Designing a Buck Regulator with the MCP34063
The chip I’m going to be talking about is the MCP34063. I really like the chip because it is very cheap, which is great for learning because you will most likely burn some up, and it’s one of the older regulator IC’s out there so there is some information floating around the internet that can help you get started. I went with TI for this because they have great customer support via their forums which is really one of the best resources out there if you are interested in learning some valuable lessons in design etc… The DIP package is also really nice and as you may have found out by looking at higher frequency switching IC’s, not everyone produces a DIP package.
If you’re interested in working with this part or learning about switching regulators I would recommend you check out the rest of this post and visit the products page where you can find a PCB to support this chip in a buck regulator design. It’s a great platform to work with and learn on and really beats a bread board!
Now for the cons: The first and probably most damaging is the low frequency that this chip runs at. It can only go up to a switching frequency of 100kHz. This is a big deal because the inductor size is driven in large part by the switching frequency. The lower the frequency the larger the inductor needed. This because a problem as output current increases because the larger the inductor usually means that it has a lower max current rating. Basically, as the inductance value increases, the max current you can put through it decreases. Another similar problem is that as the switching frequency decrease the out capacitor must increase to smooth out voltage ripple. This will mean that the inductor and the output cap will be pretty large in size. The current limitation for the internal BJT is 1.5A which isn’t too bad, but if you need more than that this will be a big problem. One way to get around this would be to drive a larger FET, but this will increase real estate requirements since parts are being added.
The switching frequency is controlled by a timing capacitor. This part can be a little bit tricky because the timing capacitor doesn’t control frequency, but rather on time of the internal BJT. This is the T_on part the duty cycle. I’ll go over this in the design example, but the idea is that the frequency is derived from the duty cycle and the duty cycle is derived from the input and output requirements along with some small component paramteres (such as the diode and internal switch saturation voltage). Once all of that is decided and after some rough calculation a value for the timing capacitor can be chosen.
The output voltage is the simplest part because it is a basic voltage divider set by two resistors. The only other big component (design wise) is the inductor, but this could also arguably be the most important. This part should be chosen based on voltage ripple and current requirements. The larger the inductor (while switching frequency is constant and output cap is constant) the smaller your ripple voltage will be. Overall I like this chip since it is so cheap and if you have the real estate it’s not bad. It also uses cheap passive components. I would definitely recommend this chip (and from TI) as a learning tool. Below are some pics so you can see a built up board with components. I will be working on some design tutorials soon.