The test design with a LED counter, UART loopback and all I/O pins are generating a 5 MHz square wave. dpa_test_rev3_0.zip
Improvements related previous board revision
Complete redesing of the power supply (improved filtering), decreased voltage step for the Vcore from 10 mV to 5mV, added support for L and LI speedgrades with lower default Vcore than 1.0 V
Price and assembly complexity has been lowered (the profi assembly company is no more required).
Components have been renumbered for better identification in BOM, the documentation is more straithforward
All components with a same value have unified packages, 0402 is the default footprint for all discrete components
Components with EOL and NRND status have been replaced
Known issues and problems
Stencil openings for MP8843 and XCL214 ICs are poor - they have to be wider a little bit.
1V8 and 3V3 are enabled by POWER GOOD signal of MP8843 - could case a problem in case of the direct Vcore via dedicated SMA.
Unperformed tests
Power OK
FPGA via JTAG lives
LEDs and buttons
UART loop back
I/O pins
I2C chain discovery Working on FPGA side only, the ISO1540 is shorted again (Replace with SI8600A).
Tuning the MP8843 via I2C
Tuning the oscillator via I2C
LNA Test
Using BGA2869 IC, any LNA with the same footprint (SOT-363) can be used
Tested by Honza Říha, observations: every piece has different behavior (different gain), but the best piece is comparable/better than the original LNA.
Tested by Matěj Bartík, observations: a new batch of 3 PCB assembled in our infrared oven, all LNAs have same properties → they can't be soldered by hand.
Would we like to have an LNA onboard on DPACard?
DPA Card Development Board Rev2.1 (with RF transistor)
Number of items: 3 1 (Miškovský, Buček, Jeřábek)
One board was dead on arrival from assembly house.