3

I made a through-hole shield based on the ENC28J60, but I haven't had luck getting the device to do anything at all.

At this point I'm afraid I'm in over my head. I've spent some time reading the data sheet and clumsily poking at it with my bus pirate with no luck.

If anyone has pointers or the time to glance over my schematic (pdf)(eagle) I would be extremely grateful.

Shield

Trygve Laugstøl
  • 1,410
  • 2
  • 19
  • 28
lolsborn
  • 133
  • 5
  • 1
    Do you have a logic analyzer? Is the ENC28J60 responding *at all* to your attempts to communicate? – Connor Wolf Nov 22 '12 at 20:14
  • 1
    Do you have an oscilloscope (what equipment do you have)? Is there any activity on the clock pins? Can you get a reading with a voltmeter? Is there voltage on Vcap pin? What's the input current of the device? – AndrejaKo Nov 22 '12 at 21:36
  • 1
    Have you looked at the Nanode? http://solderpad.com/monsonite/Nanode. I think you may need a handful of passives between the Ethernet controller and the MagJack. – vicatcu Nov 23 '12 at 02:38
  • @vicatcu If you would add this as a answer I think I will accept it. There have been a couple of other answers who have certainly found issues with my work, but I am definitely missing a number of passives which I had thought to be integrated in the magjack and it's a pretty glaring oversight. – lolsborn Nov 23 '12 at 07:21

3 Answers3

4

200nF on your crystal seems way too big. Usually 20pF would be more appropriate.

Edit: other reference designs show caps between 18pF and 33pF (here, more can be found on the interwebs). The exact cap values depend on the crystal. The only downside to having caps slightly too large is that the crystal will take a little longer to start.

Nick Alexeev
  • 37,739
  • 17
  • 97
  • 230
Jim Paris
  • 2,685
  • 1
  • 12
  • 12
  • Great catch! This just appears to be a typo on my schematic though, the actual part I'm using is 20pf http://www.digikey.com/product-detail/en/RPE5C2A200J2P1Z03B/490-3637-ND/946273 – lolsborn Nov 22 '12 at 21:20
  • 1
    @Steve918 Check the crystal with an oscilloscope. It wouldn't hurt. – Nick Alexeev Nov 22 '12 at 21:41
  • Both answers were equally valid as they found issues with my design, but I accepted Jim's since it was first. – lolsborn Nov 25 '12 at 06:49
  • @NickAlexeev An oscope often has a decent sized capacitance, it can stop an oscillator from starting by putting so much capacitance to ground. – Kortuk Nov 25 '12 at 21:53
1

The bias resistor looks wrong.

You have a 2k 1%.
The datasheet calls for a 2.32k 1%

Robert Deml
  • 3,232
  • 6
  • 29
  • 27
  • That's strange, I have an older copy of the datasheet http://cl.ly/1W3o3A2W0D1l that calls for 2k 1% – lolsborn Nov 23 '12 at 06:06
1

Per the OP's request I'm reposting this as an answer. Have you looked at the Nanode? I think you may need a handful of passives between the Ethernet controller and the MagJack.

vicatcu
  • 22,499
  • 13
  • 79
  • 155