DIARY #16
SOLVING INTERFERENCE
PROBLEMS
The Garbage Can Test
When Paul Hanson was testing Freewave DRG115
Radios on small buoys, one buoy was equipped with a highly sensitive
infrared gas analyzer set of sensors. Which system reports on
the amount of Carbon Dioxide in the air out on, and just above,
the water.
This is a view of the case with the gas spectrometer
installer in the upper right, the battery in the center, and
the data logger in the foreground. The Freewave radio was placed
in the upper left, the only space left, underneath a foam layer.
Below is the cylinder, held by Matt, into
which the air flows and the gas is analyzed.
When the Freewave radio was put in the case
with the analyzer, together with the CR10X data logger and battery,
when the shore based computer read the data continuously, a
wide amplitude sine wave was produced, overwriting the carbon
readings which should not have varied more than 5 or so measurement
units over time, but were swinging as high as 10 to 200 units.
The following picture of the PC CRT while data was being collected,
indicates the degree of interference. The line should be essentially
flat.
When the radio on the small buoy was turned
off, the interference went away.
Several times during our meetings the issue
of this interference came up. Everyone assumed that it was from
radio wave frequency propagation by the radio, and that moving
the antenna might be the solution. But everyone was concerned
also, for if the wireless devices produce interference with
the gas spectrometer there might be other cases where the data
is more subtly warped.
We set up a series of tests, wherein I first
changed the frequencies being used by the Freewaves, which normally
operate all across the spectrum from 902-928Mhz. There is a
'Hop Table' function in Freewaves which permit changing the
sets of frequencies used by the radios between those extremes.
The Hop Table is usually used to change the frequencies to conform
to different nation's spectrum rules. Such as Brazil's rules
that limit spread spectrum radios to the 902-925, rather than
928, bands.
But after going though all 6 frequency choices,
the problem did not go away.
So I tried the 'Garbage Can' test. Using
a large metal Garbage Can in the lab, I extended the distance
the radio was from its original place only 8 inches from the
circuitry of the gas analyzer, as far as the RS232 cable would
reach - about 8 feet - and put it on the bottom of the can.
With the metal of the can between the radio and the case with
the analyzer. We also carried the lap top computer running the
analytical software and the second radio out behind a metal
door. The interference problem went away.
The problem was not the radiated RF from
the antenna, but the radiated energy from the motherboard of
the radio itself, too close to the motherboard of the analyzer.
The solution is to separate the two, even
though, in the small space of the small buoy that is not easy.
If a larger radio is used, then a separate waterproof box with
a metal liner, or metal box inside is required.
We also tested the small radio, held high
above the equipment case, and by careful placement of it, eliminated
the interference.
There will be interference from some radios
with some sensor systems. Changing the physical relationships
between the radios and the sensory circuits, shielding them,
can help solve the problem.
