
Built it with 219b sw45k because I didn’t have any single-emitter lights with it. Emitter came original to the light.
Used an Emisar D1v2 2.5A linear driver, as well as an amber button. The button was a bit short but a small glob of hot glue on fixed it, though it is not very easy to feel the clicks.
The driver had to be dremeled down to 21mm from 22mm diameter and the hole of the centering ring enlarged slightly so it cleared the components.
If the driver isn’t clocked correctly it shorts, so the centering ring and driver alignment is key to it working. The flashing pads have to be roughly aligned on the same side of the glass, for the linear driver. I don’t know if a boost driver would fit, as it is slightly taller and might have different locations for the wires and pads.
The H1 is a pretty interestingly designed host, because all the soldering can be done outside the light, since everything slides in through the driver hole and the buttton slots through. IMO it as good of a beginner host to mod as the S2+, if you’re somewhat experienced with a dremel or are using a known-compatible driver that doesn’t need resizing.
8 Thanks
Switching looses are looses too. Otherwise all fets would be made with low impedence and slow ;))
Oof. Time for a forged one.
Very nice. Easiest host ever right?
Ya I guess so lol. I don’t even use this for anything real heavy duty. Stuck it back together with JB weld…see how that holds up. Can’t be any worse than it was already. Might even make it stronger
Close, the M21C-U was easier because it needed no modification other than hotglue on the button. Driver just had the wires through holes and it all buttoned up nicely.
The 21mm space for the driver is kinda weird. Especially because there’s clearly room for 22mm but it just bottlenecks right there. Just to make it harder for you I guess.
The M21C-U is 23mm, and has a pretty narrow ring so it is a perfect fit. The most annoying part was simply putting the cables through the holes. The H1 definitely could’ve been 22mm, I suggested it to Simon but I doubt it’ll get implemented.
Naw I can’t see that ever happening either.
As long as you have room to grind the driver down. Do they even sell a 17mm/20mm to 21mm adapter? I don’t think so. And who has a fine pitch 21mm die l ying around.
Remember, the H1 is e-switch. I don’t think there are many side e-switch 17mm anduril drivers at all.
The emisar driver barely fit, I had to grind it down to the holes on the outer ring. Any tighter and it wouldn’t have worked without just losing the ring, milling the driver pocket, and gluing it in.