Magnetic flow meters utilize an insulated flowtube that carries liquid flowing past an electromagnet and electrodes. The electrodes are sealed in the flowtube and make contact with the flowing liquid. The electrodes sense an electromotive force (EMF) magnetically induced in the liquid, and proportional to flow rate according to Faraday's law of electromagnetic induction.

Along with this flow EMF, undesired noise is often received from the electrodes of a magnetic flow tube. This electrical noise is indicative of undesired operating conditions associated with the flowing liquid such as poor liquid grounding connections, excessive electrical resistivity in the flowing liquid, particles in the liquid impacting the electrodes, ongoing dissolving of particles in the liquid, ongoing chemical reactions in the liquid, entrained gas bubbles and the like.

This"electrode noise,"that originates in the liquid, can give rise to measurement instability or

variability in the flow output of the transmitter that can go undiagnosed by the operator of the process plant for long periods of time.

SUMMARY OF THE INVENTION A magnetic flow transmitter with noise diagnostics and a method of noise diagnostics are disclosed. The transmitter includes a digital processor that generates spectral components of a digitized, amplified electrode signal from a flow tube carrying a flow of a liquid. These spectral components extend over a fundamental frequency that selected for a coil driver output provided to the flow tube and also extend over a frequency of at least one known potential noise source. The digital processor generates a diagnostic output as a function of at least one spectral component at or near the frequency of the noise source. The digital processor also generates a flow output as a function of at least one spectral component at or near the fundamental.

The diagnostic output indicates the frequency and amplitude of noise detected from the fluid and whether the noise is large enough to affect the stability of the flow output so that corrective action can be taken if needed.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates a magnetic flow tube and a magnetic flow transmitter ; FIG. 2 illustrates a block diagram of a first embodiment of a magnetic flow transmitter with diagnostics ; FIG. 3 illustrates a block diagram of a second embodiment of a magnetic flow transmitter with diagnostics ;

FIG. 4 illustrates a sampled waveform of spectral components of an amplified, digitized electrode signal including a flow signal and noise signals at multiple frequencies; FIG. 5 illustrates a flow chart of a first diagnostic process; and FIG. 6 illustrates a flow chart of a second diagnostic process.

DETAILED DESCRIPTION OF THE ILLUSTRATED EMBODIMENTS A magnetic flow transmitter is disclosed in which a digital processor with diagnostics detects-- undesired noise received from an electrode circuit of a magnetic flow tube. The electrical noise is indicative of undesired operating conditions associated with the flowing liquid such as poor liquid grounding connections, excessive electrical resistivity in the flowing liquid, particles in the liquid impacting the electrodes, ongoing dissolving of particles in the liquid, or ongoing chemical reactions in the liquid, entrained gas bubbles and the like.

The noise, when it is large enough, can reduce the stability of the flow output of the magnetic flow transmitter.

The disclosed transmitter includes a digital processor that generates spectral components of a digitized, amplified electrode signal from a flow tube carrying a flow of a liquid. The spectral components extend over a fundamental frequency selected for use in a coil driver output provided to the flow tube, and also extend over a frequency of at least one potential noise source.

The digital processor generates a diagnostic output as a function of at least one spectral

component at or near the frequency of the noise source. The digital processor also generates a flow output as a function of at least one spectral component at or near the fundamental.

The diagnostic output indicates whether the stability of the flow output is affected by the noise so that corrective action can be taken.

In FIG. 1, a partially cutaway view of an embodiment of a magnetic flow meter 20 is illustrated.

Magnetic flow meter 20 includes a flowtube 22 formed of low magnetic permeability material with electrically insulating liner 23, an electromagnet 24 with coils 26, a ferromagnetic core or shield 28, and electrodes 30,32. The electromagnet 24 and the electrodes 30,32 are wired to a transmitter circuit 34. In operation, the transmitter circuit 34 drives the electromagnet 24 with an electrical current, and the electromagnet 24 produces a magnetic field 36 illustrated with arrows inside the flowtube 22.

Process liquid 21 fills the flowtube 22 and flows through the magnetic field in the flowtube 22. The flow induces an electromotive force (EMF, voltage) in the liquid 21. The insulating liner 23 prevents leakage of the EMF from the liquid 21 to the metal flowtube 22. The electrodes 30,32 contact the liquid 21 and pick up or sense the EMF which, according to Faraday's law, is proportional to the flow rate of the liquid 21 in the flow tube 22.

The EMF from electrodes 30,32 is carried to the transmitter circuit 34 by leads 38 that are insulated to avoid leakage. The transmitter circuit 34 has an electrode input circuit with high input impedance to limit leakage as well.

The electrodes 30,32 are well isolated from sources of noise and leakage originating outside the flowtube 22, however the electrodes 30,32 are exposed to contact with the liquid 21 passing through the flow tube 22, and are thus susceptible to electrical noise that is present in, or conducted by the liquid 21 itself. The electrical noise is indicative of undesired operating conditions associated with the flowing liquid such as poor liquid grounding, excessive electrical resistivity in the flowing liquid, particles in the fluid impacting taie- electrodes, ongoing dissolving of particles in the liquid, or ongoing chemical reactions in the liquid, entrained gas bubbles and the like.

In FIG. 2, a schematic block diagram of a first embodiment of a magnetic flow transmitter 40 is illustrated connected to a magnetic flow tube 42. The magnetic flow transmitter 40 includes a driver 44 that provides a driver output 46 at a selected fundamental "Fo,"also called the first harmonic or the lowest frequency component of a complex wave. The magnetic flow tube 42 has an electromagnet or coil 48 that is driven either directly or indirectly by the driver output 46. Typically, driver 44 provides a one ampere, approximately square wave, current to the coil 48. Electric current in coil 48 produces a magnetic field in flowing liquid 50 in the flow tube 42. An electromotive force (EMF or voltage) proportional to flow velocity is induced by the interaction of the magnetic field and the flow of the liquid 50 according to Faraday's Law. Electrodes 52,54 pick up or sense this EMF. Electrodes 52,54 are coupled to amplifier 60 by leads 56,58.

The amplifier 60 has an amplifier input 62 that couples to the electrodes 52,54 on the flow tube 42 and has an amplifier output 64. Typically, amplifier input 62 includes high impedance unity gain amplifiers or buffers to avoid excessive leakage currents from the electrodes 52,54. As one skilled in the art would appreciate, the amplifier output 64 typically has a single ended output that is referenced to a circuit common or ground. Amplifier 60 thus typically performs a differential-to-single ended conversion function.

An analog-to-digital converter 66 receives the amplifier output 64 and generates a digital output 68. that is a series of digital words that are proportional to the amplifier output 64.

A digital processor 70 receives the digital output 68 and generates spectral components 72 of the digital output 68 that extend over both the fundamental 74 and a frequency 76 of a known potential noise source. The known potential noise source can be any noise originating in, or conducted by the liquid 50 in the flow tube as discussed above.

The processor 70 generates a flow output 80 as a function of at least one spectral component at or near the fundamental 74. Preferably the processor 70 and the driver 44 are controlled by the same clock, and then the flow output is represented by one spectral component exactly at the fundamental 74. The processor 70 further generates a diagnostic output 82 as a function of at least one spectral component at or near the frequency 76 of the noise source. The flow output 80 and the diagnostic output 82 are preferably galvanically isolated from the electrodes 52,54.

This galvanic isolation can be integrated into the transmitter 40 in several different ways. Amplifier 60 can be a isolation amplifier, the digital output 68 can be coupled through an optical isolator, or the output 80,82 can be coupled through optical isolators, for examples.

The processor 70, in one preferred embodiment, comprises a digital signal processor (DSP). The operation of the digital signal processor is preferably synchronized with the sampling done in A/D converter 66 so that delay time and response time ef- the transmitter are shorter. In another preferred embodiment, the processor 70 comprises a microprocessor that also performs other control and computing functions in transmitter 40. The use of a microprocessor system is advantageous because it allows detailed integration of the digital signal processing with other transmitter functions in a software program stored in memory in the microprocessor system. In yet another embodiment, the processor includes digital bandpass filters with center frequencies corresponding with the flow signal frequency and also frequencies selected to provide a smaller number of spectral representations at known noise or interference frequencies.

The processor 70 preferably generates the spectral components 72 in the frequency domain as a transform function of 1024 byte blocks of 24 bit digital data in the time domain data at the digital output 68. Larger blocks of memory, such as 2048 or 4096 byte blocks can be used to get narrower resolution in each data bin, improve accuracy of the noise measurement and reduce"leakage"of the flow

signal into nearby bins, especially when noise is at frequencies of less than 100 Hz. The transform function is preferably a Fourier type transform, such as a fast Fourier transform (FFT) or a discrete Fourier transform (DFT). The. time domain data at the digital output that is processed by the transform is substantially a rectangular wave or square wave with noise superimposed on the wave. The wave has substantially the same shape and fundamental Fo as the wave provided by driver 44.

The processor 70 also preferably generates spectral components 72 that are synchronized with the driver output 46. Processor 70 can be synchronized with driver 44 by use of the rectangular or square wave components present at digital output 68 and a phase locked loop. Synchronization can also be achieved by clocking both the processor 70 and the driver 44 from a common clock. Preferably, however, the processor controls the driver as explained in more detail below in connection with FIG. 3.

In FIG. 3, another embodiment of magnetic flow transmitter 90 is illustrated in which a digital processor 92 is synchronized with a driver 94 using a synchronization line 96. The embodiment shown in FIG.

3 is similar to the embodiment shown in FIG. 2 and the same reference numerals used in FIG. 3 as are used in FIG. 2 to identify the same or similar features. The synchronization line 96 synchronizes the operation of driver 94 under control of the digital processor 92 so that the substantially rectangular flow signal is processed in a synchronous manner with the correct phase. The processor 92 controls the driver 94 via line 96 to provide the desired synchronization.

When noise is of a high enough level to affect the stability of the flow output 98, the processor 92 can select, change or control the fundamental Fo via select line 100 coupled between driver 94 and processor 92. By changing the fundamental Fo, the processor 92 can move the flow signal in the frequency domain to avoid the detected noise and thus restore the stability of the flow transmitter output automatically. To improve stability when high noise levels are detected, the processor 92 can also select, change or control signal processing functions such as damping of the output signal, or select known noise rejection software routines for use by the processor 92. Noise rejection software techniques can be used to effectively increase the averaging time of the noise signals to improve stability. Algorithms resident in the noise rejection software routines differentiate noise from changes in the process variable to allow rapid response to those changes.

Signal processing techniques can also be selected from those described in the above cited pending application serial number 09/209,134. The algorithms used in the software routines can include neural networks, wavelet signal processing and other known filtering techniques.

In FIG. 4, spectral components of an amplified, digitized electrode signal including a flow signal and noise signals at multiple frequencies are displayed in an output image of a spectral analyzer. A spectrum extending from frequencies near zero frequency at 110 up to 100 Hertz at 112 is displayed. The fundamental Fo of the drive and the flow signal is at approximately 37.5 Hertz at 114 and a second harmonic

of the flow signal at approximately 75 Hertz is displayed at 116. Noise from the 60 Hertz power line appears in the liquid in the flow tube as displayed at 118. Noise from particle impact on electrodes, chemical reaction noise and the like appears as so- called"red noise","pink noise,"or"1/F noise"at 120.

In digitally processing the data shown in FIG. 4, various known spectral averaging,"focussing,"and regression techniques can be used to combine a larger series of spectral data points into a smaller sera's- of averages of small groups of adjacent data points to achieve higher repeatability and stability. The flow signal, for example can be a numerical regression of several data points around the fundamental Fo. These alternate averaging or regression techniques are especially useful in situations where the processor is not precisely synchronized with the driver, or where the processor generates more spectral data at finer spectral increments than is actually needed for calculating flow rates. Spectral data points can also be distributed in non-uniform increments for specialized noise problems, especially when noise is at less than 100 Hz.

In FIG. 5, a diagnostic method is illustrated in more detail. In using the method shown in FIG. 5 in a magnetic flow meter, a coil on a flow tube carrying a flow of a liquid is driven with a substantially rectangular wave at a repetition frequency or selected fundamental Fo. A signal from electrodes on the flow tube is amplified to provide an amplifier output. The amplifier output is converted to a digital output in an A/D converter. The digital output from the A/D

converter is taken at 140 in FIG. 5 and used as data input to a software routine 142 that can also control the A/D converter. A block of data from the digital output is stored in RAM at 144. The processor uses a portion of the block of data stored in RAM at 144 in a software routine that calculates flow rate by evaluating the data in synchronization with the coil drive as illustrated at 146. The results of the calculations at 146 are passed to an output software control routine 148 which controls the flow output 150.

A portion of the block of data stored at 144 is also passed to a spectral software routine 152 that calculates the spectral distribution of electrode noise. The noise values are diagnosed in a software routine at 154 and noise values are output at 156.

The diagnostic software routine 154 also generates text strings of recommendations for corrective action at 158. The noise values at 156 and the text strings at 158 are diagnostic outputs 160 that are passed to digital communication or transmission to a remote location, display on a liquid crystal display that is located in the transmitter or used to select a different fundamental frequency Fo to avoid the noise.

In FIG. 6, an alternate diagnostic method or software routine is illustrated. The software routine starts at 170. Data from the digital output of the A/D converter is stored in RAM at 172 and a diagnostic routine is performed on the data at 174. The diagnostic routine at 174 performs a Fourier type transform on the time domain data and generates a frequency domain representation as spectral values that extend over the fundamental frequency and known

interfering noise frequencies. Next, the spectral values are compared to stored reference values to ascertain if 60 Hertz noise is high enough to cause instability at 176. If the 60 Hertz noise is too high, the routine branches at 178 to routine 180 to transmit an alert and noise values to the customer at 182 at a remote location, display the noise value and the alert on a local display and provide text strings suggesting corrective action and control the fundamental Fo to a different value to avoid noise as illustrated at 184 or take other corrective action ; as-- illustrated at 186.

If 60 Hertz noise is not detected, then program flow branches at 188 to compare stored spectral values to reference levels stored in memory at 190. If low frequency (1/F) noise is too high, then the program flow branches at 192 to perform corrective actions described above. If low frequency noise is not too high, then program flow branches at 194 to return to the main program flow at 196.

The various methods, software routines and instructions can be stored on a computer readable medium such as a hard disc, a removable disc, a CDROM or other media for downloading or uploading to memory in the magnetic flowmeter.

Although the present invention has been described with reference to preferred embodiments, workers skilled in the art will recognize that changes may be made in form and detail without departing from the spirit and scope of the invention.