Fault isolation and recovery in a distributed control network

• US 6,147,967 A
• Filed: 05/09/1997
• Issued: 11/14/2000
• Est. Priority Date: 05/09/1997
• Status: Expired due to Term

• Current Assignee: I/O Controls Corporation
• Original Assignee: I/O Controls Corporation
• Inventors: Kuang, Michael, Ying, Jeffrey
• Primary Examiner(s): Horabik, Michael
• Assistant Examiner(s): JONES, PRENELL P
• Application Number:
  US08/853,893
• Time in Patent Office: 1,285 Days
• Field of Search: 370/216, 370/221, 370/222, 370/223, 370/224, 370/242, 370/245, 370/249, 370/226, 361/95
• US Class Current: 370/222
• CPC Class Codes: G06F 11/2242 in multi-processor systems, e.g. one processor becoming the test master G06F11/2736 takes precedence

Abstract

A control network has a bus to which is connected a master node and a plurality of slave nodes in a loop configuration. Each of the master node and slave nodes has a transceiver connected to the bus through a shunt circuit. The shunt circuit comprises a switch (such as a relay) in parallel circuit configuration with an impedance element (such as resistor). The switch is operated under control of the node and under normal conditions remains closed, thereby allowing signals to be carried freely over the bus. When a fault occurs, each of the nodes opens its switch, causing the separate portions of the bus to become isolated. The master node then instructs each slave node, starting with the closest slave node, to close its switch. When a slave node fails to respond, the fault location has been detected. The master node repeats the process for the opposite direction of the loop. Operation then continues, with the slave node nearest the fault remaining isolated. If the short circuit occurs at the master node, one of the slave nodes may take over for the master node after a predetermined wait period.

Claims

1. A control network comprising:

a common bus for bi-directional communication;

a plurality of nodes communicatively connected to said common bus, said nodes comprising

a master node, and

a plurality of slave nodes connected to said master node in series connection such that each node is connected to two adjacent nodes, said nodes collectively forming a loop;

a plurality of fault detection circuits, at least one fault detection circuit located at each node; and

a plurality of shunt circuits, at least one shunt circuit located at each node, each shunt circuit comprising a controllable relay in parallel with an impedance element;

wherein said shunt circuits divide said common bus into a plurality of bus segments, each bus segment connecting two adjacent nodes, and each shunt circuit connecting two adjacent bus segments; and

wherein said controllable relay is disposed in a closed position when in a non-fault mode such that signals can propagate around said loop over said common bus, is opened at each node when the node'"'"'s fault detection circuit detects a fault so that signals cannot propagate around the entirety of said loop over said common bus, and is re-closed by the node a predetermined amount of time later in dependence upon a relative location of the node with respect to the master node.

2. The control network of claim 1 wherein said fault detection circuit comprises a timer, and wherein a fault condition is declared when a redefined fault detection time period is timed out by said timer.

3. The control network of claim 2 wherein said timer is reset periodically upon the detection of communication signals received at the node.

4. The control network of claim 1 wherein said impedance element at a given node comprises a resistance value selected such that communication signals from said master node will propagate through no more than one impedance element when said relay at each node is opened after detection of said fault.

5. The control network of claim 1 wherein said controllable relay is re-closed upon receipt by the node of an instruction from the master node.

22. The control network of claim 1, wherein each node comprises a transceiver, said transceiver connected to the shunt circuit at the node by a second impedance element.

23. The control network of claim 1, wherein each slave node is configured to take over as the acting master node a predetermined time period after detection of said fault, said predetermined time period unique to each slave node.

6. In a control network comprising a plurality of nodes connected in a loop and communicating over a continuous bidirectional common bus, at least one of said nodes being designated as a master node and one or more of said nodes being slave nodes, a method comprising the steps of:

(a) detecting a fault condition;

(b) isolating each node from its adjacent nodes except for an impedance between each node and its adjacent nodes by breaking said common bus into a plurality of bus segments, such that a separate bus segment connects each pair of adjacent nodes;

(c) transmitting over said common bus a de-isolate control message from the master node to the nearest isolated slave node with respect to the master node;

(d) receiving the de-isolate control message at said nearest isolated slave node unless the fault condition prevents receipt of the de-isolate control message;

(e) de-isolating said nearest isolated slave node in response to receiving said de-isolate control message at said nearest isolated slave node, thereby causing the nearest isolated slave node to become de-isolated;

(f) sending over said common bus a de-isolation acknowledgment message from said de-isolated slave node to said master node; and

(g) repeating steps (c) through (f) until a nearest isolated slave node fails to receive the de-isolate control message due to the fault condition.

7. The method of claim 6 wherein step (a) comprises the step of detecting the fault condition separately at each node.

8. The method of claim 6 wherein step (a) comprises the step of generating a fault detection signal individually at each node upon the timing out of a local watchdog timer.

9. The method of claim 6 wherein step (b) comprises the step of opening a relay at each node, each relay separating a pair of adjacent bus segments.

10. The method of claim 9 wherein step (e) comprises the step of closing the relay at the nearest isolated slave node in response to receiving said de-isolate control message from the master node, thereby re-connecting said pair of adjacent bus segments.

11. The method of claim 6 wherein step (g) comprises the steps of repeating steps (c) through (f) in one direction of the loop until a nearest isolated slave node in the one direction of the loop fails to receive the acknowledge receipt of the fault test message due to the fault condition, and repeating steps (c) through (f) in the other direction of the loop until a nearest isolated slave node in that direction of the loop fails acknowledge receipt of the fault test message due to the fault condition.

24. The method of claim 10, wherein step (b) further comprises the step of separating each pair of adjacent bus segments by an impedance element.

25. The method of claim 24, further comprising the step of connecting in parallel each relay and each impedance element separating the same pair of bus segments, and further comprising the step of connecting a second impedance element between (i) the relay and impedance element separating at a given juncture between two bus segments, and (ii) a node connecting to said juncture through said second impedance element.

26. The method of claim 6, further comprising the step of substituting one of said slave nodes as acting master node in the event that none of the slave nodes receives said de-isolate control message from the master node within a predetermined time period after detection of said fault condition.

12. A control network comprising:

a common bus for bi-directional communication;

a plurality of nodes communicatively connected to said common bus, said nodes comprising

a master node, and

a plurality of slave nodes connected to said master node in series connection such that each node is connected to two adjacent nodes and each pair of adjacent nodes is connected across one of a plurality of bus segments of said common bus, said nodes collectively forming a loop;

a plurality of fault detection circuits, at least one fault detection circuit at each node; and

a plurality of isolation circuits, at least one isolation circuit located at each node, each isolation circuit allowing signals to pass unimpeded across said common bus when in a non-fault mode impeding signals across said common bus by impeding signals between two adjacent bus segments connected to said node upon detection of a fault by said fault detection circuit, and resuming unimpeded passage of signals across said common bus after receiving a de-isolation instruction from the master node over said common bus;

wherein the master node transmits de-isolation instructions sequentially to each slave node over said common bus to determine the location of said fault.

13. The control network of claim 12, wherein said isolation circuit comprises a relay in parallel circuit connection with an impedance element.

14. The control network of claim 13 wherein said relay is disposed in a closed position when in a non-fault mode such that signals can pass unimpeded between the two adjacent bus segments connected to said node, is opened when said fault is detected thereby causing the impedance element to impede signals between said two adjacent bus segments, and is re-closed a predetermined amount of time after being opened in response to said de-isolation instruction being received from the master node.

15. The control network of claim 12 wherein said fault detection circuit comprises a timer, and wherein a fault condition is declared when a predefined fault detection time period is timed out by said timer.

16. The control network of claim 15 wherein said timer is reset periodically upon the detection of communication signals received at the node.

27. The control network of claim 12, wherein each slave node is configured to take over as the acting master node a predetermined time period after detection of said fault, said predetermined time period unique to each slave node.

28. The control network of claim 13, wherein each node comprises a transceiver, said transceiver connected to the isolation circuit at the node by a second impedance element.

17. In a control network comprising a plurality of nodes connected in a loop such that a portion of a continuous, bi-directional common bus connecting the nodes passes through each node, and wherein at least one of said nodes operates as a master node and one or more of said nodes operate as slave nodes, a method comprising the steps of:

(a) detecting a fault individually at each node;

(b) in response to said fault, increasing an impedance at each node of the portion of the bus passing through the node so that signals cannot freely propagate around the entirety of said loop;

(c) transmitting a control message from the master node over said common bus to a slave node adjacent to the master node;

(d) receiving the control message at said nearest slave node unless the fault condition prevents receipt of the control message;

(e) decreasing the impedance at said nearest slave node in response to receiving said control message at said nearest slave node;

(f) sending a response message from said nearest slave node over said common bus to said master node; and

(g) repeating steps (c) through (f) for each slave node in series until a slave node fails to receive the control message due to the fault condition.

18. The method of claim 17 wherein the step of increasing the impedance at each node of the portion of the bus passing through the node comprises the step of opening a switch at each node and causing each portion of the bus to be linked to the next bus portion by a resistor at each node shunting the switch.

19. The method of claim 17 comprising the step of repeating steps (c) through (f) for each slave node in the opposite direction from the adjacent slave node, except in step (f) that the impedance is decreased at the previous node instead of the newest slave node.

20. The method of claim 17 wherein the step of detecting said fault individually at each node comprises the steps of:

monitoring at each node communication signals carried over the bus;

timing out a fault detection period at each node;

resetting the fault detection period at a node when a communication signal is detected at the node; and

asserting a fault detection condition when the fault detection period times out.

21. The method of claim 20 further comprising the steps of:

timing a master failure detection period at a slave node;

internally resetting the master failure detection period at the slave node when a communication signal is detected at the node;

asserting a master node failure condition when the master failure detection period times out; and

substituting the slave node for the master node in response to the assertion of the master node failure condition.

29. The method of claim 17, further comprising the step of substituting one of said slave nodes as acting master node in the event that none of the slave nodes receives said control message from the master node within a predetermined time period after detection of said fault.

30. The method of claim 18, further comprising the step of connecting, at each node, a second resistor between a transceiver in the node and a common terminal of said first resistor and said switch.

31. A control network, comprising:

a plurality of nodes, said nodes comprising a master node and a plurality of slave nodes; and

a bi-directional common bus, said nodes connected to said bi-directional common bus in a loop configuration;

wherein each node comprises a shunt circuit and a transceiver connected to said shunt circuit through a first impedance element;

wherein said shunt circuit comprises a controllable switch in parallel with a second impedance element, said shunt circuit at each node connecting together two adjacent segments of said bidirectional common bus;

wherein each node maintains the switch at its respective shunt circuit closed when in a non-fault mode, thereby allowing free propagation of communication signals around said bidirectional common bus;

wherein each node opens the switch at its respective shunt circuit in response to detection of a fault condition, thereby connecting each pair of adjacent segments of said bi-directional common bus together by said second impedance element of said node; and

wherein, in response to said fault condition, the master node sequentially instructs the slave nodes to close the switch of each slave node'"'"'s respective shunt circuit.

32. The control network of claim 31, wherein said fault condition comprises the failure to detect expected communication signals within a predetermined fault detection time period.

33. The control network of claim 31, wherein a fault location is detected when the master node fails to receive an acknowledgment from one of the slave nodes in response to the master node'"'"'s instruction to the slave node to close its switch.

34. The control network of claim 31, wherein one of slave nodes takes over as the acting master node upon the failure of any slave node to receive an instruction to close its switch after detection of said fault condition.