Data logging in PM Designer: data loggers, historic tables and trend graphs

Creating a data logger

Right-click Data Loggers in the Project Manager and choose Add Data Logger, or use Panel > Data Logger > Add (Section 13.1). The General page (Section 13.1.2) defines what is sampled and when:

• Read Address and Sample Size: the first word of the data block to log and how many words, up to 1,023. One sample is one copy of that block with a time stamp.
• Number of Samples: the depth of the logging buffer. The memory comes from the panel's non-volatile pool shared with recipes and alarms; the compiler tells you when it does not fit.
• Sampling Method: Timed every 1 to 65,535 seconds; Triggered once per rising edge of a trigger bit, so the PLC decides when a sample is worth taking; Clocked at fixed moments of each hour (every minute, every 5, 10, 15, 30 or 60 minutes); or Passive, for a logger that only receives data loaded from a file or written by a macro.
• Notification bits that go on when the buffer is full or a sample is taken, and a Data Logging Macro that runs after every sample to process or convert the data and store it with SAVE2LB.
• Save logged data to file automatically at an interval, in CSV, TXT or LDF format, into the default folder or a path you give, with a file name that can include the date.

The Data Item page (Section 13.1.3) splits the sampled block into named items with a data type and display format, exactly as for a recipe block, and the address it shows for each item is where the latest sampled value can be read by any object. Names are per language and become the column headers in tables and files.

Logging to files

The logging buffer is bounded; files are not (Section 13.1.4). Three ways write a logger's data out: a function button with Save Logged Data To File or Save Logged Data (.ldf file); a command-block general command (operation codes 1, 14 and 10 for TXT, CSV and LDF); or the automatic periodic save on the General page. CSV is for people and Excel. LDF is the panel's own binary format and is the one you can load back into a logger or a file buffer to display again, provided the data item definitions match and the file has no more samples than the receiving logger holds. A common pattern is logger #2 sampling the process and saving temp_logged_YYMMDD.ldf every hour, and a passive logger #1 into which the operator loads any day's file to review it on a historic trend.

Name the folder with General Setup's default-folder rules (a new folder every day or month) so that a panel never exceeds the 4,096 files per folder limit, and point the folder at a USB stick or SD card on panels that have one.

Displaying logged data

Every historic object has a Data Source: the data logger's live buffer, or a file buffer (ID 0 to 15) filled by a function button with the Load Logged Data From File operation. Up to 16 file buffers exist per panel.

Historic data table

A grid with one row per sample and one column per data item plus date and time (Section 13.2). Scroll it with scroll button groups or a scroll bar, pick which items to show on its Data Item page, and filter rows on the Query page by a time range or by a Query Processing Macro that returns QRET(1) for rows to keep. Selecting a row can write that sample's values to a block of words, which is handy for a "load these settings" button.

Historic trend graph

Curves of up to several data items against time (Section 13.3). The Curve page assigns each curve a data item, range, style and colour; the Axis page sets the time range (hours, minutes, seconds, or a dynamic range the PLC can change), major and minor divisions, grids, and the time and date formats; a cursor lets the operator pick a moment and read the values, and the Query page limits what is drawn. The trend is drawn from stored samples, so it survives screen changes and power cycles, unlike a line chart that only shows what it has read since it was opened (see Chapter 6).

Historic event table, radar and scatter charts

A historic event table (Section 13.4) lists predefined events with messages, a historic radar chart (Section 13.7) draws up to 8 curves on a polar time slice with a cursor, and a historic scatter chart (Section 13.6) plots one data item against another over a time range with point selection and dynamic ranges. All read from the same loggers and file buffers.

File buffers and macros (V4)

V4 adds macro commands that treat a file buffer as a record set, documented in the FB_REC application notes:

$U11 = 0
$U10 = LB_2_FB(1, 3, $U11)       // copy logger #1's buffer to file buffer #3 (no filter)
$U12 = FB_REC_CNT(3)             // how many records are in file buffer #3
$U100 = FB_REC(3, 0)             // oldest record into the word array at $U100
$U12 = $U12 - 1
$U200 = FB_REC(3, $U12)          // newest record into $U200...
$U300 = FB_REC_DATE(3, 0)        // date of the oldest record
$U310 = FB_REC_TIME(3, 0)        // time of the oldest record

SAVE2LB stores a converted block in a logger from a data logging macro and SAVE2LB_T (V4) does the same with a time stamp you supply, which lets a panel log data that arrived with its own timestamp, for example from a batch controller. CLR_LB empties a logger and REF_H_OBJ redraws the historic objects after a macro has changed the data.

Getting the data to a PC

• USB stick or SD card: the periodic save or a function button writes CSV files you can carry away.
• FTP: the panel is an FTP server (Panel Setup > FTP/VNC); a PC can fetch the log folder over the network with a browser or a script, and a macro can push files to an FTP server.
• Data Transfer Helper: Get logged data from a chosen logger for a time range into CSV or TXT without touching the project (Section 19.2).
• MQTT on V4 for live values rather than files (Chapter 12).

Design notes

• Log what you will look at. A 10-word block every 10 seconds for a month is 2.6 million samples; a clocked logger every 5 minutes with a daily CSV covers most reporting needs in a fraction of the memory.
• Use a triggered logger for batch records: the PLC raises the bit at the end of each cycle and the sample is the cycle's result.
• Keep the data item definitions stable once files exist; a changed layout stops LDF files loading.
• Tick Keep Previous Data > Sampled Data when downloading a revision that does not change the loggers, and save to file anyway.