Profinet Diagnostic tools

A recent thread about Profibus diagnostic made me think about Profinet diagnostic.


Are there any good tools and software out there that you use to diagnose Profinet networks?

Siemens' own Proneta. Works pretty well when you have to check a large network. You can also do IO-test on the ET200SP platform.

Proneta is a good tool but it does not give you any in depth diagnostics about your network.



Surely the same problem with poor cabling, reflections,delays and retransmission's must exist also for Profinet, or is Profinet just some much more forgiving in its design that there is less need for in depth diagnostic tools

I have been searching for the same thing, but hasnt been able to find anything.
The closest is an Ethernet cable tester. But that only test the cable itself.

Bratt said:

Surely the same problem with poor cabling, reflections,delays and retransmission's must exist also for Profinet, or is Profinet just some much more forgiving in its design that there is less need for in depth diagnostic tools

Click to expand...

One can argue that Profinet is more forgiving than Profibus since Profibus is electrically connected from end to end, it is more difficult to pinpoint an error, and an error can have more grave consequences since an error can shutdown the entire bus, rather than just one segment.

One thing that Profibus gives you that Ethernet/Profinet does not, is that you can tweak the bus parameters manually. I have on a few occasions increased the "retries" from default 1 to 3 or 5 to make the bus more tolerant to noise.

https://procentec.com/products/mercury/?content-1

We purchased this last autumn. It reports network load and discarded packets on each device, ARP requests, detects duplicate IP addresses in the network, etc. We haven't had that much use for it yet, but it has been helpful in couple of situations to track problems in the network. My understanding is that it can be used on most of the industrial ethernet fieldbuses, but so far we have used it only with Profinet.

Other plus side is, that the ProfiTrace mentioned in other thread can be connected directly to this one.

Ah, of course I should have checked the company that also makes the profitrace if they also have somehing for Profinet.

Apart from that, what you descibe sounds like it checks for "logical" errors, but not for noise, waveform or grounding problems.
For a Profinet system, does it do something that the Siemens Topology diagnostics do not do ?

One thing that Profibus gives you that Ethernet/Profinet does not, is that you can tweak the bus parameters manually. I have on a few occasions increased the "retries" from default 1 to 3 or 5 to make the bus more tolerant to noise.

Click to expand...

Dosen't changing the "Accepted cycles without data" and/or update time give you somewhat of the same functionality? This settings can be set individually for each node. Sometimes i need to change this to get the MRP functionality to work properly to avoid errors when the bus "switch" direction

Hietsu said:

https://procentec.com/products/mercury/?content-1

We purchased this last autumn. It reports network load and discarded packets on each device, ARP requests, detects duplicate IP addresses in the network, etc. We haven't had that much use for it yet, but it has been helpful in couple of situations to track problems in the network. My understanding is that it can be used on most of the industrial ethernet fieldbuses, but so far we have used it only with Profinet.

Other plus side is, that the ProfiTrace mentioned in other thread can be connected directly to this one.

Click to expand...

Thanks I will look into this piece of hardware.


Do you know if it also detects duplicate mac address? I experienced very strange network problems when 2 devices get the same mac adress. Once a programmer managed to copy another CP card with a manually set mac adress and therefor 2 devices got the same mac. Another time I managed to get the same mac as a colleague when i copied a VMvare cloned. It took sometime to pinpoint the cause of the problems in both cases.

Bratt said:

Dosen't changing the "Accepted cycles without data" and/or update time give you somewhat of the same functionality? This settings can be set individually for each node. Sometimes i need to change this to get the MRP functionality to work properly to avoid errors when the bus "switch" direction

Click to expand...

I guess it does.
I havent been desperate enough with a Profinet system to have to find where it can be tweaked. I have been desperate with a profibus system.

Btw. when I had the problem with profibus network I had to tweak it to keep it running. And it was a profitrace that finally made it possible to pinpoint the real culprit (poor earth in the entire facility).

JesperMP said:

Ah, of course I should have checked the company that also makes the profitrace if they also have somehing for Profinet.

Apart from that, what you descibe sounds like it checks for "logical" errors, but not for noise, waveform or grounding problems.
For a Profinet system, does it do something that the Siemens Topology diagnostics do not do ?

Click to expand...

You are correct, it works on the protocol level and does not measure any physical attributes. From the reports it gives it is possible to narrow down the areas with issues, but does not explicitly point out the problem.

I have to admit my exposure for Proneta is very limited, so at the moment I cannot compare these two. For the ET200SP system this is probably more limited, since from what I gathered you can track changes of input and output states with the Proneta.

Bratt said:

Thanks I will look into this piece of hardware.


Do you know if it also detects duplicate mac address? I experienced very strange network problems when 2 devices get the same mac adress. Once a programmer managed to copy another CP card with a manually set mac adress and therefor 2 devices got the same mac. Another time I managed to get the same mac as a colleague when i copied a VMvare cloned. It took sometime to pinpoint the cause of the problems in both cases.

Click to expand...

I can't say for sure how it would react in case of duplicate MAC address, since I haven't encountered that issue before. It provides a list of found devices and their MAC addresses, among other info, so it would be possible to look for it from there, but of course in larger networks it would be very cumbersome and frustrating. I might be able to test the double MAC addresses, if I'm able to visit my office in the near future.

It does highlight duplicate IP addresses, though.

A lot of the Scalance managed switches from Siemens I've used have a built in cable tester function.

There are Network Management Software (NMS) packages out there that work as something like a SCADA system for your network. They can pull data out of all the managed devices on the network, show topology, show alarms from things like packet error counts, etc. Siemens has one that is of course designed to work with Siemens PLCs in addition to standard network components.

There are errors that can arise from just about every level of the standard ISO/OSI model. https://en.wikipedia.org/wiki/OSI_model

The best way to check for Physical errors (layer 1) is to do cable testing. In theory, a bad cable between two devices in the network HIGHLY unlikely to be able to cause an error anywhere else.

From there, you are at the Data link layer (layer 2). You can have lots of errors here that get diagnosed in a ton of different ways. If you wired up a loop, it'll shut things down without any ability to diagnose. Something like Proneta would show you if there were duplicate MAC addresses as it discovers your network topology.

Proneta is also a big help at the Network Layer (layer 3). This would show you some IP based problems (duplicate IP,

From there we pretty much jump to the Application Layer (layer 7). This is where Profinet sits, and where a protocol dissector can look at the traffic between the PLC and the IO to see if something is going wrong. In my experience, Protocol dissectors have the most value when you fear one of the devices is not conforming to the standard. The Diagnostics in the PLC do a decent job of telling you what's going wrong otherwise (timeouts, etc). It might not say WHY, but it tells you what happened.

Thank you for those insights mk42!