RIO unit - Release Notes

Software in production: AMC 09.01<br\> Software released date: 2003-05-06<br\> Note: Additional features available in 08.27<br\> Note 1: We sometimes do bugfixes in older versions while working with a new version. You can’t read the list version by version, and always assume that a correction is included in the next. Be aware of high release numbers, e.g. 06.05 vs. 07.01. Check version dates, and also comments for each version. <br\> Note 2: For each software version the NVRAM version is listed. If the NVRAM version is different, the AMC board must be coldstarted, and then you must do a SendAll from AlphaPro to restore the configuration.<br\><br\>

RIO 02.13 (2001-01-25)

Release : New MCU 68HC11E1 replacing 68HC11A1.

EEPROM and Watchdog enabled on the new MCU.
Software is still compatible with 68HC11A1.

Functional changes

EEPROM protection in the new MCU removed.

The EEPROM and CONFIG are protected by the BPROT register that was not used in 68HC11A1. It is cleared during startup to enable EEPROM and CONFIG writings.

New IAR version used.

This should not affect RIO functionality, but has required some changes in the project files "cstartup.s07" and "lnk6811.xcl".

Errors Corrected

Terminal commands were unavailable after the first one.

Observation: The RIO did not respond to commands after the first one unless a hash (#) was applied in front, but this should not be necessary. Correction: The terminal command function was not prepared for the "LF" character following the "CR", and would take it as the first character of the next command.

Restart time stored as random number

Observation: A RIO that is powered up would store the reset time as a random number (often FFFF) rather than 0000. When the correct time is received from the exchange max. 20 seconds later, this time is stored as reset time. I.e. the problem is just cosmetic, but it may be confusing. Correction: The instructions storing the time were executed before the time had been initialized. Fixed.

RIO 02.12 (1999-01-27)

Release: Standard, bugfixes and minor additions

Recommended used with AMC 07.30 (AMC 07.09) or newer.

Functional Changes

Fast Reset of RCOs when no traffic on data bus

The RIO previously held RCOs active for 3-5 minutes. This may cause problems for applications which have manual backup in case the exchange fails, as there is no indication of exchange failure. This has now been changed so that no data bus activity will reset RCOs after 15 seconds. If bus activity resumes, the exchange is informed, and it will refresh RCOs from it’s internal state.

DIP switches read periodically

It’s now possible to change the switches, and they take effect immediately. Every time a switch is changed, the RIO resets to adjust for the new setting. If a terminal is connected, you can see the device number, protocol version and link mode as part of the reset printout. (If you select point-to-point, the terminal is connected to the exchange port and there will be no immediate response, but when the RIO becomes lonely after 15 seconds it will print the switch settings as a reminder).

Communication statistics logged every hour

One line with error counter plus message counter for own messages and total messages on the bus is printed. Printing one per hour makes it easy to see if communication problems occurs on certain times of the day, and if it’s dependent on the volume of RIO data protocol traffic.

Support for simple logging for the exchange

The data protocol messages STATUS_REPORT and LOG_STRING are received and printed. With AMC 07.30 it’s possible to re-direct the exchange’s log port to external devices like PCs, and the RIO has been used to test this feature and is now able to work as a log device. (.ex_profile.line_monitoring.report1_addr) Properties:

- Requires RIO to be on a multidrop bus, and there should be no normal RIOs on this bus (due to speed requirement for RCI/RCOs)
- Proper time stamp incl. seconds
- Line error messages for stations are printed in a simple format (no details)
- The event handler’s LOG action generates text which is printed here
- Log can contain RCI/RCO activity, or just exchange activity

With AMC08.00 it will be possible to create several call logs, e.g. one for each guard in a parking. You can then place one RIO near each guard!

Message terminal trace for broadcast messages

Added trace of broadcast messages. Note that a RIO bus should normally NOT operate with broadcast active, as it generates a lot of extra messages thus reducing the response time drastically. The addition was done to extend the RIOs bus analyser capacity in general.

PONG message optional parameters adjusted

02.10 had a message format which was re-considered due to some other on-going development work. As these parameters are optional and just human readable for debuging, it was decided to change quickly.

DEVICE_INFO optional parameters added

Preparations for error reporting in AMC 08.00 where the log port’s device up messages may contain device name and software version info. This info is collected by the other RIOs, and can be inspected using the terminal D (show all devices) command. Observe that this is only sent after reset, and if this RIO has been reset after the others, the D command will just show “unknown”.

Changed internal time reference from counter to RTI

The previous softwares used a free-running counter which was re-started by software to measure time. This caused some load-dependent variations. Now the processor’s built-in RealTimeInterrupt is used, giving a very steady reference.

Error dump with backtrace

When a data line error occurs, the last message is dumped along with the error message, thus making it possible to see link details even if the data protocol trace is switched off.

Test commands

Several commands used for software testing which can cause controlled data link errors to see that both the RIO and the exchange behaves correctly.

EEPROM stores trace mode and reset time

The RIO’s microcontroller 68HC11A1 contains EEPROM which becomes active with the coldstart procedure for the watchdog. This is now used to store the protocol trace mode over reset, and it logs the number of resets, plus the time of the last reset.