The AM Forum
March 19, 2024, 06:12:39 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Signal Generator Technique  (Read 11479 times)
0 Members and 1 Guest are viewing this topic.
ashart
Guest
« on: October 02, 2011, 11:08:19 PM »

Some bumming around the internet has suggested that a certain John Wilson, in May of 2002, published an article in Short Wave Magazine, explaining why one should make receiver sensitivity (S+N)/N tests by turning the generator modulation on and off, rather than by turning the modulated R.F. on and off.

Anybody have a copy of that article, a suggestion on where to get one, or a reliable rationale for Wilson's postion?

Tnx es 73 de al hart, w8vr.


Logged
w3jn
Johnny Novice
Administrator
Member

Offline Offline

Posts: 4619



« Reply #1 on: October 02, 2011, 11:17:29 PM »

My HP 8935 does it that way - turns the audio on and off at about a 1 second rate.  It does that because it needs to find the baseline noise to compare to the (signal+noise) level.
Logged

FCC:  "The record is devoid of a demonstrated nexus between Morse code proficiency and on-the-air conduct."
ashart
Guest
« Reply #2 on: October 02, 2011, 11:52:14 PM »

Thank you for the prompt response.

It's good to hear that your 8935 does it automatically.  While my signal generator needs manual operator-input to turn anything on and off, that feature of your 8935 supports the reported position of the writer John Wilson. 

However, the important question is why your 8935 does it that way.  Why is it better to get a baseline with the carrier on and the audio off, than to get the baseline with both carrier and audio off?

-al h

Logged
w3jn
Johnny Novice
Administrator
Member

Offline Offline

Posts: 4619



« Reply #3 on: October 03, 2011, 12:12:40 AM »

You're not concerned with the noise output of the receiver when it has no signal at the input.  What you're measuring is the performance of the receiver in the presence of a signal.  So, you need to measure the receiver's noise output with an unmodulated carrier to measure the "N" term.  The "signal" is actually the modulation on the carrier, because when you're measuring this, you're measuring at the audio output of the receiver. 

Typically, when doing this, you connect up the audio from the RX to the 8935, and the RF from the 8935 to the RX.  Then, crank down the 8935's RF output til you hit 10 dB (S+N)/N - that's usually the spec point.  The 8935 also will measure SINAD directly.
Logged

FCC:  "The record is devoid of a demonstrated nexus between Morse code proficiency and on-the-air conduct."
ashart
Guest
« Reply #4 on: October 03, 2011, 12:21:10 AM »

Ok, I'm quite sorry to seem like I'm nitpicking, and I understand what you are saying, but I still don't understand the "why" of it.

Why is it more appropriate to measure the performance of the receiver in the presence of signal?

Does not the RF carrier present in the absence of audio cause some quieting of the noise?  If so, is that desirable?  Why?  Are there other effects? 

-al h



Logged
WD5JKO
Member

Offline Offline

Posts: 1996


WD5JKO


« Reply #5 on: October 03, 2011, 11:00:02 AM »

Why is it more appropriate to measure the performance of the receiver in the presence of signal?
Does not the RF carrier present in the absence of audio cause some quieting of the noise?  If so, is that desirable?  Why?  Are there other effects? 

   There is always a signal on a receiver input. Could be QRN, atmospherics, thermal noise, tube/transistor noise, etc. We want to find the minimum carrier level inserted into the receiver that has detectable audio out into the speaker that rides over the noise level by some number of DB (usually 10 db when the on/off modulation is set to 30%). A receiver that scores well has to have high gain at low internal noise, and an extremely good AM detector that works well at low signal levels. Yes it is very desirable to have a carrier quiet down a receiver. The poorer the receiver is though, the more carrier it takes to get the same level of quieting. Who would listen to AM broadcast if the receiver was not quieted by the strength of the carrier.

Jim
WD5JKO
Logged
w3jn
Johnny Novice
Administrator
Member

Offline Offline

Posts: 4619



« Reply #6 on: October 03, 2011, 11:20:31 AM »


Why is it more appropriate to measure the performance of the receiver in the presence of signal?


-al h





To add to WD5JKO's comments, it would be meaningless to measure anything else.  There is nothing that the receiver, in the absence of a signal, can tell you regarding its performance except measuring the baseline noise output for a noise *figure* measurement.  Still, you need other measurements WITH a signal present to complete this.
Logged

FCC:  "The record is devoid of a demonstrated nexus between Morse code proficiency and on-the-air conduct."
vincent
Guest
« Reply #7 on: October 11, 2011, 03:52:02 PM »

If interested, I have that article by John Wilson (and others of its interesting articles) that I can scan and send by email, unfortunately I can’t post it for copyright reasons. That article named “Simply the best?” was referred to the Collins R-390A receiver and as already said in this thread was published by Short Wave Magazine in May 2002.
Logged
ashart
Guest
« Reply #8 on: October 11, 2011, 04:01:46 PM »

Vincent:  I'd be more than happy to read it. 

Please email it to al@w8vr.org.

Thank you very much for your generosity.

73 de al hart.

Logged
vincent
Guest
« Reply #9 on: October 11, 2011, 04:42:21 PM »

Hi Al,
I've just sent it to you.
Regards.
Vincent.
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #10 on: October 12, 2011, 07:38:39 PM »

Hi Al. Your question is an excellent one. Although the answers you received are on target, there is more detail that is really nice to know. So... if you ever have an opportunity, get a hold of M. Schwartz text, "Information Transmission, Modulation, and Noise. Should be available at local library. See pg.s Section 6.9. The short answer is the noise properties of an AM detector is partly due to the presence of the carrier alone. If the carrier is absent, the noise power of the detector is significantly different. The details and it is a great read are in the text.
Logged
ashart
Guest
« Reply #11 on: October 13, 2011, 03:29:14 AM »

W4AMV:  Thank you Alan, for your suggestion.  I was able to acauire a copy of the Wilson article, thanks to member Vincent, and I'll expand on the information there by reading the Schwartz writings.

73, OM.

al hart
www.w8vr.org

Logged
flintstone mop
Contributing
Member
*
Offline Offline

Posts: 5055


« Reply #12 on: October 13, 2011, 06:54:34 AM »

Another 'technique' is once you are in the ballpark of tuning and the receiver is responding to an RF input, you should keep reducing the initial output of the generator and re-do the alignment. Sometimes you get more tuning peaks through the RF and I.F. circuits and the AGC is not taking over.
I always reduce the generator for the least discernible sound in the headphones. I usually modulate the generator with a small amount of a 1kc tone and an AC volt meter on the output will display the little extra peak you are looking for.
The R390A alignment is pretty involved, but I can always bring her back to usually .8 microvolts from the BC band to 75M and at 40M-10 M I can achieve  .5microvolts.
Even 1 microvolt is enough for the HF bands......VHF and up you got to get .2 minimum
fred
Logged

Fred KC4MOP
ashart
Guest
« Reply #13 on: October 13, 2011, 03:39:24 PM »

The question I asked in this thread was, implicitly, “Why should one measure receiver sensitivity by alternately turning the modulation on and off on a steady carrier, instead of merely turning the modulated carrier on and off.”

Specifically, I asked about an article I’d heard of, that addressed that subject, written  by a certain John Wilson.  Soon thereafter, member Vincent was kind enough to send a copy to me.

I thought that the essence of the relevant portion of the article might be of interest to some of the respondents to this thread, so here’s my understanding of Wilson:


The common diode detector has a non-linear behavior.  There is a region of its characteristic curve, close to the origin, i.e., where the input voltage is low, that has a wholly different slope than the main part of the curve, which approaches a straight line.

Absent any generator signal, the diode sees only internal receiver noise.  That noise, in a well-designed receiver, is generally so low that it lies in the lower region of the diode curve.  Then, when a signal is applied, the much larger generator signal lies higher on the diode charistic curve, in the linear region.  This change in the slope of the diode curve causes inaccuracy in the measurement of the receiver’s sensitivity.

So, explains Wilson, by leaving the generator on all the time, and just turning the modulation on and off, the diode is always operating in the linear region.  Therefore, the input voltage applied through a proper impedance to the antenna terminals, for a specified (S+N)/N, is the proper measure of receiver sensitivity.


I’ve also learned from member W4AMV, of a book entitled “Information Transmission, Modulation and Noise” by M. Schwartz, that is said to contain some relevant information.  As soon as I get a copy, and if there’s any further relevant information, I’ll post it.


73, and thanks to each of the forum members who helped me here!


-al hart

al@w8vr.org  and  www.w8vr.org
Logged
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8304



WWW
« Reply #14 on: October 14, 2011, 12:23:45 PM »

I turn the audio on and off and just listen to the speaker in order to find the minimum detectable signal. Right or wrong I think its a scientific test.
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
flintstone mop
Contributing
Member
*
Offline Offline

Posts: 5055


« Reply #15 on: October 16, 2011, 07:33:16 AM »

The question I asked in this thread was, implicitly, “Why should one measure receiver sensitivity by alternately turning the modulation on and off on a steady carrier, instead of merely turning the modulated carrier on and off.”

Specifically, I asked about an article I’d heard of, that addressed that subject, written  by a certain John Wilson.  Soon thereafter, member Vincent was kind enough to send a copy to me.

I thought that the essence of the relevant portion of the article might be of interest to some of the respondents to this thread, so here’s my understanding of Wilson:


The common diode detector has a non-linear behavior.  There is a region of its characteristic curve, close to the origin, i.e., where the input voltage is low, that has a wholly different slope than the main part of the curve, which approaches a straight line.

Absent any generator signal, the diode sees only internal receiver noise.  That noise, in a well-designed receiver, is generally so low that it lies in the lower region of the diode curve.  Then, when a signal is applied, the much larger generator signal lies higher on the diode charistic curve, in the linear region.  This change in the slope of the diode curve causes inaccuracy in the measurement of the receiver’s sensitivity.

So, explains Wilson, by leaving the generator on all the time, and just turning the modulation on and off, the diode is always operating in the linear region.  Therefore, the input voltage applied through a proper impedance to the antenna terminals, for a specified (S+N)/N, is the proper measure of receiver sensitivity.


I’ve also learned from member W4AMV, of a book entitled “Information Transmission, Modulation and Noise” by M. Schwartz, that is said to contain some relevant information.  As soon as I get a copy, and if there’s any further relevant information, I’ll post it.


73, and thanks to each of the forum members who helped me here!


-al hart

al@w8vr.org  and  www.w8vr.org
I have been waiting several days after your reply to reply.......so my words don't get stupid.....
I realize what you were implicitly asking in the beginning of this thread. You were questioning the need to cut modulation on-off from a signal generator....I can read
After your quest was answered I chimed into the Generator or "tuning technique' , (which is the title of this thread!!), about reducing the RF level each time you go through the alignment of the various sections in a receiver to get the best sensitivity. You don't align a radio with a 5 microvolt signal and your done!!! Some folks here might may already know the signal to noise test of alignment and I was adding to your thread. Now maybe you didn't appreciate that part.
I notice your kinda new here and we welcome you Al, but don't start flaming folks for making suggestions or even a little poking, as I have noticed in some of your other threads. Take it down several notches.....turn your sensitivity down..... a lot.

Fred
Logged

Fred KC4MOP
Rob K2CU
Contributing
Member
*
Offline Offline

Posts: 346


« Reply #16 on: October 18, 2011, 02:07:12 PM »

You can find some text excerpts here:

http://books.google.com/books?id=oRSHWmaiZwUC&pg=PA100&lpg=PA100&dq=Information+Transmission,+Modulation,+and+Noise+AM+detector&source=bl&ots=1OCug_sh-4&sig=yYIWxidDJawTXTgRDLe4CY5CVJ0&hl=en&ei=V7KdTuShI8n10gGL5fW2CQ&sa=X&oi=book_result&ct=result&resnum=2&sqi=2&ved=0CCkQ6AEwAQ#v=onepage&q=Information%20Transmission%2C%20Modulation%2C%20and%20Noise%20AM%20detector&f=false

It basically explains why there is an increase in noise from an AM envelope detector in the presence of a carrier. Keep in mind that this happens when there is no AGC action. So, if you want to measure performance in terms of S+N to N ratio, you would be correct in using the modulation on/off technique. This effect is also exaggerated by having twice the bandwidth in AM over that of SSB. If you want to get a measure of the noise contributed in your Am detector, you can do the following.

with AGC off, set radio to SSB and measure sensitivity at 10dB S+N/N
With radio still in SSB mode, set generator to AM at 30% modulation and zero beat carrier and set carrier level so that the one Am sideband gives you 10dB S+N/N.  The carrier will be at about 16.5dB more signal than that of SSB signal. This is the simple math of SSB vs. DSB plus carrier.   .15uV in one sideband equates to a 1uV carrier with 30% modulation.
Change radio over to Am mode and determine increased carrier level required for 10dB S+N to N using the modulation on/off technique.
The difference is the added loss or noise increase due to the AM detection process.

Note that the measurement will require the AGC to be turned off. This is to ensure that no AGC action has started. Some radios have their AGC threshold set too low and start desensing the radio before an adequate signal to noise ratio has been achieved for AM.

This technique can be used to compare different types of AM detectors, if one is interested.
 
 
Logged
W4AMV
Member

Offline Offline

Posts: 672


« Reply #17 on: October 18, 2011, 02:43:44 PM »

Its great to see that this is a classic and as an IEEE re-issue. Thanks for the find. I have the original undergraduate text and it appears more detail has been added to the noise and detection chapter. I suspect that my original text was revised and this link is to that revised edition. I cannot over emphasis the hours spent in reading are re reading these chapters before the light began to click. If you can stick through the mathematics the result is quite rewarding!
Logged
ashart
Guest
« Reply #18 on: October 18, 2011, 04:36:06 PM »

Hi, Rob -

Thanks for the link.  As Alan, W4AMV, noted, I'll have to wade through the  arithmetic - fortunately, I've been there and done that sort of arithmetic, but it was 50 years ago - so it'll be slow going.

Also, Alan had previously recommended Schwartz's book to me, so my only complaint now is why you sent that link the day after I ordered the book!  :) :)  Well, now I can start to digest the stuff while waiting for the book to come.

73 Rob, es tnx agn!

-al hart
Logged
WD5JKO
Member

Offline Offline

Posts: 1996


WD5JKO


« Reply #19 on: October 18, 2011, 05:43:27 PM »



The Technical Surveillance Counter Measures (TSCM) folks deal a lot with receiver sensitivity in the presence of noise. They have a neat link relating the theoretical noise floor for a given resolution bandwidth. Here is a link to this discussion:

http://www.tscm.com/TSCM101noise.html

Remember with 50 ohm systems, -107 Dbm is 1 Uv.

So using the topic of "Signal Generator Technique", the receiver bandwidth plays a huge role in the sensitivity recorded.

Jim
WD5JKO
Logged
Rob K2CU
Contributing
Member
*
Offline Offline

Posts: 346


« Reply #20 on: October 19, 2011, 11:46:25 AM »

Sorry Al,

I don't get to check into this forum everyday...my bad.

That link was at the end of the third Google search page.
Logged
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.091 seconds with 19 queries.