************************************************************************************** Version 102D April 12, 2012 Features changed since 102B: 1. Updated MXI file for configuring higher baud rates Fixed issues: 1. Vantive 43062 Errors on receiving frames smaller than six bytes from non-XBee 2. Vantive 43167 Errors on receiving remote commands from a PC. 3. Vantive 42857 X-CTU modem configuration screen didn't show RS485 options for D7 4. Fixed various internal errors that were causing instability, especially with TCP sockets and associated sleep. Known issues: 1. The PL command is not functional in Ad-hoc mode. (Transmit power does not change when PL is changed.) 2. Associated sleep depends on DTIMs from the access point. If the access point is lost (e.g. powered off) during associated cyclic sleep, then the module will go to sleep and stay asleep until the access point is restored and the module is reset. 3. Associated sleep is much improved, but the following issues are still known: a. Sometimes the SleepRq pin must be toggled more than once to awaken the module. b. Sometimes a module that receives data from a sleeper can stop receiving data until it first transmits. c. Two modules sending TCP frames to each other while operating in associated cylic sleep cease to transmit and can require a reset to recover. 4. Data loss may occassionally occur when transmitting in transparent mode with UART baud rate set to a maximum of 921,810bps. To alleviate (reduce probabilily of an issue), it is recommended that RO be set above 3 and if using a USB DIGI development board, set the transmit and receive buffer sizes in the USB settings to maximum size. 5. Unconnected ADC lines don't read maximum voltage if another ADC line is pulled lower than .88 volts and AV = 1 (2.5 volts). This is best resolved by connecting the ADC lines of interest. ************************************************************************************** Version 102B Feb 23, 2012 Features changed since 101E: 1. SLEEP_RQ/nDTR/DIO8 (XBee pin 9) is used by default to awaken and to sleep the module in pin sleep mode, even when using the SPI. In 101E, SPI_SSEL (XBee pin 17) was the only option available for awakening and sleeping the module in SPI pin sleep. In This release, SPI_SSEL may still be used, providing the D8 parameter is not configured as a peripheral (1). 2. Like previous releases, this release allows SPI operation to be forced by holding the DOUT (XBee pin 2) low at the time of reset. But this release also allows SPI operation by enabling the SPI lines and writing them with the WR command. If D2, D3, D4, and P2 are configured as peripherals and written to non-volatile memory with the WR command, then the SPI port may be used on reset rather than the UART, providing the first input comes from the SPI. Also, the UART may be disabled with the P3 and P4 commands so that the UART is never enabled and the SPI will be used, regardless of first input. 3. Options for mark parity and space parity were removed from the NB command. 4. WEP support was added by including a new parameter of 3 to the EE command. 5. I/O samples are formatted the same as in 101E except that the API identifier is 0x8F rather than 0x82 6. Associated sleep is disabled in AdHoc mode. If associated sleep is selected in AdHoc mode, deep sleep will be used instead. 7. Broadcasts transmission when in TCP mode. If either the configuration or the API packet specifies TCP mode on a broadcast, the broadcast will occur using UDP mode instead. However, a TCP broadcast in transparent mode may turn into a TCP unicast, providing there is a connected socket to any destination. 8. The ACK bit in the the TX_64 API frame is ignored when using the broadcast address. Fixed issues: 1. Receptions from TCP port 80 are now functional, providing there is only one connected TCP socket at the time of the reception. 2. If the access point cannot be reached for 10 seconds, then the association indicator (AI) will become non-zero. Loss of the access point also causes the associate LED to stop blinking if D5 is configured as a peripheral (1) 3. A radio that receives data on its TCP listening socket may immediately transmit a packet back to the origin on that same socket. 4. Fixed problem with data getting stranded on sender when streaming data at 921,600bps. 5. Fixed problem with breaking up data into packets smaller than the maximum packet size while streaming data. (The RO parameter was not working properly.) 6. Fixed issue with sending a single byte first on a maximum size packet in transparent mode. 7. Fixed issue with wrong or missing TX_status messages when TCP data is sent to an an invalid address. 8. Fixed the following issues with associated sleep: a. If data is received in associated cyclic sleep, the module will stay awake for at least ST time. b. If module awakens with RF data from the device or if transmit and receive data are interleaved, it will no longer go permanently asleep. c. In SPI associated sleep, the SPI_ATTN line (XBee pin 19) is no longer asserted while the module indicates it is asleep. d. Disable floating inputs during sleep to minimize power usage and to make it easier to achieve target power usage during sleep. 9. Fixed multiple issues with exception faults and buffer leaks. 10. Fixed RTS flow control so that the host can stop output on the UART. Known issues: 1. The PL command is not functional in Ad-hoc mode. (Transmit power does not change when PL is changed.) 2. Associated sleep depends on DTIMs from the access point. If the access point is lost (e.g. powered off) during associated cyclic sleep, then the module will go to sleep and stay asleep until the access point is restored and the module is reset. 3. Sometimes a transmission will indicate success when it is not received at the other end. This can occur in both TCP and UDP modes, and is more likely in associated sleep mode. 4. Associated sleep normally works. However, multiple issues remain which include the following: a. Sometimes the module can get stuck permanently asleep. b. TCP socket failures are seen more often with associated sleep mode than in full power mode. c. Sometimes no TX_STATUS message occurs following a transmission. In this case, the module will not sleep. d. A sleeping radio that is receiving bursts of data sometimes remains awake permanently. e. SPI associated pin sleep fails to receive data reliably while asleep. 5. Data loss may occassionally occur when transmitting in transparent mode with UART baud rate set to maximum 921,810bps. To alleviate (reduce probabilily of an issue), it is recommended that RO be set above 3 and if using a USB DIGI development board, set the transmit and receive buffer sizes in the USB settings to minimum size. ************************************************************************************** Version 101E Sep 15, 2011 New features: 1. Added support for AV command to allow the user to specify the analog VREF to be either 1.25 volts or 2.5 volts for analog reads. 2. Added support for PD command to allow both pullups and pulldowns on DIO lines configured as inputs. 3. Added support for PWM by allowing PWM configuration on the P0 and P1 commands and by allowing duty cycle specification with the corresponding M0 and M1 commands. 4. Power supply voltage is no longer included as part of the analog sample. However, this may still be obtained using the %V command. 5. Remapped SPI_nATTN from DIO9 (pin 13) to DIO1 (pin 19) to allow DIO9 to be used as an On_nSleep pin so that SPI master can know when the module is awake and ready to accept data. This means that DIO0 (pin 20) is the only pin available for analog input when using SPI. Fixed issues: 1. Watch dog resets upon wakeup from associated sleep mode have been eliminated. 2. Associated pin sleep and associated cyclic sleep are now functional when using the SPI. 3. Error no longer occurs at the end of an active scan. Known issues: 1. It was observed while streaming 30,000 bytes of data at 921,600bps in transparent mode that some of the data was lost or got queued up on the transmitter to be transmitted with the next burst of data. 2. Lockup on TCP transmissions was observed after continuous testing over three days. Transmissions ceased and TX_STATUS messages were not received. 3. After a radio in API mode receives data on its TCP listening socket, it cannot immediately transmit data back to the originating radio on that connected socket. Rather, it must wait for TM to expire (which is one second by default) before it can send back to the originating radio. 4. The 0x82 frame type used to output I/O sampling data formats the I/O sampling data like ZigBee rather than like 802.15.4 which is the only previously existing software that uses the 0x82 frame type. ************************************************************************************** Firmware Version: 1015 Release Date: July 6, 2011 Initial release of XB24-WF module Known issues: 1. Occasional data loss was observed when streaming large amounts of data in transparent mode at 921,600 bps on the UART when using the XBIB-U-DEV. 2. When configured for associated pin sleep (using either the SPI or the UART interface), a watchdog reset may occur on a pin wakeup event. 3. Data transmission does not work when using associated pin sleep with the SPI interface. This renders this feature unusable in this release. 4. The first transmission after waking up from associated cyclic sleep fails when using the SPI interface. 5. An ERROR is reported at the end of an active scan, even though the preceding data is valid.