This is the RP2040 minimal circuit, as linked to from the RP2040 hardware guide. I thought I'd have a go at making it, just to see if I could.
Spoiler: I couldn't. Then I could.
I couldn't find the NCP1117 voltage regulator, so I switched it for the AZ1117. That's essentially the same device, only it needs a slightly different value for one of the capacitors.
I also changed their 16MB flash for a 2MB one, which was much cheaper and adequate for my purposes.
I labelled the revised board v1.1DF and had it made up by JLCPCB.
The 0402 sized components are very fiddly, but not really any different to the bigger ones. I couldn't use my preferred 3mm iron tip, I used a 1mm conical one with the iron temperature turned up to about 335C. I still don't find it easy to get heat into the joint but it does work. Some components are so close to each other it's hard to get the tweezer tips in the gaps.
I pulled a pad off the board at C12. They're a lot more fragile than bigger pads. C12 is only one of many decoupling capactitors, so I figured I could get away with this one missing.
After soldering those, the joints are greatly improved with a 1 minute blast of 230C air with a small nozzle, just to melt and reseal the surface of the solder. That removes the icicles, and any joints that are dangerously close to each other change into smooth round blobs which makes inspection much easier.
The RP2040 was initially beyond me. I only had hot air and an iron and I couldn't get it to work. Then, 2 years later, I bought a hotplate, and that does work. My process:
Tin the pads of the RP2040 footprint with a small amount of solder. Don't do the big central ground one. Heat the area with air until the solder melts and the icicles all go into neat blobs. Then repeat the process. Nothing sticks to solder like solder, so this gives a double layer, neatly blobified with the hot air.
Now add a small amount of solder to GND pad, then wick it off again. This should leave the tinest amount on the pad.
Add a thin smear of gel flux. Not too much, otherwise it boils and the IC floats around on it when it's being heated. Place the IC on the PCB using the footprint's white boundary lines as a guide. There's no way to be sure the pads and the pins line up.
Heat the hotplate to 150C. When it's ready, place the PCB on the plate. Then turn the heat up to 210C. Don't place it on too soon, the plate takes several minutes to come up to temperature, which overexposes the components to heat.
When the plate is at 210C, time 30 more seconds, then turn the heat off. Wait for the temperature to drop to 170C, at which point the solder will have solidified. Then remote the PCB with tweezers.
This approach works for me. Look at the lower edges of the IC where they meet the PCB. There should be no gap.
My friend, AM, has a bit of experience with soldering QFN-56 packages in an oven. He tried to solder one of these, plus another one he did with hot air. Both those failed as well. So I did 2, AM did 2, and none of them worked.
So it's definitely a fiddly business, but with the hotplate it can be made to work.
Derek Fountain, May 2026