2.4 DUAL LEVEL SENSING
3 - PROCESSING &
CIRCUITRY
When two trip levels are desired, for
example for high-low limit sensing, the
electrode or probe set should have two
3.1 SLOSH FILTER
distinct tiers.
A typical twin external
It is desirable to suppress rapid, multiple
detections of fluid level generated by the
surface movement of the fluid, for
electrode is shown in Figure 2-3 (they are
connected together to the sense line);
typical internal twin electrodes are shown
in Figures 2-6, 2-7, 2-9, and 2-11. The
response of a properly constructed 2-tier
probe is shown in Figure 2-3.
example in
a
moving vehicle. To
accomplish this, the QT114 incorporates a
detection
integration
counter
that
increments with each detection until a limit
is reached, after which point one of the
Dual level electrodes should have an
approximately 3:1 surface area ratio or
more from T2 to T1; that is, the surface
area at T2 should be at least 3x the
surface area of the electrode at T1. There
is no penalty for making T2 excessively
large. The high ratio is required to
overcome the QT114's decreasing gain
with increasing Cx load (Figures 4-1, 4-2).
With internal dual-level probes where T1
and T2 are substantially separated, the
intervening connection between the two
levels should be more thickly insulated, for
example with a thick plastic spacer, and
OUT lines is activated. If during
a
detection ‘event’ the fluid level falls below
the electrode level (signal rises above a 'T'
point in signal counts), the counter
decrements back towards zero. Over a
long interval the up and down counts will
tend towards either zero or the limit, with
the result being a statistical function of the
number of detections vs. non-detections. If
on average there are more detections than
non-detections, the counter will eventually
make its way to the limit value and an
OUT line will activate.
Figure 2-12 A 2-tier spiral wire
probe with ground rod
Once a detection has been established,
the counter must find its way back to zero before the affected
OUT line goes inactive, via the same process. Although the
counter has a nominal reaction time of 15 seconds, in some
cases it may take several minutes before the outcome is
resolved depending on the violence of the fluid surface. If the
fluid surface is stable however, it will only require 15
seconds to change the state of an OUT line.
Both OUT1 and OUT2 have their own independent slosh
filters. Both are enabled or disabled in unison by strap
option, pin 4, 'FILT' as follows:
any remaining internal gap inside the spacer should be filled
with silicone sealant or epoxy. This will help to prevent the
signal from rising much between the two levels, thus
preserving a crisp bi-level response like that shown in Figure
2-3.
2.5 GROUNDING CONSIDERATIONS
In all cases ground reference coupling to the fluid must be
made. In aqueous fluids, this can simply mean connecting
the metal vessel to circuit ground, or inserting a bare metal
element into the bottom of a plastic or glass vessel. The
degree of galvanic contact is not critical, so scale and
corrosion on the ground electrode are not of great concern
especially if the 'connection' to the fluid is substantial
enough.
FILT = Gnd
Slosh filter off
FILT = Vcc
Slosh filter on
FILT strapping can be changed 'on the fly'.
If direct electrical contact to the fluid is not possible, a large
piece of external metal can be bonded to the outside of the
vessel and grounded. Once this is done, the signal should be
monitored while the vessel is touched by hand; if the
grounding is sufficient, the signal will not move or will move
only slightly.
Very large vessels, even if not grounded, often do not require
additional provision for grounding since the bottom surface
area and free-space capacitance of the tank may be
sufficient for ground return coupling.
In some cases (windshield washer tanks on cars for
example) there will exist a water path to a chassis-grounded
fitting somewhere downstream of the tank, or the water path
may be labyrinthine enough to provide enough capacitive
coupling to the grounded chassis even if it does not make
galvanic contact. In these cases no further provision for fluid
grounding is required. Simple experimentation will easily
determine whether the existing amount of parasitic coupling
to ground is enough to do the job.
3.2 CALIBRATION
Both the T1 and T2 trip point values are hardwired internally
as functions of counts of burst length. Sensitivity can be
altered relative to these trip points by altering electrode size,
geometry, degree of coupling to the fluid, and the value of
Cs. Selecting an appropriate value of Cs for a given
electrode geometry is essential for solid detection stability.
The QT114 employs dual threshold points set at 250 and
150 counts of acquisition signal. The signal travels in a
reverse direction: increasing Cx reduces the signal counts;
as a result, 250 counts of signal corresponds to the most
sensitive or ‘lower’ setting (T1), and 150 the least sensitive
'upper' setting (T2).
The baseline signal count when the electrodes are 'dry'
should begin at over 300 counts or more if possible. With a
small, weakly coupled electrode the baseline signal can be
trimmed to be closer to the 250 mark with a potentiometer to
provide a higher apparent gain by closing the gap between
the baseline and T1 (see below). The spread between T2
and T1 is fixed and cannot be separately trimmed.
In the case of coaxial probes, the ground connection is
inherent in the outer cylinder and no further ground
connection is required.
Increasing Cs will increase the baseline counts, while
increasing Cx will decrease it. When optimally tuned, each
threshold point will be symmetrically bracketed by signal
- 6 -
