Enough with circuit theory for now, although I haven’t even began to touch the basics on that subject. Only in the area of feedback circuitry there are tons to know and explore about control theory, stability and various applications. These amplifiers are definitely not just useful as amplifiers, no doubt. This post revolves mostly about electronic and also mechanical considerations of the design.
Its not only an amplifier
As it was mentioned in the first post, this circuit is battery operated. Although not designed for a specific battery, it was clear that some single cell phone battery will be used. The Galaxy S2 battery was used for the initial prototype. Choosing a charging component requires a lot more experience than I have, but I gave it a go. Consulting with some friends, two components came up: BQ25120 and BQ24230. I always find the TI website easy enough to go around, so I checked them out. However both are BGA soldered. suppliedBall Grid ArraySo what is BGA and why is it horrific for a mediocre designer (or more important, a terrible solderer)? It means that the component is supplied, during the manufacturing process, with a Ball Grid Array of solder on the pads. So using expensive (or in my case, in existent) equipment, you can solder it on your well designed PCB. Looking for a new component on the website went through these logic steps: Power Management -> Battery Management Products -> Battery Charger IC. From here I guessed that Standalone Charger is the right category to go by. A short search, refining the voltage to 4.2 Volts and the packages to something I know (say, SOT-23?) obtained the BQ21040. This 6 pad component seemed simple enough, even for me. With a max. charging current of 800mA, it seemed ideal, as well.

Putting your component on the board (using free Open-Source software)
One of the significant disadvantages of using free software is the service you get. For instance, importing a SOT-23 component seems trivial enough, however in this case it took a bit of a hassle (not to mention the amplifier, TBD). First, there is no symbol! So using KiCAD, you need to start off with that. In the schematic editor, click the library editor symbol.

The library editor opens and now you are probably lost as I was. Have no fear! First, unless you feel a very specific need to make a new library, choose one of the existing ones. This is done by clicking on the ‘Select Working Library’ button

Select the most suitable one in that moment. If you can find one, make a new one. Chances are that you are not inventing a new type of component and you can stick it some where there. libcms is the compatible one, in this case. Start drawing the component with the extremely uncomfortable symbol editor, cry and draw some more. End up with something like this.

Connect it in your schematic according to the manufacturer instructions. Here there only two parameters I set. The charging current was set to ~450mA, so the resistor connected to ISET was chosen to be 1.3kOhm. The second one was the LED resistor. Chose 1kOhm, as I was using a 5mm LED – ~2.2V “fall” on the LED, leaving about 2V to go through the resistor. Hence the current that goes through the LED is ~2mA. With respect to the 450mA charging current, this is negligible. Since this doesn’t really go to market, I disabled the temperature safety mechanism (which prevents batteries from exploding) by pulling it down with a 10kOhm resistor, as defined in the datasheet.

Connectors are actually USED
This condescending title means one thing: Don’t use the cheapest\first connector you find. Consider this: in this case, we have two 3.5mm audio jacks (input and output). At least one of them (I’m guessing the output) will be connected\disconnected several times along the PCBs life. Another annoying tendency of headphone jacks is to be strained in various directions. This calls for a massive and durable audio jack. At first I chose CUIs SJ-3502-SMT. Due to availability issues in my country, I eventually went with Farnell‘s MX387GL. This is definitely a sturdy connector to go with, that will resist some torque. Again, editing a symbol and placing it on the schematic (this time into the conn library):

The same considerations were applied to choosing the USB connector. In this case I think I over killed it, but I went with an expensive Wurth electronics one, 629105150521. As the amplifier was already chosen before, that’s it with the connectors and special components.
Getting sloppy towards the end
Where did I stop thinking, then. The last part to connect is actually the battery. As phone batteries do not come with special battery housings and standard AA\AAA are out of the question. From this point on there are two choices:
1. Design a chassis that will hold the battery and proper contact springs. This is expensive, time consuming and requires some actual mechanical engineering skills. Even if you go for “simple” 3D printing, you need to actually know what you are doing.
2. Improvise and hope for the best. So for that extent I added two massive test points, to solder the battery on them. Soldering wires to battery is not only hard (as the metal doesn’t usually interact with the solder), but dangerous. Heating up the battery with a welder can cause all kinds of hell. From exploding batteries to gas leakages. Having said that, lets move forward!

Connecting it all together
It’s time to finish the schematic. There are some extra components here connected to the amplifier circuit, but they wont be discussed here. Same as programming, it is important not to be lazy with the net names and make them traceable. Hope you enjoyed the project so far. Next post will be regarding some extra design considerations of the amplifier circuit. After that, let the layout begin!
