TWI812276B - Method and system for testing the impact of noise on the performance of a hard-drive - Google Patents

Abstract

A method for testing the impact of noise on the performance of a hard-drive includes recording equipment recording the noise from a server to generate an original sound file, a speaker playing the original sound file, when the speaker plays the original sound file, the recording equipment records the noise from the speaker to generate a noise file, the processor generating a first sound pressure level spectrum according to the original sound file, the processor generating a second sound pressure level spectrum according to the noise file, the equalizer adjusting the second sound pressure level spectrum to match the first sound pressure level spectrum, the processor generating an updated noise file according to the adjusted second sound pressure level spectrum, the speaker playing the updated noise file to the hard-drive, and when the speaker plays the updated noise file, a reading rate measurement device measuring the data access rate of the hard-drive.

Description

Testing method and system for vibration and noise affecting hard disk performance

The present invention relates to a testing system and method for hard disk performance that is affected by vibration noise. In particular, it relates to a testing system and method for testing hard disk performance that is affected by vibration noise frequency and sound pressure.

Computer systems, including desktop computers, industrial computers and servers, generally use hard disks to access data. However, as the storage density of hard disks increases, the sensitivity of hard disks also increases. As the heat generated by the CPU becomes more and more obvious, the fan also needs to increase its speed to dissipate heat, resulting in more and more noise and vibration. The hard disk is often affected by vibration noise and reduces performance (for example, causing hard disk read and write problems). If the computer system's central processing unit operates at a faster speed, it will not be able to access data normally, which will affect the operation of the computer system. In the worst case, the hard disk will be damaged, which may cause important problems for customers. Significant loss of data. The existing method of testing the sound pressure sensitive frequency of a hard disk is to use single-frequency tones of different frequencies to play back the hard disk, measure the performance of the hard disk, and evaluate the sensitive frequency of the hard disk. However, it cannot reproduce the real sound pressure received by the hard disk in the server, and it lacks consideration of the impact of multi-frequency sound ranges. Because the source of the single-tone tone is a signal generator and lacks a feedback correction mechanism, the sound pressure adjustment of the single-tone tone is not accurate enough, and there will be errors in the measurement of hard disk performance.

The embodiment provides a testing method for the impact of vibration noise on hard disk performance, which is used in a testing system. The test system includes recording equipment, equalizer, processor, speakers, hard disk and read rate tester. The ways in which vibration noise affects hard disk performance include the recording device recording the vibration noise emitted by the server to generate the original audio source file, and the speaker playing the original audio source file. When the speaker plays the original audio source file, the recording device records the vibration noise emitted by the speaker to generate the vibration noise. source file, the processor generates the first sound pressure level spectrum (SPL) based on the original sound source file, the processor generates the second sound pressure level spectrum based on the vibrating noise source file, and the equalizer adjusts the second sound pressure level The accurate spectrum is consistent with the first sound pressure level spectrum, the processor generates an updated vibration noise source file based on the adjusted second sound pressure level spectrum, the speaker plays the updated vibration noise source file to the hard disk, and when the speaker plays the hard disk The read rate tester measures the data read rate of the hard drive while the disc is playing the updated vibration noise source file.

The embodiment further provides a testing system for hard disk performance affected by vibration noise, including a recording device, a speaker, a processor, an equalizer, a hard disk, and a read rate tester. The recording equipment is used to record the vibration noise emitted by the server to generate the original audio source file. The speaker is used to play the original sound source file. When the speaker plays the original sound source file, the recording device records the vibration noise emitted by the speaker to generate the vibration noise source file. The processor is used to generate a first sound pressure level spectrum based on the original sound source file, and generate a second sound pressure level spectrum based on the vibration noise source file. The equalizer is used to adjust the second sound pressure level spectrum to match the first sound pressure level spectrum, wherein the processor generates an updated vibration noise source file according to the adjusted second sound pressure level spectrum. The read rate tester is used to measure the data read rate of the hard drive when the speaker plays the updated vibration noise source file to the hard drive.

100: Hard drive performance testing system

10: Recording equipment

20: Speaker

30:Computer

32: Processor

34: Equalizer

36:Filter

40:Hard disk

50:Read rate tester

200,400:method

S202~S220, S502~S522: steps

1A and 1B are schematic diagrams of the hard disk performance testing system of the embodiment.

FIG. 2 is a flow chart of the operation method of the hard disk performance testing system of FIGS. 1A~1B.

3A and 3B are schematic diagrams of the first sound pressure level spectrum and the second sound pressure level spectrum.

4A and 4B are schematic diagrams of a hard disk performance testing system according to another embodiment.

FIG. 5 is a flow chart of the operating method of the hard disk performance testing system of FIGS. 4A to 4B.

FIG. 6 is another schematic diagram of the spectrum of the second sound pressure level.

Figure 7 is a schematic diagram of the linear relationship between the sound pressure and gain of the speakers in Figures 1A~1B and Figures 4A~4B.

1A and 1B are schematic diagrams of the hard disk performance testing system 100 of the embodiment. The test system 100 is used to test the hard disk performance affected by vibration and noise. The test system 100 includes a recording device 10, a speaker 20, a computer 30, a hard disk 40, a reading rate tester 50, a processor 32 and an equalizer 34. The processor 32 and the equalizer 34 may be independent devices, or in embodiments, the processor 32 and the equalizer 34 may be included within the computer 30 .

The equalizer 34 may be a component including a large number of resistors, inductors, and capacitor circuits. The sound signal is an analog signal, and the electric waves it generates are all alternating currents. Different frequencies of alternating current will produce different impedances (reactance) for different capacitors and inductors, so that the amount of conduction at each frequency will be different, thereby achieving The purpose of equalizing the pitch. In embodiments, the equalizer 34 may also be software used to adjust the digital audio signal. Digital sound signals are usually generated from analog sound signals through Fast Fourier Transform. In addition, the hard disk 40 in the embodiment can be fixed in a rack to facilitate testing.

The purpose of the embodiments of the present invention is to use the real vibration noise received by the hard disk in the server as the playback sound source, so as to better present and evaluate the actual impact of vibration noise on the performance of the hard disk. The detailed operation method is described below.

FIG. 2 is a flow chart of the operating method 200 of the hard disk performance testing system 100 of FIGS. 1A-1B. How-to 200 consists of the following steps:

S202: The recording device 10 records the vibration noise emitted by the server to generate the original audio source file.

S204: The speaker 20 plays the original audio source file.

S206: When the speaker 20 plays the original sound source file, the recording device 10 records the vibration noise emitted by the speaker 20 to generate the vibration noise source file.

S208: The processor 32 generates a first sound pressure level (SPL) spectrum based on the original sound source file.

S210: The processor 32 generates the second sound pressure level spectrum according to the vibration noise source file.

S212: The processor 32 compares the first sound pressure level spectrum and the second sound pressure level spectrum.

S214: The equalizer 34 adjusts the second sound pressure level spectrum to match the first sound pressure level spectrum. After the equalizer 34 adjusts the second sound pressure level spectrum, the processor 32 may compare the first sound pressure level spectrum with the second sound pressure level spectrum. If the second sound pressure level spectrum is not consistent with the first sound pressure level spectrum, the equalizer 34 can be used to adjust the second sound pressure level spectrum. In this way, the iterative method is used to play back the sound source and modify it so that its second sound pressure level spectrum matches the vibration noise actually felt by the hard disk 40 in the server.

S216: The processor 32 generates an updated vibration noise source file according to the adjusted second sound pressure level spectrum.

S218: The speaker 20 plays the updated vibration noise source file to the hard disk 40.

S220: When the speaker 20 plays the vibration noise source file to the hard disk 40, the read rate tester 50 measures the data read rate of the hard disk 20.

3A and 3B are schematic diagrams of the first sound pressure level spectrum and the second sound pressure level spectrum. The vertical axis in the figure is the sound pressure level (Sound Pressure Level), ranging from 15 to 85dB. The horizontal axis is frequency, and the usual measurement range is 100~20000Hz. It can be seen from FIG. 3A that the second sound pressure level frequency spectrum of the vibrating noise source file does not overlap with the first sound pressure level spectrum of the original sound source file. The equalizer 34 is used to adjust the sound pressure level of the second sound pressure level spectrum at each frequency, so that the second sound pressure level spectrum and the first sound pressure level spectrum can be perfectly matched. The quantities are consistent, as shown in Figure 3B. The processor 32 may generate an updated vibration noise source file according to the adjusted second sound pressure level spectrum. When the speaker 20 plays the updated vibration noise source file to the hard disk 40, the read rate tester 50 can measure the data read rate of the hard disk 20, thus more accurately simulating the actual vibration experienced by the hard disk 40 on the server. noise and its impact.

4A and 4B are schematic diagrams of a hard disk performance testing system 400 according to an embodiment of the present invention. The difference between the hard disk performance testing system 400 and the hard disk performance testing system 100 of FIGS. 1A-1B is that the hard disk performance testing system 400 may further include a filter 36 . Filter 36 may be a stand-alone device, or in embodiments, filter 36 may be included within computer 30 . The remaining component configurations are the same and will not be described here.

FIG. 5 is a flow chart of the operating method 500 of the hard disk performance testing system 400 of FIGS. 4A-4B. How-to 500 consists of the following steps:

S502: The recording device 10 records the vibration noise emitted by the server to generate the original audio source file.

S504: Speaker 20 plays the original audio source file.

S506: When the speaker 20 plays the original sound source file, the recording device 10 records the vibration noise emitted by the speaker 20 to generate the vibration noise source file.

S508: The processor 32 generates the first sound pressure level spectrum according to the original sound source file.

S510: The processor 32 generates the second sound pressure level spectrum according to the vibration noise source file.

S512: The processor 32 compares the first sound pressure level spectrum and the second sound pressure level spectrum.

S514: The equalizer 34 adjusts the second sound pressure level spectrum to match the first sound pressure level spectrum. After the equalizer 34 adjusts the second sound pressure level spectrum, the processor 32 may compare the first sound pressure level spectrum with the second sound pressure level spectrum. If the second sound pressure level spectrum is not consistent with the first sound pressure level spectrum, the equalizer 34 can be used to adjust the second sound pressure level spectrum. In this way, the iterative method is used to play back the sound source and modify it so that its second sound pressure level spectrum matches the vibration noise actually felt by the hard disk 40 in the server.

S516: The filter 36 filters the adjusted second sound pressure level spectrum to filter out higher sound pressure in at least one specific frequency band. The filter 36 is then used to filter the adjusted second sound pressure level spectrum to filter out the higher sound pressure in at least one specific frequency band, so that the sound pressure ratio of the second sound pressure level spectrum that is higher can be measured. Performance impact caused by hard drive 40.

S518: The processor 32 generates an updated vibration noise source file according to the filtered second sound pressure level spectrum.

S520: The speaker 20 plays the updated vibration noise source file to the hard disk 40.

S522: When the speaker 20 plays an updated vibration noise source file to the hard disk 40, the read rate tester 50 measures the data read rate of the hard disk 20.

The filter 36 is an electronic component composed of resistors, inductors and capacitors. The filter 36 only allows specific frequencies to pass and can block audio signals of other frequencies. In embodiments, the filter 36 may also be software for filtering digital audio signals. Digital sound signals are usually generated from analog sound signals through Fast Fourier Transform.

FIG. 6 is another schematic diagram of the spectrum of the second sound pressure level. The vertical axis in the figure is the sound pressure level. The actual vibration noise affected by the hard disk 40 in the server is usually 50~90dB. The horizontal axis is frequency, ranging from 100 to 3000Hz. In the picture, there are peaks around 100~300Hz, 800~1200Hz, 1600~2100Hz and 2800Hz, which are relatively high sound pressures. After using the filter 36 to filter out the peaks in these bands, the processor 32 can generate an updated vibration noise source file according to the filtered second sound pressure level spectrum. When the speaker 20 plays the updated vibration noise source file to the hard disk 40, the read rate tester 50 can measure the data read rate of the hard disk 20, so that the relatively high sound pressure at a specific frequency can be actually measured against the hard disk 40. Disc 40 performance.

Table 1

Table 1 is an example of the read rate performance of the hard disk 40 under different environments. The measured read rate of Hard Drive 40 in a server using Nidec fans is 47.6% of the original read rate (ie, factory specifications). In an environment where the speaker simulates Nidec fan vibration noise, the measured read rate of the hard disk 40 is 64.9% of the original read rate. In an environment where the speaker simulated Nidec fan vibration noise and the high-pitched pressure was filtered out, the measured read rate of the HDD 40 was 69.5% of the original read rate. Because there are other vibration noises and interferences in the server, there are more factors that affect the performance of the hard disk 40, causing the hard disk 40 read rate to drop even more. The AVC fan also presents the same situation, and the embodiment will not be described in detail here.

FIG. 7 is a schematic diagram of the linear relationship between the sound pressure and gain of the speaker 20 in FIGS. 1A to 1B and 4A to 4B. In the figure, the vertical axis is the sound pressure of the speaker 20 and the horizontal axis is the gain of the speaker 20 . In the gain range of -50dB to -15dB, the relationship between the sound pressure of the speaker 20 and the gain is linear, that is, as the gain increases, the sound pressure also increases linearly. When the gain exceeds -10dB, the speaker 20 enters a saturated state. Even if the gain is increased, the sound pressure cannot be increased. A gain exceeding 30dB will exceed the load of the speaker 20, causing its sound pressure to become smaller or even unable to produce sound. Through the linear relationship between the sound pressure and gain of the speaker 20, the most suitable sound pressure and gain of the speaker 20 used in the hard disk performance test system affected by vibration and noise can be found, thereby increasing the accuracy of the test system.

In this embodiment, the server can be used for artificial intelligence (English: Artificial Intelligence, referred to as AI) computing, edge computing (edge computing), and can also be used as a 5G server, cloud server or Internet of Vehicles server.

In summary, embodiments of the present invention provide a testing system and method for determining the impact of vibration noise on hard disk performance. It can more accurately reproduce the real sound pressure received by the hard disk in the server and consider the influence of multi-frequency sound range. The embodiment has a feedback correction mechanism that can accurately adjust the sound pressure, making the measurement of hard disk performance more accurate.

The above are only preferred embodiments of the present invention, and all equivalent changes and modifications made in accordance with the patentable scope of the present invention shall fall within the scope of the present invention.

200:Method

S202~S220: steps

Claims (6)

A testing method for the impact of vibration noise on hard disk performance, used in a test system. The test system includes a recording device, an equalizer, a processor, a speaker, a hard disk and a reading rate tester. The vibration noise The test method for the impact of noise on hard disk performance includes: the processor measures a linear relationship between a sound pressure of the speaker and a gain of the speaker; the processor generates a sound pressure effective range based on the linear relationship; the recording equipment records a The vibration noise emitted by the server is used to generate an original sound source file; the speaker only plays the original sound source file within the effective range of the sound pressure; when the speaker plays the original sound source file, the recording device records the vibration noise emitted by the speaker to Generate a vibrating noise source file; the processor generates a first sound pressure level spectrum based on the original sound source file; the processor generates a second sound pressure level spectrum based on the vibrating noise source file; the equalizer adjusts the The second sound pressure level spectrum is consistent with the first sound pressure level spectrum; the processor generates an updated vibration noise source file according to the adjusted second sound pressure level spectrum; the speaker plays the sound to the hard disk Update the vibration noise source file; and when the speaker plays the updated vibration noise source file to the hard disk, the read rate tester measures the data read rate of the hard disk. The test method for the impact of vibration and noise on hard disk performance as described in claim 1, wherein the test system further includes a filter, and the test method for the impact of vibration and noise on hard disk performance further includes the filter adjusting the adjusted second sound. The pressure level spectrum is filtered to filter out high sound pressure in at least one specific frequency band. The test method for the impact of vibration noise on hard disk performance as described in request 1, further including the processing The device compares the first sound pressure level spectrum and the second sound pressure level spectrum. A test system for the impact of vibration noise on hard disk performance, including: a hard disk; a recording device used to record the vibration noise emitted by a server to generate an original sound source file; a speaker used to play the original sound source file, wherein When the speaker plays the original sound source file, the recording device records the vibration noise emitted by the speaker to generate a vibration noise source file; a processor is used to generate a first sound pressure level spectrum based on the original sound source file, according to The vibrating noise source file generates a second sound pressure level spectrum, measures a linear relationship between a sound pressure of the speaker and a gain of the speaker, and generates a sound pressure effective range based on the linear relationship; an equalizer, For adjusting the second sound pressure level spectrum to be consistent with the first sound pressure level spectrum, wherein the processor generates an updated vibration noise source file according to the adjusted second sound pressure level spectrum; and a first reading A rate tester is used to measure the data reading rate of the hard disk when the speaker plays the updated vibration noise source file to the hard disk; the speaker only plays the original sound source file within the effective range of the sound pressure. . The test system for vibration noise affecting hard disk performance as described in claim 4 further includes a filter for filtering the adjusted second sound pressure level spectrum to filter out high-frequency components in at least one specific frequency band. sound pressure. As claimed in claim 4, the test system for hard disk performance affected by vibration noise, wherein the processor further includes comparing the first sound pressure level spectrum and the second sound pressure level spectrum.

Images