RX Noise Bridges Belrose_1988 05

RX Noise Bridgesls it better to build your own RX noise bridge or buy one ofprofessional decide.

qualitf These performance results will help you

By John S. Belrose, VE2CVARRL Technical Advisor Tadoussac Dr Aylmer, PQ JgJ 1G'1, Canada

Appel, WB6AWM/5.! Cliff proposed an alternative method of measuring the impedance in an antenna system; however, it apparently did not provide an

he accuracy of an RX noise bridge has been questioned by Clifford J.

@fi(A)

2273.6 xr : ----F- -

(Eq 4)Comparing Expected Bridge Performance Several bridges are commercially available, and a homemade one is described in The ARRL Antenna Book.2 Each bridge employs different values for Cl and C2. The reactance-measuring precision of the bridge is a function of the bridge used and the accurary of its calibration. Suppose we can adjust the reactance dial and read it to a precision that is better than +5 degrees in dial rotation. The null is much sharper than this. The results of my calculations, listed in Table l, show what this means in terms of reactance variation for a frequency of I MHz. The sensitivity of the bridges to measure reactance neru resonance can be compared. Clearly, the MFJ bridge should be able

improvement

in

accuracy

Although it is clear that these little bridges are not intended for accurate measurement,they are useful devices for measuring the

for

him.

resonant frequency

of antenna systems.

This article compares the expected performance of various bridges, and comments on operational experience with two of them. General Description

the bridge circuit. The remaining two

generator and an RF impedance bridge. Two arms of the bridge are driven equally by the noise generator through a trifilar broadband ferrite transformer. One transformer winding couples noise energy into

Fig lA is a diagram of the RX bridge. It contains a wideband

noise noise

windings are arranged so that each is an arm of the bridge. Cl and Rl complete onearm; C2 and the unknown compose the remainder of the bridge. The terminals labeled RX connect to a detector-any communications receiver tuned to the frequency of interest. Fig lB shows another noise bridge configuration, with emphasis

Fig 1-At A, the RX noise bridge comprises a wideband noise genrator and an RF impedance bridge. B is a noise bridge redrawn to show the trifilar winding.

to measure reactance near resonance with the best accuracy, but the maximum value of inductive reactance for which the bridgebalances

bridges. This is particularly a problem pF, and our unknown is a pure resistance. The bridge will balance at midscale, and the effective capacitance lange is +70 pF. The effective capacitance values are read

is less than that of the other

when the bridge is used at higher frequencies, since the reactance values in Table I are calculated for a frequency of I MHz. For other frequencies, these values are

on the trifilar winding. The value of C2 is chosen so that at balance (when a non-reactive load is present) Cl tunes at about midscale. The unknown reactance can be capacitive or inductive, within a certain range of values. This is true because the reactance of the unknown element effectively adds or subtracts from the capacitive reactance of C2. Cl's dial is calibrated in terms of the difference in capacitance. Calibration is in picofarads, and plus or minus with respect to the setting that nulls with C2. Forexample, assume we are using a Palomar RX bridge where Cl : 140 pF, C2 : 70

from the dial. Capacitive or inductivebe calculated from the equations below:

divided by the frequency15

in MHz. At

reactance is obtained from a curve that can

MHz, the maximum inductive reactance

^c

:

2"tc,I27ttc2

I

that can be measured using the MFJ bridge is 20 ohms. This bridge, however, has a "range expander," which, in effect, shunts

2rf (C2 2rf (C2 +

C6)

the unknown with a 2fi)-ohm resistance.This has the effect of increasing the maxi-

(Eq l)C6)

xL =

where C6 : x dial reading. If C6 is calibrated in picofarads, and frequency is measured in MHz, the appropriate equations for the Palomar RX noise

(Eq 2)

mum range of impedance values that can be measured. A better way to measure large values of inductive reactance is to place a known-value capacitor in series with theunknown. This procedure helps to preserve

the inherent accuracy of the bridge. Practical Use of RX Noise BridgesSince the null

bridge are:

of a noise bridge is very

rNotes appear on page 39

v ^c

2273.6 159,155 ----J---f?dlet(Eq 3)

sharp, some skill is required to use the bridge. When tuning for a null, listen with earphones and tune carefully. It is better to start with low receiver gain. Switch the

34

trsT-

Table 1 Comparison of Calculated Bridge Sensitivityc1 RX Noise Bridge(pF)

c2(pF)

Reactance change for small dial change about midscale (a5o) -81.41 +72.542.8/ + 39.0

ARRL Antenna Book

PalomarMFJ

Heathkit

250 't20 140 70 1G300 180 7-120 68

-

13i1.5/ + 'l 19.4

120.5/+ 109.25

-

6630 5684 7073 5617

xc

Maximum Reactance (approx) xL 690 1150 300 1014

impedance (Zy). We are interested in finding the inverse, that is, we measure the input impedance and we want to determinethe load (or antenna) impedance. There is a trick to using the standard transmissionline equations to get what we want without

feed lines are rarely exactly Yz \,. Another way to determine Z" is to use standard transmission-tne equations and a calculator. Transmission-line equations, in general, are designed to calculate the input impedance (Zi) for a known load

'RF attenuator into the circuit if your receiver has one. Because the n and xcontrols interact, they must be adjustedalternately until a deep null is obtained. The best way to do this is to tune the n dial in

not be measured if the received level of noise and interference is high. There is,however, a method to make measurements under such conditions, providing an antenna matching system is available. Prefera-

noise generator. Antenna impedance can-

having to derive new equations or write a program for the calculator. For example, my HP4ICV calculator contains a firmware program to calculate input impedance for a known load impedance. This program takes into account a lossy line. For the reverse procedure, do as follows:

small increments starting from zero. For each increment of R, slowly rotate the x dial over its range. Repeat until a null is found, then fine-tune the bridge for a very deep null at normal receiver gain. The bridge noise level of 59 can be reduced to Sl. From my experience with the bridge, I find it is trickier to use than a professional impedance bridge. Table 2 compares measurements made with two commercial RX noise bridges. Two 75,/80-m antenna systems were use{: a drooping dipole and a half-delta loop. The impedance at the input of the feeder coaxial cable was measured at three fre-

bly, the matching network is a T match with a roller inductor, so an exacl matchcan be obtained. Adjust the tuner for an SWR of I : l, then disconnect the network without touching the dial settings. Next, terminate the tuner input port (transmitter) with a 5G.ohm load, and connect the bridge to the tuner output port (antenna) using the shortest possible leads (coaxial couplers). Now, measure the impedance seen by the bridge. Recall that a matching network is a conjugate matching device. Therefore, if the bridge, connected as above, indicates R + JX, the antennaimpedance is R -.7X. This is a somewhat tedious way to measureantenna impedance, but accurate results are obtained because the bridge is nulled under

l)

impedance

Suppose the measured input is Zin : Zm /trt, where m

designates rneasured values;

3) If this calculated input impedance is Z'in /0'in, then the real load impedance 21 (or since our load is an antenna, Z) is Zy - Za : Z'io, and 0t = 0^ = d'in t 180o, so that 0u is between minus 90" and plus

of your feeder coax, if it had been terminated, by Zt : Z^ /0m -180";

2) Calculate Z'io, the input impedance

on each side of resonance. From these measurements, the SWR was calculated from Eq I and 2 of The 1988 ARRL Hondbook, arld compared with the measuredSWR.3

quencies: at resonance and one frequency

no-noise conditions. Measuring Antenna Impedance So far, we've discussed how the RX noise

The Palomar bridge seemed to be the easiest to use. The null was more easily found, and the controls felt more positive. The box in which the Palomar bridge is housed is solid (compared with the more flexible mini-cabinet that houses the MFJbridee). The difference in bridge sensitivity

bridge works, and commented on

put. Therefore, at the %-tr resonant frequency, the input impedance (26) will bea very

Let's use this method to obtain the impedance of my half-delta loop, which is fed by a short length of Belden 8214 coaxial cable, at 3650 kHz. First, determine the length of the feeder, if it was not previously measured. To do this, short the antenna end of the feeder and measure the input impedance. Recall that a Zz-\ transmission line reflects the load impedance to the in-

90 ".

na

opepational use of the bridge. We've also observed some measured values of anten-

system impedance. These values,

for measurement of small reactance valuesis evident from the data in Table l. The impedance values marked with an asterisk (*)

in Table 2 for the Palomar bridge were clearly a reactive load, but the reactance values measured by the MFJ bridge are too small to measure with the Palomar bridge. In effect, the bridge signal source is a

transmission line reflects the load impedance to the input. In practice, however,

however, represent the impedance at the feeder input (Zil, and not the antenna impedance (2"). To determine 2", we must know the characteristics and electrical length of the feeder coadal cable. Then, a Smith Chart can be used to determine the antenna (or load) impedance. Alternatively, a %-\ feed line can be used; a half-wave

low pure resistance (zero reactance). To determine this frequency, estimate the feeder length and calculate what this frequency might be. For this calculation, the phase velocity factor for the coaxial cable

must be known

8214 coax). Use the RX noise bridge to measure the input impedance of the shorted coaxial cable. If the frequency is too

(V. =

0.78 for Belden

low, the input impedance will be a capacitor in series with a low resistance. If it is too high, the input impedance will be an inductor in series with a low resistance. At the resonant %-\ frequency, the input im- tpedance will be a very low pure resistance. For our circuit, this frequency was found to be 15.9 MHz. Therefore, the length of coax is: L 150/15.9 (0.78) = 7.36 meters

Table 2 Comparison of Measurements By Two Commercial RX Noise Bridges for 75l80-m Antenna SystemsDipole

3900 3550 3650 3800

lreS (kH! Palomar MFJ 3650 85 + p 95 +p 3750 55 +p. 53 +j10lmpedance

50 +760 45 +frO

Calculated SWR Palomar MFJ Measured 1.7 1.9 1.6 1.1 1.2 1.1 3.1 3.3 2.4

slvR

=

=

x

7.36/0.3048

=

24.15 feet

The impedance of our half-delta loop at 3650 kHz is determined by:Z1o (measured)

Hall-Delta Loop

-p. 70 -j464s

30 +7O- 21 -ja.6

45 -i14 72 -j56

1.6 1.1 2.3

2.s 1.4 2.6

2.6 1.5 2.9

47.13 / - 11 .28" To use the transmission-line progr:rm with an HP4ICV calculator, we need to know the impedance of the coaxial cable(continued on page 39)

:

= 45 - jl4

May

1988

35

Paragon. It does some of the basics-like AGC and CW keying-very well. Most of the things I didn't like are subjective. The Paragon is certainly worth your consideration if you're in the market for a high-

performance transceiver.

It's also worth

considering Ten-Tec's well-deserved reputation for excellent customer service.

East, Sevierville, TN 37862,500-Hz CW filter,

Manufacturer: Ten-Tec, Highway

4lltel

615453-7172. Price class: Paragon, $225; Model 960 power supply, $230; Model 285

$70.

F-F

Fig 4-50".,r"1 display of the Ten-Tec Paragon transmitter output during phase-noise (continued trom page 35) testing. Power output is 100 W at 3.5 MHz (A) and 14 MHz (B). Each vertical division is 10 dB; each horizontal division is 2 kHz. The scale on the spectrum analyzer on which these photos were taken is calibrated so that the log reference level (the top horizontal (2" = 50 ohms), the velocity factor (given line on the scale.in.the photos) represenls--9_0 d_Bc/Hz and the baseline is -1.40 dBc/Hz. above), and the loss parameters, Phase-noise levels between -60 and -140 dBc/Hz may be read directly from the photo- whcre "':'* := graphs. The carrier, which would be at the left edge of the photographs, is not shown. ihdse photographs show phase noise at frequeniies 2to 2b kHz-offset'from

RX Noise Bridges

th;';;;;r.

*

tttl'iffiiilt,)tffii*

has worn

off, you

can take time to go over

the Detailed Operating Instructions section to find out the function and proper use of each front- and rear-panel control and jack.

modification for using a 9-V NiCd battery for backup. According to the manual, you

can keep the NiCd battery adequately charged if you use the Paragon at least 34hours per week. This modihcation, which is factory installed in Paragons with serial

l0-8S./cm. These values are for a frequency of f = 3650 kHz. They were provided byBelden. Hence, using the program we can calculate Z'in for a load impedance Z'y: Z'r: 47.13 /-l't.28-180' (Eq 5) (Eq 6) Z'. : 67.2 /174.13"

G=

=8x

condubtance/unit length

Hookup The Installation section of the manual gives a brief description of what you needto hook up the Paragon. It's pretty straightforward. For starters, you need a 13.8-V power supply capable of delivering about 20 A. Wiring information is given in case you don't use Ten-Tec's matching Model 960 power supply. The Installation section recommends a low-impedance dynamic or

numbers higher than 395, requires asoldering iron, a 2.2-k0, %-W resistor, and about 15 minutes of your time. It's well

and

so

worth doing. 0perating Impressions I enjoyed using the Paragon over a period of several months. In addition to casual CW and SSB operating and DXing,

Z" = 67.2 /174.t3-180" (Eq 7) : 67.2 /-5.9'= 66.8- j7 ohms(Eq 8)This impedance is close to that expected for

a half-delta loop.Notes

electret microphone (high-impedance mics-above 25 k0-won't work), although text accompanying the mic connector wiring diagram later in the manual says that the microphone circuit has been designed for high- or low-impedance mics with at least 5-mV output.back up the microprocessor and RAM when power is removed. The battery backup allows the Paragon to retain memory and frequency information, as well as clock,/ calendar time and date. Battery installation requires removing the top cover and pluggng in a connector. Batteries not included..,

I used it in several SSB and CW contestsincluding the phone ARRL NovemberSweepstakes. I like the panel layout and the way most of the features work. I initially had trouble getting on the "right

J. Appel, "How to Measure Antenna lmpedance," QEX, Jan 1987, p 3. zJ. Hall, ed., The ARRL Antenna Book3M. Wilson, ed., The 1988 ARRL Handbook,(Newington: ARRL, 1987), p 1G2.degrees (Newington: ARRL, 1983), 14th edition, p

1C.

1$t0.

frequency" during CW operation. The

Before using the Paragon, you are encouraged to hook up a 9-V battery to

instructions on using the spor switch are confusing, and the peak S-meter reading does not correspond to zeto beat. In fact, the manual suggests using the wide filter in

John S. (Jack) Belrose received hb BASc and

MASc (EE)

conjunction with the seor switch; this is not acceptable on a crowded band. Aftersome experimenting and after "working myself" on a second transceiver in theshack,

British Columbia in 1950 and 195i,, respectively, and his PhD in Radio Physics from Cambridge University while in England on an Athlone Fellowship. From I95l to 1953, he was

from the Univenity of

I discovered how to

get on frequency.

equipment when I'm away from the shack, and my first 9-V alkaline cell lasted months.) The new manual includes information on a

Paragon connected to a 13.8-V dc source (even with the power switch turned off). This means leaving your power supply turned on all the time. The original manual indicates that an alkaline battery will provide about 150 hours (6.25 days) ofbackup if power is removed. (fhis l5Ghour rating is very conservative. I unplug my radio

According to the manual, the battery backup is not necded if you leave the

Don't ask me to try to put it into words!The receiver held up well except duringpeak phone-contest periods when there were

many strong sigrrals on the band. During those periods, I had difficulty finding openspaces on the band and had difficulty copying weaker signals. The problem seems to be receiver phase noise, rather than front-

end overload. Use of the attenuator made a noticeable improvement in signal copy. The general-coverage receiver works fine. AM selectivity is good, thanks to the 6-kHz crystal filters. All in all, Ten-Tec has a fine radio in the

employed by Canado's Deferce Research Board, and worked in LF communications at the Radio Propogation Laboratory,in Ottawa. Since 1957, Jack has been with the Communications Research Centre, Dept oJ Contmunications, in Ottawa, where he b carrently Director oJ the Radio Propagation Laboratory. Dr Belrose b Conadian Panel Coordinator for the AGARD Electromagnetic Propogation Panel ond is Choirman of an Intbrim Working Party of the CCIR (Study Group 6). He b an ARRL Technical Adiiso; and has been a licensed amateur since 1948. His former call signs were VE7QH, VE3BLW and VE2SA. He has been licensed as VE2CV since Nov I96L fack and his wife, Denise, hove three children. Jack's hobbies are Amateur Radio (particularly antennas and radio communications technology), photography, conoeing, swimming, touring/camping by tent troiler and exercising EiFI his dog, Rufus.

May

1988

39