Content originally posted in LPCWare by profanum429 on Fri Apr 03 18:34:27 MST 2015
We have a LPC4088 QSB + associated Experiment Base Board. We've been running a LPCOpen based network application on one of them for the past 2 months without issue and ordered three more QSBs + base boards. These three arrived and all exhibit the same issue; when we do a cold start we'll see the PHY immediately think it has a connection (polling the PHY status initially returns 0x02 then constantly 0x3C immediately after). This behavior happens wether a cable is plugged in or not. Pressing the reset button results in normal expected behavior that our original setup produces.
Removing power completely and doing another cold start brings back the original problem which can be solved by the user doing a reset. Doing a NVIC_SystemReset does not fix the issue nor does connecting a GPIO to the reset line and pulling it high in software. I'm assuming this is an issue with the LAN8720 needing to have a 100 uSec reset time and my software reset strategies don't produce nearly that amount of time.
I also see this same issue using the tcpecho_sa sample project from the 2.10 version of LPCOpen for the LPC4088. I'm wondering if anyone else has seen this issue? I'm thinking we might of gotten a set of QSBs (the ones with the problem all have the same lot and date code, different from the working one) that might have a defective component? Next step will probably be monitoring the power signal going to the LAN8720 on startup on a bad board and a good board but that will have to wait until Monday.
