First post here. I just got into Arduino, with a Mega 2560 R3 and an Ethernet Shield W5100 R3.
When the two are connected, any attempts to load any sketch (even blink) time out. (Win7 64 bit, Arduino sw v 1.0.5, no install issues or driver issues in Device Mgr). I've tried to press the reset button on the Ethernet board or Arduino to no avail.
Curiously:
NO problems at all loading sketches onto the Arduino board itself, without the Ethernet shield attached.
NO problems loading an Ethernet function sketch like the Web Server in the Examples section, powering everything down, connecting the Ethernet shield, powering the boards up. The sketch runs just fine / Ethernet / http functions work (I can see the web page served by Arduino).
So it appears the Ethernet Shield is only interfering with the upload function. Any way I can fix this?
Some more observations on this problem. With the shield removed, there is no problem uploading a sketch, even a sketch that sets the ethernet shield as a server. The shield can then be installed, the server functions, and you can even read from the usb port the data that is being serial printed. You cannot however then upload anything via the usb until the shield is again removed.
I tried the suggestion of setting pin 4 as high to shut off the SD card function on the ethernet shield, to no avail.
I also tried resetting the shield just after compiling and during upload, still no function.
I may run the shield off the mega board, and connect it only with wires. Hopefully, I can simply cut power to the shield to restore usb function.
Thanks for looking into that Crossroads. I'm pleased that it appears to work, but annoyed that I will have to pull the shield off every time I need to update code, which is frequent. I also have one of those "other" adapters, that uses a different chip and is connected only by wires. Sorry, I'm at work and can't remember the chip name. J60 something. I haven't tried it yet. I'd hate to lose compatibility or function in the future.
So jumpers between cards seems to be the way to go for now - seperate Reset so you can control it to hold the board in Reset while you program. There's not too many actual signals being used.