Known Issues

REV1 Board

The position of the CTS and VCC pin are swapped on the serial connectors. This doesn't affect anything unless you try to power the board from the serial connector by jumpering JP1 or JP2. It can be fixed by crossing the CTS and VCC wires running from the FTDI adapter to the Serial connector on the board, or simply leaving JP1 and JP2 unpopulated and powering the board from an external supply via the RC2014 bus.
The DTR trace near the edge of the board runs too close to the Serial 1 pads. On some boards, the solder mask near the ground pin does not cover the DTR trace, so a solder bridge can short them together. This can prevent the AVR from automatically resetting when programmed by avrdude and can damage the serial to USB adapter. This issue occurred on two of the three OSH Park-manufactured boards that I received. Other boards built by Seeed Studio using the same design files did not exhibit this problem. The problem can be fixed by cutting the Serial 1 DTR trace on either side of the GND pad and soldering a bodge wire from the DTR pad of the Serial 1 connector to the DTR pad of C7.

The position of SPIA0 and SPIA1 on the 74HCT139N are swapped. This has been fixed in software by swapping the definitions for SD_SEL and AUX1_SEL in spi.h. The BOARD_REV variable in the Makefile controls which SPI addresses are used. For BOARD_REV 1 and 2, the swapped addresses are used. For BOARD_REV 3, the originally intended addresses are used.
It is possible for the serial connection to provide enough leakage current to keep the SD card powered and therefore it may not be reset properly on power up. If the main RC2014 power switch is turned off while the serial adapter remains connected, the SD card may not work when the RC2014 is turned back on. To fix this, either momentarily disconnect the serial adapter to remove all power to the board, or eject and reinsert the SD card to remove power and reset it.
Newer features requiring outbound IORQ control are not supported on an unmodified REV1 or REV2 board.

J4 (IO WAIT ADDR) jumper must be shorted in the far right position (all addresses). Other settings will not work correctly. Originally the intention of this jumper was to allow wait states to be introduced only for a subset of the IO addresses that the AVR is meant to handle; however, since REV3, z80ctrl the AVR is connected directly to the IORQ line so it sees the signal even if a wait state is not generated and will attempt to check the address bus to see if the address requested is one it should handle. If wait states are not enabled for that address, the Z80 will have already moved on, and the address on the bus is no longer valid. The AVR may erroneously attempt to handle the IO request for the invalid address and cause bus contention.

The REV5 board was a redesign and had several problems that have been fixed in REV6. Use REV6 instead.