Router takes Coordinators place

Hi,

1.What Sample application are you using?

2. Are you using a PRO project for sample application devices?

3. What are the compile options you have enabled for your ROUTER and COORDINATOR?

LPRF Rocks the World

I have the same question.

For my application I am using a CC2430 (128Flash) with ZStack 1.4.3-1.2.1 running in it.

As per our requirements, the PAN ID for our wireless network is set to a specific value (i.e. 0x0123).

How can we then remove the feature of the ZStack for PAN ID conflict resolution so that the PAN ID for a coordinator device will not increment when there is a conflict.

P.S. For our project we have an "aggregator" so we require that the coordinator must always have the same PAN ID.

Grant,

A few ideas for you to investigate:

1. Do have the capability to packet sniff?

if you do: 

2. Do you have the NV_RESTORE feature enabled?

if you do:

Please look for the following messages a coordinator realignment and orphaning, NWK rejoin and rejoin response between C and ED

LPRF Rocks the World

thanks for your attention,

I am not using PRO feature. 

in router

-DRTR_NWK                      // Router Functions

coordinator       

-DZDO_COORDINATOR                  // Coordinator Functions

-DRTR_NWK                          // Router Functions

i think there is a compile option SOFT_START  in old stack (1.4.3)   

Grant you can try this option.

but in stack 2.1   ,  it is not supported .[:(]

Hi LPRF Rocks...

I have the capability to packet sniff and enabled NV_RESTORE.

Looking at the messages this is what I observe:

Netwrok topology (1 Coor, 1 router, 1 ED)

Step 1: turn on Coor

  - See beacon request with no response

Step 2: turn on Router

  - See beacon request from router with response from Coor (indicating it is the coor with PAN ID 0x777)

  - See association and data request from router and both acknowledged by Coor

Step 3: turn off Coor and then turn it back on

  - See beacon request from Coor

  - See beacon response from router (indicating it is a router with PAN ID 0x777)

Step 4: turn on ED

  - See beacon request from ED

  - See beacon response from router (indicating it is a router with PAN ID 0x777)

  - See beacon response from Coor (indicating it is a Coor with PAN ID 0x778)

  - ED joins router

From the sequence I can see that when the coordinator was turned on again, it detected that a router with PAN ID 0x777 is already present. I believe that when this happened the Coordinator used the next PAN ID which is 0x778. Now, from my experience the previous ZStack-1.4.2-1.1.0 did not have this similar behavior. And I believe this feature is just recently added in this release as indicated in the README file (C:\Texas Instruments\ZStack-1.4.3-1.2.1\Documents\CC2430\README CC2430 Full.txt) which I quote:

"When a Coordinator starts a PAN and detects a duplicate existing PAN ID, it now calls ZDApp_CoordStartPANIDConflictCB() [in ZDApp.c] to assign a different PAN ID. "

Now, I would like to know if it is possible to remove this "fix"?

Otherwise, can you suggest and better alternative to workaround this issue.

Really appreciate any input on this.

P.S. As for SOFT_START, I'm not sure how that will help as it only allows the router to start as coordinator when it does not detect any. For your system the coordinator is the aggregator so all dat must go towards it and not any other wireless device (i.e. router) in the network.

Thanks,

Grant

Hi Grant,

SOFT START was a side track.

I guess the case or system scenario you are testing for is a Coordinator Failure, and network restore. Right?

Normally, this is a procedure that would need be handled by application level and ZDO on each device.

1. How do you detect a Coordinator Failure in your network?
2. How is this handled by the nodes that are connected to the Coordinator?

Normally it would be a orphaning or rejoin procedure started. The idea of the NV_RESTORE is to bring back the network to its orginal comfiguration.

Please verify that your coordinator as Z-Stack is restarted, tasks are initialized, and that your Coordinator as ZDAPP_Init is started will trigger the following event in the ZDOInitDevice(..)

From Z-Stack:

  if ( networkStateNV == ZDO_INITDEV_RESTORED_NETWORK_STATE )

    {

You can bypass this procedure by holding a specific button down to bypass the NV. If you do this your PAN coordinator should start with the PAN ID set in the config file.

Have you tried and test this with your coordinator by doing a reset and run? 

LPRF Rocks the World

Hi,

i'm implementing an application upon the TIMAC 1.2.1 and i have the following problem:

how can i permit all nodes (except the coordinator) to relay the beacons?

in my code there is these functions in osalTaskAdd:

MAC_InitDevice();

MAC_InitCoord();

(i'm using non beacon model)