PT107293B - METHOD FOR DETERMINING THE DELAY BETWEEN MEASUREMENTS IN TWO OR MORE SPECTRUM ANALYZERS OR POWER METERS - Google Patents

Abstract

THE OBJECTIVE IS TO MEASURE THE DELAY BETWEEN THE INSTANT MEASUREMENT START (A8) BETWEEN TWO OR MORE SPECTRUM ANALYZERS BASED ON THE POWER MEASURED IN EACH ONE USING A PROPORTIONALITY RELATIONSHIP BETWEEN THE TWO GRANDEZES. THIS METHOD OF TIME MEASURE INVOLVES INJECTION IN EACH OF TWO ASKED SYNCHRONIZED AND IDENTICAL ASK (AMPLITUDE SHIFT KEYING IN ENGLISH) SIGNS, EACH MODULATED BY A 50% SQUARE DUTY CYCLE IN ENGLISH. THIS METHOD IS VALID FOR CONTINUOUS CURRENCY (CC), INJECTING AN ASK (RADIO FREQUENCY (RF) / MICROWAVE SIGNAL) A SQUARE WAVE IS INJECTED AND THE POWER IS MEASURED BY A DC POWER GAUGER WITH SHOT.

Description

DESCRIPTION

Method for determining the delay between measurements on two or more Spectrum Analyzers or Power Meters

Technical field of the invention

Electronics (Instrumentation), Radio Frequency, Microwave Summary of the invention

The present invention relates to the measurement of the time difference (A8) between the start of the power measurement point (in a scan) on two or more hardware, software or spectrum triggered spectrum analyzers (ASs). mixture. The difference between the delay times of these shots is the determined quantity. These delay differences may be due to the ASs own hardware or software or because the triggering signals are already out of sync. If there is sufficient temporal accuracy in the ASs (or triggered power meter), it can be used to measure the delay difference in trips due to an external circuit. This can be used, for example, to measure delay differences in transmission lines.

Thus, it allows to measure the degree of synchronism between simultaneous measurements of ASs. This can be extrapolated for any simultaneous measurement with other radio frequency, microwave or direct current (dc) power meters.

General Description of the Invention Objective is to measure the delay between the start of measurement (A8) at two or more ASs (a single triggered RF / Microwave power meter may be used) based on the power measured at each. It is assumed that the command to make measurements is made in both cases by software software on both or hardware on both or mix (different for each). This time measurement method involves injecting into each AS two identical Amplitude Shift Keying (ASK) synchronized signals (the result of modulating a carrier by a 50% service factor square wave). If ASs are located remotely from each other, ASK signals must be synchronized in some way, for example by GPS. If co-located the ASK signal can be divided by a signal splitter or splitter (B3). The method allows for the correction of the time difference calculation based on the measurements in case ASK signals do not have the same measured power due to ASs calibration errors or there is no symmetry in the signal divider. The calculated signal representing the time difference shows good accuracy due to its flatness, which happens if the measured time period (after being calculated) is much larger than the instrument's time inaccuracies and when the calculated time is of the period order of the ASK signal divided by 4. The method is also applicable for Direct Current (DC) with a positive square wave (between positive and zero) with 50% service factor and with DC power meters with firing. The whole theory is equally applicable.

This method requires a group of measurements commonly referred to in the Sweep ASs manuals.

Description of the Figures

Figure 1. Cognitive Radio Measurement Campaign Scenario. Al - Base Station, A2 - Obstacles, A3 - Sensor (AS /) responsible for P (t, s ₂ ), A4 - Sensor (AS ₂ ) responsible for P (t, s ₂ ), A5 - Computer. The distance between sensors is less than 200 meters.

Figure 2. Measuring Device 1 (one), Bl- Laptop, B2 - Signal Generator (ASK Generator), B3- Signal Splitter, B4, B5 - Spectrum Analyzers, B6 - Ethernet (Dual Ethernet Card) Connection ), B7 - SMA Cable, B8 Universal Output Serial Bus USB-powered and controlled card TTL, B9 - Cable connected to Spectrum Analyzer trigger input, B10 - USB Cable

Figure 3. Measuring Device 2 (two), Bl - Laptop, B2 - Signal Generator (ASK Generator), B3 - Signal Splitter, B4, B5 - Spectrum Analyzers, B6 - Ethernet (Dual Ethernet Card) Connection ), B7 - SMA Cable

Figure 4. Measuring device 3 (three), Bl - Laptop, B2 - Signal Generator (ASK Generator), B3 - Signal Splitter, B4, B5 - Spectrum Analyzers, B6 - Ethernet (Dual Ethernet Card) Connection ), B7 - SMA cable, B8 - USB powered and powered chart from TTL digital outputs, B9 - Cable connected to Spectrum Analyzer trigger input, B10 - USB cable

Figure 5. ASs Rectangles at a Single Point Service Factor

- Time windows Measurement in two of scan measurement, 50% Sinusoid - ASK signal.

Figure 6 .. Case where any l.

gives approximately zero to

Figure 7. Calculation of the difference between the time points at the start of the power measurements (one scan) between two ASs (Software Triggering)

Figure 8. Result of calculating the difference between the time points from the start of the power measurements (one scan) between two ASs (Hardware Trigger).

Figure 9. Zener Diode and Schottky Voltage Limiting Diode connections at both ends of the RG58 trigger signal transmission cable. 11,12,15 - Schottky Diodes, 16 - Zener Diode, 13,14 - RG58 cable termination. LI - USB Powered Chart Side TTL Digital Outputs, L2 - Spectrum Analyzer Side, VA - 5Volts USB Powered Chart TTL Digital Outputs

Detailed Description of the Invention

Figure 2 shows an example of a device for indirectly measuring the difference (in seconds) between time points from the start of measurements (A8) of two ASs (B4 and B5). In this device (Figure 2), the triggering is done by hardware in one AS and software in another. The measurement start time instant in hardware triggered AS is deterministic and in software start AS the measurement start time instant is non-deterministic. Figure 3 shows the device with two software-triggered ASs (triggered by programming commands transmitted over Ethernet) and Figure 4 shows another device with two hardware-triggered ASs via the trigger input (Trigger), and a Card controlled and powered by Universal Serial Bus (USB) TTL digital outputs (B8). Programmed by USB by Laptop (Bl). In each AS an ASK signal of equal powers (not really equal due to asymmetries in the signal splitter) is injected and synchronized. The measurement is made with center frequency equal to the carrier frequency and SPAN zero (this AS parameter, with this value implies that all power measurements of a scan are made at the same center frequency). SPAN is a bandwidth in which the center frequency of the scan measurement points is scattered and equally spaced. The resolution bandwidth must be such that it must include in excess the ASK signal bandwidth (which can be gauged from the signal display on the AS, with appropriate span and measurement bandwidth). Figure 5 shows the ASK wave with two power measurement time windows each corresponding to a scan point in each AS.

At each point of the scan AS measures the energy and then divides by the measurement time of that point to calculate the power.

We will initially assign to the measurement period at each scan point (T _{sw ρ} , ΆΊ), half of the ASK signal period (T _ASK / 2, A9). Note that in each AS the scan time must be the number of scan points times this time (for SPAN zero the minimum scan time is zero [7, page 76]). Therefore it is not necessary to take into account transient effects on the scanning time. In this case, the time difference modulus of the ASs (given in Figure 5 by, A8) is directly proportional to the Energy difference modulus between them (as shown in Figure 5). In this perfect scenario, with T _{sw P} ⁼ '^ ASK ^ r θπι each point of the scan (possible number of points per scan on ASs 501, 1001, 2001, I ...

points, each section has I points) corresponds to the energy p ^l _ p ^{l r} ASy ^r AS ₂ gives the same result (^

T ^J ASK ·>

corresponds to the energy T _ASK P _ASK ):

T sw_p pí ¹ AS _} ~ ^p k T,

P ^{l 1} AS1 _p ^{l r} AS ₂

2P

- ¹ ASK

TP ¹ ASK ¹ ASK

P _AS are the powers made on a logarithmic scale as measured dBm (1) (linear, each

ASK although P _AS ^ is measured to pass to Watt) at each scan point i AS (ASi and AS ₂ , B4 and B5). Pask θ is the signal strength jected at each AS. The other variables are shown in Figure 5. It may happen that the relative phase of measurements with the ASK signal is that of Figure 6 and in this case the power difference is zero. One way of solving this problem is to make the measurement time at each scan point slightly greater than the half-time of the ASK signal such that the ment (o at the two value of the slightly measured points otherwise than the maximum is the phase). relative change will change over time points between start time instants ASs continues to be given by Equation 1). Equation 1 along the scan index i changes from zero to a maximum, as it happened before they remained constant, the value of the difference between time points.

Pask can be measured directly from each section (varrimenT

T 'y ^ASK to) (because ¹ _{sw p}

). There is a measured energy equality relationship

T

SW_p

MAX (P _A5i ) and the energy of the ASK wave giving p

¹ ASK

T MAX (Pi _Ç) ^p ~ ^sw ie {l, ..., Zp

T ¹ ASK (2) where MAX (P _A5i in each section does not represent the maximum power measured here forward

ASi (0 AS chosen is {1, ···, ^} will be omitted from the equations.

The difference between the calibration errors in the ASs and the attenuation difference in the signal splitter (unbalanced) can be compensated (It is interesting to compensate one AS against the other since we do not have the actual value of the magnitude) by introducing the following power gain. in AS ₂

MAX (P ^)

MAX (Pb)

Substituting Equation 2 into Equation 1, introducing Equation 3 into Equation 1 gives

ΜΑΧ (Ρ ^) 2 (4)

I ^^ P'as) cancels AS ₃ calibration errors if the error is modeled by a gain.

Method Accuracy

Given that the relative error of function F is given as a function of the errors of its variables by & F

(5)

So considering a great precision in the variables involving time and we have

P ^{l 1} AS _? / \ (£ * 22 ^_ ^ 21) (6) y P ^l τ

Ay _ ¹ ASK__AS] / _ \ ¹ ASK max (AJ ^{11 12} 2 max (AJ being -Relative Measurement Error of Pas ₁ , £ * 12 - Relative Error of MAX measurement ^ P _A s ^, ^ * 21 - Error Relative of ^ AS ₂ measurement, ^ * 22 -Relative of MAX measurement

Note that in the case of the same AS, the relative error may be of the same signal and in this case there is an error canceling effect. Assuming the standard deviation of the relative power error measured in the ASs is £ _r (best case [5], [6] typically 1.5% / 100, but found websites reporting errors of 0.4%) and for the worst case and £ * 12 ^{- -} £ * 11 ^{- -} £ * _r and £ * 2i ^{- -} £ * 22 ^{- -} £ * r then (7) ^£ r ^ ASK pi ¹ ASj p ¹ ' ^r AS ₂ ^ MAX (pjJ ⁺ MAX (p, _r = 2 k JJ

The relative error

T ^± ASK ~ QT T ^- Δ with ^x Δ as measured (A8) in measurements). But there is an effect ρ; _5ι μαχ (ρ; ₅₂ ) -ρ ^ μαχ (ρ ^) (8)

Theoretical J given by Equation 8, for difference of start times of the me is approximately 5% (for 1.5% of real can of the error or greater.

Lsk ^of 2 increase accuracy decreases. T \ _SK I 2 Entely large so expected split-time. For a T _{ASK to} measure delays until error is if the relative minor accuracy (A9)

2T ^{z 'IJ} to (A9) must be chosen suficia measuring instants much difference to ensure pre (A7) given but this method can not and T ^"511" _P

T -T -T ^± A ¹ ASK ^± sw_p phase terminated between ASK signal and scan points). However it is t _a <Ut _a „ _k -t)

Δ 2 \ ASK sw_p) (Only if there is a recommended measurement range that deduces

T sw _p _ ^ ASK 1 _ | _ l

δ> 5

J <15.

determines the time period of the

By

100) (9)

This set of ASK wave equations as a function of the time we want to measure. Equations 9 The time to be measured should be less than approximately one quarter of the ASK wave period.

& is a parameter that will define the percentage by which the scan time at one point is more than half of the ASK wave period. As defined in Equations 9 it is recommended to be between 5 and 15.

These simultaneous conditions, expressed in Equations 9, ensure that the waves presented in the Results section have sufficiently wide tops.

Results

Figure 7 is an example of such a measurement in the scenario of two software-triggered ASs (B4, B5) (scenario Figure 3). Represents four sections of 501 scan points each. The parameters are T _{sw p} = 21ms, T _ASK = 40/775 'θ P _ASK = 0.44 // W (-33.56dBm) (Power is not very relevant. However, it must be at least 1000 times (30dB) above noise level and take into account non-linearities of instruments). The two ASs (B4, B5) used in the measurement were of the same brand but of different generations (Rhode & Schwarz, models FSP40 and FSQ8). The signal generator (B2) was Rhode & Schwarz, SMU200A. As you can see, the time difference (A8) is given by the maximum that does not change during a section. The actual relative error can be considered small due to the flat top of the waves of Figure 7.

Figure 8 shows the scenario measurements of two hardware-triggered ASs (B4, B5) (scenario Figure 4). The USB powered TTL digital output Card (B8) has been configured to trigger across two TTL Input / Output lines in the same Push-Pull (Byte - 5V, Low - 0V) port. The USB powered and TTL digital output Card (B8) is connected by two 90-meter RG58 (B9) cables to the ASs trigger inputs. Both cable terminations are protected with Schottky Diodes (11,12,15) and a Zener Diode (16) to limit the signal above and below. The USB-powered, TTL digital output Card is protected by two Schottky Diodes (II, 12, model BAT85S) because there is access to the power pins. On the AS side is protected with a Schottky Diode (15) and a Zener Diode (16, model BZX79-C5V6) because there is no access to the higher ASs supply voltage (see Figure 9). The trigger voltage level on both ASs was 1.4V. This value will have to be taken into account for applications where the time difference between the start of measurements has to be calculated accurately.

T

It was used - “- = 100 // 5 for a measure of time difference of approximately 8 // 5 ¹ . Despite the lack of flat tops of the waves it was found that the measured value has more repeatability than the case of software firing (see the tops between sections). The lack of flat tops is due to the fact that the time difference is much smaller and the assumption of ASs time accuracy is no longer valid and for the measured time (8 // 5) and period of the ASK signal the Equation 8's relative error gives a much larger value (around 45%).

This method gives the relative delay between ASs but not what triggers first. This can be found in the case of software triggering by introducing a delay (software, a fraction of the total delay) in the trigger command of one of the ASs. If the average delay increases then this is the latest AS. Otherwise it is the early AS.

Conclusions

A new method to determine the time between start times of measurements of two or more ASs was presented. The method may have applications beyond the objective needed by the authors to find the best triggering methods to achieve synchronism between spectrum occupancy measurements for cognitive radio. The measurements show much better results than the theoretical error if the measured time period is much larger than the instrument time inaccuracies and if the measured time is approximately the ASK signal period divided by 4. This can be explained by because in determining the delay there may be an error canceling effect and the accuracy of the measurements is better than the specifications of the ASs.

Application Examples

In the Cognitive Radio fusion study it is necessary to make synchronous power measurements within a couple of hundred meters between measuring devices in order to measure shadow conditions (Shadowing).

This is the tracking of measurements on a single station [1] in GSM bands. Figure 1 shows the scenario of the Cognitive Radio measurement campaign. (A5) is the power measured in ASj. The rectangles show the time windows of measurement. For measurements to be useful (to merge later) they need to be synchronous across two or more ASs. With the proposed method is found the difference between time instants (A8) .For good timing is necessary that 'T _{sw sw fusion} _ _ f ~ _us ion measurement time period in each scanning point and then to fuse. Do not confuse with the measurement time at each point in the method to determine (A8).). In [1] the peer scan period was 4.62ms or equal to the time frame of a GSM frame. 0 time between the beginning of the ASs measures (T _A, A8) 8 // 5 is a fair degree of synchronization to a scanning time per point at the GSM frame. The difference between the start of measurements may be due to differences in processing time between ASs and lack of synchronization of commands sent by the computer via Ethernet Cables (B6), ie in the case of two software trips. In the case of hardware trips the delay may be due to different delays in response to tripping by ASs and different delays in tripping introduced by circuits outside the ASs. The proposed method allows to evaluate various approaches to obtain synchronism. Previous work in this context has been done with GPS synchronization [2] which does not allow flexibility in the choice of scanning time and therefore does not allow continuous measurements without time lapses.

Measurement of time is usually done back to back at the MAC layer in computer networks and involves considerable technological resources [3], [4]. In this case the method is simplified due to the fact that the original system includes ASs and thus the delay can be calculated by precision power measurements.

This invention may give rise to Application Notes by AS brands such as Rohde-Schwarz, Agilent, Tektronix, etc. A group of companies that may be interested in industrialization are manufacturers of delay measures and delay differences in transmission lines. Absolute delays on transmission lines can also be measured if a TTL I / O line is connected directly to the trigger input of an AS (or power meter) and another TTL I / O line is connected through the line measuring the delay. . This delay can be used to measure the length of transmission lines. The generalization of the idea to a square wave and Continuous Current Power Meters (instead of ASK and ASs signals) broadens the idea's applicability and many more companies may be interested. A Continuous Current Power Meter has a simpler technology that broadens applicability.

Description of the ASs Hardware Firing Experiment

1. Mount the device in Figure 4.

2. At the ends of the RG58 Cables (B9 of Figure 4), each 90 meters long, make the connections of Figure 9. The model RG58 cable loop is connected to 0V on the USB-controlled portion of TTL digital outputs. (13). On the ASs side the mesh is connected to the outside of the input (14) input connector (ground).

3. Manually program the signal generator (B2, R&S, SMU200A, R & SOSMU-B9 / -B10 / -B1 options) to generate an ASK wave with the desired period by varying the Symbol / Bit rate in the baseband block. Larger periods of the ASK carrier modulating square wave can be obtained with Data Patterns (by selecting Patterns in the baseband window) of consecutive ls, followed by consecutive Os. The period of the ASK wave can be determined by Equations 9. Not knowing at the outset the time to be measured may have to try several periods of the ASK signal until it can be measured. You should try to choose T _ASK so that you can over-measure the expected time.

Then if you change T _ASK a small percentage and give a different measured time, you have not yet found the right value. Program in ASK modulation, ASK Modulation Depth = 100%, rectangular baseband filter with internal clock and internal data. In the radio frequency section set the carrier frequency, and the carrier level such that the signal level (total power, encompassing the entire ASK signal bandwidth) on the ASs is more than 30dB above the noise level. that, however, do not enter the saturation of ASs inputs.

4. Run Labview program on laptop (Bl) to program ASs (B4, B5. Rhode & Schwarz, FSP40 and FSQ8 models) with the appropriate scan time (T _sw p times number of scan points), number sweep point (this parameter can only take certain values determined by the manufacturer), RMS measurements, SPAN = 0, ASK signal carrier frequency, measurement bandwidth (before manually measuring with an AS, with non-zero SPAN ASK signal bandwidth to determine measurement and video bandwidth (3-10 times the measurement bandwidth) and put the ASs into hardware trigger standby. These settings are made through two Ethernet cables (B6, STP - Shielded Twisted Pair in English, direct connection without repeaters), one to each ASs, with more than 90 meters (<100 meters) each. On the computer side is a dual Ethernet card (with two ports, lOOBaseT).

5. Following the Labview program toggle the (two) TTL Output lines (same byte) of the USB-controlled Card from TTL digital outputs (B8) from approximately 0V to approximately 5V. Both ASs should take measurements due to tripping (signal through cables RG58, D9). The voltage level at which the ASs trigger must be equal on both and be on the up (Positive Edge) or down (Negative Edge) on both.

6. Read the measurements (floating point scan group) of the two ASs over the Ethernet connections to the computer and write to disk.

7. Calculate (4) for each scan point to obtain a section of Figure 8 (using all points). You can use the same computer that reads the measurements of ASs to do the calculations and generate the graphs with a program like Matlab.

8. Measures may to be made with more than 2 ASs (no ASs) but the calculations are made with at measures of 2 ASs. This generates the shots at n> 2 TTL outputs ( the divi-

signal will have n outputs as well), read the measurement sets with n Ethernet cables connected to the n ASs. But the calculations are made with 2 sets of measurements (2 to

2) from 2 ASs. These calculations can be repeated with the 2 ASs measurement sets of ns to cover various combinations of 2 ASs groups.

REFERENCES [1] L. Mendes, L. Gonçalves, and A. Gameiro, GSM Downlink Spectrum Occupancy Modeling, at IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC'11), (Toronto, Canada), 10-14 September 2011

[2] M. Wellens, J. Riihijãrvi, M. Gordziel, and P. Mãhenen, Evaluation of Cooperative Spectrum Sensing Based on Large Scale Measurements, at Third IEEE International Symposium on New Frontiers in Dynamic Spectrum Access Networks (DySPAN), (Chicago , Illinois, USA), 14-17 October 2008.

[3] B. Ngamwongwattan and R. Thompson, Measuring One-Way Delay of VoIP Packets without Clock Synchronization, at IEEE Instrumentation and Measurement Technology Conference (I2MTC 2009), (Singapore), 5-7 May 2009.

[4] A. Hernandez and E. Magafia, One-Way Delay Measurement and Characterization, at Third International Conference on Networking and Services (ICNS 2007), (Athens, Greece), 1925 June 2007.

[5] Agilent PSA Series Spectrum Analyzers - Data Sheet, [6] R&S FSQ Signal Analyzer Specifications, [7] Fundamentals of Spectrum Analysis, Christoph Rauscher, Volker Janssen and Roland, Sixth Edition, Rohde & Schwarz, October 2008, 2015

Claims (1)

1. A method for determining the time between measurement starts between two or more Spectrum Analyzers - B4 and B5 - or Power Meters characterized by: The. ASK Signal Injection - Amplitude Shift Keying, in English - with a period determined by the set of Equations E4 from a time to be measured, and with a 50% service factor and 100% Radio Frequency - or Microwave modulation depth, as a test signal in both or more Spectrum Analyzers - AS - or simpler RF Input and Microwave Power Meters; or alternatively by the Square Wave Signal Injection with period determined by the set of Equations E4 from a time to be measured, with 50% service factor as a test signal on two or more direct current input power meters; B. Generating the Universal Serial Bus USB-Commanded, Letter-Triggered Signals from TTL-B8 digital outputs that will propagate through circuits, such as RG58 cables, which introduce propagation delays until ASs trigger input ; W. The equation E1 is calculated by means of a calculation device which allows the difference in the delays of the climbs only or the descent only in the trigger inputs of the ASs or power meters to be determined by converting a Power difference into a time difference, from the measurements, taking into account calibration errors between the measuring devices and power asymmetries due to the signal divider - B3 -; d. execution, by means of a calculation device, of Equations E2 and E3 which allow an estimate of absolute and relative errors respectively (of these time differences). ASK max (p; _5i ) 2 = ² _r T ¹ A, (E3) (E4) δ> 5 J <15 ^± A _i > ² Δ - Time difference to be calculated at the measurement point ie the real time that will be approximated to the maximum along r, respectively. T _AS k - ASK wave period P _AS1 - Scan Point Power i on Spectrum Analyzer 1 P _ASi - Scan point power r in Spectrum Analyzer 2 T _{sw p} - Scanning time at a measurement point - Represents the maximum power measured in each section in AS _X P _A s ₂ ) - Represents the maximum power measured in each section in AS ₂ Ο - It is a parameter that will define the percentage by which the scan time in a point is more than half of the ASK wave period. As defined in Equations E4 it is recommended to be between 5 and 15. By Equations E4 the time to be measured should be less than approximately one quarter of the ASK wave period. G _Á Takes the value max (p; _Si ) / max (p; _S2 ) - Estimation of the absolute error in calculating the time difference at point i. £ _r - standard deviation of the relative power error measured in ASs.