‘signal. stimplest fequen eresolution Chapter 3 Signal Processing Objectives ‘+ Itisveryimportant to understand the operation af the vibration analyzer you areusing tocallect vibration data, 1+ Thesettings ae very important and canhave a HUGE effect on the quality and Usefulness ofthe data you ae collecting, and therefore on the success ofthe program. In this section we wil discuss the key settings and some of the mathematical relationships.Introduction {tis very important to understand the operation ofthe vibration analyzer you ae using to cole. bration data. The setings are veryimportant and can have aHUGE effect onthe quality and Usefulness ofthe data you are collecting, and therefore on the success of the program. In this section we wal discuss the key settings and some of the mathematical relationships. The Category I ibration analyst must understand the basi settings ofthe vibration analyzer to ensure routine measurements are performed correctly (the Category Il analyst should take responsbity for the settings used). “This chapter discusses what realy happens to the signal from the transducer; the processes related ta making the measurement from the perspective of capturing and manipulating the signal to produce the data we need, Itinvolves digitizing the signal and producing the spectrum, Issues related to that are covered such as ites, sample rates, resolution and windowing. A quick overview Inside the data collector] analyzer the “analog” signal from the sensor is manipulated ina variety of ways and “digitized” based on our settingsin order to provide usable data for analysis such as: waveforms, spectra, rms readings and more. We need to understand the settings and what happens “inside the box” so that we get good, repeatable data that contains the information we can use to determine the condition of the machine. five do not use the right settings, we will miss dues that reveal the nature and severity of fault condition. Figure 31 Data collector settings When you buy a digital camera it also has a ot of settings to choose from such asthe number of pixels, aspect ratio, Stop and flash settings. Ifyou take a picture at night with the flash off you can immeciately tell that you had the wrong setting asthe picture does not come out well Unfortunately, the same isnot true with vibration data. You may have the settings set incorrectly and you might not realize itt ‘7010 Nets situa Alias rs=r—3 ‘minumobusnsiue comse Fegire32 Manppatingthe data ‘Step One Power the sensor The ICP accelerometer requires power sent tot. The signal from the sencar ian electrical signal -an analog signal ‘Step Two Digtize the signal ~The Analyzer fs a digitalinstrument so the data must be converted into digital signal sa the analyzer willhave numbers to work with. Figure 33-Theanalog signal must be tized so theanayzrhasrumbarsto work with ‘There are a number of considerations: ‘© Does the voltage from the sensor match the input range ofthe analyzer? ‘+ How quickly should the signalbe sampled? {+ How many samples are kept? ‘Step Three ~ Process the signal- generate other data that can be worked with: spectra, ms ‘overallreadings, envelop spectra Considerations the processing: Fikering Integration Windowing AveragingcHA NAL PROCE Filters ‘When a transducer is mounted on a machine, the electrical output isa continuous analog signal representing the vibration at that location. Itisimportant to understand what fers do and how they are used in the field of vibration, “There are four types offers: Low pass ters: et low frequencies passthrough ‘+ Band pass ters: ft frequencies within aband to passthrough + Band! stopfters: blocks frequencies within band from passing through | High pas fiters lets high frequencies passthrough. “The band stop fiter isnot used as frequently in vibration analysis so we willnot cover it inthis Filter characteristics “Two types of fiters are common used in the field of vibration, Digital and Analog. Digital fiters are achieved with spectal “digital signal processing” (DSP) chips or software, [Analog fiers ae created with electronic components such as capacitors and resistors. Many analyzers stilhave some analog fier. “There are two issues to be concerned about: + Thefier characteristics (the cutoff) + Setting Times Ideally, the filters would block the unwanted frequencies ike putting up a brick wallto stop them (see Figure 3-4). This would provide a clean break and keep out all unwanted signals. However, thisis not the case.= = el to + een roe Fae art ec Fite Fgure4-laa ter design wouldBlock out all unwanted frequencies Inrealty, especially with analog filters, there a region where some frequencies willbe attenuated, but not blocked. There are other ways the datais massaged to get the results needed.CHAPTER 9 - SIGNAL PROCESSING into tion long enough to allow the circuit Filter examples Figure 36s spectral data from a compressor. This datais unfiltered data from 0 to10 kHz.CHAPTER 9 - SIGNAL PROCESSING PAGE 3-7 Figure 3-6-Data from amachine, nite applied Figure 37s the same data but witha Low Pass fiker applied Its letting the low frequencies ‘through while blocking the high frequencies. Low:pass filters are by far the mast common fter type, earning wide popularity in removing alias signals and for other aspects of data acquisition and signal conversion, For alovepass filter, the pass-band extends from DC (o Hz) the frequency specified, and the stop-band les above the specified frequency. Feure37-Same data but witha Low Pasfiter applied. Band: pass filters transmit only those signal components within a band around a center frequency. An ideal band-pass iter would feature brick wall transitions at fL and tH, rejecting, allsignal frequencies outside that range. Band-pass fiter applications include situations that requite extracting a specific tone, such as atest tone, from adjacent tones or broadband noise See Figure 38 Figwre38-Gond Pas fiteraple. it effectively bloks the high and low Frequencies. ‘S7OTeNonassiaie—Alighiswesenee——SSCSC~S~*~*~SCS*C COBand Stop (sometimes called band reject or notch) filters transmit all signals except those between specified frequency ranges. These filters can remove a specific tone -such asa 50 oF 60 H2line frequency pickup from other signal. Figure 9- Band Stop fiterblcks the frequencieslna partiadar frequency band, Ina High Pass filter, the pase-band ies above a specified frequency, while the stop-band resides below that point. Feure310- High Pass ter bocksal quencies below the specie frequency. High pass fters are normally used in early bearing wear detection. Manufacturers have their names for this feature: Spike energy detection, HFD, and ethers. It blocks the high amplitude, lower frequencies to enable the scaling to adjust to the low amplitude levels of early bearing ‘wearin the higher frequencies Fiters are usedin a number of applications inside the data collector. Four ofthe most obvious application are anttafasing fters, integration (converting the acceleration signalto veloc), tracking filters, and fkersusedin high frequency bearing analysis (demodulation, PeakVue, acceleration enveloping). Key points 4 Fiters are used in many pacesin vibration analysis. 1 When you select certain settings on your data collector you ae actually setingup diferent fers. ‘+ A Category analyst needs tobe able to recognize the thee types of fters (not the band top). 2010 Mobs iste Al nohs resaved onwe collect data with an accelerometer we are measuring acceleration. Ifwe wish to view or analyze the data in velocity units we have to convert the signal from the accelerometer via Integration can be performed on the analog signal before it s digitized, or on the digital signal. Some analyzers give you the option. 4 Analog integration: ©The very low frequency signals are amplified. We must use a high-pass fer to remove the low frequency signals. The digitized time waveform willbe in units of velocity. © Digitalintegration: 0 The digitized time waveform willbe in units of acceleration Itis then integrated in Ue software, Remember thatthe phase shifts by go degrees when we switch fram acceleration to velocity and by another 90 degrees when weintegrate from velocity to displacement. Acceleration leads velocity which leads displacement, Also recallthat acceleration accentuates higher Arequencies (above 2,000 #2), velocity fs good at mid frequencies and displacement is better at low frequencies. “S 7oi0 Netios stale Alias ws=Ted ‘mioumobusinsiue com