The Shocking Truth About the Frequency Domain

Transcription

1 The Shocking Truth About the Frequency Domain Presenter Herb Schueneman Founder, Board Chairman Moderator Edmund Tang Lab Manager, CPLP Professional September 2017

2 Herb s Bio 1940 s; starting out 1950 s; foundation building 1960 s; learning and watching amazing change 1970 s; climbing the ladder 1980 s; incredible growth 1990 s; pain and peaks 2000 s; descending the ladder 2010 s; passing the torch, amazing grace 2

3 Herb s Bio (cont d.) WESTPAK, Inc. Chairman of the Board, current President & CEO, Launched 1986 Pre-WESTPAK Employers Western Electric, Baxter Labs, Clorox Co., Lansmont BS, Michigan State University MBA, Northern Illinois University Raised on a farm in Michigan 3

4 Definitions Dynamics the motion of bodies under the action of forces Time Domain Analysis of time series data with respect to real time (rather than frequency) Frequency Domain Analysis of signals with respect to frequency (rather than time) Vibration conversion A spectrum 4

5 Scope Dynamics The Time Domain The Frequency Domain - Resonance - Diagnosing Frequency Response Case Studies 5

6 Start Here: VIBRATION A sub-set of Shock or is it the other way around? Why? Random Vibration A series of polarized excursions about a reference that vary randomly in amplitude and frequency over some period of time Sine Vibration Similar to Random, however the amplitude normally does not vary Frequency is controlled by the operator 6

7 Sine Vibration Time Domain Period Frequency Domain Frequency (f n ) = 1/period 7

8 Sine Vibration (cont d.) Where does it come from? 8

9 Random Vibration Time Domain Frequency Domain 9

10 Spectrum What is is? A spectrum can be described as: - A frequency domain representation of a time domain event. Simple! 10

11 Time Domain It s Real We live in the TIME DOMAIN as we Age (passage of time) Travel (miles/hour) Earn pay ($/hour) Describe a mechanical shock event Amplitude vs. time 11

12 Vibration All laboratory vibration is generated and controlled in the TIME DOMAIN Amplitude vs. time 12

13 Vibration We communicate with the vibration machines (give commands and receive the feedback) through the FREQUENCY DOMAIN Frequency vs. time 13

14 Vibration in the Test Laboratory Question How does the vibration controller convert the INPUT SPECTRUM we give it (Frequency Domain) to obtain the TIME DOMAIN signal needed by the (displacement controlled) vibration machine? 14

15 Vibration in the Test Laboratory (cont d.) Answer For sine (digital), the vibration controller synthesizes a sinusoid with at least 8 control points for each cycle. More difficult For sine (analog), vibration controller simply uses a function generator and varies the frequency and amplitude to achieve the desired results. Easy to do. 15

16 Vibration Response - Feedback Resonance identification!! This is - or should be - the primary reason for vibration testing in the first place. Extra credit: Why is this so? Answer WEBINAR Vibration-related damage during transit and KEY POINT! distribution is unlikely except when product resonances are excited. The problem becomes much worse when the transportation vehicle and/or the package system itself amplifies vibration input (resonates) at product natural or critical frequencies. 16

17 Vibration Response - Feedback (cont d) We measure response in the TIME DOMAIN, but we report and display it in the FREQUENCY DOMAIN. Extra credit: Why? Answer: Because it s easier!!! 17

18 Resonance So what is resonance, anyway? Let s define: SDOF Natural Frequency Resonance Fundamental (frequency) Harmonics 18

19 Resonance (cont d.) Single Degree of Freedom (SDOF) 19

20 Diagnosing Frequency Response EXAMPLE #1 20

21 Diagnosing Frequency Response EXAMPLE #1 cont d. A. What do we know? Primary Resonance = 14 Hz, amp 14X Secondary Resonance = 8 Hz, amp 9X All harmonics are 1X B. Potential Problems Both primary & secondary resonances are in the dense transportation envelope band (4-16 Hz) Since this test specimen was a crated assembly on a cushioned pallet, we can assume one of the peaks was the pallet resonance and the other was the product structural resonance. There should be a 1-octave separation between these. There is not. 21

22 Diagnosing Frequency Response EXAMPLE #1 cont d. C. Likely Results The truck carrying this packaged product will likely have an average acceleration input of perhaps.25 G RMS in the frequency band of 8 to 14 Hz. The package will take this input and multiply it by a factor of 9 at its resonant frequency of 8 Hz. This results in an average acceleration input of 2.25G's (.25 X 9) through the package and into the product. The product will take this "input" of 2.25Gs at 14 Hz and multiply it by a factor of 12 resulting in a net response of 27 Gs 14 times a second and likely damage before the product is delivered to the customer. 22

23 Diagnosing Frequency Response EXAMPLE #1 cont d. THAT s what we can learn from this spectrum! 23

24 Ratio (G/G) Diagnosing Frequency Response (cont d.) EXAMPLE #2 9) Cham ber Gasline 9) Chamber G Hz 0.10 Gasline/Control 12.2 G/G Hz Gasline/Control G/G Hz Frequency (Hz) line/control 4.52 G/G 24

25 Ratio (G/G) Diagnosing Frequency Response (cont d.) EXAMPLE #2 cont d ) Chamber Gasline 9) Chamber G A. WHAT DO WE KNOW? Primary Resonance = 16 Hz, Amp = 12X Secondary Resonance = 24 Hz, Amp = 4X Harmonic at 49 Hz, Amp = 4X Hz 0.10 Gasline/Control Hz Gasline/Control 12.2 G/G G/G Frequency (Hz) 78 Hz line/control 4.52 G/G B. POTENTIAL PROBLEMS: Primary resonant frequency has an amplification that may result in fatigue-related issues Harmonics are amplified at relatively high levels Damping is minimal Vibration attenuation doesn t occur below 60 Hz 25

26 Ratio (G/G) Diagnosing Frequency Response (cont d.) EXAMPLE #2 cont d ) Chamber Gasline 9) Chamber G 1.00 C. Likely Results: Scuffing Hz 0.10 Gasline/Control Hz Gasline/Control Fatigue-related damage to sensitive components 12.2 G/G G/G Frequency (Hz) 78 Hz line/control 4.52 G/G 26

27 Ratio (G/G) Diagnosing Frequency Response (cont d.) EXAMPLE #3 14) Chamber Mount 14) Chamber M Frequency 12 Hz 14) Chamber Mount/Control G/G Frequency (Hz) 27

28 Ratio (G/G) Diagnosing Frequency Response (cont d.) EXAMPLE #3 cont d 14) Chamber Mount 14) Chamber M A. WHAT DO WE KNOW? Primary Resonance = 12 Hz, Amp = 5X No significant secondary resonances No significant harmonics Frequency 12 Hz 14) Chamber Mount/Control G/G Frequency (Hz) B. POTENTIAL PROBLEMS: None! (good damping, reasonable natural frequency, etc.) C. LIKELY RESULTS: Successful shipment 28

29 Damping (Simplified ) 29

30 Multiple Degree of Freedom Systems MDOFs Damping vs. Destructive Interference 30

31 Multiple Degree of Freedom Systems (cont d.) MDOFs are the Real World Much more complex than a SDOF model Every engineer knows this! Learn from SDOF examples because they are simple and understandable. Then apply this to MDOFs in the real world. 31

32 Solution Case Study #1 Description: Large open-frame product with many distinct components fastened to it Monitored component was a heavy control box Package was a floating base system with 2 thick PE foam isolation 32

33 Solution Case Study #1 (cont d.) Bare product resonance search test data monitored on the control box. fn = 8 Hz, Amp = 23X 33

34 Solution Case Study #1 (cont d.) Test data: PACKAGE resonance search showed fn = 8 Hz, Amp = 55X!!! 34

35 Solution Case Study #1 (cont d.) A. What do we know: Product and package both resonate at 8 Hz 8 Hz is in the center of dense transportation envelope band (4-16 Hz) Both product and package are lightly damped at resonance B. Potential Problems: Product fatigue Excessive scuffing Loosening of fasteners Etc. 35

36 Solution Case Study #1 (cont d.) Conclusions from vibration testing: This product has virtually no chance of surviving transportation The product itself is perhaps un-shippable in its current format The package will actually destroy the product. 36

37 QUESTIONS 37

38 More Questions Later? Submit webinar and test questions to WESTPAK via our website at 38

39 Next Webinar Topic Sample Size Rationale For Medical Device Package Validation Presenters: Andrew Bevil Engineering Services Manager WESTPAK Steven Walfish, Ph.D. President Statistical Outsourcing Services Date: Wednesday, December 13, 2017 Register at: 39

40 About WESTPAK Two Locations San Jose Laboratory San Diego Laboratory 83 Great Oaks Boulevard Roselle Street San Jose, CA San Diego, CA Contact Us 40