Pages: [1]   Go Down
Author Topic: Ethernet shield 3.3V conflict  (Read 2588 times)
0 Members and 1 Guest are viewing this topic.
Offline Offline
Newbie
*
Karma: 0
Posts: 4
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Hi everybody,

I see in the ethernet shield diagram, http://arduino.cc/en/uploads/Main/arduino-ethernet-shield-05-schematic.pdf, that the 3.3V generated by the on board linear regulator is also connected to the power supply connector.

This cause two voltage regulators (one on the ethernet shield an the other on the cpu board (either Mega or Uno)) to be connected in parallel, and it seems to me this is not a good thing !

Wath do you think?

Logged

0
Offline Offline
Faraday Member
**
Karma: 16
Posts: 2857
ruggedcircuits.com
View Profile
WWW
 Bigger Bigger  Smaller Smaller  Reset Reset

It's not a problem to connect two linear regulators in parallel (of the same voltage!) Theoretically one will be just slightly higher than the other (due to manufacturing variations) and will supply all the current while the other will supply no current. In practice, due to resistance of PCB traces and such, there will be some current split between them.

But it is not a problem.

--
The Rugged Motor Driver: two H-bridges, more power than an L298, fully protected
Logged

Offline Offline
Newbie
*
Karma: 0
Posts: 4
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

I appreciate your optimistic view of the issue.

Anyway, do you think there is same reason to adopt a design technique that relies on the values of pcb's tracks resistance? 
Logged

0
Offline Offline
Faraday Member
**
Karma: 16
Posts: 2857
ruggedcircuits.com
View Profile
WWW
 Bigger Bigger  Smaller Smaller  Reset Reset

No. The +3.3V supply coming from the Duemilanoves and earlier boards came directly from the FT232 device and was limited to only 50mA, not protected, etc. so it's best to not even count on it being there. If anything I'd recommend a design technique that uses a single regulator on the shield itself, derived from the 5V supply.

On the Uno's you have a proper 3.3V regulator so you could ask it to share the load current by playing tricks with resistance, but why solve a problem that doesn't exist?

--
The Aussie Shield: breakout all 28 pins to quick-connect terminals
Logged

Offline Offline
Newbie
*
Karma: 0
Posts: 4
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

Thanks for your answer, RuggedCircuits.

Changing subject, by the way can you explain me why in the ethernet shield there is an inverter to enable the SPI interface ony when it is selected, i.e. when SS is low?
In other words, why SEN is not just pullupped?

Thanks in advance.

Marco
Logged

0
Offline Offline
Faraday Member
**
Karma: 16
Posts: 2857
ruggedcircuits.com
View Profile
WWW
 Bigger Bigger  Smaller Smaller  Reset Reset

I am not sure but I have a feeling that the WizNET chip does not "play nice" like other SPI chips do. They are supposed to hold the MISO line floating when not selected but I seem to remember that this chip does not and always drives it either high or low, thus conflicting with any other SPI devices you may have hooked up.

But I could be wrong.

--
The Rugged Motor Driver: two H-bridges, more power than an L298, fully protected
Logged

Offline Offline
Newbie
*
Karma: 0
Posts: 4
View Profile
 Bigger Bigger  Smaller Smaller  Reset Reset

You are right RuggedCircuits   smiley  : the file SPI.zip, downloadable from this page

http://www.wiznet.co.kr/Sub_Modules/en/product/product_detail.asp?Refid=99&page=1&cate1=5&cate2=7&cate3=26&pid=1011&cType=2#tab

describes exactly what you said about the w5100 SPI.

Thanks again.

Marco
Logged

Pages: [1]   Go Up
Jump to: