0000 0000 0110 0001

LM2596 Step Down (Buck) Switching Voltage Regulator (PSU I)

Down a big rabbit hole measuring nano-currents I recently found myself wondering if it is time to DIY or BUY a Power Supply Unit (PSU) for the workshop. I'd like to be able to dial up reliable/variable power of course, but I'm also looking for stable (linear, non-ripply) power for microcontroller projects where dependable voltage and current limiting is expected/implied for most applications. My current options (pun intended) are not so fancy in that regard.

Ideal therefore is some sort of buck/boost IC leading off to linear regulators as per the following back of the envelope drawing.

I'll probably end up using a buck/boost module feeding in from a typically strong AC/DC converter (e.g. a laptop supply perhaps operating around 20V at 3-5A). I am looking specifically at the ZK-4KX buck/boost unit for this "stage one" of the PSU - which I'll test separately in the next couple of weeks.

In the meantime I've dragged out some LM2596 buck converter ICs billed in the datasheet as "a monolithic integrated circuit ideally suited for easy and convenient design of a step−down switching regulator (buck converter). It is capable of driving a 3.0 A load with excellent line and load regulation."

We have looked at the LM2576 before, and seriously for my application there is no appreciable difference, but for the sake of the blog and channel I thought I'd drop in the LM2596 just to show them working.

The LM2596 can be configured to drop down the supply voltage, let's say from 20V to 8V, and then this could feed into an L7805 to give a flatline 5V for any hungry μC that needs a feed.

The breadboard revealed that the LM7805 chips might be a little shonky, so I ended up with my old favourite the HT7850! 

Unfortunately the current draw of these little guys is not really up for PSU work, so I'll need to spend some time looking at all of my LM7805 chips, or maybe an alternative.

Next in the PSU project I will be testing the ZK-4KX in combination with linear regulators and then I'll put the whole shebang together in a 3D box and get out (maybe) of this rabbit hole.