MH6
Posts: 18
Joined: Wed Aug 26, 2015 2:26 am

Diagnosing problems with LAN9514 and Ethernet Jack

Sun Mar 27, 2016 2:02 am

I recently built a custom Pi CM board, and everything seems to be working OK except for the ethernet jack. I am utilizing a LAN9514 IC to split the CM's USB line into an ethernet port and 4 USB ports.

I have an eth0 interface available, but the Ethernet port's LEDs do not light up when a cable is plugged in. And I can't connect to the internet of course. I've verified the cable works with a regular Pi.

The only difference I've noticed in the startup logs between my board and the regular Pi is the eth0: link up line:

Code: Select all

[   24.221960] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
[   25.820904] smsc95xx 1-1.1:1.0 eth0: link up, 100Mbps, full-duplex, lpa 0xC1E1
...
on my board, this line is not present:

Code: Select all

[   19.365597] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
[   85.621659] Adding 102396k swap on /var/swap.  Priority:-1 extents:1 across:102396k SSFS
...

At this point I'm not sure if its a hardware issue or software issue. mii-tool reports "eth0: link down". Continuity seems good on the traces between the LAN9514 and magjack, and for kicks I tested the LED inidicators on the magjack manually and they lit up fine.

I am using the default drivers included in Raspbian Wheezy V7 (from nightly build 03/26/2016)

ifconfig shows no RX or TX traffic, and no assigned IP address:

Code: Select all


eth0      Link encap:Ethernet  HWaddr b8:27:eb:d9:5d:f9
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)
Is there anything else I can try to see if I can pinpoint the issue?


Here is the LAN portion of my schematic:
Image
http://i.imgur.com/ljolAcI.png

W. H. Heydt
Posts: 9899
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Sun Mar 27, 2016 3:27 am

I recall back when Pis were new (so new they weren't really available) that a batch had to go back to the factory because the wrong kind of RJ-45 jack was installed. Something about built-in "magnetics".

MH6
Posts: 18
Joined: Wed Aug 26, 2015 2:26 am

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Sun Mar 27, 2016 10:30 pm

W. H. Heydt wrote:I recall back when Pis were new (so new they weren't really available) that a batch had to go back to the factory because the wrong kind of RJ-45 jack was installed. Something about built-in "magnetics".
The J3011 ethernet jack I am using has built-in magnetics. Same as the one the LAN9514i reference design uses.

MichaR
Posts: 16
Joined: Thu Mar 10, 2016 5:49 am
Location: Berlin / Germany

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Tue Mar 29, 2016 5:29 pm

Hi,

I also did it with the reference schematic for the LAN9514 and it works without any problems. The difference I can see between your schematic and the reference is, that you did not connect the JTAG-Pins except TDO to VDD and on of the Test-Pins is not connected. Atached is my schematic I use. Don't be confused, my board is powered via PoE so I use a normal RJ45-jack and an external transformer to be able to access all 8 pins of the ethernet connection.

best regards
Micha
eth.png
eth.png (34.53 KiB) Viewed 4004 times

MH6
Posts: 18
Joined: Wed Aug 26, 2015 2:26 am

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Tue Mar 29, 2016 5:54 pm

MichaR wrote:Hi,

I also did it with the reference schematic for the LAN9514 and it works without any problems. The difference I can see between your schematic and the reference is, that you did not connect the JTAG-Pins except TDO to VDD and on of the Test-Pins is not connected. Atached is my schematic I use. Don't be confused, my board is powered via PoE so I use a normal RJ45-jack and an external transformer to be able to access all 8 pins of the ethernet connection.

best regards
Micha
eth.png
Thanks for your reply.

I recently submitted my design to Microchip's complimentary LANCheck service, and they did warn me about the JTAG pins being unconnected. Other (working) CM boards on this forum left them unconnected as well though, so I'm not sure if that's truly a problem or not. I've already corrected it in Revision B of my design though, just in case.

Microchip didn't see any major glaring errors, but they did give me a lot of minor recommendations, which I am following in Revision B.

I'm beginning to think my problem may be user error, in that maybe I've shorted something somewhere or have a pin that isn't properly connected and I just can't tell. I would think it'd be all or nothing though, and the USB portions work. My second board attempt failed miserably (LAN9514 chip didn't work at all as far as I could tell). My third and final attempt will probably be later this week. If that fails too, then I'll probably get some place in China to assemble Revision B...


If you don't mind telling me, what crystal did you use for the LAN9514? Apparently it needs a drive level of 300-500uW, and the crystal I'm using maxes out at 300uW. Though I'd imagine if that were the problem, USB wouldn't work either, and it does. I'll still be replacing it just to be on the safe side.

Also, did you have to do any kind of software related setup or did the default drivers work out of the box?

Thanks!

MichaR
Posts: 16
Joined: Thu Mar 10, 2016 5:49 am
Location: Berlin / Germany

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Wed Mar 30, 2016 4:47 am

Hi,

the XTAL I use is a ceramic SMD type: http://www.reichelt.de/Crystals-SMD-Cer ... ANGUAGE=EN

That one seems to be out of specs a little bit (only drives 100uW), but works fine on several boards.

If you soldered the chip with it's QFN package by hand, probably you have an soldering problem. I got them too. Because the diepad is really big, it's really easy to shorten one or more pins with ground, if you use hot air for soldering the diepad. You should check every pin for shortage against ground and it's left and right neighbor. If you use addtional flux paste based on zinc chloride (i.e. EDSYN ULF-10) you should wash the PCB with alcohol (isopropanol). For a project in the company we had strange effects with that type of flux with some PCB's stop mask paint - especially if the paint was not the standard green one. There was a chemical reaction between the flux and the paint which caused a low resistance between 10 and 100 Ohm between pads with low distance. After cleaning the PCB with alcohol the effect was gone.

There was no need to change any driver. The reference schematic http://ww1.microchip.com/downloads/en/D ... 14_sch.pdf works out of the box here.

Micha

Btw: the working PCB: http://www.pcbway.com/project/share/Ras ... mount.html ;)

MH6
Posts: 18
Joined: Wed Aug 26, 2015 2:26 am

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Thu Apr 14, 2016 6:42 pm

MichaR wrote:Hi,

the XTAL I use is a ceramic SMD type: http://www.reichelt.de/Crystals-SMD-Cer ... ANGUAGE=EN

That one seems to be out of specs a little bit (only drives 100uW), but works fine on several boards.

If you soldered the chip with it's QFN package by hand, probably you have an soldering problem. I got them too. Because the diepad is really big, it's really easy to shorten one or more pins with ground, if you use hot air for soldering the diepad. You should check every pin for shortage against ground and it's left and right neighbor. If you use addtional flux paste based on zinc chloride (i.e. EDSYN ULF-10) you should wash the PCB with alcohol (isopropanol). For a project in the company we had strange effects with that type of flux with some PCB's stop mask paint - especially if the paint was not the standard green one. There was a chemical reaction between the flux and the paint which caused a low resistance between 10 and 100 Ohm between pads with low distance. After cleaning the PCB with alcohol the effect was gone.

There was no need to change any driver. The reference schematic http://ww1.microchip.com/downloads/en/D ... 14_sch.pdf works out of the box here.

Micha

Btw: the working PCB: http://www.pcbway.com/project/share/Ras ... mount.html ;)

Good information, thanks.

I finally got around to making board #3, which works. So I'm not sure what I messed up on the first two, I could find no bridges or shorts to ground. Since #3 works it's most likely a problem with me being terrible at building them and not a problem with the design itself, so I'm a little more confident now that I should be able to get 10 or so working boards assembled over in China when I finish the next revision.

My LAN9514 chip does run pretty hot to the touch though. Could be due to the fact I only have 9 thermal vias. Microchip recommended 36 in a 6x6 grid, which I've done in Revision B. Other than high temps, everything is working perfectly now.

Thanks for your help.


It's alive!
Image

LulaNord
Posts: 1
Joined: Fri Mar 03, 2017 5:32 pm

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Fri Mar 03, 2017 5:56 pm

Hi...i am a new user here. As per my knowledge the difference I can see between your schematic and the reference is, that you did not connect the JTAG-Pins except TDO to VDD and on of the Test-Pins is not connected. Atached is my schematic I use. Don't be confused, my board is powered via PoE so I use a normal RJ45-jack and an external transformer to be able to access all 8 pins of the ethernet connection.

pcb prototyping
Last edited by LulaNord on Wed Mar 22, 2017 8:25 pm, edited 1 time in total.

marcosartore
Posts: 14
Joined: Thu Jan 15, 2015 3:07 pm

Re: Diagnosing problems with LAN9514 and Ethernet Jack

Tue Mar 14, 2017 4:57 pm

Hello,
I wish to ask you guys who succeeded in interfacing LAN9514 with the Compute Module 3 if the JTAG connections are just important for proper biasing the device or if they need to be used after the board is populated, i.e. I wish to understand if this ethernet chip works as-is or need to be somehow programmed before utilization. In addition, I wish to understand if each chip already owns a MAC address or if you must purchase it separately and store into the EEPROM.

Many thanks in advance,
and best regards...
Marco

Return to “Compute Module”