I want to build my own TA driver, only with amc. First I wanted to make a pcb for TA firmware, but then I discovered the bistro HD. And there was a problem with the capacitor. I can not find it anywhere in size 0805, and if it is already quite expensive. I found one cheaper, but it is in size 1206 and the plate begins to do a little tight. Can you recommend some (and preferably with free shipping) capacitor?
In addition, please check the correctness of the scheme.
I’m assuming you are referring to C42 since its the biggest capacitor on your schematic. How about soldering two smaller 0805 capacitors on top of each other (parallel configuration)? Unless the height is also a problem. Any reason you have to have a 47uF size. How about 22uF? Here in the U.S. 47uF, 10v or 6.3v can be purchased at Digikey or Mouser. No free shipping there. When needing a small priced item, I wait until I need more stuff.
It must be 47uF
Well, that explains why it needs to be 47uF, but not why it needs to be a tantalum cap. Why wouldn’t this 0805 47uF cap work? http://uk.farnell.com/murata/grt21br60j476me13l/cap-mlcc-auto-x5r-47uf-6-3v-0805/dp/2672188
Chatika , the capacitor mentioned by you is a decoupling capacitor for the supply voltage , and tantalum are known to have lower current loses and better stability in time than electrolitic ones. But is not a must ! The SMD ceramic capacitor , as the one Mike C suggested , is PERFECT for what you need , even better than the tantalum ...Go for it , and good luck with your venture , nice to see that the DIY didn't die yet ...
In Build Instructions Flintrock gave a link to the tantalum capacitor . That’s why I thought it was required.
Interesting driving current selection scheme with 3 output lines. I like this approach, I guess you aim to avoid PWM at the output, isn't it Chatika vas Paus?
Now that I'm here, I'd like to discuss about PWM output filtering via LC filter. If PWM frequency is moderate (2KHz max to be safe I think) the 7135 regulators can work nicely without losing regulation early, this means a very low PWM-free moonlight mode could be achieved by attaching an inductor in series with the regulator(s). 
Cheers :-)
I’m no expert, not by a long shot, but don’t low-pass filters in that frequency require rather large inductors?
Low-Pass Filter a PWM Signal into an Analog Voltage @ allaboutcircuits.com 
Me too no expert there, and I can peek up there the answer is a massive inductor.
Since Chatika vas Paus seems to be willing to use multiple driving lines, one of them could be used for moonlight mode with a very low power tiny regulator or current limiting resistor.
Cheers ^:)
P.S.: mmm, TA stands for Texas Avenger driver, a fourth line would also mean slight firmware modification, doesn't it? Well anyway.
Well, I can’t speak for Flintrock’s firmware and drivers, but I use ceramic 47uF caps for my OTSM caps and they work fine, I get over 10 seconds of accurately measurable off time with them. On the other hand the MCU I use has different power mode options, but still, 47uF ceramic cap should be fine for any OTSM capable driver available here in BLF.
Of course. The more modes without pwm, the better.
There is some prototype firmware. In the description of the bistro HD there is a short description of “4-channel-dual-switch”
That’s great news. Maybe in a few months 
(it’s hard to fit 10 AMC on 17mm pcb), when I put together the first prototype I will let you know.
I managed 12 of them on 17mm, and can turn any number of them on. They are grouped 1,2,4,4 and one which is dedicated for PWM between the 350mA steps 
12!? 
But how? I have a problem with 10. Did you use attiny in the qfn case? Could I see how more or less pcb looks like (PM)?
No need for PM. This design is (for me) soon to be obsolete. I’m moving away from 7135s. MCU is indeed QFN package.
Edit: I can point out that I have KiCAD installed. It enables four layer routing. I was unable to do this with two layer design.
So many questions …
Where is the place for + batteries?
What do these letters mean at the holes?
And I do not understand. Why everyone, or at least the majority do not like amc-based drivers?
Good Lord! 
I've been sparingly touting with this here for some time already.
May I ask what sort of regulator do you plan on adopting?
Cheers ^:)
The AMC7135 cannot switch on/off fast enough for fast-PWM (≈6KHz is maybe too fast already), loses regulation early and efficiency. Also, driving current is non-adjustable.
Cheers ^:)
I don’t know anything. All my plan for rubbish.
The FET drivers seem perfect, but I like to know what the maximum current is of LEDs. FET pushes into the diode as many current as possible. The maximum mode depends on the battery used.
Correct me if I’m wrong.
Some time ago I found this thread:
https://budgetlightforum.com/t/-/35542?page=1
It’s old, but it seems to be still up-to-date
Well, a driver with a bigass MOSFET for turbo plus an LC damped smaller MOSFET would rock I believe. The LC damped MOSFET could be PWM switched on/off fast enough (various tenths of KHz) for effective smoothing of the output to the emitter with adequate inductor/capacitor pairs. 
Cheers :-)
Oops, I should have explained that too. I don’t have a photo of this particular driver built, but it’s designed for stiff copper wire through the + hole which is bent over the MCU. I’ve used this technique successfully. On this old ATtiny85 based driver in the photo I have two wires soldered together. It works fine as long as you have battery clearance (I always did), but I quickly realized that a single strand of wire was enough:
I do have a 17mm version with a proper battery contact plate for a spring, but that one only has 10 x 7135s.
As it’s a QFN MCU there is no clip to flash firmware. As I’m constantly developing and testing firmware I want to be able to update firmware in my lights. The holes are for what I call acupuncture style flashing:
The letters are just so I know which hole is what
Well, I do like them. I’ve been using them for a long time, they are cheap and easy to use. But for me I’ve always wanted to make a constant current driver with fully adjustable modes by firmware with no PWM. For no real reason other than it’s a fun hobby. I’m not totally rid of PWM though, I have to use it for moonlight as the lowest constant current I have been able to get is 20mA. So for moonlight mode I use 16 bit PWM on that 20mA, with that I can get as low moonlight as anyone could ever want.
Heh, I know… Well, not so sure about “sparingly touting” part 
I haven’t seen many alternatives though, at least not for electronics noobs like myself who doesn’t know much about electronics and doesn’t read on Russian forums. It’s really hard to beat 7135s in terms of price and ease of use.
Easy to use digipot controlled constant current regulators. I have a few to evaluate. Not the best choice in terms of efficiency, and for sure not cost effective, but easy to use for a noob like myself. I just wanted to design my own constant current driver, not really caring about price or efficiency, it’s just a hobby anyway. The real electronics guys and gals would probably laugh at my choices of regulators, but it’s all for the fun of developing