summary: power supply and chassis

[Martin Czech]

::::	>2. Enclosure
:::	>Use a metal enclosure for your device. This will damp rf
:::interference.
:::	>Try to have as few holes in the enclosure as possible. Otherwise
:::	>rf signals may sneak in. 
:::
:::While this is definitely a good choice, I'd say it's not the only one.
:::I have quite some of my circuits in wooden enclosures, and I'm not
:::plagued with RF interference at all. (But then I try to take a lot of care
:::about the circuit layout - I won't have sensitive high impedance parts
:::go over long wires for instance.)
:::I'd say a metal enclosure is not always necessary, but *if* you have
:::a metal enclosure, then grounding it (with a C at least) is a must.
:::

@ least one person seems to have read my mail ;->

Juergen, this is true. Otherwise our bench test setups would not work
either.  It was about a good design that works under any condition.

Conditions:

But I live near (r=5km) to a very large antenna, I guess AM radio
(some MHz).  The railway is r=20m away, so there is strong 16,333 Hz
(yes sixteen and a third) and also railway radio interference. The
railway signals are so strong that they stop my heart rate monitor from
working! The next big tv antennas are only 20km away. This is because I
live near to the black forest mountains, so the have to have some extra
antennas there on the top.

So the air is full of all kinds of waves where live.  I had radio in my
guitar setup, and also in filter circuits.  Of course this depends on
how long the lines are.

This is scary, you sit at home, 23:00, everthing dead quiet, soldering
some resistors in, and then suddenly you hear voices (silent but audible)
comming out of your circuits, or your guitar gear!!!

The interference depends on weather, time of day, sun activity etc etc...


btw.:

I forgot to mention that if the PSU is in an extra case (which is not a
bad idea) the secondary power leads from there should get dissipative
ferrite cores  or should be wound on those cores in order to damp any
line bound rf interference and to not let any rf in via cables or lines.

Computer shops carry those, for mains line "cleanup".
They can be clipped on the cable, even after construction.

Our signal lines inputs should get RC filtering for the same reason.
This can be seen e.g. in the ssm2217 data sheet.

Interference is line bound, or due to muddy lines in 90% of all cases,
as far as I experienced. Very little comes through the air.

Lines are so nice antennas, rf just can't resist to creep in ;->

m.c.