Fw: Re: Fw: Re: [Test-Equipment] Spectrum Analyzer BW measurement

Sat, 31 Jan 2004 09:59:40 -0800

Fascinating.. Let me try to say what I read below and understand about 
all this.

  1).  The display width is increased by the analyzer bandwidth with the 
increase getting larger for large signals. Somehow I understand this for 
a few minutes and then lose it again only to recover a few minutes 
later. Tough concept.

  2). The display amplitude is the integral (average) of the energy in 
the filter at any give instant. So the correlation between the filter 
bandwidth and the display is dependent upon the frequency slope of the 
energy, making it very hard to correct for a different filter bandwidth.

  3). The analyzer sweep speed gets in to the act, by producing a 
sampling of the process and large errors if the speed is too high.

WBob

Gary Schafer wrote:
> Kees,
> 
> Another thing that may help with the filter widths. Look at the 
> unmodulated signal again and then trace that curve on a transparent 
> paper including the db scale.
> 
> When looking at a signal with that same filter, put the transparent 
> curve on the screen but UPSIDE-DOWN. You can then correlate how far out 
> a particular signal is going to be on the filter skirt due to the 
> strength. You will see that a weak signal will fall in the narrower part 
> of the filter but a stronger signal (by x db) could occupy more of the 
> filter.
> 
> A signal that is around 3 db out of the noise, which may be 50 to 60 db 
> below the main signal, is going to have the response of the filter at 
> the -3 db points. so it is going to be narrow at weak signals.
> 
> This works if the weak signal is separated from the large signal. In 
> other words if the curve or slope of the signal is not continuos from 
> the stronger signal. Not so it looks to be one signal but separate big 
> signal and separate small signal.
> 
> This is where the higher resolution (narrower filters) come in. You can 
> separate the weaker signals to know that there is something there rather 
> than just one strong signal that looks wide because of the filter skirt.
> 
> Once you know that you have a separate signal way out from the main 
> signal then you can use the inverse curve of the filter to figure out 
> how much the filter is widening the signal.
> 
> But again, if that separate signal (has its own curve on each side of 
> it) is down near the noise then the -3 db filter bandwidth will be 
> pretty close.
> 
> 73
> Gary  K4FMX
> 
> 
> 
> [email protected] wrote:
> 
>> Say Jim, since you also have a Marconi TF-2370......
>>
>> What do you get on your's when measuring the BW of the internal 10Mhz
>> signal at -3dB, -25dB, -45dB and -55dB when using the "2" (normal) and
>> "3" (wide 10x) position ?  
>> I get:
>> -3dB 500Hz BW and 4.7Khz BW (+4.2Khz)
>> -25dB 1.6Khz BW and 12.6Khz BW (+11Khz)
>> -45dB 2.4Khz BW and 21.6Khz BW (+19.2Khz) -55dB 3.2Khz BW and 30Khz BW 
>> (+26.8Khz)
>>
>> Some of the above are "a little difficult to measure" but it basically
>> holds to the "wide" position being 10x the "normal" position, which it
>> should be. But what this indicates to me, is that (me) measuring a real
>> 20Khz BW signal with the 500Hz RBW may measure 23Khz BW at -55dB, but
>> measuring the same signal with a 5Khz RBW may measure 45Khz BW at -55dB.
>> That is what I see when measuring these transmitters ....around 17Khz in
>> the "normal" 500Hz RBW position and roughly 42Khz BW in the 5KHz RBW
>> "wide" position. I just want 1) to make sure my SA is OK and 2) how do I
>> correlate to a -55dB, 3Khz RBW specification.
>> If I were a rough estimating man, I'd say 3Khz is roughly 6x "normal", so
>> about 18Khz plus the 20KHz signal, so at -55dB a real 20Khz BW signal
>> would measure approx 23Khz with a 500Hz RBW, 38Khz with a 3Khz RBW, and
>> 45Khz with a 5Khz RBW. How does anyone get any good correlation data with
>> that scheme ??? I sure can see why you really should just go back to the
>> Old Rule of reducing the RBW/BW number down to a few % for good BW data
>> and let 'er rip.
>>
>> I was asked if this was a paying job. No, I'm early retired and really
>> like this stuff. The transmitters being worked on are Radio Control units
>>
>> for the club which I just decided to look at and maybe stir up some
>> interest with the appliance operators. But I need some more education
>> first.
>>
>> 73 Kees K5BCQ  
>>    
>> --------- Forwarded message ----------
>> From: [email protected]
>> To: [email protected]
>> Cc: [email protected],[email protected]
>> Date: Fri, 30 Jan 2004 15:53:43 -0600
>> Subject: Re: Fw: Re: [Test-Equipment] Spectrum Analyzer BW measurement
>> Message-ID: <[email protected]>
>>
>> Kees and group,
>> I would guess it depends on the shape factor of the filter.  The Marconi
>> uses crystal filters, should be very steep skirts!  This correlation
>> issue is  an ongoing problem when you look at specifications closely. 
>> Typically, these issues are ignored or at best overlooked.
>> Comments?
>> Regards from Arkansas,
>> Jim
>> An interesting side note, I joined this group when I started
>> troubleshooting my Marconi SA.  I just plugged through and found the 10
>> ICs that had given up the ghost in the logic unit and the uncountable
>> open via connections.  Took months of off and on effort.  Now I am
>> collecting test equipment to check power levels. 
>> On Fri, 30 Jan 2004 14:57:27 -0600 [email protected] writes:
>>
>>> Wait a minute, that "actual BW is roughly the measured BW minus the RBW
>>> applies only to the -3dB point". The -55dB point of the internal
>>> oscillator, when measured with a 500Hz RBW is 3.2Khz and when measured
>>> with a 5Khz RBW it's 30Khz. Approx a 10x relationship which is the
>>> specified RBW ratio. So now how do I correlate to a 3Khz RBW measured on
>>> a different SA ?
>>> 73 Kees K5BCQ  
>>
>>
>> _______________________________________________
>> Test-Equipment mailing list
>> [email protected]
>> http://mailman.qth.net/mailman/listinfo/test-equipment
>>
> 
> 
> 
> _______________________________________________
> Test-Equipment mailing list
> [email protected]
> http://mailman.qth.net/mailman/listinfo/test-equipment
> 
> 
> 