A prototype SBC design has two 100 pin LQFP TI MCU's soldered directly to the PCB. Any suggestions on a test clip to attach to them? I have found some at QFPTestClips but they are extremely expensive. Considering putting a 25 pin header on all four sides of the MCU, but this may turn into 100 little antennas. Any suggestions.

 

One of the downsides of hi density packaging.

Does it have a JTAG interface so that you could do test vectors thru it ?

Or build your own bed of nails, but that will take some time and effort.

These days its almost cheaper to add a second MCU into design and have it test/stim
the DUT. And do it thru a USB interface on test/stim chip.


Regards, Dana.

 

I've done layouts similar to what you're talking about using PICs. Those are uControllers so it's only IO at the pins i.e. not fast bus signals. YMMV. The old Microchip ICE2K/4K emulators used a similar arrangement to connect to the target board and Microchip still uses it for the 'PIM' modules which allow you to change the processors on the Explorer 16/32 and other development boards. Those are uControllers so it's only IO at the pins i.e. not fast bus signals. YMMV.

When it wasn't practical to use the breakout pattern, I would frequently prototype the circuit on a development layout PCB with layout sensitive stuff as close to production as possible with the uC breakouts, test points, mounting holes for standoffs etc. as desired for development, then lay out the the production board as close to that as physical size allowed without the extras. I found using those 'mule' boards to be a good compromise between getting a reasonable prototype environment while avoiding lots of micro-connections on what would be a production form factor. Sometimes, you can add a few high-density header patterns on a temporary PCB layout extension and make plug-in PCBs that have an array of testpoints for easy connection to scope probes, logic analyzers etc.

Connectors for this one are Samtec SMS-121-1-G-S socket strips.

EDIT: @danadak posted while I was typing.
Good luck and welcome to AAC!

 

Thanks John and Dana, your comments are helpful.

Both MCU's currently have JTAG headers which was my initial test / debug plan - just use that. But it would be nice to be able to examine individual pins. My thought was to bring out all 100 pins, but I really don't need to do that, could just do the I/O port pins, and not include any of the higher speed pins, clock, etc. Good suggestions for interim PCB setup more for testing.