Ethernet shields causing DUE to output 5V from 3.3V pin

Thank you for your consideration and thoughts on this! Any experience or ideas greatly appreciated!

In testing 3 different types of Ethernet shields with the DUE I've found that with all three cause the DUE to output 5V from the 3.3V pin. The IOREF pin is the same...5V when the shield is on. If I remove the Ethernet shield it is all back to 3.3V.

This was measured at the 3.3v and gnd pins on the DUE. This is just the DUE with the Ethernet shield and occurs whether the DUE is powered by the USB or a 9V wall wart. I tested and found this problem with the W5100, W5200, and W5500 shields. All 3 have the R3 layout with SDA/SCL pins.

  1. Does anyone else have this issue from their Ethernet or other shields?

  2. What type of problems would or could this cause?

Does the 5.0V coming back to the DUE from the shield cause it a problem? It doesn't seem to?
It "seems" to run fine with just the DUE and shield.

My external circuits are fine with 5.0V because I can also run them with the UNO but I'm wondering if SPI or I2C connections or possibly something else coming back to the DUE will not possibly cause damage to it.

  1. I do have a TL431 voltage reference circuit that I use with my Uno to provide 3.3v voltage reference that I think I could modify to get things back in line. I'm thinking that the 5.0V that is coming out of the DUE and create a circuit to drop it to 3.3V to power my other circuits so that the control lines would all be back to 3.3V...but I'm thinking that this could sink a lot of current to ground.

Thank you!

UPDATE:

I found something interesting here ... it appears that these ethernet shields, rather than using the ioref pin to determine what arduino (uno or due or whatever) the shield was on, were just interconnecting the 5V pin and the ioref pin on the shield. Somehow this apparently causes the 3.3V pin to output 5V.

I'm not sure who came up with that idea or what they could have been attempting to accomplish.

For what it's worth I tried clipping the ioref pin off the shield and the DUE is now outputting 3.3Vs again and everything seems to be working. Of course, this will eliminate the possibility of adding another shield (that uses the ioref) on top of the ethernet shield but I can live with that.

Does anyone have a better solution?

Which ethernet shield are you using? The official Arduino shield does not show the 5V pin connected to IOREF on the schematic.

Tim,
Again, MANY THANKS for your help!

I'm not sure if or where you can even get an official Arduino ethernet shield anymore. Many of the clones available seem to be pre-R3 without SCL/SDA or ioref. Three of the shields I have been testing with appear to be R3 and do have SCL/SDA and ioref but of course the 5V problem.

While the W5100 has the ioref pin correctly labeled, the other two shields mislabel it as 5V and I just visually reconfirmed that they've cross connected it with the SPI 5V. The first thing I did was to just bend the ioref pin out before putting on the shield. When that worked well, I cut it off. As I mentioned before in my previous topic, I had to trim all of the pins to get the shield to seat correctly so that the SPI would connect properly.

The W5100 I've been testing with now seems to be working well with the ioref hack:

http://www.ebay.com/itm/161406767490?_trksid=p2057872.m2749.l2648&ssPageName=STRK%3AMEBIDX%3AIT

I haven't tested everything yet with the other two but they definitely had the 5V problem as well.

The W5200 shield -

http://www.ebay.com/itm/131603255867?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

The W5500 shield -

http://www.ebay.com/itm/291495860309?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT