Geekzone: technology news, blogs, forums
Guest
Welcome Guest.
You haven't logged in yet. If you don't have an account you can register now.


541 posts

Ultimate Geek
+1 received by user: 125

Subscriber

Topic # 201678 29-Aug-2016 17:51
Send private message

We have a number of proprietary network devices deployed to different sites which utilise an ASIX AX88796 LF 100BASE-TX NIC.

 

About half of the devices each connect directly to an HP Laptop's integrated Realtek PCIe NIC using a Cat6 Ethernet patch cable.

 

The other half each connect directly to a Dell Micro PC via a D-Link Dub 1312 USB 3.0 NIC using a Cat6 Ethernet patch cable.

 

On a number of the D-Link connected devices we have begun to notice an issue which appears to be progressively getting worse whereby the when the device is turned on first thing in the morning it can take up to 30min for it to be detected (as a LINK / connection) by the PC network adaptor.

 

During this time the device LINK is more likely to be detected by certain NICs I've tested than others.

 

Adaptors more likely to detect a LINK:

 

  • HP integrated Realtek PCIe Ethernet
  • Dell integrated Intel I217-LM Ethernet

Adaptors less likely to detect a LINK:

 

  • D-link Dub 1312 USB 3.0 Ethernet Adaptor
  • TPT-8020A Cable Tester

This appears it could be a fault with the device which I hypothesise could be due to failing caps on the board; possibly causing an initial voltage drop which is tolerated by some NICs (integrated) better than others (USB).

 

I would like to obtain some quantifiable data as to why some NIC will connect with the device when it's in this state and others wont until after some warm-up time and am looking for a way to measure the TX/RX voltages etc of the device NIC and all PC NICs for comparison.

 

Any ideas what tests could be carried out to determine the connection quality of the NIC (rather than a LINK / NO LINK result by my current tester) or other issues to consider?

 

(I don't currently have access to a scope but may be able to get hold of one if necessary.)

 

 


Create new topic
Meow
7911 posts

Uber Geek
+1 received by user: 3930

Moderator
Trusted
Lifetime subscriber

  Reply # 1619242 29-Aug-2016 17:56
Send private message

Are these connected to a switch with power management? Have seen this kind of thing with some DLINK "Green" switches.







541 posts

Ultimate Geek
+1 received by user: 125

Subscriber

  Reply # 1619259 29-Aug-2016 18:20
Send private message

michaelmurfy:

 

Are these connected to a switch with power management? Have seen this kind of thing with some DLINK "Green" switches.

 

 

No, device is connected directly to PC.

 

I have tried disabling all power management features I could find on the D-Link USB NIC.

 

This morning I had one take about 30min for a LINK to come up even after reboot.


3115 posts

Uber Geek
+1 received by user: 1207

Subscriber

  Reply # 1619361 29-Aug-2016 22:06
Send private message

Does using a really long Ethernet cable compared to a short patch cable make any difference? And does using a crossover cable compared to a straight through wired cable make any difference?








541 posts

Ultimate Geek
+1 received by user: 125

Subscriber

  Reply # 1619412 30-Aug-2016 08:12
Send private message

Aredwood:

Does using a really long Ethernet cable compared to a short patch cable make any difference? And does using a crossover cable compared to a straight through wired cable make any difference?



The systems we're having issues with use a 1m Dynamix Cat6 cable (the others use a 3m unshielded Cat6 cable).

I haven't tested yet using longer cables however I now see anecdotes about issues with shorter cable lengths.

Crossover cables didn't make a difference.

1984 posts

Uber Geek
+1 received by user: 133

Trusted

  Reply # 1619879 30-Aug-2016 20:19
Send private message

Ethernet is supposed to be 10m or 15m for certain applications, so maybe the NIC is expecting a structured cabling network. Presumably the device doesnt use PoE since you connect to laptops, but have you tried setting them up on a network and potentially leaving them on 24/7?





Qualified in business, certified in fibre, stuck in copper, have to keep going  ^_^



541 posts

Ultimate Geek
+1 received by user: 125

Subscriber

  Reply # 1620106 31-Aug-2016 10:33
Send private message

webwat:

 

Ethernet is supposed to be 10m or 15m for certain applications, so maybe the NIC is expecting a structured cabling network. Presumably the device doesnt use PoE since you connect to laptops, but have you tried setting them up on a network and potentially leaving them on 24/7?

 

 

No POE. Devices work fine 24/7 if left running. Just an issue during "warm-up" which appears linked to colder weather.

 

 

 

Does anyone know how the NIC physically determines whether a LINK is present on the cable?




541 posts

Ultimate Geek
+1 received by user: 125

Subscriber

  Reply # 1624987 7-Sep-2016 11:16
Send private message

Still interested in understanding more about how the link layer is detected and/or negotiatiated as further testing reveals it's a multi-factor issue;

 

a) When I replace the AX88796-based NIC in our proprietary device with a new one the issue is resolved; suggesting a faulty NIC which appears more prevalant at colder temperatures.

 

b) However I've also found a BIOS update for the Dell OptiPlex which "improves compatibility with USB devices" (i.e. the D-Link USB to Ethernet adapter) and installing this likewise seems to resolve the issue.

 

c) Lastly I've discovered that the OEM manufacturer who's chip is used in the D-Link Dub 1312 has released a driver update in April which also fixes some relevant bugs. D-Link however are yet to package the driver so I'm having to test an unsigned version which I've had to manually modify the VID, PID and device descriptors on to work with the DUB 1312 and it appears to also show improvement in connection time. It's proving difficult however to get D-Link to provide a signed version.

 

 

 

This reinforces my hypothesis that there is some sort of compatibility issue or fault with some of my proprietary AX88796-based NICs (which I would still like to test & quantify) which is either tolerated or exacerbated dependant upon the host NIC / drivers.


Create new topic

Twitter »

Follow us to receive Twitter updates when new discussions are posted in our forums:



Follow us to receive Twitter updates when news items and blogs are posted in our frontpage:



Follow us to receive Twitter updates when tech item prices are listed in our price comparison site:



Geekzone Live »

Try automatic live updates from Geekzone directly in your browser, without refreshing the page, with Geekzone Live now.



Are you subscribed to our RSS feed? You can download the latest headlines and summaries from our stories directly to your computer or smartphone by using a feed reader.

Alternatively, you can receive a daily email with Geekzone updates.