Chip JN5169 hanging in Runing

[1039521082]

We find that JN5169 is hanging under some unknown circumstances. We control JN5169 reboot every 1min, after rebooting about 5hours, JN5169 board report ERROR: Extended status 82 and ERROR: Extended status 89 over 10 times, and then hang. When hanging, JN5169 no print, no timero runing, no DIO working, no communicating with host via uart1. After lasting about 8hours, we power JN5169 board again, and problem disappears.

We guess that JN5169 is hanging, and watchdog on chip doesnot work.

My environment:

1Host+1Coord+16Router, and topo is star

Our code is based on AN1229, with SDK 4170_v1840

I have some questions:

1.What is the clock source of watchdog on chip? systick, gptimer, orwakeup timer?

2.What is the reason of 10 times ERROR: Extended status 82 and ERROR: Extended status 89?

3.Is there some reason that SDK could be hanged without watchdog reboot?

4.What is the most relative reason of this?

[mario_castaneda]

Hi Xuanpu,

Our code is based on AN1229, with SDK 4170_v1840

The coordinator that you are working on is from the JN-AN1229? The routers that you have, are these from 1229?

What is the size of the routing table?.

Did you capture the packets in the air?

Regards,

Mario

[1039521082]

Hi Mario，

My Coordinator code is from AN1229, and Router is device based on Ti chip CC2538. Both Coord and router is using Zigbee PRO.

And my table size is:

Mac Table Size 102

Active Neighbour Table Size 102

Address Map Table Size 1

Broadcast Transaction Table Size 20

Child Table Size 51

Discovery Neighbour Table Size 8

Nwk Fc Save Count Bit Shift 10

Route Discovery Table Size 6

Route Record Table Size 8

Routing Table Size 51

The scene didnot appear in my place, so I did not capture air packets

[mario_castaneda]

Hi Xuanpu,

What are the modifications that you did in the Coordinator?

The coordinator only has the endpoint 0 and the custom cluster. Could you please try to use the JN-AN-126 as a reference?

Regards,

Mario

[1039521082]

We add about 100000 lines code into AN1229 Coord, mainly handle signal, sample data, transfer data, handle uart data etc. We use endpoint=10 and cluster is privite. We could not use 1216 as reference due to our code line

[mario_castaneda]

Hi Xuanpu,

 Is it possible that you could share your sniffer log? the packets in the air

What is the Maximum Number of Simultaneous Data Requests that you have set?

By default, the value is 5 without ACK and with ACK is 3.

Regards,

Mario

[1039521082]

Hi Mario, 

Question occurred weeks ago, I can't find ubiqua log,sorry.

Our ZC, ZED and ZR use API ZPS_eAplAfUnicastAckDataReq when sending unicast data. Our network traffic is about 25 frame per second. So I increase Simultaneous Data Requests size. 

Maximum Number of Simultaneous Data Requests 10

Maximum Number of Simultaneous Data Requests with Acks 8

Best regards,

Xuanpu

[mario_castaneda]

Hi Xuanpu,

Just to confirm, did you solve your issue?

Could you please provide the sniffer log? I am not sure if your Zigbee network has a lot of that Data Requests, it could give us an idea of this issue.

Regards,

Mario

[1039521082]

Hi Mario,

I have not find the reason. We handle the issue by adding a reboot way in host device.

I can not really provide sniffer log.

Our Zigbee network does not use data request by setting ZED rxonwhenidle=1 in device announce

PS, this issue does not occur again these weeks. We now guess  there is possible issue caused by EMI?

[mario_castaneda]

Hi Xuanpu,

Are you printing the status of the ZPS_eAplAf UnicastAckDataReq?

After changing the Maximum number of simultaneous data requests with Ack's

Did the device reset again?

Regards,

Mario

[1039521082]

Hi Mario,

I see the return value of unicast send, it is right. 

Reset is made by ourselves, that is not a problem.

We wonder why Coord could hang with no watchdog reset?

[mario_castaneda]

Hi Xuanpu,

I wanted to confirm that return value, the coordinator is looping for some packets that he is waiting for.

I am not sure how are you implemented the ZPS_eAplAfUnicastAckDataReq.

Is it possible that you could share your sniffer log? It will help.

Regards,

Mario

[1039521082]

Hi Mario,

I really don't have sniffer log. Question occurred weeks age, and not in my scene. 

After question occurred, I remade question, but question not appear.

Could you provide some guessing reason?

[mario_castaneda]

Hi Xuanpu,

Could you provide some guessing reason?

I could provide better feedback if you provide the sniffer log.

However, some recommendations.

 1) Mac Table Size = 102

  2) Active Neighbour Table Size + Address Map Table Size = 100 ( suggest that keep Active Neighbour table size as default value 26 in case no fragmentation for link status message, then Address Map Table Size can be set to 102-26=76)

  3) Routing Table Size can be set to the same value (76) as that of Address Map Table Size.

Regards,

Mario

[1039521082]

Hi Mario,

We have 3 connection topo, that is,

1. 50ZED connect ZC with star topo, all is JN5169

2. 50ZR connect ZC with star topo, ZR is one-hop cc2538 device and ZC is JN5169

3. 50ZR connect ZC with mesh top, all is JN5169

Could my ZC satisfy 3 topo at the same time with your recommendatory config?

Thank you

Best Regard

Xuanpu

[mario_castaneda]

Hi Xuanpu,

The 3 topologies connected to your Coordinator.

You need to be careful about the memory that you have. Keep in mind that the coordinator will save the short address, mac address and the key for each device.

If you have routers in your network, you will have a mesh topology.

Please look at the https://www.nxp.com/docs/en/user-guide/JN-UG-3113.pdf B.5.3 MAC Address Table

Regards,

Mario