Yeah, OSHPark does 4-layer. Richard made one of his MTN-MAX buck drivers in 4-layer. I don’t know if he still sells that one. But, he was getting them from OSHPark. The price is a bit higher and the lead time is longer compared to the regular 2-layer boards.
OSH Park does four layer boards. They don’t support buried or blind vias though. The lead time is shorter than the two layer boards with extra copper thickness, and not noticeably longer than regular two layer boards.
Oh yeah, I forgot that they don’t do blind vias. But, even if you have to do a via all the way through the board, that still takes up less space on the top and bottom than a complete trace. The vias can be made small and placed under components or in some other out-of-the-way place. Still, if the layout can be made to fit onto a regular two-layer board with a 17mm diameter and no air-wires, that would be better.
Cool, great way to add copper
Very nice indeed. I just saw this thread after finished reading your PM. I’ve been looking for a compact 20mm buck boost driver for some of my projects too (non flashlight). Sweet, now we have a good programmable driver for those 144AM.
- Clemence
Mike C asked me to share the libraries I used in EAGLE to design my first attempt of the 17mm board,
so here are the devices I created myself: My Library
If you need the ATTiny84a footprint as well, I suggest downloading the Sparkfun library here: Sparkfun Library
If you want me to design the MAX5424 as well, just ask for it
Thanks for sharing Schoki!
loneoceans: When you have made these boards, have you used stencils or applied solder paste by hand? I have had a few shorts on MCU pins when using default MCU footprint for my 841 when applying by hand. The default mask pad size is bigger than the actual pin and overlapped with the other pins. I made it better by customizing the footprint so the mask pad size is exactly the same is pin pad size. However, I’m now using stencils I made with an even smaller cream pad size for the pins.
I think he soldered it by hand, pre tinning the pads and then reheating them holding the parts with tweezers
Looking good! The 20mm version would be perfect to put into a Convoy L2 with short battery tube.
Laying out the board on 4 layers would certainly not only allow better performance but also allow better routing. However specifically avoided a 4 layer stack to keep the PCB costs as low as possible. Adding blind or buried vias significantly increases cost so I usually try to avoid them as much as possible. I'll explore a little with the 17mm version when I have time though it looks to me that there should be no problems making a 2-layer 17mm version.
I didn't bother fabricating a stencil for this since I was expecting to only make a very small number of boards, so I soldered them all by hand with a combination of reflow and solder paste applied by hand using a toothpick (typically I don't get any bridges if the right amount of paste is used, or I fix it using solderwick), and solder the rest by hand under magnification with a tiny tip soldering iron. Like you mentioned soldermask expansion needs to be adjusted. I think I used 1 or 2mil expansion, depending on the fab I go with.
“Improved lowest level of CC brightness (just around 1mA - true ‘firefly’ mode)
Modes for my own firmware: 1 = firefly <1mA, 2 = low ~180mA, 3 = mid ~600mA, 4 = ~1.8A (~1000 lumens), 5 = 3A (turbo)
Measured efficiency of 92% at 3A (~6.5V) output”
Can this driver correctly work with triple LEDs, 8-9V ?
17mm 9V/1A for 219C triple will be very interesting driver.
(and <=5mm height, for E2L host)
Yes it will work just fine with some small adjustments (e.g. to make sure it doesn't try to drive 9V at 3A, adjust components to make it more efficient st the desired drive currents etc).
3A - huge overload for N219C (with serial connection), ~1-1.3A current in (turbo) mode will be nice.
Wanted 17mm (and 5mm height) driver .
fine modes may be, for example :
1 = firefly 1mA,
2 = low ~30-40mA,
3 = mid ~100-130mA,
4 = high ~400mA,
5 = turbo ~1-1.3A
3A is a huge overload for the 219C???
I have run them up to 12A and they still worked fine. Check my sig for a test of them, they didn’t even reach peak lumens until ~6A. Although around 4-5A is ideal IMO.
thanks, but I tell about triple of 219C with serial connection, for every-day usage (not about 1-time tests).
And, for serial connection, 3x219C & 1A will require 8-9W, for 3A it will be ~27Watts of power, and ~36-38W for 4A.
30-35W in small case - overheat in few seconds.
Ok, so you are saying that 3A is too much for the host, not for the 219C LED’s themselves. That makes more sense.
Besides heat from the LEDs, running at 3A at 9V output is also impractical from a single 18650 cell drive in most flashlights as well. The 27W output will require at least 30W input or more, which for most 18650 cells will require about >=10A input current which will be just about too much for the driver to handle safely. Contact and switch resistance will also lead to significant losses. However, 9V at 2A or so is completely OK (at least for the GXB20) :)
loneoceans, when you plan to produce 17mm drivers ?
I didn’t see this in the OP. Curious what are the limits of this driver when it comes to power. I assume that the true limits lie in the input current?
I assume this could boost power to an 12v XHP35 as well with some component changes, what kind of output current would be possible?
Also, did you ever say what components you are using for this driver? I never saw a parts list, curious what boost IC you are using along with the rest of the parts.
If this could be shrunk to 17mm an xhp35 S2+ could be a really interesting EDC light (or xhp50.2 for that matter).