• 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Ethernet trouble(Hardware failure?)
#1
I have ethernet trouble with a certain netPI device(MAC addr: B827EB538466)
During a data transfer through ethernet connection, the connection may lost suddenly.

In syslog, I found a log message that "usb usb1-port1: disabled by hub (EMI?), re-enabling".
Since the recorded time of this log message is well corresponded with the time of connection lost, I suppose this is the root of this problem.
Based on the fact that the ethernet on the LAN9514, which is a microchip on raspberry pi, runs from usb, it is reasonable.


Since this problem does not occur with another netPI device(MAC addr: B827EBEFDC49), I suppose some hardware failure occurs this problem.
However, I am not sure.
So, Please tell me more if you know about this problem.

Best regards,
  Reply
#2
Hello,

yes the used microcontroller within netPI device is the LAN9514 chip which is responsible for both the USB communications and the Ethernet communications the same time.

So let me ask the following detail: do you have anything connected to the USB ports of your netPI?

Thx
Armin
You never fail until you stop trying.“, Albert Einstein (1879 - 1955)

  Reply
#3
Dear Armin,

Thank you for your reply. I will answer your question.

Nothing is connected to the USB ports of both of my netPIs.

Best Regards,
  Reply
#4
Ok so there is no current consumption over USB ports which could influence your netPI powering.

I searched on the Internet about "disabled by hub (EMI?), re-enabling" message and I found a lot of hits.

Are you sure this log message came during netPI runtime and not from the beginning during the standard bootup sequence? I have not analysed the netPI standard message log after a fresh restart sequence ever ... so I don't know what messages are all traced by default.

If you say you have "ethernet trouble" ... which extract trouble is this then? Is netPI no more available/reachable over Ethernet any more? Or is this just a temporarily trouble of a short time? What do you need to do with your netPI to recover from this state? Re-powering it or does it recovery by itself after a time?

Then a question to the installation position of your netPI devices? Is this a productive system under real machinery conditions with EMC and EMI? Or are your two netPIs just running in more laboratory environment?

Thx
Armin
You never fail until you stop trying.“, Albert Einstein (1879 - 1955)

  Reply
#5
Dear Armin,

Thank you for your research. I will answer.

>>Are you sure this log message came during netPI runtime and not from the beginning during the standard bootup sequence? I have not analysed the netPI standard message log after a fresh restart sequence ever ... so I don't know what messages are all traced by default.

I am not sure. However, I think it came during netPI runtime and not from the beginning during the bootup sequence.
The attached file is the syslog which contains the message below. I think this log message is on netPI runtime.
<3>1 2021-06-16T02:31:05+00:00 NTB827EB538466 kernel - - [meta sequenceId="1"] [  399.529904] usb usb1-port1: disabled by hub (EMI?), re-enabling...


>>If you say you have "ethernet trouble" ... which extract trouble is this then? Is netPI no more available/reachable over Ethernet any more? Or is this just a temporarily trouble of a short time? What do you need to do with your netPI to recover from this state? Re-powering it or does it recovery by itself after a time?

It is a temporal trouble of a short time(several seconds). However, I use camera devices through ethernet, and the temporal lost of ethernet connection causes fatal error for the camera devices. Therefore, what I need is not the way to recover, but the way to not cause this temporal connection lost.
However, this temporal connection lost occurs for only a certain netPI device in my environment. So, if this is a problem on a certain device, such as a hardware physical matter, I can buy another netPI device and use it instead of the deteriorated one. In contrast, if it is not unique problem for this certain netPI device, I have to deal with this problem.
So, what I am most concerned about is whether this problem is more likely to be a unique problem for a certain device, or not.


>>Then a question to the installation position of your netPI devices? Is this a productive system under real machinery conditions with EMC and EMI? Or are your two netPIs just running in more laboratory environment?

Our product is something like a productive system under real machinery conditions, not in laboratory environment.
However, this connection lost problem also occurs on my office desk. There are some wifi networks, but it is not so harsh environment.


Best Regards,


Attached Files
.zip   syslog_usbError.zip (Size: 40.86 KB / Downloads: 0)
  Reply
#6
Ok, what I can do is looking to my personal netPI and analyse the log file and compare it to your log file.

As next I am asking the development team when Linux or driver will be reporting "disable by hub"

I have one more question about the type of OS you are using on netPI. Is it the original netPI OS and original SD card from Hilscher, or are you using any other OS like Raspbian OS for your task.

We need to narrow down the problem. I still do not believe that netPI has an internal hardware problem. So next to the pins + and - have you also grounded the netPI well with the 3rd pin of the power connector? Also the device has a clamp to fit on the DIN rail ... is this DIN rail put to ground potential as it is usual for industrial installations?

You say you have trouble with the Ethernet connection ... is it possible to put an Ethernet switch between the netPI device and the rest of you LAN network. Of course just for test purposes only? How often is the Ethernet connection to netPI lost per day? During night times, during day time? During intensive production times?

You say you have two netPI devices ... are both netPIs used in similar environment so that you can say both applications are the same and the whole setup can be compared to each other for both netPI?
You never fail until you stop trying.“, Albert Einstein (1879 - 1955)

  Reply
#7
Dear Armin,

Thank you for your support. I will answer.


>>So next to the pins + and - have you also grounded the netPI well with the 3rd pin of the power connector? Also the device has a clamp to fit on the DIN rail ... is this DIN rail put to ground potential as it is usual for industrial installations?

No, the netPIs are not grounded with the 3rd pin.


>>You say you have trouble with the Ethernet connection ... is it possible to put an Ethernet switch between the netPI device and the rest of you LAN network. Of course just for test purposes only?

I'm sorry but I cannot understand the purpose of this test. However, it is possible. I will use GS108PE-300JPS as a ethernet switch. Is it all right?


>>How often is the Ethernet connection to netPI lost per day? During night times, during day time? During intensive production times?

I have not counted it yet. However, The connection lost usually occurs within 30 minutes after the start of ethernet communication.
Our product is not intended to running constantly. When needed, the product get upped. So, there is no specific time the problem occurs.
 

>>You say you have two netPI devices ... are both netPIs used in similar environment so that you can say both applications are the same and the whole setup can be compared to each other for both netPI?

I think they are in very similar environment, whose netPI firmware versions are the same, their LAN settings on Control Panel are the same, whose containers on docker is built by the same image.
Usually, they use another power source machine, but I have tested them by using the same power source machine. However, the problem only occurs for the specific netPI.
The test application is a windows console application which sends files repeatedly to a netPI with scp, through ethernet connection. So there is no test program on netPI. Therefore, there is no application difference.

The only difference which I have realized is that the bad netPI is surrounded four side by metal box, and near to camera device(During my ethernet test, the cameras are powered off), on the other hand, the good netPI is not surrounded and not near to camera device.


Best Regards,
  Reply
#8
Well to your question about "I'm sorry but I cannot understand the purpose of this test" ... since 25 years with hilscher I collected also experiences with malfunctions of devices caused by EMC problems conducted over the cable. So if you say that the Ethernet connection is instable ... and 30 minutes is really a short time in my eyes ... then it is reasonable to think also about an EMC disturbance that might be distributed over the Ethernet cable causing the trouble. So my idea was to put just for a test a simple Ethernet switch in between the main Ethernet cable and the netPI and see if the problem is still happening. So in my eyes any Ethernet switch would do a good job for this test.


You say the malfunctioning netPI is installed in a box ... how about to exchange this netPI with the good working netPI while changing the application. In this case if the good netPI is still working good then indeed the other netPI is the reason for the trouble and needs to be exchanged.

Thx
Armin
You never fail until you stop trying.“, Albert Einstein (1879 - 1955)

  Reply
#9
Dear Armin,

I detached the malfunctioning netPI from the metal box, and use ethernet switch between this malfunctioning one and windows PC. However, the ethernet connection lost still occurred within about 20minutes.

On the other hand, I attached the good working netPI to the metal box and use ethernet switch between this good working one and windows PC. The ethernet connection lost did not occur for this netPI for 1hour.

Therefore, it appears that the problem is unique for malfunctioning one.
However, there is not only the possibility of hardware physical matter, but for the possibility of wrong settings on the malfunctioning one. 
Since I do not know the way to compare all the settings on netPI, but for just looking the Gateway Manager pages, I cannot be sure if their settings are the same or not.
Is there any other good way to compare their settings?

I can reinstall the netPI firmware to the malfunctioning one. If some wrong settings are the causes of this problem, reinstall for malfunctioning one may solve the problem. Should I try reinstalling?

Best Regards,
  Reply
#10
Hi there,

there is no "user setting" or any "wrong setting" I can imagine that is causin and Ethernet trouble after a running period of 20 minutes or similar.

But let me ask you about the 2 MAC addresses of the two units printed on the product label. I have to compare them from their revision. The revision decides on the type of SD card the unit is produced with. Else the netPI device has never ever been changed in any other details since 5 years. So the hardware itself is stable.

If you want you can send me the two MAC addresses also as private message.

Thx
You never fail until you stop trying.“, Albert Einstein (1879 - 1955)

  Reply


Possibly Related Threads…
Thread Author Replies Views Last Post
  Hardware watchdog in a container Armin@netPI 0 1,838 January-7th-2021, 07:53 AM
Last Post: Armin@netPI
  netPI RTE + remote access + IO data access on ethernet/ip MGharat 1 2,147 December-24th-2020, 01:43 PM
Last Post: Armin@netPI

Forum Jump:


Users browsing this thread: 2 Guest(s)