RCM goes into the weeds during network startup

Rabbit
1

I’m using RMC6700/50 boards and DC10.66. When I run network software (e.g. Samples/tcpip/tftp/tfptclnt.c) with the module on the development board everything works fine. However, when I move the RCM to my target board the debugger reports “error While Debugging: Timeout while waiting for response from target”. I’ve traced the code to lib/rabbit4000/tcpip/ip.lib line 907 (pkt_init()):

   	ifte->ncd->ioctl(ifte->state, PD_INITIALIZE, ifte->instance, i);	// Reset interface and tell it the i/f number

I can’t seem to get any deeper into the code. I think this is a preloaded pointer into an assembly routine?

My target board implements the ethernet phy per TN266, figure 3 (RCM7600 Ethernet Interface Recommandations). I also have an FT232R chip wired and programmed like the one on the development board. Could any of the other I/O lines be tripping things up?

2

I found the problem on the 3.3v power rail. When the network initializes it starts drawing about 200ma more. This caused a drop to 2.6v, and reset the processor. 220uf fixed the problem.