Issue 140 March 2002
Spy-Size Event Logger

Event Group

This last group is the working registers for the Event Logging mode of the DS1678. The first eight registers get written following the first event of the mission (logging session). They are a copy (time stamp) of the RTC registers reflecting the exact time of the mission’s first event. The time each event occurs is based on this start time. Time between events (the logged data) is the elapsed time from the last event. Logged data consists of a two-byte elapsed time count. Each count is equal to the configuration set in the control register (bits 4 and 5, see Figure 4). This makes a minimum count of one equal to 1 s, 1 min., or 1 h.

Naturally you want to pick the resolution that best fits your minimum time between events. Events will be logged even if they occur at a rate quicker than the chosen resolution. This would be data logged with an elapsed time count of zero. For instance, if you chose hour increments of the elapsed time counter and events happened every 15 min., you might see three logs of zero elapsed time counts and one log with one count. Independent of logging, which consists of saving and clearing the elapsed time counter (ETC), the RTC is in charge of incrementing the ETC on a time match determined by the duration interval bits in the control register.

On the other hand, you must also pay attention to maximum times, which are 65,535 times the resolution. Therefore the maximum ETC for the second’s duration interval is 65,535 s, or 18.2 h. If your event happens once per day, the elapsed time count will overflow (and roll over to zero). So again you can see the importance of choosing a resolution that fits both your minimum and maximum times among events.

Following the time stamp registers (48–55) is the event zero elapsed time from last event counter 2-byte register (56–57). At the beginning of the mission, this register pair is written with zeros. This designates the first event and the start of a mission. If the RO bit is set (allowing rollovers), this register pair will hold an elapsed time count since the last event (event 1024) after the rollover has occurred. (This is one way of determining a rollover. But, as indicated, an ETC of zero is possible.)

The next three registers (58–60) hold a 3-byte count of the number of events that have been logged since the mission began. Ordinarily this register would increment to 1024 and the logger would be full of data. However, you have no real idea when this might occur, so even if you’ve configured the logging to stop at 1024 events (the maximum), the DS1678 can continue counting events even if it can no longer log them.

So far I haven’t discussed the logging memory, in fact, you may have noticed it isn’t shown in any of the previous figures. That’s because you don’t have direct access to the logging memory. You can’t write to it, only the event logger can. This prevents data tampering. Any attempt to access the data halts the mission and locks in the data. After it’s been halted, the only way to begin a mission is to clear the memory and start from scratch.

Following the ETC register pair is an address register pair. This register pair (63–64) points to the next logging memory address or the location where the ETC will be stored on the next event. Initially it will point to address zero (2047 being the last logging memory address prior to rollover). The 2-byte ETC count will be stored at the pointer at the next event and the pointer will be incremented (two consecutive bytes for each 2-byte ETC value). When RO equals one and a rollover occurs, not only is the ETC stored in register pair 56–57, which was zero, but a new RTC sample is stored in the time stamp registers (48–55). Events will continually be logged, overwriting the oldest data with new.

Even though you are prevented from writing to the logging memory, there must be a way to retrieve the logged data. The last three registers (65–68) hold the key. The first two registers are an address pointer pair enabling the user to read the logging memory. The last register will hold the data pointed to by the first two address registers. The address pointer will automatically increment for multiple reads to the data register (43), so you can skip having to continually write the address to read multiple data bytes. When you’re pointing to the address of interest, you can read as many multiple bytes as you wish. Note multiple byte reads are not supported in the simple serial-to-I2C converter in this project, so you’ll have to update the address for each read.