SA update to version v 6.2.3.3 : OFDM Module.

SA update to version v 6.2.3.3

Long enough time ago SA users SA had paid attention that the copy of the selected on the sonogram fragment visually looks shorter than the original. This interesting phenomenon, entirely linked with property of uncertainty of FFT transformation.

The matter is that transition from the time area into the frequency area, which provides FFT transformation, excludes possibility of exact definition of position of concrete frequency component by the time. FFT only "says" that the component is entering into the block, but where is the component located in the block is unknown.

It is the very old and standard problem. It has no exact decision, but we counted possible to lower a sharpness of this phenomenon. As the basic problems arise within the small blocks, when the resolution on time on

the sonograms is rather good.

This is realized in the current version.

We just make replacement of one problem by another. If in the old versions, copied by the sonogram fragment, physically had such duration which had been displayed at the moment of copying. That in the current version, the physical duration is longer than it is specified at copying.

We have considered possible to go on such exchange, for two reasons.

In the first, users confuses that the copied fragment visually looks shorter, and this situation demanded actions to regulate it.

In the second, the work on the sonogram, assumes visual perception, and it is better to see the copied fragment the way it has been selected, neglecting real accuracy. Even more so further it is necessary to work with the fragments in the area of the sonograms displaying.

The mode of a signal review has been added by the users offer.

The review is provided by FFT window with the size in 1024 counts. The primary goal of the mode it is to pass part of the signal with preamble or another problem segment (by having placed the start point in the necessary position), before the correlation triangle search will begin.

In the review mode, FFT window moves on the signal loaded in OFDM module, by the slider which is shown on the picture.

After key parameters of the signal are gotten, it is possible to define the detailed parameters of the preamble or other fragmen, by moving symbol-by-symbol backwards on the part passed earlier.

The detailed example on the real signal is described in the article OFDM analysis in SA version 6.2.3.3

Also the problems concerned users bug reports are removed.

Using SA to measure and correct sound card digitizer errors.

Using SA to measure and correct sound card digitizer errors.

This article had been written completely on the basis of letters sent by SA users.


All sound cards and A/D converters have some clock error. This is especially true if converters are commercial and cheap ones, like PC sound cards and similar. Professional and expensive converters exhibit a much better clock stability and jitter.


To correct this error, one must know nominal parameters of the signal under test. If these parameters are known, it is quite easy to correct digitizing clock error using SA.


The SA method of “Correction of BR” is quite good for this job, but perhaps using resampler as data input is a better procedure.


The correction factor will have to be measured for any digitizing speed and mode.
Bear in mind that for cheap cards, the speed can vary due to various factors, so if high precission measurements are required, a new calculus should be carried out.


For a good measurement, a quite big signal is needed.

This method is useful for PSK,FSK,MFSK and other modulations. To use it with OFDM, some more operations should be carried out using SA.


Ideally, an external signal like GPS 1000 Hz or a signal from a high end signal generator will provide the best results. Also, a radio timing signal should be good enough.


The best way to understand the subject is an example.
A well known signal as Stanag-4285 (sampled at 8000 sps) will be used to show the method.




We know 4285 has a nominal speed of 2400 sps. Since a frame is 256 symbols, the frame time must be 106,666666 mS. This is the value that should be obtained if using the VMW feature of SA when signal structure is perfectly vertical.





As we can see, the mesured value is 106,69241.


Correction factor =measured value/nominal value=106,69241/106,66666= 1,000241
Error PPM=(correction factor-1)* 1000000= (1,000241-1)*1000000= 241 PPM.


Real Digitizing speed= 8000*1,000241 = 8001,931
Real modulation speed (Br) = 2400/1,000241 = 2399,421739


Measured SA Br= 2399,41


Now, lets go to correct the signal using calculated BR in SA.

The signal parameters are almost perfect, so we can save it and after this procedure, we can be quite sure the new signal will be demodulated using any comercial demodulator.

Also, we know the correction factor for the used card in that speed.

As soon as parameters of a sound card in the record channel are received, then it is possible to realize precise measurements of replay( reproduction) channel for sound cards, which have separate record/reproduction channeles.

You can perform these measuremens simply by having connected an output of such card with an input, playingback a known/synthesized signal/file. Measurements should be realized by the technique described above. As the error of the channel of record is already known, it can be easily considered, and there will be only an error of the channel of reproduction/replay/playback.

PSK: CIS-3000

CIS-3000

PSK-8, Br-3000.

PSK

The record of this signal is here. Send by RadioKoteg.

Author: SergUA6

Band Width ~3400 Hz

Low Range ~600 Hz, in this example only!

Baud Rate 3000 Hz

n-Ary (PSK/MPSK) PSK-8

Carrier frequency ~2313 Hz, in this example only!

ACF Preamble ~ 85 ms

RX mode SSB

Sonograms

pic.1 Generall view

pic.2 General structure of the signal

Diagrams:

pic.3 8th harmonic

Pictures:

pic. 4 Phase constellation

pic.5 ACF preambles

CIS-3000 - With very high probability this signal has the Russian origin roots. More and more often there are signals which are leaving the limits of standard phone channel. This signal represents that kind of signals. The spectrum of this signal occupies 3400 Hertz. The maximum technical speed 3*3000 = 9000 bps.

SA update to version 6.2.3.0

SA Update to version 6.2.3.0.

Changes has been added in OFDM module operation.

Algorithm of search of a correlative triangle has been also changed. Current implementation of this algorithm successfully enough copes with the short signals.

Function of obtaining of exact value of clock frequency is updated. Now besides obtaining of the clock frequency, the task of obtaining of an optimal position of the signal by frequency is solving. It does automatically translates the triangle in the necessary positive polarity. All that demands some time, so no need to be nervous:-)

Some pauses (till 3-5 seconds) in operation of "Get Br" function are possible, The pauses depend on many factors:

from sampling rate of a signal

from its position on frequency

from the sizes LS and LG etc.

In generall, the program solves wide enough range of tasks in an automatic mode. From the user it is required only to situate the triangle in "a good" position and to call function "Get Br", and it is not important if there will be the positive triangle or negative, the program itself will translate the triangle into the necessary type, through search of optimal shift by frequency.

It is necessary to note that shift by frequency isn't precise because precision isn't necessary, while it is very close to optimal value, from the point of view of the program certainly.

All these changes and additions are preliminary preparation of the module for the decision of the primary task: obtaining LS and LG values of concrete signal, within the limits of our ideas, which were described in the article Parameters of OFDM signals.

Some minor problems had been found out by users and solved by us.

The more detailed information of operation with OFDM signals in version 6.2.3.0 will be described in separate article.

Good luck!

MFSK: WinMor

WinMor, mode 2*MFSK-4

Amateur's mode. Br - 46.875

Here is the record of the signal. Send by JenyaT.

Author: SergUA6

Band Width ~500 Hz

Baud Rate 46.875 Hz

Count of Carriers 4 in one group and 4 in second group

Step between carriers 46.875 in group, 93.75 between groups Hz

RX mode SSB

Sonograms:

pic.1 General view

pic.2 Detaltization. Two groups of MFSK-4

pic.3 Detalization of the groups. Space between the groups.

Diagrams:

pic. 4 Start

pic.5 Start detalization

Pictures:

pic.6 Graphic spectrum of the signal

WinMor, mode 2*MFSK-4 - it is the new amateure mode (atleast it is announced in 2010 year), which had been created by radio amateurs. The protocol has a lot of transmission formats. In the current example of the signal, two MFSK-4 channeles are used. The signal is bright enough, so there shouldn't be problems with it's identification.

SA update to version 6.2.2.8

SA update to version v 6.2.2.8

This is planned updating.

The new element in the general options is added.


Internal management of markers is reconsidered. Appearance and the size of the markers has been changed.

It is problematic enough to surely catch the markers by mouse with the thickness in 3 pixels, on the big monitors. Although it is required not so often, it may cause some inconveniences.

Correct independent work of two and more Phase Plane modules is realized. Some users actively use this possibility, and in old versions it worked not correctly.

We also prepeared the demo clip, which shows how to work with the bottom panel of tools. In this clip, we finish the review of the basic SA working window, and in The further we will countine to make demo clips about work with concrete modules.

All demo clips are recommended to be started on monitors with the permission not less then 1024х768 in mode Normal size 96 DPI.

The list of demo files of own SA format are located in special section of the site => Video examples on use SA.

Good Luck.

SATCOM Satellite, US INTEL - 260.41, UHF (B)PSK-2, Br - 260.41

SATCOM Satellite, US INTEL - 260.41, UHF

(B)PSK-2, Br - 260.41

Here is the signal. Send by RSCAN.

Author: SergUA6

Band Width ~520 Hz

Baud Rate 260.47 Hz in this example, exact value 260.41

n-Ary (PSK/MPSK) (B)PSK-2

ACF ~1900 ms preamble, ~368 ms data/message

RX mode SSB

Sonograms:

pic.1 General view

Diagrams:

pic.2 General structure of the signal

pic.3 Phase constellation

Illustrations:

pic.4 ACF

SATCOM Satellite, US INTEL - 260.41, UHF - is the very wide spread signal, which can be met in the channeles of satellite group SATCOM. This signal is encrypted.

SA update to version 6.2.2.5

SA update to version 6.2.2.5

Automatic synchronization of phase added in FSK-2 dem module.

The presence of two degrees of adjustment (in phase and frequency) for setting/selecting accurate clock for many presents certain difficulties. As miminum the phase synchronization can be achieved fairly quickly and qualitatively, in automatic mode.

This is implemented in the current update.


If the automatic phase is ON(enabled by default) setting the exact value of the clock frequency is greatly simplified. In addition, in most cases, automation will work better and error free.

Nevertheless, the possibility of a transition to a fully manual mode is preserved, sometimes it's necessary.

Automation mode helps to significantly over short and very short signal/fragments.

Algorithm for automatic phasing is not sensitive to the long series of zeros or ones at the beginning of the fragment, it usually is the problem seriously. This situation handled correctly, however, should not take very long series. Also not recommended to take when copying or selection, start part of signal with noise, this can completely disrupt the work of the algorithm.

Although algorithm was developed an for signals FSK-2, he copes with C4FM, which greatly facilitates the identification of this regime and etc.

In general, the automatic phase synchronization, when set properly of clock frequency. Guarantees a correct bitstream and greatly simplifies the module FSK-2 demodulator in SA.

Generally, issues guarantees the correctness of the bitstream is a separate issue, which many prefer to sidestep/not see. :-)

Good Luck!

SA update to version 6.2.2.4

SA update to version 6.2.2.4

Under a 64-bit OS (Vista, W7) are errors in the previous version.

The current version of SA 6.2.2.4 passed rigorous testing, all identified problems are eliminated. This version very stable and sustainable for both 32-bit OS, and 64-bit.

Good Luck.

Update SA to version 6.2.2.0/1

Update SA to version 6.2.2.0/1

In this version added the ability to copy the various fragments of the signal from WF module.

Copy fragments signal from the sonograms, can not operate sufficiently exactly. This is not a problems of SA, it is a fundamental limitation FFT.

The fact that the FFT, in general case , makes a transition from time domain into the frequency domain, and such a thing as time loses its meaning. Within the FFT block does not have time, and to say exactly where in that block is one or another component of the spectrum is impossible. This leads to the fact that selecting/separating out a segment of the signal at the sonograms, there is always uncertainty within the block FFT. Sometimes this quantity is large enough. Hundreds of milliseconds or few seconds and more, depending on size of block FFT.

Copying a fragments of the signals as waveform is much more accurate, but has own problems. Impossible to say exactly, from the image of waveform, exist/included a some frequency components in the selected segment or no.

Although the selection of segments from by waveform and is much more accurate, nevertheless, also not an accurate method. Accuracy is limited to one or two samples. In general, the problem is really accurate measurements and actions, is surprisingly/unexpectedly deep and multifaceted task.

Also introduce the possibility of transferring a fragment of the signal directly in the FSK-2 dem, and solves the problem of obtaining "clean" bit stream, in many difficult cases.

A small example of obtaining a "pure" bit stream from the rather complicated to demodulate the signal.






As usual, apart from updates, fixed bugs found by users. And including minor corrections to improve the work of the various modules and the whole program.


Good Luck!

Update SA to version 6.2.1.8

Update SA to version 6.2.1.8

In this version, we fixed an error in the function of BRC (Baud Rate Correction) discovered by users, and introduce/added a two modifications.

The first.


Sometimes, it is easier to work with a relative or even abstract concepts, because no matter the specific values, but important that they fit into a certain framework. This applies to the oversampling and shifting of signal by frequency.

Usually, it is required that a signal was in the middle of the spectrum, and the sampling frequency was not much excess.

In this version You can select the frequency shift or frequency oversampling, just selecting the desired point from the list. In some cases, it greatly reduces the time and effort. Possibility of manual input for precise actions, of course, also saved.

The second

VMW module can potentially provide extremely high accuracy, at the request of Spanish colleagues, we insert in the current update stroboscopic method of measurements.

The idea is that, if the data in the record is sufficient, You can get very precise parameters of the signals.

However, there are problems, large amounts of data require very large computing and time.

Given the characteristics of the module VMW, we just calculate and to display every 1 st, 2 nd, 4 th etc lines/rows in the visualization VMW. For all the simplicity and elegance of this method, it provides exellent/perfectly accuracy of measurement.

This is one of the realizations of the stroboscopic measurement method. And like all such methods, it provides increased accuracy with increasing period gating, in this case by rows.

In the screenshot above, perfectly clear that the old method to accurately measure the period of the sine (exactly 1000 Hz) is extremely difficult, error of 20 nanoseconds is not noticeable. In really, an error and in 100 nanoseconds is very difficult to visually detect the slope of the lines is enough small.

A new method for gating each 1024 lines, clearly and accurately detect the difference in a nanosecond! Actually the accuracy is increased compared with the first/old method in 1024 times.

Need to be cautious, as stroboscopic methods for periodic signals may give a false and sometimes very misleading results, due to skip of parts of signal. To prevent this, you should always start measuring without gating. And gradually increasing the accuracy to measurement via of gating.

Good Luck!

SA Update to version v 6.2.1.6

SA Update to version v 6.2.1.6

In this version speed of calculations, at search of a correlative triangle in OFDM module, is increased non less then 30%.

The algorithm of resampling of records is completely reconsidered.

Earlier we marked, that resampling algorithm in SA though provides precision accuracy on sampling rate, nevertheless distorts Amplitude frequency characteristic/Phase frequency characteristic (AFC/PFC) of signals. Though, it does not have the great importance in overwhelming majority of cases, as distortions affect the upper part of a spectrum. But the problem exists, we knew about it, and we promised whenever possible to solve it.

The problem is most brightly shown at resampling of OFDM signals, as OFDM signals are very sensitive to this sort of distortions.

In current version 6.2.1.6, resampling function works practically with absolute accuracy, in reasonable limits of course.

It is most easier to show the difference on the sonograms.

1

An old algorithm: there are distortions AFC/PFC on the upper channels of OFDM signal, which are visible even on sonogram.

New algorithm: there are no distortions.

2

An old algorithm: resampling of the record with sampling rate 68 mHz, on 60 mHz. As anti-alias filters aren't used, manual pre-processing is demanded, for removal of unnecessary components of the spectrum.

New algorithm: an excellent quality. Anti-alias filters are used by default. Although it isn't the main criteria of qualitative resampling, nevertheless it strongly simplifies operation with resampling function , excepting manual pre-processing.

Also the minor bug, which were found out by our users are solved.

Good luck!

FSK: "Avaris" - domestic(Russia) system of special communication on VHF

"Avaris" - special-radio service(VHF), FM-GFSK, Br-9600.

Author: SergUA6

Band Width ~12000 Hz

Baud Rate 9600 Hz

RX mode Source I/Q record

The record of fragment of work of "Avaris" system.

Sonograms:

pic.1 General view

Diagrams:

pic.2 Signals demodulation

Pictures:

pic.3 General structure of the signal

"Avaris" - domestic(Russia) system of special communication. In a basis of this system, lies using of special repeaters under similar mark "Avaris". Repeaters provide expect the basic functions, also functions of recording coded/encrypted speech messages and their subsequent transfer to concrete addressees or group of users. In this system, and not only in this, speech is digitized by means of the delta modulation. The received digital stream is ciphered and transferred, in this case, GFSK modulation is used.

SA Update: RRC filter is added.

SA update to version 6.2.1.4

Three new elements are added:

1.In the module of form viewer: median filter had been replaced by new RRC filter. By default alpha = 0.5

2.RRC filter is also added in the module of FSK-2 demodulator.Alpha=0.5 by default. If RRC filter is in OFF state, then the filter, which is analogous to the main filter, works in the signal's form viewer module.

3. The following elements are added in the module of FSK-2 demodulator:

1) elements of image vertical scaling

2) an indicator of filtration process

3) the buttons of filtration start and back into initial state.

4) the cutoff range(frequency) in RRC filter can be specified from 5 Hz to Fd/2 Hz. where Hd is sampling rate. In usual(special, soft) filter cutoff range can be specified only from 30Hz to Fd/3, in the force of realization specifications.

Special attention is focused on saving of signal's fase before filtration and after filtration. The phase do not change

guaranteed.

Also some minor bugs, which had been found by users, are correct.