Sd Card

description14 papers

group1 follower

lightbulbAbout this topic

An SD card (Secure Digital card) is a non-volatile memory card used for storing digital data in portable devices. It is widely utilized in cameras, smartphones, and tablets, offering various storage capacities and speed classes to accommodate different data transfer needs.

lightbulbAbout this topic

Key research themes

1. How do flash memory-specific algorithms and data structures address the unique challenges of SD card storage?

Flash memories used in SD cards present distinct constraints — notably the limited number of erasures per block and the requirement to erase large blocks instead of individual bits. This theme focuses on how specialized algorithms, data structures, and controller architectures are designed to handle wear leveling, reduce erasures, and manage flash-specific storage operations to optimize performance and durability of SD cards.

Algorithms and Data Structures for Flash Memories

by Kumar Amit

2019

Key finding: The paper identifies the inherent limitations of flash memories — like block-level erasures and finite erase cycles — and surveys specialized data structures that support efficient, not-in-place data updates, wear leveling,... Read more

articleView Paper downloadDownload

A NAND Flash Memory Controller for SD/MMC Flash Memory Card

by Lan-Rong Dung

2024, IEEE Transactions on Magnetics

Key finding: This work designs and implements a NAND flash controller with a t-error-correcting BCH ECC tailored for NAND flash bit error patterns. It introduces a code-banking mechanism and built-in defect management and wear-leveling... Read more

articleView Paper downloadDownload

Refactored Design of I/O Architecture for Flash Storage

by Sungjin Lee and

2015, IEEE Computer Architecture Letters

Key finding: By removing redundant flash management layers—the flash translation layer (FTL) and traditional file system—this paper proposes a refactored I/O architecture where the file system directly manages NAND flash storage,... Read more

articleView Paper downloadDownload

keyboard_arrow_downShow more

2. How can SD cards be integrated with embedded systems for specialized applications combining storage and device functionalities?

This theme explores deploying SD cards within embedded systems such as DSP-based CNC machines and minimal hardware voice playback devices. It focuses on hardware-software integration strategies, device driver implementation, and application-optimized use of SD cards to handle large data files or storage of compressed audio with minimal decoding hardware.

Designing the SD Card Device Driver for File Management of DSP Based CNC Machine

by Sunil Morade

2022

Key finding: This paper develops a cost-effective embedded system solution that uses the SD card interfaced via SPI on a TI DSP Launchpad to expand memory for storing large CNC toolpath (.NC/.CNC) files. The dual-core microcontroller... Read more

articleView Paper downloadDownload

Embedded Voice Reconstruction with Minimal Hardware using CVSD Modulation

by Abdurrahman Özlem

2022

Key finding: The study proposes a low-cost embedded voice playback system using CVSD modulation on an 8-bit microcontroller that stores compressed voice data on an SD memory card. This approach eliminates expensive dedicated decoders and... Read more

articleView Paper downloadDownload

3. What types and risks of residual sensitive data remain on second-hand SD cards, and what forensic approaches analyze this?

This theme deals with digital forensic investigations focusing on residual or remnant data on used SD cards. It encompasses techniques for data recovery, the sensitivity of recovered data, implications for data privacy and security, and the importance of proper data sanitization before sale or reuse.

A 2011 investigation into remnant data on second hand memory cards sold in Australia

by Krishnun Sansurooah

2013

Key finding: Through acquisition and forensic analysis of 119 second-hand memory cards, this study found that many cards still contained highly sensitive personal and confidential data, despite being sold as used. It highlights end-users’... Read more

articleView Paper downloadDownload

Analysis of Digital Information in Storage Devices Using Supervised and Unsupervised Natural Language Processing Techniques

by Lucas Hernández

2024, Future Internet

Key finding: This work applies machine learning and NLP techniques to automate forensic content analysis on digital storage devices including SD cards. It improves evidentiary value extraction by uncovering relationships between document... Read more

articleView Paper downloadDownload

All papers in Sd Card

Sign Language Translator Using Hand Gloves

by Harshada Rahul Patil

2024

This paper presents a prototype Indian Sign Language (ISL) translation system. The sign language glove seems to be a very useful tool to aid in communication with the deaf. Our implementation of the glove is going to deal with the 26... more

Every Sign Language isa well-structured body gesture; every gesture has meaning assigned to it. Sign Language helps with the communication for deaf people [3]. The block diagram of Sign Language Translator Usin; Hand Gloves is as shown below in Fig 1.

Fig 2. Manual Alphabets of Indian Sign Language (twoemu.com)

Fig 3. Unidirectional Flex Sensor The Flex sensors (Fig. 3) are bend sensors that changes its resistance depending on the amount of bend on the sensor. They convert the change in bend to electrical resistance; the more the bend, the more the resistance value increase. They are usually in the form of a thin strip from 1”’ to 5” long _ that vary in resistance; they could be made ina unidirectional or bidirectional form [1].

T he impedance buffer in the Basic Flex Sensor Circuit is a single sided Operational Amplifier, used with these sensors because the low bias current of the Op-Amp reduces error due to source impedance of the flex sensor as voltage divider (Fig. 5). Suggested Op-Amps are the LM358 or LM324 [1].

Fig 6. Basic Flex Sensor Circuit Output for different values of flex sensor. We have implemented this basic flex sensor circuit in Orcad software for different values of R1 as 10kQ, 25kQ, 40kQ & obtained the output voltage VO which can be represented graphically as shown in Fig 6.

descriptionView Paper arrow_downwardDownload

Built-in Electric Car Diagnostic Interface

by Zdeněk Fajkus

2023

Import 11/07/2012Cílem této diplomové práce je navrhnout řešení pro diagnostické rozhraní elektromobilu a realizovat diagnostický modul. Vzhledem k povaze zařízení se jedná o práci, ve které hraje velkou roli komunikace a množství... more

descriptionView Paper arrow_downwardDownload

INTERNATIONAL JOURNA L OF ENGINEERING SCI ENCES & RESEARCH TECHNOLOGY Accelerometer Based Gesture Driven Embedded System for Differently Abled

by ganesh more

2023

This document proposes an idea of an automated system based on accelerometer which can be very helpful for the differently abled patients. The limited num ber of pre

descriptionView Paper arrow_downwardDownload

Data Transfer without using Internet or Bluetooth

by Vineeta Soni

2023

This paper is about sending Data without using Internet or Bluetooth because these days Smartphones acquiring the whole world via Android os where one uses apps to fulfill their needs. But for sending SMS or Data either they will have to... more

descriptionView Paper arrow_downwardDownload

Embedded Voice Reconstruction with Minimal Hardware using CVSD Modulation

by Abdurrahman Özlem

2022

In embedded systems, the playback of pre-recorded voice data requires special hardware and/or dedicated decoding software. The introduction of high capacity flash memories makes it possible to use higher bit rates and very long playback... more

Figure 1 — Typical Voice Playback Section A basic embedded voice circuitry consists of the data storage media (preferably SD card), the decoder and the digital-to-analog converter (DAC), as shown in Figure 1. Data stored in the non-volatile memory is normally encoded (compressed) in order to save space. Audio compression techniques are almost always lossy; i.e. the reconstructed audio is not exactly the original. The loss in the quality of the reproduced signal depends on the algorithm

architecture [1]. The frame clock is generally 64 x bit clock; by assuming that the bit clock is at 44100 Hz, the frame clock will be at 2.8224 MHz. A synchronous data transfer at those speeds is probably beyond the capability of a low cost microcontroller and must hence be performed by a separate uP/DSP.

DM, in contrast to other coding techniques, uses 1-bit (2- level) quantization. The data can directly be transmitted over a communication channel without any need to shift the digital N-bit data. It is also proven to be much more tolerant to high Bit Error Rates (BER) caused by channel noise. Because of these features, it has been ideal for wired or wireless telecommunication systems.

DM (also known as Linear DM or LDM) uses a quantization with a fixed step size. A fixed step size has two drawbacks; it makes high granular noise, inducing low Signal-to-Noise Ratio (SNR) with small input signals, where step size is too large (Figure 4); or it causes slope overload on large input signals, indicating too small step size (Figure 5). Therefore, it has very small dynamic range for optimum performance. Figure 4 — Granular Noise

DM with a variable step size is referred as Adaptive DM (ADM). Being first proposed in 1970 [12], CVSD is a variant of ADM. In addition to point-to-point communication, CVSD is commonly used in digital voice recording/messaging and audio delay lines, as well as in tactical secure (encrypted) communications (refer to MIL-STD-188-113 at 16/32 kbps and Federal Standard 1023 at 12 kbps) [4].

For this adaptation, a typical CVSD encoder adds some extra blocks before the integrator (Figure 6). The input of the integrator (consecutively the step size) is no more a constant level (“O” or “1”), but now altered by a Pulse Amplitude Modulator (PAM), which is driven by a Syllabic Filter according to an overload algorithm. The input pass-band filter rejects any non-voice component, to increase SNR of the system. It generally accepts 300 Hz ~ 3 kHz for low data-rate implementations. At high sample rates, the band of the filter can be widened or the filter can even be omitted completely.

characteristics of the overload generator, syllabic filter, PAM and the integrator must be identical to produce the correct signal. So again, the analog signal at the output of the reconstruction integrator of the decoder is theoretically the same as the signal at the reference input of the encoder comparator. The low-pass filter (LPF) at the integrator output will eliminate most of the quantizing noise, especially if the clock rate of the bit stream is an octave or more above the bandwidth of the input signal [3]. The LPF is essential for bit rates below the audible range, and generally, the lower the bit rate, the better the filter must be.

The 3-bit shift register monitors the digital output for an overload, i.e. it fills with all 1’s or all 0’s. This overload condition is also called coincidence. When overload occurs, the gain or slope of the integrator (step size) is meant to be too small. The overload signal charges a LPF, called the syllabic filter, which controls the gain of the reconstruction integrator via the PAM. The higher the frequency of coincidence, the greater will be the ramp amplitude. This is basically a companding (compressing/expanding) algorithm, which significantly increases the dynamic range, compared to fixed-step size DM (Figure 8). A typical compression ratio (max/min step size) is 16:1 [5]. Figure & — Signal-to—Noise Performance

The reconstruction integrator (also called principal integrator) is of the exponential (leaky) type to reduce the effects of digital errors via the communication channel. Figure 9 demonstrates the reconstructed output that tracks the input.

Four-bit companding has proven to be most effective for clock frequencies (f,) greater than 32 kbps [4]. With 4-bit algorithm, 8 clock periods will be necessary for two coincidences, one for the raising and the second for the falling side of a sine wave. This will produce a signal with a duty cycle of 25% at the syllabic filter input. So the maximum input frequency the system can handle will be equal to f. / & for large signals. For very low input levels, the maximum bandwidth will extend up to f,/6 (Figure 10). Figure 10 — Frequency Response

Figure 11 — SNR Improvements The second measure is to insert a logarithmic amplifier between the syllabic filter and PAM. This amplifies the syllabic filter output logarithmically, enabling big step size changes for small modulation changes, which highly widens the companding capability of the system. The modulation can then be kept within a smaller range around 8.5%, by vastly increasing the usable dynamic range. The resulting SNR curves are shown in Figure 11 [7]. Curve c shows the “classic” CVSD performance, curve b with the incorporation of the double-pole integrator, whereas curve a additionally implements the logarithmic compander. That way, an SNR of 30 dB can be achieved at 32 kbps over a dynamic range of 45 dB.

The implemented decoding hardware is depicted in Figure 12. Overload generator is omitted here since it can be carried out in software. Note that the output LPF is also removed; since the clock rate is well beyond the audible range, the bandwidth of the output amplifier will sufficiently filter the inaudible 96 kHz component.

microcontroller. If the port of the microcontroller is open- drain, a pull-up resistor (4.7kQ) will be necessary. In order to refrain from a negative supply, the op-amp is fed only from +12V; in this case +5V becomes the analog ground. MLA is hence in inverted logic. The op-amp is in voltage-follower configuration to feed the next section. The transfer equations are presented below. Note that all voltages are with respect to +5V and thus the output will be negative.

The amplifier (Figure 15) ensures that the syllabic filter output remains near the desired modulation level of 8.5% throughout the practical input dynamic range. The voltage- swing at the line MLA is 5 V; hence the diode center voltage will be 0.43 V, at a forward current of 22 uA. For a dynamic range of 52 dB (400-fold change in the diode current), the diode voltage varies only about + 0.15 V, corresponding of a modulation change of about + 3%. In the encoder, the feedback resistor is set to 3.3kQ such that the peak output at maximum (11.5%) modulation will be 30% (1.5 V). The smaller feedback resistance (220Q) in the decoder decreases the peak output level to a more suitable value (0.1 V).

PAM is a +1 multiplier (Figure 16). It either outputs the positive or the negative of the input level, in accordance with the Non-Return-to-Zero digital input signal (NRZ) coming from the microcontroller. BF245C is a JFET which acts as a bipolar switch. When closed, the op-amp behaves as inverting, otherwise as non-inverting. The internal on- resistance of the JFET is around 100Q, so the off-resistance is set externally (by the parallel resistor) to 10kQ such that the geometric mean equals to the input resistance (1kQ). This causes a small decrease in the modulator gain, together with some non-linearity distortion especially at higher input levels.

The circuit in Figure 18 is a complete CVSD decoder which combines the syllabic filter, logarithmic amplifier, pulse amplitude modulator, reconstruction integrator and the power amplifier. It needs a single 5V supply.

Figure 19 — Gain versus DC volume control

Figure 21 shows the voltage on the integration capacitor (100nF) parallel to the pot, which converts the digital input NRZ into an analog signal. This is in fact the compressed form of the original signal with a very limited dynamic range, similar to an AGC-applied voice signal.

Figure 20 — Original Input / Decoded Output Signal

Figure 23 — Logarithmic Modulation Signal

In SPI mode, the data direction on the signal lines is fixed and the data will be transferred in byte oriented serial communication [11]. The command frame from host to card is a fixed-length packet as shown in Figure 25. After a command frame has been sent to the card, a response to the command will be sent back from the card. Because the data transfer is driven by the serial clock generated by host controller, the host controller must continue to read data, send a OxFF and get received byte, until a valid response is detected. The response should be sent back within the command response time (NCR). The specification contains a rich command set, of which the following ones are necessary for our application:

Figure 26a — Decoding Software Flowchart for SD (Tightly Interlaced)

This loosely-interlaced (1:7) arrangement will save 43% memory as compared to the previous tightly-interlaced (1:1) method. Figure 26b exhibits the distinct part in the decoding algorithm. Appendix 4 is the sample decoder software fot 89S8253.

Figure 27 — Embedded Playback Software Flowchart

This is the CVSD encoder software, which simulates the codec hardware depicted in Section 3. The valid input is an uncompressed audio file with WAV extension. So the existing audio file (generally in MP3 format sampled at 16 ~ 44.1 kHz) must first be converted to PCM format by an additional audio utility (such as Lame, Audacity or Goldwave). WAV2SES will re-sample the input data at the encoding clock rate (e.g. 96 kHz). No input filter has been implemented; the audio can be filtered through the auxiliary utility whenever needed. The output of WAV2SES is the aforementioned bit-stream, packed into a file with the extension SES (“voice” in Turkish). Syllabic filter time constant, reconstruction filter’s first and second pole time constants, logarithmic amplifier parameters, PAM gain, overload algorithm bits and the interlacing mode can be selected freely. It can also normalize the input audio signal. WAV2SES uses the formulae given in Section 4. The Overload Algorithm is implemented as per Figure 28. Bits represents the shift-register size and the recommended value is 4.

This way, the slope closely tracks the target step-size. In practice, tl he target vo tage should be about 4 higher than that of the envelope detector to minimize slope overload. This makes the slope coe increase t the slope he granular overload. T depends on the input seems to fficient equal to 133%. Larger values noise, whereas smaller values enlarge he optimum value for the buffer length signal; for azans, a number of near 800 be appropriate, which corresponds to a look-ahead time of 8 ms @ 96 kbps; for Quran recitations in contrast, a length of around 500 proves to be more suitable (5 ms @ 96 kbps). The target step-size decay time is adjusted by the software as to be same with the look-ahead time. The prediction equations are presented below:

be in discrete steps, but rather be curved through the 2" order integrator. Figure 30 — 96 kbps, OdBm, unfiltered

Figure 33 — 96 kbps, -40dBm, filtered Figure 32 — 96 kbps, -40dBm, unfiltered

We have additionally tested the system with a full- amplitude 10 kHz wave to verify the frequency response (compare Figure 10). The simulation (Figures 38-39) shows that the codec can successfully reproduce a 0 dBm signal at 10 kHz, where the slope (step-size) is 10-fold compared to | kHz. The encoder output of the logarithmic amplifier in Figure 15 increases now to 4.7 volts to generate this much steeper ramp. Figure 38 — 96 kbps, OdBm, unfiltered with 10 kHz input,

The system also behaves well at low frequencies, up to DC. Figures 41 and 42 show the full-amplitude DC-offset response of the codec; filtered and unfiltered, respectively. Here, the input is supplied with +5 volts with respect to the analog ground (+5V). Figure 41 — 96 kbps, unfiltered, with +5 V DC input

The notched appearance of the output wave should not mislead the reader; since the input is not a_ physical microphone but a digitally sampled audio-file instead; our original input is not a perfect sine wave either. The highest sampling rate of an audio file is generally 44.1 kHz (CD quality). A 10 kHz pure sine wave sampled at 44.1 kHz is depicted in Figure 40 for comparison.

Figure 39 — 96 kbps, OdBm, filtered, 10 kHz input

The index table will occupy the first 100 sectors. Eact sector of 512 bytes holds info for 64 records as below: The information for each record consists of the start address and length (in terms of sectors). Both are 32-bit entries.

The SD card is partitioned by the manufacturer, in order to be recognized as a logical drive by a computer operating system (OS). This is accomplished by writing a Master Boot Record (MBR) to the very beginning (sector 0) of the memory space (physical address). MBR contains the partition table, which holds the partition entries for each logical drive. Each partition entry possesses the 4-byte Logical Block Address (LBA) of the first absolute sector in the partition. When formatted by the OS, this first absolute sector of every ogical drive will contain the File Allocation Table (FAT). Considering the complex structure of any FAT, we prefer not to format the card by the OS; instead we write the audio data directly as image. A simple index table will be added to the top if multiple recordings exist, as follows:

[V.| / G check assures low frequency compensation (including DC). This predictive measure seems to be effective indeed to decrease the slope overload and thereby the noise level. The table below contains comparison data of the sample azan file (Figure 29) for different configurations: The selected predictive buffer length is 800 and the slope coefficient 1.33, as explained above. Analysis shows that the SNR increases by 4 dB for 1:1-interlaced coding and 2 dB for 1:7 interlacing. Maximum Overload Length and Distortion also become significantly reduced. Nevertheless, the results can differ from sample to sample.

Signal-to-noise ratios for various conditions are simulated through WAV2SES. Five different frequencies covering the whole audible spectrum have been examined. The SNR filter is of Moving Average (MAF) type and its stop-band frequency is adjusted to one-tenth of the sampling rate (9.6 kHz for 96 kbps). The first two columns display the results obtained by using the standard design values, for 1“ and 2" order integration, respectively. The latter columns show the optimized cases where the integration time constants are adjusted for each frequency.

descriptionView Paper arrow_downwardDownload

Data Transfer without using Internet or Bluetooth

by vineeta soni

2022

descriptionView Paper arrow_downwardDownload

INTERNATIONAL JOURNA L OF ENGINEERING SCI ENCES & RESEARCH TECHNOLOGY Accelerometer Based Gesture Driven Embedded System for Differently Abled

by Ricardo Amol

2022

This document proposes an idea of an automated system based on accelerometer which can be very helpful for the differently abled patients. The limited num ber of pre

descriptionView Paper arrow_downwardDownload

Data Transfer without using Internet or Bluetooth

by vineeta soni

2022

— This paper is about sending Data without using Internet or Bluetooth because these days Smartphones acquiring the whole world via Android os where one uses apps to fulfill their needs. But for sending SMS or Data either they will have... more

descriptionView Paper arrow_downwardDownload

Data Transfer without using Internet or Bluetooth

by vineeta soni

2022, ( IJSER.org ) - International Journal of Scientific & Engineering Research

descriptionView Paper arrow_downwardDownload

Utilizando módulos para cartões SD em conjunto com outros dispositivos SPI

by José Daniel Soares Bernardo

2021, Blog Eletrogate

Os módulos para cartões de memória SD utilizados em sistemas microcontrolados como Arduino, ESP32, PIC, etc., são necessários geralmente quando se deseja armazenar dados, como em um data logger (registrador de dados). O protocolo de... more

descriptionView Paper arrow_downwardDownload

Smart Dialing System Using GSM During Emergencies In Absence Of Cellular Networks

by Radhakrishna K

2020

The emergency situations which includes the road accidents and other perilous incidents like cardiac arrest, results in high risk if not treated well within Golden Hour. In emergency medicine, the term ‘Golden Hour’ refers to the time... more

descriptionView Paper arrow_downwardDownload

Advanced Android Application to Locate and Track Mobile Phones(AdAALT

by J4R - Journal for Research

2016, Journal 4 Research - J4R Journal

Smart phones play an important role in our day to day life. So misplacing or losing of mobile phones can have a great impact in our life. Hence it is necessary to develop an application in all smart phones to locate the lost phone or to... more

descriptionView Paper arrow_downwardDownload