EMI with VHF radio

Hello.

I'm having the following problem with mini2440:

It causes interference with a VHF Motorola radio mounted on a truck. I'm
powering the board with an OKI-78SR switching PSU  plus an inline LC filter
(100uH + 2x 0.1 uF ceramic caps) inside a metal box. It only causes
interference on certain channels (ex: 144.0000 MHz) and it decreases
putting the box away from the antenna coax cable, but it still interferes.

What steps could I make to reduce this problem? I know there's a whole
theory behind EMI, but I would like to know if someone had similar problems
or how to make effective improvements on shielding so the problem is
minimized.

Thanks.

Got a spectrum analyser?  If the mini2440 is producing a discrete frequency
at 144MHz then there are probably other discrete frequencies being produced
as well . . . related to clock frequencies used on the mini2440 dev board.

If you are only concerned about one particular frequency maybe changing the
master clock frequency to something else might be something to investigate.

I would really look at spectrum analyser output or failing that do a full
spurious analysis using all clocks on the mini2440 board.

Sadly, I don't have a spectrum analyzer. OSC doesn't reveal much since
there's a lot of noice on the VHF radio band (135+ MHZ) anyways. 

I don't know if the noise is on discrete bands, but it's very evident in
some bands, where the base noise is low, and there isn't much difference on
other bands, it could be that the noise on that bands is higher than noise
produced by the board.

Some additional info:

If I connect the board in a lab to a different PSU than the radio, there
isn't noise unless I get the box close to the antenna  (50cm or less).

Metal box doesn't make much difference in this aspect (it is properly
grounded with the board).

Adding LC filter helps reduce noise when I connect the board to the same
PSU as the radio, but still present (I tried adding additional filters but
effect doesn't vary).

Putting a Ferrite on the power cable at box output doesn't make a
difference.

Thank you for the help dave.

Could you either take a screenshot of what the noise looks like on an
oscilloscope or a sound recording of what the noise sounds like?

.1uF may not be a good bypass at VHF.  Are they chip capacitors?  If you
want to filter out band-specific noise I looked at some additional low-pass
filtering on the supply.

A picture of your added filter box could reveal why it is not filtering
enough at VHF.

We had some micro2440 modules tested for EMI and discovered that the LCD
circuits in the module produce enough noise in the VHF band as to affect
the sensitivity of nearby VHF receivers.

The noise is always present, but its frequency varies depending of the type
of image being presented on the LCD and is generated by the micro2440 as it
does not go away if you happen to disconnect the LCD module.

The only reliable solution we found was to redesign the board we connect
the micro2440 to and hide all LCD signals under a ground plane. We also had
to put the whole thing into a grounded metal case.