In a method for measuring flow by means of an ultra sonic flow
meter, an ultra sonic signal is transmitted in an upstream and a
downstream direction. This is measured upon receipt of an ultra
sonic signal and the measurement stopped at a stopping point. A
first series of transmissions is made, and the starting time of
each transmission is incremented or decremented until a time difference
between the upstream and downstream signal is inside a reference
band. Flow is then calculated in accordance with the time measurements.
What is claimed is:
1. A method for measuring flow by means of an ultra sonic flow
meter, the method comprising the steps of transmitting an ultra
sonic signal in an upstream and a downstream direction, initiating
a time measurement at a starting time upon receipt of the ultra
sonic signal, and stopping the time measurement at a stopping time
making a first series of transmissions where the starting time in
each transmission is incremented or decremented until a time difference
(.DELTA.t) between the upstream and downstream signal is inside
a reference band (.DELTA.t.sub.span) and calculating the flow based
on the time measurements.
2. A method according to claim 1 including the further step of
generating a second series of transmissions following the first
transmissions, where the starting time in each transmission is incremented
or decremented until the time period (DS_RUNUP) between the starting
time (DS_START) and the stopping time (DS_STOP) is approximately
equal to a multiple of half the time period of the ultra sonic signal,
but preferably equal to the time period.
3. A method according to claim 2 including the further step of
generating a third series of transmissions following the first and
second series of transmissions, incrementing or decrementing the
starting time in the third series of transmissions until a difference
between a calculated average transmission time and a transmission
time reference value determined on the basis of the media temperature
is inside a reference band.
4. A method according to claim 3 wherein the incrementation or
decrementation during the third series of transmissions is done
in steps of a timely resolution (t.sub.sig) that is approximately
equal to the period of the received signal.
5. A method according to claim 1 wherein the incrementation or
decrementation is done in steps of a timely resolution (t.sub.res)
defined by the resolution of the microcontroller.
6. A method according to claim 1 wherein following the starting
time, the time measurement is stopped at a first positive zero crossing
of the received ultra sonic signal or at a first negative zero crossing.
7. A method according to claim 1 wherein following the starting
time, the time measurement is stopped at a first positive zero crossing
following a negative zero crossing of the received ultra sonic signal,
or vice versa.
8. A method according to claim 1 wherein the starting time in
the first transmission in the first series of transmissions is a
fixed value (DS_START_INI).
9. A method according to claim 2 wherein the reference band (.DELTA.t.sub.span)
is delimited by a maximum difference time (.DELTA.t.sub.max) and
a minimum difference time (.DELTA.t.sub.min), the width of the band
being smaller than the time period (t.sub.sig) of the ultra sonic
10. A method according to claim 9 wherein a safety margin is added
to the reference band, the safety margin corresponding to at least
the timely resolution (t.sub.res) of the incrementing or decrementing
steps, and preferably the sum of a resolution and the duration of
a liming noise (t.sub.n).
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is entitled to the benefit of and incorporates
by reference essential subject matter disclosed in Danish Patent
Application No. PA 2002 01018 filed on Jun. 30 2002.
FIELD OF THE INVENTION
The invention concerns a method for measuring flow by use of an
ultra sonic flow meter. More specific, the invention is directed
to a method of measuring transmission times of, and time difference
between, an upstream and a downstream ultra sonic signal.
BACKGROUND OF THE INVENTION
The time difference between an upstream and a downstream ultra
sonic signal is proportional to the flow, and used in time of flight
ultra sonic meters as a measure of the flow. If the time difference
.DELTA.t becomes longer than the duration of the period of the ultra
sonic signal, an exact detection of the time difference becomes
difficult due to the signal periodicity. In order to avoid this
problem, known solutions provide detection circuits that are practically
independent of the extent of .DELTA.t, i.e. the detection circuit
makes measurement possible on ultra sonic flow meters, which have
a .DELTA.t longer than the period of the signal. An example of such
a prior art detection method and circuit--also called a trigger--is
described in the following, where the envelope of the upstream and
downstream signal play a significant role.
The basic purpose of a trigger in a transmit time ultrasonic flow
meter is to "point" out the time of arrival of the ultrasonic
signal. This is used to measure both the difference between the
upstream and downstream transmission time and to measure the two
transmission times. From these values the flow Q can be calculated
according to (1):
.DELTA..times..times. ##EQU00001## where .DELTA.t is the difference
time, t.sub.1 and t.sub.2 the transmission times and "k"
is a constant dependent on the geometry of the tube. If the media
is known, the measurement of the two transmission times can be replaced
by measuring the media temperature and calculating the sound speed
C from knowledge of the variation of the speed of sound with temperature:
Q=k.DELTA.tC.sup.2 (2) where Q is the flow, k is a constant, .DELTA.t
the time difference and C the sound speed.
FIG. 1 is an illustration of the receive signals--the first arriving
is the result of the sound pulse travelling in the flow direction,
1 and the second is the result of the sound pulse travelling against
the flow direction, 2. In the following the term zero crossing will
be used, in practical implementations this will be signal zero (the
middle of the range of voltage used in the implementation) or some
value either a little over or under the signal zero. Still referring
to FIG. 1 the basic problem is to trig or initiate the time measuring
circuit with the "same" zero crossing in the upstream
and the downstream sound pulse, otherwise a wrong .DELTA.t is measured.
P1 and P2 are to be imagined as same zero crossings because each
have a distance of 31/2 periods from reception of the sound pulse.
Also indicated in the Figure is the period t.sub.sig of the sound
signal and the time difference .DELTA.t.
FIG. 2 shows how a prior art ultra sonic flowmeter uses the envelope
of the ultrasonic signal to achieve a zero crossing detection that
is independent of the length of .DELTA.t. The incoming signal (S1)
is rectified (B1) and the result is (S2). This signal (S2) is fed
through a band pass filter with non-minimum phase behaviour (B2).
Non-minimum phase systems have the transient property that their
initial direction of response is in the opposite direction of the
final value--as a consequence, if the filter parameters are chosen
appropriately, the output of the filter (B2) will have a well defined
zero crossing indicating the receive time. Furthermore this zero
crossing will be independent of the amplitude of the receive signal.
The signal on the output of the filter is seen as (S3). The zero
crossing of the signal (S3) is detected by the zero cross detector
(B3), this signal (S4) is arming the zero cross detector (B4). After
arming the zero cross detector (B4), the next positive or negative,
dependent on the actual implementation), zero crossing in the original
receive signal (S1) is detected by (B4) resulting in the signal
(S5). The time where the signal (S5) changes from low to high is
measured relative to the time of the transmit burst (or relative
to another time with a known relation to the transmit time). If
the time between the zero crossing of S3 and the following zero
crossing of S1 is very short, there is a risk of detecting two different
zero crossings of S1 due to random noise. To avoid this situation
it is detected if the two zero crossings are too close, and if this
is the case, the transmit signal is inverted--and hence the receive
signal. The consequence of the inverted receive signal is that the
previously very short time difference between S3 and S1 is now close
to one half period of the receive signal. One can chose to measure
transit time on the signal zero crossing (S5) or on the zero crossing
of the signal (S4). After having calculated a time as described
above for an upstream signal, the same procedure is used on the
downstream signal. From these two times, a difference time is established
and the flow Q calculated.
The described detection method works well in systems were the span
of .DELTA.t is unknown. This is the case for a general purpose ultra
sonic flow meter as the one described above, which are used for
a variety of tubes having different diameters. This type of ultra
sonic flow meters must be able to cope with a very wide span of
.DELTA.t. However, in some systems, the span of .DELTA.t is limited
by fluid velocity and/or the mechanical arrangement of the ultra
sonic transducers which means that the ultra sonic converter can
be designed according to other and less demanding principles. Such
a limitation in .DELTA.t is the case, if the two ultra sonic transducers
mounted in the tube are very close to each other. It will then be
known that .DELTA.t e.g. will have a maximum value of e.g. 1 .mu.s.
Further, a drawback of the prior art design described above is the
relatively extensive and thus costly use of electronic circuitry.
Another weakness of the method is the dependence on a stable signal
envelope. If for instance a single pulse in the receive signal has
a lower amplitude due to electrical noise or particles/air bubbles
in the liquid, the envelope form changes, and consequently a wrong
.DELTA.t will be calculated.
Based on the foregoing, the object of the invention is to provide
a detection method which is realized in a simpler way and with fewer
electronic components, and still gives a reliable statement as to
the difference in transmission time, .DELTA.t.
SUMMARY OF THE INVENTION
The basic idea of the invention is that of trial and error. During
the first series of transmissions, multiple transmissions are performed,
each transmission consisting of an ultra sonic signal in the downstream
and upstream direction. The starting time of time measurement in
the first series of transmissions is based upon a good estimate
of when the incoming ultra sonic pulse arrives. If .DELTA.t is inside
the reference band, the starting time chosen was appropriate. If
not, a second transmission is launched, but this time the starting
time is incremented or decremented by an amount, thus raising the
chances of success. These transmissions are repeated until .DELTA.t
is inside the reference band. The ideal case is the one in which
the initial estimate of the starting time corresponds to the optimum.
In this case, only one transmission will be performed. The method
is especially applicable in ultra sonic flow meters having small
.DELTA.t's. Advantageously, the inventive trigger method makes it
possible to dispense with much of the trigger hardware, hereby lowering
the cost. Compared to FIG. 2 electronic circuits B1 and B2 can
be omitted. The method has the further advantage, that it is simple
to implement and very robust. Thus, the method is practically independent
of signal distortions, because it is based on the detection of zero
Following the first series of transmissions and having placed .DELTA.t
within the reference band, a second series of transmissions can
be launched. Though the method of using a first series of transmissions
suffices to make a correct calculation of .DELTA.t and flow Q, an
improvement is reached by introducing a second series of transmissions
sequentially following the first series. The second series remedy
problems of variations in the travel time and thus changed zero
crossings of the received ultra sonic signal due to varying temperatures
of the liquid media. The second series of transmissions is repeated
until the time period between the starting time and a stopping time
of the time measurement is approximately equal to a multiple (1
2 3 . . . ) of half the time period of the ultra sonic signal.
However, a full time period of the ultra sonic signal is preferred
as reference value. The use of half the time period as reference
can be preferred if a time circuit with no unlinearities is used.
The idea of this second measure in the inventive method is to keep
the average time from the starting time to the stopping time constant,
hereby positioning the starting and stopping times in an optimum
position independent of temperature changes in the media.
Advantageously, the first and second series of transmissions can
be supplemented by a third series, where the starting time is incremented
or decremented until a difference between a calculated average transmission
value and a reference value is inside a reference band. The reference
value is determined as a function of the media temperature, and
the temperature is either measured or calculated.
The incrementation or decrementation of the starting signal in
the third series of transmissions is preferably done in steps with
a resolution in time that approximately corresponds to the period
of the received signal. A slight variation in the period of the
received signal is normal due to differences in temperature on the
transducers, thus +/-40 kHz on a 1 MHz transmitted signal is to
be expected, but in practice this has no influence. Thus a resolution
departing by a small amount of the period time can be used.
In relation to the first transmission series, the timely resolution
used in positioning the starting time corresponds to the timely
resolution of a digital control unit.
The time measurements, initiated at the starting time, may be stopped
at the first positive or negative zero crossing of the received
ultra sonic signal, but this demands the use of fast electronic
circuits. It is preferred, that the time measurement is stopped
at the first positive zero crossing following a negative zero crossing
of the received ultra sonic signal, because it allows the currents
and voltages in the electronics to settle, thus avoiding unlinearities.
Of course the order can be changed into stopping on a negative zero
crossing following a positive.
It is preferred, that the starting time in the first transmission
in the first series of transmissions is a fixed value which may
be stored in a memory. The fixed starting time is chosen to start
approximately in the middle or in the first half of the receiving
train of pulses, which ensures a signal with sufficient amplitude
The reference band of the first series of transmissions is delimited
by .DELTA.t.sub.max and .DELTA.t.sub.min and the width of the band
is smaller than the time period of the ultra sonic receive signal.
Preferably, the reference band is narrowed by adding a safety margin
to the reference band. The safety margin consists at least of the
timely resolution, but may also comprise a duration of a timing
noise. By incorporating a safety margin into the reference band,
increased stability of the method is achieved.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention will now be described on the basis of the figures,
FIG. 1 is a time-amplitude diagram of an upstream and a downstream
signal travelling in a media in a tube and generated by ultrasonic
FIG. 2 shows a prior art detection scheme used for finding .DELTA.t.
FIG. 3 is a time-pulse diagram according to the invention.
FIG. 4 shows the reference band of the time difference .DELTA.t
according to the invention.
FIG. 5 shows a flow chart according to the invention
FIG. 6 is a block diagram of the preferred embodiment of the invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
The following preferred embodiment describes a trigger scheme for
ultrasonic flow meters used for measuring media with a known relation
between the sound speed and the temperature of the media, and in
which the difference time, .DELTA.t, between signals transmitted
upstream and downstream is limited to be within one period, t.sub.sig,
or less of the received ultrasonic signal (i.e. within approximately
1 .mu.s for the most commonly applied ultrasonic transducers).
Referring to FIG. 3 at both the upstream and downstream receive
signal, the following sequence is executed: 1) At some instant a
signal DS_START, starts the time measurement system. The time of
DS_START has a known relation to the transmit time, and must be
within the body of the receive signal. 2) The measurement system
is stopped by DS_STOP, which is the first positive going edge following
a negative going edge in the receive signal. By using the negative
going edge to qualify the positive going edge as the stop signal,
the time measurement system is always assured to have a minimum
measurement time of one half period of the receive signal, thereby
avoiding otherwise possible nonlinearities of the measurement system
and arming logic, when measuring time intervals close to zero.
FIG. 3 shows four different trigger cases, 1 4. For clarity, only
the digitized signals of the ultrasonic receive signal bursts (RX_UP
(dashed line) and RX_DOWN) are shown. The signals shown correspond
to positive flow, i.e. flow in the direction where the downstream
signal is faster than the upstream signal. For each of the four
cases, the resulting input signal (DS_RUNUP) to the time measurement
system is shown. The positive going edge of the DS_RUNUP signal
corresponds to DS_START, and the negative going edges corresponds
to the DS_STOP signals generated for each up- and downstream measurement.
In cases 1 3 the correct .DELTA.t is measured as indicated by the
.DELTA.t arrows. Taking case 1 as example, after generating the
send signal in the down stream direction a timer is started at time
t.sub.a (initially at the time DS_START_INI) and stopped t.sub.b
when a negative flank is followed by a positive flank in RX_DW.
Now the downstream transmission time has been measured.
After generating the send signal in the upstream direction a timer
is started at time t.sub.a and stopped at t.sub.c when a negative
flank is followed by a positive flank in RX_UP. Now the up stream
transmission time has been measured. Subtracting the downstream
time from the upstream time returns a positive value of .DELTA.t.
However, in case 4 the timing of DS_START results in an erroneous
measurement of .DELTA.t because the counter is not stopped until
t.sub.d. .DELTA.t becomes negative as indicated with the arrow pointing
in the opposite direction of the arrow in case 1. Thus, as different
zero crossings of the upstream and downstream signals are compared
to each other, this measurement must be discarded.
The valid range for DS_START in the figure is denoted with letter
A, and the invalid ranges are denoted with letter B. As the flow-rate
increases, the A-ranges decreases, and the B-ranges increase correspondingly,
leaving only the start position in case 2 as valid when the displacement
between the measurements has reached the maximum value of one signal
period. In order to be able to distinguish between valid and invalid
measurements, the span of .DELTA.t must be limited to less than
the signal period, t.sub.sig. The following is observed: The maximum
measurement range: .DELTA.t.sub.span=.DELTA.t.sub.max-.DELTA.t.sub.min<t.sub.sig.
At positive flow an erroneous measurement results in a measured
.DELTA.t of: .DELTA.t.sub.meas=.DELTA.t-t.sub.sig. At negative flow
an erroneous measurement results in a measured .DELTA.t of: .DELTA.t.sub.meas=.DELTA.t+t.sub.sig.
However, when repositioning the starting time, there are some practical
limitations to take account of. First, the DS_START instant can
only be adjusted with a minimum resolution of t.sub.res as indicated
in FIG. 3 typically dictated by the clock resolution of a microcontroller.
A typical resolution is 250 ns for a microcontroller running a 4
MHz clock speed. Second, the peak to peak timing noise, t.sub.n,
as measured on the qualifying (negative going) edge of the receive
signal (RX_UP/RX_DOWN), relative to DS_START (the noise is partly
thermal noise from the circuitry, and noise induced by flow fluctuations).
This leads to the following practical constraints on the reference
TABLE-US-00001 At positive flow, .DELTA.t > 0: .DELTA.t.sub.max
< t.sub.sig - (t.sub.res + t.sub.N) At negative flow, .DELTA.t
< 0: -.DELTA.t.sub.min < t.sub.sig - (t.sub.res + t.sub.N)
FIG. 4 shows the relations between .DELTA.t.sub.measured, .DELTA.t.sub.max,
.DELTA.t.sub.min, t.sub.res, t.sub.n and the true .DELTA.t. The
unbroken curve corresponds to successful measurements of .DELTA.t,
and the dashed curve corresponds to erroneous measurements of .DELTA.t,
.DELTA.t.sub.error. The boxes t.sub.res+t.sub.n represents the above
mentioned inequalities. Results are skipped if they are outside
the limits of .DELTA.t.sub.min and .DELTA.t.sub.max. Considering
the case of .DELTA.t.sub.max as the limiting factor in FIG. 4 it
is seen that the limited resolution of DS_START, t.sub.res and the
noise t.sub.n allows for a similar amount of negative flow .DELTA.t.sub.min,
and vice versa. If .DELTA.t.sub.min is the limiting factor, the
opposite will be the case.
The inventive trigger scheme has three different levels of action,
which all perform adjustments on the DS_START value used for the
following measurement. Before each pair of upstream and downstream
measurements are performed, DS_START is calculated as: DS.sub.--START=DS.sub.--START.sub.--INIT+L1+L2+L3.
DS_START_INIT is the initial value of DS_START. L1 L2 and L3 are
the adjustment results for each trigger level. The initial value
Level 1 of the trigger scheme is the basic trigger functionality
which ensures that .DELTA.t is measured on the same zero-crossing
(relative to the signal start) in the upstream and downstream receive
signals. The level 1 mechanism is as follows: If .DELTA.t.sub.measured
is outside the interval [.DELTA.t.sub.min; .DELTA.t.sub.max], L1
is incremented in steps of t.sub.res, following the sequence: L1=(0-1-2.
. . (n-1)-0-1-2 . . . ) x t.sub.res, where n =round(t.sub.sig/t.sub.res).
I.e. L1 sweeps one period t.sub.sig of the receive signal.
The consequence of using only level 1 of the trigger is that, as
the time of flight varies with media temperature, the signal zero-crossing
used for measurement changes.
Level 2 of the trigger is only invoked after passing level 1 without
adjustments. The purpose of level 2 is to keep the average time
from DS_START to DS_STOP (i.e. DS_RUNUP) constant within the limits
given by t.sub.res. By choosing t.sub.sig as the target time for
the average value of DS_RUNUP, the starting point DS_START is kept
in the optimal position (corresponding to case 2 in FIG. 3), where
the distance in time to the error trig ranges B is as long as possible.
The level 2 mechanism is as follows: Calculate the average of the
measured upstream and downstream DS_RUNUP times. If the average
DS_RUNUP differs from t.sub.sig by more than .+-.t.sub.adj, L2 is
adjusted up or down accordingly, in steps of t.sub.res. The limit
for adjustment t.sub.adj is based upon t.sub.res, and should be
bigger than t.sub.res/2 because an adjustment smaller than t.sub.res
in unwanted manner would increase the distance to the optimal point.
With the level 2 part of the trigger scheme implemented, the trigger
point will track the received signal, as long as there are no signal
drop-outs. To be able to track the signal in all cases, the third
level of the trigger is invoked.
Level 3 of the trigger is only invoked after passing level 2 without
adjustments. The purpose of level 3 is to track a certain zero-crossing
in the receive signal, at all media temperatures. The measured media
temperature and the relation of sound speed to media temperature,
is used to calculate the expected average transmission time from
transmit to receive.
The level 3 mechanism is as follows: The directly measured average
transmission time (DS_START+DS_RUNUP-SIGNAL_OFFSET(*)) is compared
to the expected transmission time from the temperature measurement.
If the comparison is outside the limits of .+-.t.sub.sig/2 L3 is
adjusted up or down accordingly, in steps of t.sub.sig.
((*) SIGNAL_OFFSET is the distance from the receive signal start
to preferred signal zero-crossing).
Besides the result of tracking a constant zero-crossing in the
signal, level 3 gives the extra benefit, that it is now possible
to use the measured transmission time in the flow calculation instead
of using the transmission time measured indirectly via temperature.
This allows the temperature measurement to be rather crude without
inflicting the flow measurement. A flow chart describing the three
levels is found in FIG. 5.
FIG. 6 is a block diagram of the preferred embodiment 14 of an
ultra sonic flow meter. A generator 4 generates an exitation pulse
for a front end 5. To this circuit ultra sonic transducers 3 are
connected. From front end 5 the ultra sonic receive signals are
passed to a comparator 6 which converts the analog signals into
digital signals. A first flip flop 7 and a second flip flop 8 receives
the digital pulses. The second flip flop 8 is armed by the output
of the first flip flop, and the first flip flop 7 is armed by the
signal DS_START coming from an arming logic 13 placed in a microcontroller
14. Inputs to the arming logic is the difference time .DELTA.t,
two transmission times (t1 t2) and the media temperature. The difference
time and the transmission times are fed from a time measurement
circuit 10 which measures the width of DS_RUNUP, which is the difference
between the starting signal DS_START and the stopping signal DS_STOP.
DS_STOP (negated) is the output of the second flip flop 8 and fed
to AND circuit 9. Box 12 shows the set of parameters that are used
in the arming logic for the generation of the starting time DS_START.
Track select 15 selects the set of transducers to use.
Due to the low, .DELTA.t unbalances in the generation of the upstream
and downstream signals must be avoided. Such differences may occur
if two different electronic send and receive circuits are used,
where component tolerances cause differences in the group run time
of the signals. Even small differences will have a huge error impact
on ultra sonic systems using small .DELTA.t. In order to overcome
this problem a circuit as described in DE 100 48 959 A1 is used
as front end 5. This circuit uses one and the same electronic components
for send and receive, thus neither component tolerances nor temperature
differences influence on .DELTA.t.
The span of the typical reference band in this embodiment is -20
ns up to 600 ns. If the measured .DELTA.t falls outside of this
span, the measurement is skipped.