|
Sending signals into and out of
the RF Shielded test enclosure presents some
contradictory issues. After all, the enclosure is
designed to shield out any RF signals from passing
across its shielded walls - think of it as a barrier
which doesn't allow anything to pass across. But, what
good is the enclosure if we cannot perform any tests on
the device which is inside? We need to measure the
device's response to our stimulus and this means we will
need some sort of way of getting our desired test
stimulus signal across the enclosure's barrier.
Unfortunately, most tests signals are some sort of high
speed data and actually are nothing more than a form of
modulated RF signal!
Interfacing these data formats
requires thoughtful consideration of the data speeds and
RF frequency of the device under test. Here's the
problem: we wish to allow serial data (which can be
considered a form of modulated RF energy) into and out of the enclosure
while still stopping undesired RF signals from entering
or exiting the box. If our data signal is fairly slow,
we can use a simple capacitive filtered connector such
as the popular DB-9 style serial connector. Internally,
this connector has feed-through capacitors on all pins.
These capacitors appear electrically as capacitance from
the data pin to ground. Their performance at RF
frequencies is excellent, thus their popularity. Simple
PCB adapters can be used to connect to the DB-9 and
adapt it to USB style connectors, 3.5mm audio connectors
or even RJ style LAN/Phone connectors.
Capacitance values are available
ranging from 1000 pf to as small as 10 pf. Note that
this capacitance will be 'hung' directly on the data
line and your device will be driving not only your
desired interface device but also this capacitance to
ground. While this capacitance will load the data lines,
there is a trade-off regarding acceptable serial data
attenuation/distortion and acceptable RF filtering
attenuation.
For example, 9600 baud RS-232
serial data results in a serial data stream of
approximately 19.2 KHz, so hanging a 1000 pf capacitor
across one of the data lines would in effect load the
line with 8.3 K ohms to ground. this was calculated
using the formula for capacitive reactance:
Where: XC is
capacitive reactance in ohms, f is frequency in Hertz,
and C is capacitance in farads. In most
cases this extra loading will not result in any
appreciable attenuation or distortion of the 9600 baud
data. We can use this same sort of reasoning when we
consider other serial data interface filtering, making a
reasonable tradeoff between amount of capacitance and
data speed. In general we should use the most
capacitance possible while still maintaining an
acceptable amount of loading on the data lines, this
provides the maximum amount of RF filtering.
Let's examine this effect a bit
further. Note that the capacitance appears to your
device as a load on your data line and also that your
device has a driving impedance - called it's source
impedance. This source impedance combines with the
filtering capacitance and forms a simple R-C low pass
filter.
If the driving source
impedance is low, the capacitance effect will be less
than if the source impedance is high. Look at the R-C
filter as a simple resistive voltage divider with the
source impedance being the series R component and the
filtering capacitance being the parallel C component. As
we saw earlier, the capacitor has an impedance which is
a function of the frequency applied to it, in accordance
with the reactance formula below:
As we noted above, a 1000pf capacitor has an impedance of
8.3K ohm at 19.2KHz, so if the device has a source
impedance of 8.3K also, the net effect of the DB-9
connector's filtering is a reduction of the signal by
one-half! (the voltage divider having equal resistances
of 8.3K dividing the signal by 1/2).
We just looked at an RS-232 data
signal, what about video? Video is generally understood
to be composed of frequency components up to 4 MHz and
usually 75 ohm as the impedance of the devices used with
it. Thus the source impedance is 75 ohms and any video
passing through a filtered connector will have its
capacitance 'hung' across the video line. If we use a
1000pf filtered connector, we can analyze how much
attenuation this capacitance will have for the video.
The capacitor's impedance at 4 MHz will be:
1/((2*pi)(4e6)(1000e-12) [note the scientific notation:
4e6 means 4 times 10 to the 6th power: 4,000,000. It is easier to
use scientific notation rather than mess with all those
zeros!] after calculating, we see that the capacitor (at
4 MHz) looks like a 39.8 ohm resistor. Combined with the
75 ohm source impedance, we see that the final signal
coming across this filtered pin is reduced by about 2/3rds, leaving us with only a third of the original
signal. Also note that this is at 4 MHz - at lower
frequencies, the loss is less, so we end up with a
severely rolled-off response!
This essay is not meant to be
the end all in data filtering for RF Shielded test
enclosures, but it does serve to make an engineer
appreciate the nuances of getting test signals across
the RF shielded barrier. It is nothing more than basic
2nd year theory from your electronic engineering
education!
|
