1
\$\begingroup\$

Context:

I have designed a custom pcb utilizing the ENC28J60 ethernet chip to be used with raspberry pi. The board has 5V in which powers both the ENC28J60 (through a AMS1117-3.3 LDO regulator) and the pi. Power and internet is received from a POE splitter which has RJ45 and microUSB 5V out.

This design works fine during initial testing, but when the modules are left to run 24/7, after a few weeks of operation some of the modules loose connection indefinitely (turning it on and off again does not fix the problem). This only occurred on roughly half of them, the rest work fine.

Initially I thought that the problem was caused by the failure of the AMS1117-3.3 LDO regulator, however upon testing more modules, it has become apparent, that the failure of the LDO mentioned in my previous post was a result of user error and was not diagnosed in the rest of the modules - they have correctly functioning regulators providing 3.3V (but ethernet is not detected).

The problem seems to lie within the ENC28J60 chip itself, as upon changing just the chip with a new one, the boards once again detect when an ethernet cable is plugged in and the yellow and green rj45 led's light up. Soldering the old chips on to the donor boards which were functioning before the chip exchange, results in the donor boards no longer detecting the ethernet cable.

So this leads me to think that the ENC28J60 chips are getting damaged internally resulting in them functioning only partially (the chip is recognized on the RPi, but when ethernet is connected the rj45 led's do not light up and ipconfig does not show the ip address field at all). The symptoms are consistent on those which failed, but as a whole the problem only occurs on roughly half of the batch of modules and only after 1-5weeks of constant operation. All of the modules initially functioned as intended and other (not broken) modules have been functioning 24/7 for multiple months, however they do have occasional disconnects, though this can be resolved with a hardreset of the module, after which it functions fine again. It is unclear whether these disconnections are related to the larger issue of those that fail entirely (and only work again only by replacing the enc chip), as there are other factors at play in the network which could cause them to disconnect.

I have researched common enc28j60 problems and it seems that while it does have its kinks, in theory when it works, it should work for a long time and not break in such a way that it does in my case.

My current theory is that the poe splitters provide an unreliable voltage and thus eventually damage the enc chips, however I do not have a way of testing if this theory is true (atleast that I can think of without specialized tools). Is this a viable assumption? Or is it possible that I have received a batch of faulty enc chips?

Edit1 - added enc28j60 schematic: enc28j60 schematic

\$\endgroup\$

1 Answer 1

0
\$\begingroup\$

Probably not your problem, but as this is a custom PCB, worth checking the polarity of the Rx lines.

As detailed in item 7 of the errata sheet, the support for autopolarity is broken. The reason I raise this, is that I have seen issues where this behaviour has changed over time, with initially working parts getting stuck in non-working state which no amount of power reset would fix. In my case fixing the polarity (which I'd got wrong) made the problem go away.

\$\endgroup\$
2
  • \$\begingroup\$ Thank you for your reply. I have edited my post to include the diagram of the enc chip, but as far as I can tell, the polarity is correct. \$\endgroup\$ Commented Sep 27, 2023 at 15:23
  • 1
    \$\begingroup\$ It was a bit of a long shot, but thought worth sharing just in case \$\endgroup\$
    – colintd
    Commented Sep 27, 2023 at 16:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.