Geri Dön

Çift ton çok frekanslı işaretlerin üretilmesi ve çift ton çok frekanslı işaret alıcısındaki sayısal filtrelerin tasarımı

Dual tone multifrequency (DTMF) signal generation and realization of digital filters in the DTMF receiver

  1. Tez No: 14185
  2. Yazar: ERSOY CEYLAN
  3. Danışmanlar: PROF.DR. AHMET DERVİŞOĞLU
  4. Tez Türü: Yüksek Lisans
  5. Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
  6. Anahtar Kelimeler: Belirtilmemiş.
  7. Yıl: 1990
  8. Dil: Türkçe
  9. Üniversite: İstanbul Teknik Üniversitesi
  10. Enstitü: Fen Bilimleri Enstitüsü
  11. Ana Bilim Dalı: Belirtilmemiş.
  12. Bilim Dalı: Belirtilmemiş.
  13. Sayfa Sayısı: 134

Özet

ÖZET Bu tezde, sayısal osilatörlerin Çift Ton Çok Frekans, (Dual Tone Multifrequency: DTMF) işaretlerinin üretiminin bir uygulaması ve sayısal filtreler İle DTMF işaretlerinin belirlenmesinin uygulaması verilmiştir. Bu amaçla, sayısal yöntemle istenen frekansta işaret üretimi incelenmiş ve sayısal filtrelerin analog filtrelerden analog -sayısal dönüşümünü gerçekleyen dönüşüm teknikleri ile elde edilmeleri incelenmiştir. DTMF işaretlerini üreten ve bu işaretlerin tanınmasında kullanılan, sayısal filtreleri gerçekleyen, TMS32010 işaret işleyicisine ilişkin birleştirici dilde programlar geliştirilmiştir. Temel DTMF alıcısında gerekli olan A tipi kod, lineer kod dönüşümlerini sağlayan, kırpıcı elemanını gerçekleyen, filtre çıkışlarını inceleyen, karar verme ve zamanlama kontrolünü sağlayan programlar verilmiş, bu yapıların çalışma prensipleri açıklanmıştır. Ayrıca sonlu uzunluklu kelime ile çalışmaktan ileri gelen hatalar incelenmiştir. Cv3

Özet (Çeviri)

DUAL TONE MULTI FREQUENCY CDTMF3 SIGNAL GENERATION AND REALIZATION OF DIGITAL FILTERS IN THE DTMF RECEIVER SUMMARY In this thesis a generator which generates dual tone multi frequencies (DTMF) and a receiver which perceives dual tone multi frequencies are realized. Two types of telephone systems are being used. The first one of these is pulse system telephones. In this system when the subscriber presses the button or dials the number» the telephone system opens-closes the telephone line as many times as the number dialed. The second type of telephones are DTMF telephones. When subscriber presses the button of a DTMF telephone, a signal with two different frequencies is generated on the line. The DTMF frequencies can be divided into two groups. These groups are known as LOW group and HIGH group frequencies. LOW group frequencies are 697 Hz, 770 Hz, 852 Hz, and 941 Hz. HIGH group frequencies are 1209 Hz, 1336 Hz, 1447 Hz, and 1633 Hz. When any button on the telephone is pressed, a dual frequency with this frequency components is given on the line. A digital -signal processing CDSPO pP can handle Touchtone CDTMF> dialing and decoding over telephone lines. If a computer system already has a DSP fjP and A/D and D/A converters in place, then the system can decode DTMF signals and any Touchtone telephone can serve as a data entry terminal or a remote-control console. The only cost for these DTMF enhancements is additional program space in the juP's ROM. We can generate the DTMF signals in two ways. In the first method, we can realize a DTMF tone generator which consists of a pair of programmable, second order harmonic oscillators. The sample-generation rate of the oscillators determines the total harmonic distortion of the output. Cvi3The higher the sampling rate» the more nearly exact the signal will be. The second method consists of a fast direct lookup table and an enhancement of this linear interpolation aproaches. In the first lookup table method» the sin values for N angles which are uniformly spaced around the unit circle are stored in a table. A sin wave is generated by stepping through the table at a constant rate» wrapping around at the end of the table whenever 360 is exceeded. This method generates the sequence: SCModCkxDELTA,ND3 for k=l, 2, 3,... where C a, b3 -remainder of the division a/b when this quotient is computed as an integer. The frequency, f, of the sine wave depends on two factors: Cİ3 The time interval between successive samples, t C23 The step size, DELTA f is given by the equation DELTA ixW tHz3 where t is expressed in seconds. To satisfy the Nyquist criterion there must be at least two samples generated in each sinusoid period. This requires that DELTA < NX3 There are two sources of error in the table lookup algorithm which cause harmonic distortion: C1D Quantization error is introduced by representing the sine table values by 16-bit numbers. C25 Larger errors are introduced when points between table entries are sampled. This occurs when DELTA is not an integer. When DELTA is an integer, quantization is the only error source and total harmonic distortion is extremely small regardless of table size. CviiDLinear interpolation method uses the values of two consecutive table entries» as the end points of a line segment. Sample points for parameter values falling between table entires assume values on the line segment between the points. This algorithm is based on the linear approximation SinC360CI+DD/N> £ SinC360xI/KD+DxCSinC360xCI+13/ND -SinC360xI/hD3 £ StIJ+DxCSCI+13-SCm where N is the sine table length I is an integer such that 0x is the normalized input signal C between -1 and 13 H is the compress! on +parameter SgnCxD is the sign C-O of x A-Law companding is defined by the equation FCxD' { for 0 filter is M -k E b z“r. YCz3 k=0 * HCz:> Kzb R k=l where HCz3,YCz3, and XCzD are the z transform of hCtO,yCxO and xCnZ>, respectively. Three different network structures often used to realize HCz3 are the direct form» the cascade form, and the parallel form. Direct -Form IIR Filter: For convenience, it is assumed that M=N. In direct form, coefficients of the network can be obtained directly from the difference equation describing the network. Direct form II has the minimum number of delays. It requires the minimum number of storage registers for computation. This structure is advantageous for minimizing the amount of data memory used in the implementation of IIR filters. CixDCascade-Form IIR Filter: The z transform of the unit -sample response of an IIR filter may also be written in the equivalent form, HCz:>rs n =î =g- k=l l-alkz -c^z where the filter is realized as a series of biquads. Therefore, this realization is referred to as the cascade form. The difference equation for cascade section i can be written as dA C nD =yA _± C rO +«. ^ d± C n-1 D +0^ d. C n-23 yt C tO =fi0± d± C nD +ft± ^C n-1 D +ft^ d± C n-23 where i=l» 2,...,N/2 y._1CnD= input to section i d.CrO= value at a particular delay node in section i yCn3= output of section i y_Cn3=xCnD «sample input to the filter y.^=yCn3= output of the filter Parallel -Form IIR filter: In this case, HCzD is written as M-N N^2 y Tik2”1 HCzD=E c z *+ E - - Et =g- k=0 k=l l-«lkz ^o^z 2 If M=0 Digital filters are designed with the assumption that the filter will be implemented on an infinite precision device. C3ÖSince all processors are of finite precision» it is necessary to approximate the ideal filter coefficients. This approximation introduces coefficient quantization error. For narrowband II R filters with poles close to the unit circle, longer word lengths may be required. The worst effect of coefficient quantization is instability resulting from poles being moved outside the unit circle. The effect of coefficient quantization is highly dependent on structure of the filter and the wordlength of the implementation hardware. The cascade and parallel forms implement each pair of complex conjugate poles separately. As a result» the coefficient quantization effect for each pair of complex conjugate poles is independent of the other pairs of complex conjugate poles. This is generally not true for direct form filters. Therefore the cascade and parallel forms of IIR filters are more commonly used than the direct form. In this study, as signal processor» TMS32010 chip was utilized. The reason for utilizing this signal processor is that the rate for instruction execution is quite high. When TMS32010 signal processor is operated at 20 MHz» the instruction execution time is 200 nano seconds. In the subject of digital signal processing» two points are quite important. The first one of these is the rate problem and the second one is having sufficient memory area. The program developed in order to generate DTMF signals is divided into two sections. The first section contains the signal generating programs and the second companding programs. Also to receive DTMF signals is divided into two groups of programs. The control programs and the filter programs. The digital filter programs are locked to certain frequencies; low, high, and band pass filters. It is possible to separate the digital filter programs in the DTMF receiver into four blocks; input filter, low pass filter block, high pass filter block, and band pass filter block. The input filter is a high pass filter which does not pass 50 Hz and 450 Hz special frequencies. Cxi>The cut off frequency is S50 Hz. The second block filter is a high pass filter block which passes high group frequencies. The cut off frequency is llOO Hz. The higher the degree of filter blocks, the better the precision of filter is. The time however» required for the execution of filter programs is increased. The microprocessor may not have sufficient time for processing the signals. For this reason» the filter degrees which are not too high but sufficiently suppresses the undesired frequencies are preferred. Moreover, in case filter degrees are high, much more addition and multiplication operations are required. This fact may cause overflows in the program, and misleads the filter outputs. Also The higher the degrees of filter, the larger memory area will be required. So, it is intended not to keep the degrees of the filters too high. The third filter block is a low pass filter block which passes LOW group frequencies. The cut off frequency is 1000 Hz. The fourth filter block is composed of band pass filters. Each of these filters corresponds to one DTMF frequency. The digital filters with known cut off frequencies are first calculated in the s domain. Then these filters are transformed into the z domain by using a transformation relation. Using a bilinear transformation method causes some disadvantages. As the cut off frequencies get closer to the sampling frequency, the transformation non-linearity increases. For this reason the cut off frequency is first established for the digital filter. CxiiD

Benzer Tezler

  1. Tez No

    142597

    Yeni Cami'nin akustik açıdan performans değerlendirmesi

    Evaluation of the acoustical performance of the New Mosque

    EVREN YILDIRIM

    Yüksek Lisans

    Türkçe

    Türkçe

    2003

    Mimarlıkİstanbul Teknik Üniversitesi

    Mimarlık Ana Bilim Dalı

    PROF. DR. SEVTAP YILMAZ DEMİRKALE

  2. Tez No

    421100

    Kesirli mertebe tek kutuplu sistem modeli için tam sayı mertebe PID kontrolör tasarımı

    Design of integer order PID controller for fractional order single pole system model

    ERHAN YUMUK

    Yüksek Lisans

    Türkçe

    Türkçe

    2015

    Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik Üniversitesi

    Kontrol ve Otomasyon Mühendisliği Ana Bilim Dalı

    PROF. DR. MÜJDE GÜZELKAYA

  3. Tez No

    633581

    Design and finite element analysis of 100 ton double girder overhead crane

    100 ton kapasiteli çift köprülü köprü tipi krenin tasarımı ve sonlu elemanlar yöntemiyle yapısal analizi

    GÜRKAN TAŞDEMİR

    Yüksek Lisans

    İngilizce

    İngilizce

    2020

    Makine Mühendisliğiİstanbul Teknik Üniversitesi

    Makine Mühendisliği Ana Bilim Dalı

    DOÇ. DR. SERPİL KURT HABİBOĞLU

  4. Tez No

    771365

    Kars yöresine ait ata buğdayı Kavılca ve modern ekmeklik buğdaydan hazırlanan un ve ekmeğin fonksiyonel özelliklerinin ve sağlık etkilerinin incelenmesi

    Investigation of the functional properties and health effects of flour and bread prepared from Kavılca wheat grown in Kars and modern bread wheat

    ELİF FEYZA AYDAR

    Yüksek Lisans

    Türkçe

    Türkçe

    2022

    Gıda Mühendisliğiİstanbul Teknik Üniversitesi

    Gıda Mühendisliği Ana Bilim Dalı

    PROF. DR. BERAAT ÖZÇELİK

  5. Tez No

    342908

    Performance assessment of a binary cycle geothermal power plant

    Çift akışkanlı bir jeotermal güç santralının performans değerlendirmesi

    MURAT KARADAŞ

    Yüksek Lisans

    İngilizce

    İngilizce

    2013

    Enerjiİzmir Yüksek Teknoloji Enstitüsü

    Enerji Mühendisliği Ana Bilim Dalı

    PROF. DR. HÜSEYİN MURAT ÇELİK

    PROF. DR. GÜLDEN GÖKÇEN AKKURT

    PROF. DR. MACİT TOKSOY

Geri Dön