Hello The Net:
Upon the suggestion of Domenico, I8CVS, I conducted some tests to compare SINAD to Noise figure testing.
Noise figure testing was performed using a HP8970B with 346 noise source on a ARR preamplifier, one of their switched units, model SP144VDG. Performance was optimized at 144 MHz. for a noise figure of 0.9 dB and gain of 25 dB.
The preamp was added before the IC-910H, which was set to 144.000 MHz in a SSB (USB) mode.
SINAD , SIgnal + Noise + Audio Distortion testing allows a complete receiver to be tested from the RF input port, thru the IF's and demodulation to audio.
For SSB mode testing I injected a low level RF carrier, (~ -120 dBm), no modulation, into the preamplifer with RX. I tuned the RX frequency to 1000 Hz less than was actually injected.
example: for a RF of 144.000 MHz, I tuned the RX, in a SSB (USB) mode, to 143.999 MHz. SINAD instruments use a 1000 Hz tone for further processing.
Connecting the audio output to a SINAD meter, in my case a Helper Instruments, Inc. model Sinadder Linear 5, I was able to get a SINAD reading from the meter.
While monitoring the SINAD meter, I tweaked the preamplifier tuning and could not improve the original SINAD measurement.
Conclusion: If the noise figure is optimized, the SINAD measurement will also be optimized.
An additional test was performed, adding a fixed 10 dB BNC attenuator after the preamplifier and before the RX. The SINAD measurement did not change. This indicated that at least 10 dB of excess gain was present and can be eliminated. It is always best to minimize system gains to maximize the ability to handle strong signals and to reduce higher order intermodulation distortion. Minimize system interstage gains until the system SINAD sensitivity or the noise figure is slightly degraded.
In this example, if I was able to reduce 10 dB of excess gain, this would result in lowering the 3rd order intermodulation distortion by 30 dB . !!!
I hope this helps folks to consider SINAD testing of their system, to verify performance.
Stan, W1LE FN41sr Cape Cod
Stan makes a good point here, especially for FM sensitivity. I worked in the 2-way business for a long time, and typically a sinad measurement would be something like the value of microvolts at the antenna input for 20 dB of quieting of the receiver. It is basically a measurement of FM receiver sensitivity. It will also show mis-aligned IF strips and detector, but for the sake of arguement in this case, lets say the receiver is functioning normally. The goal should be the minimum amount of signal input at the antenna connector quiets the receiver 20 dB.
The way it works is this: an FM modulated carrier is injected into the receiver with a modulating frequency of 1000 Hz (at typically around 30% modulation). A sinadder is simply an audio level meter that has a deep notch at 1000 Hz. Therefore, any audio that it sees that is not on 1000 hz is read as noise (or distortion). As you increase signal strength, the noise drops off, and when the signal voltage reaches 1/10th of the unsquelched noise with no signal present, the amount of signal injected into the antenna connector is read. That would be the signal necessary to quiet the receiver 20 dB.
A typical reading for sinad from a typical amateur (barefoot) rig today is such that .25 - .35 microvolts will quiet the receiver 20 dB. It will be even lower with a preamp in line.
I am not certain that a sinad reading would be of much use on an SSB receiver. Typically a similar measurement for an SSB receiver would be MDS, or "Minimum detectable Signal", which would be the amount of signal that is injected into the antenna connector that produces a faint but detectable signal in the receiver.
Basically, SINAD and noise figure measurements produce the same result. They both are a measurement of receiver sensitivity, however sinad is reliable only for an FM receiver. A sinad meter is very easily constructed, and can be purchased much more inexpensively than a noise figure meter, so therefore, if your receiver is capable of FM reception, you can tune up your system by using a sinad measurement. The only difference that I can really think of between the two measurement systems is that noise figure is an absolute measurement that is applicable to all modes of operation, whereas sinad is only applicable to FM. If, however you place a generator at the preamp input, and monitor in the FM mode as you make adjustments to your preamp, you will also be tuning for best signal to noise, which will correspond to lowest noise figure. You just will not know what THAT value is, unless it can be calculated, which I do not know the formula if it is possible to do so.
If I needed to tune a preamplifier for absolute best noise figure in a labrotory environment, then, yes, I would tune with the help of a noise figure meter, but for most amateur purposes, I see no reason why one could not tune up using sinad as the criterion.
Michael Heim Chief Engineer, Forever Broadcasting New Castle PA WKST WJST WWGY 814-671-0666 Chapter Chair, SBE-122 ARS KD0AR
--- On Tue, 12/9/08, Stan W1LE stanw1le@verizon.net wrote:
From: Stan W1LE stanw1le@verizon.net Subject: [amsat-bb] SINAD Vs. Noise Figure Testing To: AMSAT-BB@amsat.org Date: Tuesday, December 9, 2008, 3:03 PM Hello The Net:
Upon the suggestion of Domenico, I8CVS, I conducted some tests to compare SINAD to Noise figure testing.
Noise figure testing was performed using a HP8970B with 346 noise source on a ARR preamplifier, one of their switched units, model SP144VDG. Performance was optimized at 144 MHz. for a noise figure of 0.9 dB and gain of 25 dB.
The preamp was added before the IC-910H, which was set to 144.000 MHz in a SSB (USB) mode.
SINAD , SIgnal + Noise + Audio Distortion testing allows a complete receiver to be tested from the RF input port, thru the IF's and demodulation to audio.
For SSB mode testing I injected a low level RF carrier, (~ -120 dBm), no modulation, into the preamplifer with RX. I tuned the RX frequency to 1000 Hz less than was actually injected.
example: for a RF of 144.000 MHz, I tuned the RX, in a SSB (USB) mode, to 143.999 MHz. SINAD instruments use a 1000 Hz tone for further processing.
Connecting the audio output to a SINAD meter, in my case a Helper Instruments, Inc. model Sinadder Linear 5, I was able to get a SINAD reading from the meter.
While monitoring the SINAD meter, I tweaked the preamplifier tuning and could not improve the original SINAD measurement.
Conclusion: If the noise figure is optimized, the SINAD measurement will also be optimized.
An additional test was performed, adding a fixed 10 dB BNC attenuator after the preamplifier and before the RX. The SINAD measurement did not change. This indicated that at least 10 dB of excess gain was present and can be eliminated. It is always best to minimize system gains to maximize the ability to handle strong signals and to reduce higher order intermodulation distortion. Minimize system interstage gains until the system SINAD sensitivity or the noise figure is slightly degraded.
In this example, if I was able to reduce 10 dB of excess gain, this would result in lowering the 3rd order intermodulation distortion by 30 dB . !!!
I hope this helps folks to consider SINAD testing of their system, to verify performance.
Stan, W1LE FN41sr Cape Cod _______________________________________________ Sent via AMSAT-BB@amsat.org. Opinions expressed are those of the author. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: http://amsat.org/mailman/listinfo/amsat-bb
The original SINAD was a 10 dB figure which was developed at least a couple of decades after the 20 dB quieting figure. The "proper" way to measure the 20 dB quieting point was to use an AC voltmeter utilizing a cupric oxide rectifier (which was what most of the "olde tyme" VOM units used - i.e. Simpson 260). With the squelch of the FM receiver "wide open" the voltmeter was placed across the speaker and the volume control adjusted for a convenient reading (1 volt was very common). Then an on-frequency unmodulated signal was applied from a signal generator having an accurately calibrated attenuator and the signal level was adjusted to the point where 1/10th the voltage was recorded on the VOM. Since a reduction in voltage by a factor of 10 represents a 20 dB reduction this point was referred to as the "20 dB quieting" point.
Most experienced two-way radio technicians could actually make a 20 dB quieting measurement "by ear" and did not require the use of an AC voltmeter. When verified by someone watching a voltmeter while the technician did the measurement by ear virtually always resulted in a measurement that was well within 1 dB which was "close enough for government work" for field measurements.
Using a well calibrated dB meter (i.e. one made by Hewlett Packard) across the speaker will usually not result in the same reading as that done with a "plain old VOM" using the simple cupric oxide rectifier. The 20 dB quieting measurement was done by field technicians who often did not have such fancy test equipment as a calibrated dB meter.
An easy to make and a very meaningful sensitivity measurement is the LDS (least discernible signal) which involves turning on the BFO, tuning the receiver for maximum signal strength, and then reducing the signal level (using a well calibrated attenuator on the signal generator) to the point at which the signal can just be detected by ear. The reading in microvolts from the signal generator is the LDS. Of course reducing the bandwidth does affect the LDS. By doing an LDS measurement you are measuring the weakest signal that the receiver is capable of receiving through the noise generated within the receiving system.
Glen, K9STH
Website: http://k9sth.com
--- On Tue, 12/9/08, Michael Heim kd0ar@sbcglobal.net wrote:
From: Michael Heim kd0ar@sbcglobal.net
Stan makes a good point here, especially for FM sensitivity. I worked in the 2-way business for a long time, and typically a sinad measurement would be something like the value of microvolts at the antenna input for 20 dB of quieting of the receiver. It is basically a measurement of FM receiver sensitivity. It will also show mis-aligned IF strips and detector, but for the sake of arguement in this case, lets say the receiver is functioning normally. The goal should be the minimum amount of signal input at the antenna connector quiets the receiver 20 dB.
I am not certain that a sinad reading would be of much use on an SSB receiver. Typically a similar measurement for an SSB receiver would be MDS, or "Minimum detectable Signal", which would be the amount of signal that is injected into the antenna connector that produces a faint but detectable signal in the receiver.
In this modern day, soundcard programs display very weak signals much below what can be heard. The method described below can be modified using a soundcard spectrum program whereby you can lower the signal on the display to equal the noise floor and thus achieve very close to S+N/N = 0. Measuring the signal in dBm that produces this level will give one a good measure of sensitivity.
For example, my SDR-IQ is calibrated in dBm so I can see this directly. Typically with 2.1 KHz wide SSB the SDR-IQ has a noise floor of -130 dBm at 28-MHz into a 50-ohm load. On an antenna you see a much higher noise floor that shows the amount of noise the RF environment adds.
When I use a 144/28 MHz convertor before the SDR-IQ, the absolute reading is likely shifted by the gain and NF of the convertor. Use of a calibrated signal generator will give a good benchmark for sensitivity. One must insert a test signal using a directional coupler to know what actual sensitivity is when connected to an antenna.
73, Ed - KL7UW
At 01:53 PM 12/9/2008, Glen Zook wrote:
The original SINAD was a 10 dB figure which was developed at least a couple of decades after the 20 dB quieting figure. The "proper" way to measure the 20 dB quieting point was to use an AC voltmeter utilizing a cupric oxide rectifier (which was what most of the "olde tyme" VOM units used - i.e. Simpson 260). With the squelch of the FM receiver "wide open" the voltmeter was placed across the speaker and the volume control adjusted for a convenient reading (1 volt was very common). Then an on-frequency unmodulated signal was applied from a signal generator having an accurately calibrated attenuator and the signal level was adjusted to the point where 1/10th the voltage was recorded on the VOM. Since a reduction in voltage by a factor of 10 represents a 20 dB reduction this point was referred to as the "20 dB quieting" point.
Most experienced two-way radio technicians could actually make a 20 dB quieting measurement "by ear" and did not require the use of an AC voltmeter. When verified by someone watching a voltmeter while the technician did the measurement by ear virtually always resulted in a measurement that was well within 1 dB which was "close enough for government work" for field measurements.
Using a well calibrated dB meter (i.e. one made by Hewlett Packard) across the speaker will usually not result in the same reading as that done with a "plain old VOM" using the simple cupric oxide rectifier. The 20 dB quieting measurement was done by field technicians who often did not have such fancy test equipment as a calibrated dB meter.
An easy to make and a very meaningful sensitivity measurement is the LDS (least discernible signal) which involves turning on the BFO, tuning the receiver for maximum signal strength, and then reducing the signal level (using a well calibrated attenuator on the signal generator) to the point at which the signal can just be detected by ear. The reading in microvolts from the signal generator is the LDS. Of course reducing the bandwidth does affect the LDS. By doing an LDS measurement you are measuring the weakest signal that the receiver is capable of receiving through the noise generated within the receiving system.
Glen, K9STH
Website: http://k9sth.com
--- On Tue, 12/9/08, Michael Heim kd0ar@sbcglobal.net wrote:
From: Michael Heim kd0ar@sbcglobal.net
Stan makes a good point here, especially for FM sensitivity. I worked in the 2-way business for a long time, and typically a sinad measurement would be something like the value of microvolts at the antenna input for 20 dB of quieting of the receiver. It is basically a measurement of FM receiver sensitivity. It will also show mis-aligned IF strips and detector, but for the sake of arguement in this case, lets say the receiver is functioning normally. The goal should be the minimum amount of signal input at the antenna connector quiets the receiver 20 dB.
I am not certain that a sinad reading would be of much use on an SSB receiver. Typically a similar measurement for an SSB receiver would be MDS, or "Minimum detectable Signal", which would be the amount of signal that is injected into the antenna connector that produces a faint but detectable signal in the receiver.
Sent via AMSAT-BB@amsat.org. Opinions expressed are those of the author. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: http://amsat.org/mailman/listinfo/amsat-bb
Michael Heim wrote:
Stan makes a good point here, especially for FM sensitivity. I worked in the 2-way business for a long time, and typically a sinad measurement would be something like the value of microvolts at the antenna input for 20 dB of quieting of the receiver. It is basically a measurement of FM receiver sensitivity. It will also show mis-aligned IF strips and detector, but for the sake of arguement in this case, lets say the receiver is functioning normally. The goal should be the minimum amount of signal input at the antenna connector quiets the receiver 20 dB.
http://www.repeater-builder.com/tech-info/20dbquieting.html
Often times folks building repeaters and/or doing other weak-signal (contesting) work at VHF and above have great documentation and information that applies just as much to satellite operation.
It's always worth "popping our heads up" from our own little RF worlds and looking around a bit once in a while. (Like Prairie Dogs.)
I agree with you, a SINAD measurement on an SSB receiver is kinda not the right idea. It'll show you something, but what... I'm not quite sure. It CAN be used for a comparison measurement for changes, though.
Thus, the best thing about learning how to make a SINAD measurement, measuring or at least calculating Noise Figure, or WHATEVER measurements is this:
You can make CHANGES and see if you made things better or worse.
Why is that important?
It changes your hobby from subjective ("that sounds better") to quantitative ("I made my home VHF setup 3dB more sensitive on VHF today"), which is a sign you're trying harder than the majority to have a great receiver setup and antenna system.
Another useful measurement is to find the receiver's sensitivity while injecting the test signal while connected to an outside antenna, after doing it without... to see how local noise/intermod/etc., affects your receiver.
A home-made "Iso-Tee" set up for a known amount of loss on your test setup, or a directional coupler with a known amount of loss are needed for that test.
Nate WY0X
As far as I know SINAD is only used for FM sensitivity measurements, whereas the older standard that is used for AM is 20-dB quieting. Of course SSB is another form of AM. But as an (relatively) inexpensive tool for tuning RF stages for maximum sensitivity (which should coincide with minimum NF) the SINADer should work fine. Most SSB VHF receivers also include FM mode. Merely, tune for 12-dB SINAD and it should result in the best sensitivity in for SSB, as well.
If you are wanting to know the absolute noise figure then you will need to measure it using a NF meter or other standard method. But since our goal is to achieve the best sensitivity, whatever NF that results in, it seems to me that SINAD can be used for adjustments.
I have performed so many FM alignments over the years, I can tell when I hear 12 dB SINAD without use of the meter with pretty good accuracy. I perform probably 500+ of these a year. BTW state of the art FM receivers run about 0.125uV these days (most ham VHF radios are this sensitive). AT UHF they are often 0.25uV. The old standard of 25-years ago was 0.35uV for VHF and 0.50uV for UHF. MOSFet front-ends on commercial radios are pretty good. A good GasFet preamp will achieve a NF of 0.5 dB which is about -145dBm at SSB bandwidths. I don't convert dBm to uV much as my equipment has both scales on them, but I know that 0.25uV = -119 dBm FM.
73, Ed - KL7UW
At 12:48 PM 12/9/2008, Michael Heim wrote:
Stan makes a good point here, especially for FM sensitivity. I worked in the 2-way business for a long time, and typically a sinad measurement would be something like the value of microvolts at the antenna input for 20 dB of quieting of the receiver. It is basically a measurement of FM receiver sensitivity. It will also show mis-aligned IF strips and detector, but for the sake of arguement in this case, lets say the receiver is functioning normally. The goal should be the minimum amount of signal input at the antenna connector quiets the receiver 20 dB.
The way it works is this: an FM modulated carrier is injected into the receiver with a modulating frequency of 1000 Hz (at typically around 30% modulation). A sinadder is simply an audio level meter that has a deep notch at 1000 Hz. Therefore, any audio that it sees that is not on 1000 hz is read as noise (or distortion). As you increase signal strength, the noise drops off, and when the signal voltage reaches 1/10th of the unsquelched noise with no signal present, the amount of signal injected into the antenna connector is read. That would be the signal necessary to quiet the receiver 20 dB.
A typical reading for sinad from a typical amateur (barefoot) rig today is such that .25 - .35 microvolts will quiet the receiver 20 dB. It will be even lower with a preamp in line.
I am not certain that a sinad reading would be of much use on an SSB receiver. Typically a similar measurement for an SSB receiver would be MDS, or "Minimum detectable Signal", which would be the amount of signal that is injected into the antenna connector that produces a faint but detectable signal in the receiver.
Basically, SINAD and noise figure measurements produce the same result. They both are a measurement of receiver sensitivity, however sinad is reliable only for an FM receiver. A sinad meter is very easily constructed, and can be purchased much more inexpensively than a noise figure meter, so therefore, if your receiver is capable of FM reception, you can tune up your system by using a sinad measurement. The only difference that I can really think of between the two measurement systems is that noise figure is an absolute measurement that is applicable to all modes of operation, whereas sinad is only applicable to FM. If, however you place a generator at the preamp input, and monitor in the FM mode as you make adjustments to your preamp, you will also be tuning for best signal to noise, which will correspond to lowest noise figure. You just will not know what THAT value is, unless it can be calculated, which I do not know the formula if it is possible to do so.
If I needed to tune a preamplifier for absolute best noise figure in a labrotory environment, then, yes, I would tune with the help of a noise figure meter, but for most amateur purposes, I see no reason why one could not tune up using sinad as the criterion.
Michael Heim Chief Engineer, Forever Broadcasting New Castle PA WKST WJST WWGY 814-671-0666 Chapter Chair, SBE-122 ARS KD0AR
--- On Tue, 12/9/08, Stan W1LE stanw1le@verizon.net wrote:
From: Stan W1LE stanw1le@verizon.net Subject: [amsat-bb] SINAD Vs. Noise Figure Testing To: AMSAT-BB@amsat.org Date: Tuesday, December 9, 2008, 3:03 PM Hello The Net:
Upon the suggestion of Domenico, I8CVS, I conducted some tests to compare SINAD to Noise figure testing.
Noise figure testing was performed using a HP8970B with 346 noise source on a ARR preamplifier, one of their switched units, model SP144VDG. Performance was optimized at 144 MHz. for a noise figure of 0.9 dB and gain of 25 dB.
The preamp was added before the IC-910H, which was set to 144.000 MHz in a SSB (USB) mode.
SINAD , SIgnal + Noise + Audio Distortion testing allows a complete receiver to be tested from the RF input port, thru the IF's and demodulation to audio.
For SSB mode testing I injected a low level RF carrier, (~ -120 dBm), no modulation, into the preamplifer with RX. I tuned the RX frequency to 1000 Hz less than was actually injected.
example: for a RF of 144.000 MHz, I tuned the RX, in a SSB (USB) mode, to 143.999 MHz. SINAD instruments use a 1000 Hz tone for further processing.
Connecting the audio output to a SINAD meter, in my case a Helper Instruments, Inc. model Sinadder Linear 5, I was able to get a SINAD reading from the meter.
While monitoring the SINAD meter, I tweaked the preamplifier tuning and could not improve the original SINAD measurement.
Conclusion: If the noise figure is optimized, the SINAD measurement will also be optimized.
An additional test was performed, adding a fixed 10 dB BNC attenuator after the preamplifier and before the RX. The SINAD measurement did not change. This indicated that at least 10 dB of excess gain was present and can be eliminated. It is always best to minimize system gains to maximize the ability to handle strong signals and to reduce higher order intermodulation distortion. Minimize system interstage gains until the system SINAD sensitivity or the noise figure is slightly degraded.
In this example, if I was able to reduce 10 dB of excess gain, this would result in lowering the 3rd order intermodulation distortion by 30 dB . !!!
I hope this helps folks to consider SINAD testing of their system, to verify performance.
Stan, W1LE FN41sr Cape Cod _______________________________________________ Sent via AMSAT-BB@amsat.org. Opinions expressed are those of the author. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: http://amsat.org/mailman/listinfo/amsat-bb
Sent via AMSAT-BB@amsat.org. Opinions expressed are those of the author. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: http://amsat.org/mailman/listinfo/amsat-bb
----- Original Message ----- From: "Stan W1LE" stanw1le@verizon.net To: AMSAT-BB@amsat.org Sent: Tuesday, December 09, 2008 9:03 PM Subject: [amsat-bb] SINAD Vs. Noise Figure Testing
Hello The Net:
Upon the suggestion of Domenico, I8CVS, I conducted some tests to compare SINAD to Noise figure testing.
Noise figure testing was performed using a HP8970B with 346 noise source on a ARR preamplifier, one of their switched units, model SP144VDG. Performance was optimized at 144 MHz. for a noise figure of 0.9 dB and gain of 25 dB.
The preamp was added before the IC-910H, which was set to 144.000 MHz in a SSB (USB) mode.
SINAD , SIgnal + Noise + Audio Distortion testing allows a complete receiver to be tested from the RF input port, thru the IF's and demodulation to audio.
For SSB mode testing I injected a low level RF carrier, (~ -120 dBm), no modulation, into the preamplifer with RX. I tuned the RX frequency to 1000 Hz less than was actually injected.
example: for a RF of 144.000 MHz, I tuned the RX, in a SSB (USB) mode, to 143.999 MHz. SINAD instruments use a 1000 Hz tone for further processing.
Connecting the audio output to a SINAD meter, in my case a Helper Instruments, Inc. model Sinadder Linear 5, I was able to get a SINAD reading from the meter.
While monitoring the SINAD meter, I tweaked the preamplifier tuning and could not improve the original SINAD measurement.
Conclusion: If the noise figure is optimized, the SINAD measurement will also be optimized.
An additional test was performed, adding a fixed 10 dB BNC attenuator
after
the preamplifier and before the RX. The SINAD measurement did not
change.
This indicated that at least 10 dB of excess gain was present and can be eliminated. It is always best to minimize system gains to maximize the ability to handle strong signals and to reduce higher order intermodulation distortion. Minimize system interstage gains until the system SINAD sensitivity or the noise figure is slightly degraded.
In this example, if I was able to reduce 10 dB of excess gain, this would result in lowering the 3rd order intermodulation distortion by 30 dB . !!!
I hope this helps folks to consider SINAD testing of their system, to verify performance.
Stan, W1LE FN41sr Cape Cod
Hi Stan, W1LE
Tank you very much to share with us your very interesting report. I have this letter copied in a file for record.
73" de
i8CVS Domenico
participants (6)
-
Edward Cole
-
Glen Zook
-
i8cvs
-
Michael Heim
-
Nate Duehr
-
Stan W1LE