This application note shows the IO update performance of the PLC IO Interface for AB 1771 Remote IO to a DeltaV Serial Card and also to the VIM Network Gateway.

VIM Interface Tieback timing (w/15 racks of IO)

This graph was produced using a customer database with the VIM ModbusTCP interface and PIO flash version 1.2.10. Mimic was used to simulate the DeltaV IO cards. AB 1771 RIO simulator running on a PC (this provides the same update process and interval as actual IO).

Most tieback time’s fell slightly below 400 or 500 ms. Base time is based on the DeltaV module scan rate (and order), VIM update rate, and RIO bus scan rate. Variation was due to PC tieback process and communications interval variations.

DeltaV tieback module read IO from a serial card input “landing” module and wrote the outputs to another serial card output “landing” module. This requires three DeltaV modules; each is running at 100 ms. Input GESTERC17-6-1-DI/I-060-3 maps to rack 15 DI 3, GESTERC17-6-1-DO/O-062-4 maps to rack 15 DO 4.  The tieback is monitored in the simulation PC rather than using the timing functions in this module.

The response time variation is based on four asynchronous process scans:

1. DeltaV module scan rate, 3 modules each running at 100 ms (300 ms total).
2. VIM (rate running at about 10-30 ms for 15 racks.)
3. VIM to PIO module (10 MB Ethernet connection)
4. RIO bus communications (about 105 ms for 15 racks)

Grouping of tieback time is significantly better with ModbusTCP (435+-48 n= 1503 [range 300-561]) than RS485 (943+-156, n=5659 [range 481-1482]) with the Digester database. There are two visible bands of tieback timing, the second band is due to the instances where the asynchronous updates to the DeltaV miss an input scan and therefore are processed 100 ms latter.

RS485 Serial Card Interface Tieback timing (w/15 racks of IO)

This graph was produced using a customer database with the DeltaV programmable serial card (series 2) interface and PIO flash version 1.1.17. Mimic was used to simulate the DeltaV IO cards. AB 1771 RIO simulator running on a PC (this provides the same update process and interval as actual IO).

DeltaV tieback module read IO from a serial card input “landing” module and wrote the outputs to another serial card output “landing” module. This requires three DeltaV modules; each is running at 100 ms. Input GESTERC17-6-1-DI/I-060-3 maps to rack 15 DI 3, GESTERC17-6-1-DO/O-062-4 maps to rack 15 DO 4. The tieback is monitored in the simulation PC rather than using the timing functions in this module.

1. DeltaV module scan rate, 3 modules each running at 100 ms (300 ms total).
2. DeltaV serial card (rate running at about 300 ms for 7 racks on each port.)
3. Serial card PIO module scan (500K per port, 2 ports)
4. RIO bus communications (about 105 ms for 15 racks)

The speed and grouping of tieback time is significantly worse than with ModbusTCP (943+-156 n= 5659 [range 481-1482]) versus (435+-48 n= 1503 [range 300-561]). Five distinct bands of group values are visible, these represent the various permutations of the 4 asynchronous process scans that make up the values.

VIM Interface Tieback Time vs IO racks

Compare Tieback time to number of IO racks communicating. IO was simulated on a PC using a RIO card. This provides the same update process and interval as actual IO. The test was run with the RIO bus set to either 115K or 230.4K. All racks were processed on one VIM, all serial card datasets configured on one port (2) under one device. The other port was disabled. With ModbusTCP tieback time is independent of number of racks from 1 to 6 (10 at 230.4K), while with the serial card there is a direct correlation between number of racks and length of average time. With ModbusTCP connection, the maximum ranges from about 200 to 400 ms, while with RS485 this starts at 200, but climbs to 1200 ms.

VIM Interface Tieback Time vs IO racks (two ports)

Compare Tieback time to number of IO racks communicating. IO was simulated on a PC using a RIO card.This provides the same update process and interval as actual IO. All racks were processed on one VIM, all serial card datasets configured on two ports under one device each. Started with all datasets configured, removed datasets as no longer required for rack access. Disabled port when it was no longer necessary. Did not include Global or BT datasets.

Tieback Time (n=200)

Serial Card interface

Compare Tieback time to number of IO racks communicating. IO was simulated on a PC using a RIO card. This provides the same update process and interval as actual IO. All racks were processed on one DeltaV programmable serial card (series 2), all serial card datasets were configured on two ports (running at 500K) under one device each. Started with all datasets, removed datasets as no longer required for rack access. Disabled port when it was no longer necessary. Did not include Global or BT datasets.

Tieback time (ms, n=200)