[En-Nut-Discussion] Ethernut 5.0 Schematic Preview

[Harald Kipp]

Alain M. wrote:
> 1) I believe this is a small error: R74 should probably pulled up, to 
> "V3.3" sames as IC27, this way an inactive "PMCS\" would disconect IC7 
> pin4 from PA0

The idea was to allow SPI communication (mainly programming the m168) 
without wasting another port bit. But when looking into this again, I 
agree that this had been a bad idea. I'll change it to a 3.3V pull-up.


> 2) RS232 port:
> 2.1) RN1 and RN3 should be connected between the conector and the 
> capacitors. This way they can have two functions: a) dump oscilations 
> due to reflection in the cable, b) protect IC1 from spykes, acting as a 
> filter. I believe that in the current position they have no real 
> advantage, CMIIAW.

The intention is not to protect Ethernut from the outside world, but to 
protect the outside world from Ethernut. When doing the first EMI tests 
with Ethernut 2, we experienced radiation on the serial cable. Adding RC 
filters helped. Using ferrite beads might be better, but RC is much 
cheaper and easier to get.


> 2.2) Capacitors CN1C and CN2A can be too big for high speed 
> cominucations. I understand by the naming of the components that this is 
> some kind of 4 capacitor pack. IMHO they sholud be smaller and 
> removeable, note that chip specifications are for 150pF worst case 
> including cable capacitance.

Using different caps for Rx/Tx is indeed a good idea.


> 2.3) IC1 code is probably wrong: I cannot find ZT3243, but I found ST3242...

It's a Zywyn part
http://www.zywyn.com/
Despite their horrible web page, we have very good experience with this 
manufacturer.

> 2.4) ST3243E (not ST3243C) have "ESD protection ±8 kV" so IMHO input 
> network is not nescessary

Zywyn offers ESD protected level shifters only. As stated above, the 
reason for filtering is EMI suppression.

Many thanks,

Harald