[Tutorial] Z-wave diagnostics with PC Controller and Zniffer

[petergebruers 1,261]

Ah. My account and access predates the release of the 700 series, maybe that is the reason.

 

Can you try to download the old Z-Wave SDK for series 500 (blue button) via this page?

 

Please login or register to see this link.

 

EDIT 2019-11-18

 

Seems you need to download and accept:

 

"Z/IP Gateway Software Development Kit", this is

 

Please login or register to see this link.

 

Then as I said:

 

Register an account
Download Simplicity studio
Run it and select you want to de Z-Wave development and want Z-Wave SDK
You'll notice "request access" along the way.
 
In Studio there is a tools button, hidden under that is newest Zniffer and PC program.

 

The Zniffer firmware is in the older version of Zniffer, as @tinman already posted:

 
Zniffer

Please login or register to see this link.

 
Programmer

Please login or register to see this link.

Edited November 18, 2019 by petergebruers

[alex88 70]

17 minutes ago, NorthernRough said:

Thanks @petergebruers - this was the way I did it. 

But if I click @tinman's link to get the hex file, it also gives me the "Insufficient Privileges" issue.

Can't find a way around it.

 

PM me your email and I'll send you the exact files I used. They worked the first time without any issues.

This is probably because I go them from @robmacdirectly  

Edited November 17, 2019 by amilanov

[robmac 173]

21 minutes ago, NorthernRough said:

Thanks @petergebruers - this was the way I did it. 

But if I click @tinman's link to get the hex file, it also gives me the "Insufficient Privileges" issue.

Can't find a way around it.

A few people have had the issue since they updated the site. I think they do eventually grant access but it sometimes takes a while.

 

Happy to send the original files that have the firmwares .

 

8 minutes ago, amilanov said:

PM me your email and I'll send you the exact files I used. They worked the first time without any issues.

This is probably because I go them from @robmacdirectly  

 Sorry was reading down not up. 

[tinman 1,025]

i will check it with my Silabs contact, give me some days for update. 

[tinman 1,025]

Zniffer can be downloaded from here (scroll down, on right side Zniffer and Programmer links - with Firmware - are available and working for me)

 

Please login or register to see this link.

 

 

[tnesheim 8]

Hi!

I have finally made myself a Zniffer using this great thread. I have also tried to read all posts to understand...Looking in the SRC-column I see one device(157) making a lot of traffic. So, I thought that this number then would correspond to the device-ID in HC2. And by typing <myIP>/fibaro/en/devices/configuration.html?id=157 I would be able to "open" the "device tile" and hopefully do some adjustments? Am I right so far?

 

The problem is that the device is powered off and I have disabled the "tiles" in HC2 but it still generates traffic...this tells me that I'm doing/thinking something wrong? Please advice!

 

[petergebruers 1,261]

10 minutes ago, tnesheim said:

So, I thought that this number then would correspond to the device-ID in HC2. And by typing <myIP>/fibaro/en/devices/configuration.html?id=157 I would be able to "open" the "device tile" and hopefully do some adjustments? Am I right so far?

 

"Node ID" is not "Device ID". You will find Node ID on the "advanced" page of a device.

 

You can use this script to get a list of node id -> device id mappings

 

Please login or register to see this attachment.

 

Outputs a list like this

 

Please login or register to see this code.

So look for 157 on the left, then use your URL but with the master or slave device ID. You should be able to recognize that slave device (the master is rarely used or rarely named properly).

 

[tnesheim 8]

Thank you very much for the answer!

 

DeviceID...NodeID...guess I sleept in class:-)  So, I put this LUA-code into a VD and debug it for output?

[petergebruers 1,261]

10 minutes ago, tnesheim said:

So, I put this LUA-code into a VD and debug it for output?

Erm. It is a "Scene".

[tnesheim 8]

27 minutes ago, petergebruers said:

Erm. It is a "Scene".

Once again thanks! Think I´m getting old...

[petergebruers 1,261]

14 hours ago, tnesheim said:

Once again thanks! Think I´m getting old...

That's OK, you'll also get wiser. In general, and also when it comes to diagnosing Z-Wave  Did you find out what is going on with device 157 (making a lot of traffic as you have said)? You can post Zniffer questions here if they are "general" but maybe start a new topic if you have a specific issue.

[tnesheim 8]

Yes, the device with NodeID 157 was a Fibaro Double Switch that I probably missed or forgot to press save when I changed register 53 and 57 to the recommended valu... a really chatty device. I also found a two Aeon Multisensor that needed to quiet down due to settings. Now the z-wave traffic is often quiet for a few seconds and then some device reports every now and then....and of course when my door and motion sensors register movement. Sometimes there are some bursts of about 10 "telegrams". I have already minimized power/energy reports, as well as polling.
 

I guess I can do some more work on removing "jumps", but most devices talks to my controller 1-to-1. I guess that most things are ok with the Z-wave radio traffic. Next step is to look closer at my VD and Scenes...even though the CPU-load is on avarage below 10-15% some redesign according to best pratice is probably wise. 

 

In general is it possible to say something about the metrics of a healty z-wave network? 

 

[petergebruers 1,261]

2 minutes ago, tnesheim said:

Yes, the device with NodeID 157 was a Fibaro Double Switch that I probably missed or forgot to press save (...)

That is good news, seems like you've already found, and handled, some issues...

 

3 minutes ago, tnesheim said:

In general is it possible to say something about the metrics of a healty z-wave network? 

I would say, as a rule of thumb, "lower traffic is better" and "Z-Wave can handle burst of 10 commands per second". By "command" I mean something like "turn on a relay". That number might seem low but when you look at Zniffer you'll probably see that it depends or routed versus direct connection, that devices may report status, all leading to many more packets than you would have guessed from that simple "turn on". So while you theoretically can have "a packet + ACK per 20 ms" you would want to stay away from that theoretical limit. Also keep in mind, power reporting for example records data in your HC2 database (or influxdb or whatever ...) and when you actually count the number of updates per day, you might find some surprising numbers (I have seen users with 30.000 (thirty thousand) records per day.

 

Does this help?

[tnesheim 8]

Basically less is more:-)  Even more optimization of the z-wave radio traffic is indeed possible.

 

[tnesheim 8]

 

.

Edited February 12, 2020 by tnesheim

[Guest]

What does it mean when I have 50(113) in source and 110(1) in Dst?

Device 50 routed via 113 via 110 to 1 which is HC3?

It does not match with whats visible in Data as it says Routed:(113) - 50->110 - 32 - (1)

 

(It does not make sense that device 113 should go via 50 via 110 via 32 and finally to 1 as that will go zig-zag in my house on almost outer ends...)

 

Btw: anyone found a method to possibly export the list to excel to see which devices has the most telegrams?

 

Please login or register to see this attachment.

Edited March 4, 2020 by tortho

[tnesheim 8]

Sorry, that kind of route denotation I have not seen before. My have been simpler, and it was always possible to see/guess the route. 

 

I can assume you have several humor in your mesh network. Try to «fix network». In my HC2 I usually do this on the source node, and after a try og two the jumps have usually disappeared.

Edited March 8, 2020 by tnesheim

[Moucha 2]

Hello, I apologize for my English.
Can I ask someone for advice?
I am a Fibaro amateur, I made the network from the ground up myself.
180 physical devices, 2x HC2, 2x HCL
Following the instructions on this forum, I tried to optimize the network.
I started Sanity Check, and checked and reset almost all the equipment.
According to Zniffer, he found a few spam devices and will reduce network traffic by about 40-50%.
Could I ask you to watch my Zniffer record, or advise me if I can improve something?

The network now runs very reliably and very fast without delays and errors.

Thank you

Please login or register to see this attachment.

[alex88 70]

It looks like you have about 4 frames a second which is a little high.

 

I checked your explorer frames and you don't have many so that isn't the cause of the frame rate.

 

You need to check for devices that spam, for example:

- Device 59 (D3995E63) was sending over 2300 frames. You need to check this urgently.

 

EDIT:

I'm interested to understand why you have 4x controller for 180 physical devices. I have heard reports for other people that once they tune their network they have had better results after removing their HCL and keeping just 2x HC2.  For example, my setup has over 200 physical devices with 2x HC2.

 

Having a deeper look, I'd also check your settings on device ID 73 (D3995E63), 82 (EE89E9B8), 54 (D3995E63), 176 (D3995E63), 114 (D3995E63), 182 (D3995E63), 160 (D3995E63)

Edited April 29, 2020 by amilanov

[Moucha 2]

Good day

thank you very much for your answer
explanation of the device
59 - Smart Implant - motion sensors in the company garage - there is a lot of movement
73 - the same
54 - the same
114 - FGMS-001 - company garage

I don't think motion sensors will affect it (

82 - gas boiler (consumption measurement - corrected, thank you)
176 - UBS - pool temperature measurement - fixed
160 - DSW - fixed
182 - Steinel IS 140-2 - movement in front of the company garage

Why 4 units :))

there are 4 separate apartments in the house,
2 large with a large number of devices
2 small with a small number of devices
  it's probably a mistake but I wanted each apartment to have its own unit, later I merged them for easier operation. It's probably not good, but the idea that I'm redoing it - ufffff (

Thank you again for your willingness