That’s the the 8HSOP packaged device is the LED2001PHR from ST, the buck controller Matthaus used for this board. It’s the brains behind the buck section as well as part of the brawn - it has a transistor (FET I guess) inside of it, I think that’s what probably blew but I stand ready to be corrected.
The fact that you couldn’t get the buck section to light an LED by itself is probably good info for Matthaus.
Well, I just ordered 3 more of em along with the necessary caps and resistors I don’t have more of. And more ATiny13A’s as well as diodes for FET drivers while I was at it. Always something!
If that is the buck controller, what does the inductor do? I really want to save those as they had to come from the UK and were among the most expensive single items on the board.
I’m so used to my relatively large battery boards and stuff on OSHpark that there was some stigma in my mind about the price - I forgot that a little bitty board like this would be so inexpensive. I went ahead and ordered 3 boards for myself just now. I’m in the process of picking out the absolute minimum I need to get one buck section up and running at a variety of currents. I think getting 0.025Ohm, 0.1Ohm, and 0.2Ohm resistors give me the option to do 0.5A, 1.0A, 1.5A, 4A, 4.5A, and 5.0A (those last two being out of spec of course).
EDIT: hopefully someone else will chime in on what the inductor does in a buck regulator, but the following is my understanding. The inductor resists or slows down changes in current. When the buck circuit is first turned on the current increases from zero and the inductor makes that happen more slowly. Once the controller sees enough voltage across the sense resistor (meaning there is enough current) it switches the circuit off, but the inductor prevents the current from dropping quickly. Eventually the current (again, measured with voltage drop across the sense resistor) falls below a set point despite what the inductor is doing. Then the buck controller switches the circuit on again and the process repeats.
Oh man....Dale. lol. You're not having much luck!!!! I'll be sending you a 'care package' shortly.
V1 layout is VERY different to V2/3. Some of the wire pads are in completely different places. You definitely do NOT want to bypass the sense resistors so connecting LED- to GND is not a good idea. I included the GND pad to help people ground the driver itself via a separate wire (instead of relying on press fit or soldering the board in place). Wight has managed to deduct quite a lot from that one picture - his reply in post #118 is pretty much spot on.
I think the missing modes may have something to do with the PWM frequency. I really think 9.4kHz is too high. If an IC cannot react fast enough to a PWM signal (turning the IC on and off) it either never gets started, or never turns off. This has the effect of looking like it's either completely off, or stuck on. What's happening is that the PWM signal activates the IC, and the IC starts, but then the PWM says "turn off" so the IC then turns off. Except it was still trying to start up.
AMC7135 chips are VERY fast. They can handle fast PWM. From experience, this is an exception rather than the norm. I would take the working Knucklehead and reflash the same firmware (do not recompile) with the fuses suggested by wight. I'm not sure if you ever fully understood what he was saying. When you flash the ATTiny, you probably either manually enter a command line in windows, or execute a batch file. The command you enter will look something like this:
avrdude -p t13 -c usbasp -u -Uflash:w:BLF-VLD.hex:a -Ueeprom:w:BLF-VLD.eep:a -Ulfuse:w:0x79:m -Uhfuse:w:0xed:mIf you do not manually enter that sort of command, but instead just double click something, then its probably a batch file. Edit that batch file and you'll see the same sort of line.
Wight was suggesting you change the -Ulfuse part to -Ulfuse:w:0x65:m. No need to recompile anything, just re-flash the MCU.
Sorry if it was obvious. I just can't see anything in this thread that states you ever made those changes.
I remain convinced that you do actually have a working board, the firmware is just not playing nice. One of the popped buck drivers can be put down to skipping the sense resistor. I have no idea why the first one failed.
I’ll try the re-flash with the fuse changed and see what that does. Been pulled back reading a book…had to step away or keep making the same mistakes. 
Edit: I re-flashed with the fuse changed to 0x65. In attempting to fire up an XM-L2 on a Noctigon, mounted in an HD2010 pill will a chunk of copper inside…I got a flash of light and the wire slipped. Tried again and got light sparks on the cell at contact. That’s usually a sign of a short, so do I hard connect and watch for the buck converter to go up in smoke? Again?
That sounds positive, now I’m really looking forward to building one!
Here’s what I plan to get from Mouser to build the top section. I got a little sloppy to save some cash on that 22uF cap, all tolerances of that are much more expensive at Mouser than Digikey. Cart for top section without coil. In addition to that I’ll need a coil, I was looking at either IHLP2525EZER3R3M01 because it seems to be a close fit on specs and size or 74437349033 from Wurth which is a little large physically but seems like a decent electrical match and is cheaper. Actually both of those components look like they could be hand soldered, I’m not sure the Coilcraft inductor can be.
The coilcraft inductor has 2 large pads underneath the component, not accessible with a soldering iron. And the buck converter with it’s 8 side pins also has a pad underneath, occupying a large percentage of the bottom of that component.
She be working! 
3.84A in Turbo, I think the low is too low, the emitter barely glows. Supper time, gotta run, I’ll retry after supper and solder some connections, put a switch on it. 
Both inductors spec wise are good, but you will struggled to fit that Wurth one on the PCB as C1 and C3 sit very tight on L1.
As in all modes? Was this after a reflash?
I think I can make it fit OK. It’s not as big as it seems, the pads stick out the sides a little. Plus it’s got deep pads so I can offset it towards C1 a bit (although I may have to cover the vias that are under the inductor). Thanks for confirming that the specs are OK.
Does the maximum DC resistance matter as far as we know? Or are we really just concerned with inductance and current handling?
It might have a minor effect on output (provided it's not too much more than the rated coilcraft product), but inductance and how much current it can handle are the most important factors.
Yes, I reflashed with the 0x65 fuse and it works, but low is too low and doesn’t. Out of the available settings from 0-255, I have it set at 1,4,25,120, 255 and moon isn’t engaged. So the 4 needs to come up some to work with this driver. I do have 3 modes though out of the 4, and they look like this…
Running 2 18650 cells (Sanyo UR18650FJ)
Lo - -
Med 2.478V .06A
Hi 2.770V .70A
Turbo 3.390V 3.49A
The cells arent’ fresh anymore of course, with the 2 cells measuring 8.31V
Eureka!
Thanks. In that case I’m also picking up Bourns’ SRP7030-3R3FM as well for testing. It’s more dimensionally similar and pretty cheap.
That’s great! Wait to see what Matthaus says, but you could probably stack a 0.2Ohm resistor on top of the 0.025Ohm one and get your output current closer to the 4.0A range. That might be more of a bandaid fix because I can’t guess why it would be low in the first place.
Blind.
But really With 3 cells at 12.48V I got 2.468V at .08A, 2.770V at .69A and 3.338V at 3.51A. Works fine! I’ll bump that 4 to a 6 and re-flash it, see what it does…
Edit: Rewrote the UI with a setting of 6 in place of the 4, flashed it, and now there are 4 levels, with a very dim low and the same 3 above that as before. No moon, as it were, but this low is that kind of low. Meter doesn’t really show it in the amperage mode. Best I can tell it’s about .02A, with 2.434V. Getting difficult to read meters or what I type, keep looking into the light. It’s so pretty!
Great news, Dale.
EDIT: If you're going to try and go above spec, I'd seriously consider heat sinking the cr*p out of the buck IC. Like thermally gluing a small chip-set heat sink or copper disk or something directly to the IC. I also wouldn't go too far as the inductor and sense resistors are rated for what should be 4A output. They are over-specced of course, but it's unknown territory.
I am curious if the other inductors will work. ~$10 for one inductor is pretty high.
Wait, who’s paying $10 for one inductor? Coilcraft says that they have 11,039 in US stock at $1.85 each, is their shipping unreasonable or something?