Breadboard-able LPC1114 ARM Cortex-M0 Followup

Around the beginning of the year I posted this article¬†about my idea for a breakout board for NXP’s LPC1114 ARM Cortex-M0 microcontrollers after being inspired their announcement for a DIP packaged Cortex-M0 micro. Earlier this month I received the PCBs for my design and built two of the boards using my reflow oven.

In the above pictures I didn’t populate the power LED, but everything else is in place and tested. The LPC1114 QFN package was actually a pleasure to work with using solder paste and a reflow oven. I didn’t have a small enough tip for my solder paste which meant I had a little too much solder in places but that was easily cleaned up with a fine tipped iron. The most obvious place where there is too much solder is around the crystal oscillator on the right, and the 0805 capacitors on the left of the board.

After building two of the boards there were a few things I was happy with, and a few more disappointments. Firstly due to the pin layout of the LPC1114 chip in the QFN package I could make this breakout board quite small and still keep a fairly large ground plane. The small size was a goal of mine especially considering I wanted to be able to plug this board into a standard breadboard. It was almost something I would use!

Now on to the disappointments: My idea presented in my other post about reset and boot-loader buttons was a failure. I had an idea of using small edge mounted buttons to reset and enter the boot-loader mode on the LPC1114. While the idea worked in terms of functionality the implementation failed in terms of ergonomics. Pressing the buttons to load code onto the chip was an unpleasant experience which I think cost too much in usability. My next gripe which I thought might be a problem is that I still need to hook up an external device (In this case a FTDI USB to UART chip) to reprogram the LPC1114. Lastly the silkscreen on the board was blurry! Although this was my fault for making the size too small. It was legible, but it is nice to be able to read which pins you are connecting without squinting.

I am quite happy with building these breakout boards as a learning experience. I do plan on making a second revision which incorporates what I have learned. I plan on doing away with the buttons and instead adding a small FTDI USB to UART chip in a QFN package. This can also be used along with lpc21isp software to automatically reset and enter the ISP boot-loader. As an added bonus this would add USB serial functionality to the board without adding anything extra for easy debugging and prototyping. The external crystal could also be eliminated  with the pins being broken out so that any crystal could be used. This would offset some of the cost added with the FTDI chip and USB connector, especially considering the LPC1114 has a 12MHz internal oscillator.

I will post schematics and layouts for the updated version 2 of my breadboard-able ARM microcontrollers once I have them ready to send off.

Dead bug soldering SMD.

Dead bug soldering of surface mount components can be a pain, but in some cases it is absolutely necessary. Many components with be based on a standard footprint (SOIC-8 for example), however occasionally parts will have their own footprint that is impossible to buy a breakout board for.

I wanted to test a MEMS I2S microphone made by Analog Devices which had a very interesting footprint. The eight square pads had a 5 mil pitch, however there is the round ground pad around the microphone port that needs to be connected. In this case I opted to flip the chip over and superglue it down to a different breakout board I had. The two following pictures are how it turned out.

 

I first fluxed all the pads, then added solder to both the breakout board as well as the pads on the microphone. Because the round ground pad was adjacent to another ground pad I bridged the gap with solder. After cleaning my iron and applying more flux I positioned single stranded wire ( I had removed the insulation with wire strippers) with tweezers and applied heat. The final product looks a little messy close up due to excess flux. The white haze is from the superglue.

After all the connections were made the I2S interface worked like a charm. I will post a tutorial on using I2S with the LPC1769 a little later in the month.