Up one level
Back to Home
Receiver Sensitivity Measuring Methods
By Robert W. Meister WA1MIK
There seems to be two common methods of measuring FM receiver sensitivity: 20 dB quieting, and 12 dB SINAD (SIgnal plus Noise And Distortion). Most people like one method or the other, and strongly defend their choice. This article will examine the equipment and procedures I used, and compare the results.
Back in the "good old days" when FM radio was wide-band (15 kHz deviation or more), the 20 dB quieting method was used most often. Receivers needed to be accurately and thoroughly tuned. The complete IF (Intermediate Frequency) section could be aligned, and the service manuals actually gave you full and complete procedures to adjust everything in the radio. That level of detail doesn't exist any more. Many radios are considered to be throw-away items. The factory does the alignment, manuals are less and less detailed, and what adjustments are present, are done with software. Some people would look upon this as progress.
With the level of integration present in modern communications equipment, receivers hear things a lot better and tend to have wide-band front ends, but the average person can no longer align them.
There are several other good articles about measuring and understanding FM receiver sensitivity elsewhere on the Repeater-Builder web site.
Most test equipment measures signals relative to ground, and one of the input lines is very likely to be at ground potential. Be careful when attaching things to the radio's loudspeaker lines. On a radio such as a Motrac, Motran or MICOR, one side of the speaker is grounded. On a MaxTrac one side of the internal speaker is connected to ground through a 1 ohm resistor. On a Spectra, Mitrek, or GM300 mobile neither side of the loudspeaker can ever be grounded or you will let out the audio amplifier's lifetime supply of magic smoke. If you are testing a non-grounded speaker output you need to use an audio isolation transformer. You can get one from Motorola (part number 2580188B01), from a high-end car stereo shop, or from web/mail order suppliers like Crutchfield. Just tell them you need an 8-ohm to 8-ohm isolation transformer that can handle about 2 to 3 watts.
The radios that preceded the Mitrek in the Motorola mobile product line mostly used push-pull audio PA decks and audio output transformers, and drove the speakers via a hot wire to ground. The Mitreks and many later radios eliminated the audio transformer and they run the speaker as the push-pull load directly... both sides floating off of ground. This quirk is significant on your workbench: any audio test equipment that connects to the speaker (such as you would use to make a quieting measurement) must be on the far side of a 3.2 ohm or 8 ohm 1:1 audio transformer. Motorola has a low-power one as part number SLN6435 in their test equipment catalog, and they also include a higher power one in every tabletop base station (part number shown above) since common wire-line remote controls expect ground-referenced audio to drive the remote speakers. You can order either one as a spare part from Motorola.
I used an Agilent E4430B synthesized 1 GHz signal generator as the source of RF. The modulation was the internal 1000 Hz sine wave oscillator set for 3.0 kHz FM deviation.
My indicating instrument for 20 dB quieting was a Fluke 189 true RMS digital multimeter. This meter has the ability to display AC millivolts and dBm at the same time. The only downside to a DMM is that the update rate tends to make the displayed value jump around when noise is a significant part of the input signal. I also used the AC voltmeter on the HP 334A distortion meter, and the two were within 0.5 dB on all readings.
I used an HP 8901B Modulation Analyzer to measure SINAD as well as distortion. It has a digital display but can only measure SINAD or distortion at 400 Hz or 1000 Hz. The distortion can be viewed as a percentage or dB below the reference level (negative values). The SINAD function merely shows the absolute value of the distortion's dB values.
I used an HP 334A Distortion Meter for the alternate SINAD method. This unit covers 5 Hz thru 600 kHz and can measure distortion well below 0.1% (more than 60 dB down). Other HP models such as the 331A, 332A, and 333A can also be used. Distortion can be viewed as a percentage or dB below the reference level.
I had the chance to borrow an IFR 1200SS service monitor to measure SINAD. This particular unit has both an analog meter and a digital readout and measures directly in dB. At 30 dB SINAD signal level, this unit indicated 3.2% distortion and my HP 334A measured 3.4%. I'd consider that "close enough for government work."
I tend to use dBm for my RF signal levels. If you prefer microvolts, a handy two-page conversion table can be found in this 72kb PDF file
How SINAD Works:
A SINAD meter, or function in a service monitor, consists of an AC voltmeter and a narrow audio filter tuned to 1000 Hz. Some units can also generate their own tone. An RF signal modulated with a 1000 Hz tone is fed into the receiver. The detected audio level is measured (SIgnal plus Noise And Distortion). The filter removes the 1000 Hz signal, and the resulting audio is then measured (Noise And Distortion); the difference between the two values is the SINAD number in dB. Some units may use an automatic gain control to set the reference level; others may rely on two meters or a switch to measure the pre-filter and post-filter levels; still others just use a differential voltmeter circuit to determine the difference between the two levels. Whichever method is used, the difference between the signal levels with and without the 1000 Hz tone filtered out gives you the dB SINAD value.
SINAD measurements take into account any non-linear operation, irregularities, or deficiencies in the IF, detector, and audio stages of the receiver, mainly because these can all increase the distortion present in the audio signal. It also requires that the modulating equipment and signal have a lower distortion level than the receiver you're trying to measure. Tuning a receiver for best SINAD value may take longer and require more equipment, but it will also get the most performance out of the entire system.
It is critical that the 1000 Hz filter and the 1000 Hz tone be accurately tuned to each other. Many SINAD meters give you the ability to adjust one or the other. If your SINAD instrument generates its own tone, feed that into the meter's input and adjust the filter tuning for the lowest meter indication (best SINAD value). The absolute tone frequency is not important (within 1%) as long as the notch filter is tuned to it.
20 dB Quieting Method:
I've always used the 20 dB quieting method for measuring sensitivity, but not for the absolute "factory spec" value. I tend to use it for comparison with other similar receivers. My reason for liking this method is simple: it requires very little test equipment, and even that doesn't have to be very sophisticated.
Connect a signal generator to the receiver's RF input port, and an AC voltmeter to the loudspeaker. On a MaxTrac, I use the handset audio signal present at the front panel MIC jack. The official EIA specification says to use the speaker terminals. Whichever is used must provide de-emphasized audio.
With no signal present, open the squelch and measure the AC voltage. If your AC meter has a dB scale, note that reading. On a 450 MHz MaxTrac, I usually measure about 700 millivolts or -1 dBm (relative to 1 milliwatt at 600 ohms) on noise.
Send an un-modulated RF signal into the radio, starting at the lowest level possible. As you increase this signal level, the receiver noise will start to decrease. This is the "quieting" effect that FM is so well known for. Adjust the RF level until the AC voltmeter reads 10% of, or 20 dB lower than, the value you had with no signal. The amplitude of the RF signal is now the 20 dB quieting sensitivity of the receiver. On a 450 MHz MaxTrac, the AC voltage would now be around 70 millivolts or -21 dBm, and the sensitivity is usually around -118 dBm (just under 0.3 microvolts).
You can often measure the 20dB Quieting sensitivity just by listening to the received audio on the radio's loudspeaker. As you increase the RF level, the noise will decrease, and so will the crackling and popping sounds. When you get to the point that you hear one pop or crackle every few seconds, that's usually the 20-22 dB Quieting point. You can use this to quickly compare receivers or preamp performance.
12 dB SINAD Method:
Connect a signal generator to the receiver's RF input port, and a service monitor (or other piece of equipment capable of making SINAD measurements) to the loudspeaker. On a MaxTrac, I use the handset audio signal present at the front panel MIC jack. The official EIA specification says to use the speaker terminals. Whichever is used must provide de-emphasized audio.
Set the RF signal source for a strong signal modulated with a 1 kHz tone set for 3 kHz deviation (60% of the receiver's normal bandwidth). For 2.5 kHz wide systems, you would set the deviation to 1.5 kHz. If your SINAD meter can generate its own tone, use that as a modulating source. Adjust the display device for a reference level if necessary.
Start decreasing the RF signal level. The receiver noise will simultaneously increase. Continue lowering the signal until the meter reads 12 dB (or 25%). The amplitude of the RF signal is now the 12 dB SINAD sensitivity of the receiver. On a 450 MHz MaxTrac, the 12 dB SINAD sensitivity is usually around -120 dBm (about 0.22 microvolts).
Alternate SINAD Method:
If you don't have a SINAD meter or a service monitor with that capability, you can make a similar measurement with an audio distortion meter. This device consists of a tunable notch filter and an AC voltmeter that can measure the audio level with or without the filter in the circuit. By notching out the fundamental audio tone, the residual noise and distortion can be measured. Note that this is exactly what a SINAD meter does. Not surprisingly, the procedure is nearly identical.
Connect a signal generator to the receiver's RF input port, and an audio distortion meter to the loudspeaker or handset output, just like you would do with a SINAD meter.
Set the RF signal source for a strong signal modulated with a 1 kHz tone set for 3 kHz deviation (60% of the receiver's normal bandwidth). For 2.5 kHz wide systems, you would set the deviation to 1.5 kHz. Adjust the distortion meter for its reference level of either 0 dB or 100%, then adjust the tuning for minimum distortion level. It should be possible to get the distortion below 3% (30 dB down).
Start decreasing the RF signal level. The receiver noise will simultaneously increase. Continue reducing the signal until the meter reads 25% distortion (12 dB down). The amplitude of the RF signal is now the 12 dB SINAD sensitivity of the receiver. On a 450 MHz MaxTrac, the SINAD sensitivity is usually around -120 dBm (around 0.22 microvolts).
I used a 450 MHz MaxTrac as my subject receiver because it was the nearest radio to my test bench. I used the handset output signal on the MIC jack. Technically, I should have connected my equipment to the loudspeaker terminals, but since the handset jack is the detected and muted audio at the top of the volume control, the only circuitry that was not included was the audio power amplifier.
I did try several measurements with my equipment connected to the speaker terminals. This had a very slight effect on the distortion, but it was negligible, under 5%. I therefore stand behind my decision to use the handset output on the MaxTrac.
I measured the distortion at several audio frequencies at both 3 kHz and 5 kHz deviation. I found that the lowest distortion was obtained with an RF level of -91 dBm, so I left it there. I also found that slight tuning (+/- 2 kHz) of the carrier frequency made a drastic difference on distortion at the higher deviation level.
|3 kHz||5 kHz|
I also measured the distortion while adjusting the RF frequency up and down by as much as 7 kHz. This table shows that there is a slight imbalance in the IF tuning on this receiver.
|400 Hz||1000 Hz|
I varied the carrier frequency at two different signal levels (20 dB quieting and 30 dB quieting) to see how much of an effect that had on the quieting measurement. The values in the table are dB Quieting:
|20 dB||30 dB|
I measured the various quieting and SINAD levels and summarized them below. A 1 kHz audio tone at 3 kHz deviation was used for all the tests except quieting. All negative values are the RF signal level in dBm required to reach the specified dB quieting or SINAD level. The column headings list the equipment model number used to measure Quieting or SINAD, as described in the Sensitivity-measuring Equipment section above.
Summary of Measurements and Conclusions:
I found very little difference between the readings I got with the SINAD and the distortion meter. The values were so close, the variations could be due to equipment temperature and rounding errors (I only recorded signal levels in whole dBm values). The numbers I got with the 20 dB quieting measurement were also extremely close to the SINAD/Distortion values, certainly within +/- 1 dB of each other. It was nice to see that the distortion and SINAD values also matched the recently calibrated IFR service monitor.
The 20 dB quieting values showed less sensitivity to signals being off frequency. This method also did not take into account any problems in the IF, detector, or audio stages of the receiver. In fact, if the audio amplifier had a lot of distortion with tones, this measurement would not have shown that.
The SINAD and distortion values were very sensitive to the carrier frequency. Slight errors here were immediately detected by the equipment. Any problems with the audio stages would have reduced the SINAD values, requiring a higher RF signal to compensate.
It was never my intention to force any method of measuring sensitivity upon the reader. I merely wanted to show the various techniques and compare the results. It would appear that one could measure absolute sensitivity using any of the test procedures and equipment mentioned in this article with fairly accurate results. For true receiver performance, the SINAD method takes into account more of the radio's circuitry, but if all you want to do is compare sensitivity, for example, with and without a preamp, the 20 dB quieting method is just as reliable. Of course, if the preamp adds a lot of noise to the received signal, the quieting measurement could be affected and give a false reading. This would have a similar worsening effect on the SINAD value as well. I still like the 20 dB quieting method because I find it easier to determine by ear: when I've got enough signal to eliminate all the audible crackles, that's 20 dB quieting. Also, 12 dB SINAD is a fairly noisy signal and rather touchy; 1 dB of RF input level change can make a 3 dB difference in the SINAD value.
I recorded six files, all at 11 kilobits per second (8-bit mono) sampling rate, at several RF input signal levels, using the handset output pin of the MIC connector on my 450 MHz MaxTrac. Each file contains about five seconds with no signal, followed by about five seconds with a signal. These are all .WAV files about 115kb in size. When you click on a file it will be downloaded to your system and whatever application your computer is configured to use to play .WAV files will open up and play them. As we (the repeater-builder.com staff and myself) don't know your computer, we can't help you if they don't work. Before this page was made public we had Windows, Macintosh and Linux systems test the prototype page and the files work on all of them without error.
12 dB Quieting: -122 dBm
12 dB SINAD: -120 dBm
20 dB Quieting: -118 dBm
20 dB SINAD: -118 dBm
30 dB Quieting: -109 dBm
30 dB SINAD: -109 dBm
Credits and Acknowledgements:
Thanks go to Rob Barba for driving 30 minutes each way to my house with his IFR 1200SS service monitor one evening so I could spend 10 minutes acquiring data for this article.
Thanks also go to the various people who read the article and offered constructive comments.
Of course, thanks always go to Mike Morris WA6ILQ of the Repeater-Builder staff for converting this article to HTML.
The author can be contacted at: his-callsign [ at ] comcast [ dot ] net.
Back to the top of the page
Up one level
Back to Home
Article text © Copyright 2006 by Robert W. Meister WA1MIK
This web page, this web site, the information presented in and on its pages and in these modifications and conversions is © Copyrighted 1995 and (date of last update) by Kevin Custer W3KKC and multiple originating authors. All Rights Reserved, including that of paper and web publication elsewhere.