[En-Nut-Discussion] Strange reset behavio (sligthly OT)

[Ole Reinhardt]

Hi all,

Sorry for asking a question a bit out of topic.

I'm just building a complex system (automatical test engine for
automobile parts) that consists of about 12 µC Boards running NutOS. All
of these boards are connected by a CAN bus (SJA1000 connected to the
ATMega128) to eachother.

The main power is switched on by a relais. This relais is chattering?
(in german: "das Relais prellt"). So without any precautions there a lot
of transients on the main power line for about 2 ms after switch on of
the mainpower.

My problem: Some of my boards does not reset properly. I don't have a
reset controller (like MAX811) or something similar. Ther's ony a R/C
combination (C = 100nF, R = 100kOhm) to generate a reset signal.

Like this:

         -------              |  |
VCC -----|  R1 |-----+--------|  |------- GND
         -------     |        |  |
                     |
                     |         C1
                  /reset

My first consideration was that the transients on the main-power prevent
the boards from booting properly. So I build a L/C combination to block
these transients:

Power supply / relais                       Backplane / Boards

        L1          C1
+ ---/\/\/\/\-------+-------------------------------- +5V
                    |
                    |
                 =======

                 -------
                    |
                    |
- ------------------+-------------------------------- GND

L = between 50µH and 100µH (from my hobby box)
C = elko, 470µF

Everything now seems alright. Main Power rises a bit above 5V and then
goes down to 5V, but no more fast transients. Beside this nothing was
better. Some boards still does not boot properly.

No I changed the above circuit a bit (in my eyes nonsene, because I now
have a L / C oscilator and it should make the transients even worse):

Power supply / relais                       Backplane / Boards
                
+ ------------------+--------------------------------- +5V
                    |
                    \
                    /
                    \
                    /      L1
                    \
                    /
                    |
                    |
                 =======

                 -------   C1
                    |
                    |
- ------------------+-------------------------------- GND

WOW! Now it works. But as I supposed the transients on the main power
are even worse and overshooting (undershoting) the 5V a lot. 

In my eyes this works by accident, but I'm searching a REAL solution.

Is there anybody who can give me an advise what could be wrong?

Regards,

Ole Reinhardt


-- 
kernel concepts    Tel: +49-271-771091-14
Dreisbachstr. 24   Fax: +49-271-771091-19
D-57250 Netphen    E+ : +49-177-7420433
--