RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer

Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

Thank you for the very informative post.
It is a real pleasure to see such old gear in action.

By the way, there is no match between a modern FFT analyzer and previous (sweeped) analog spectrum analyzers.

Sweeped spectrum analyzers basically see 'one spectrum frequency at the time' and the overall analysis time, in case of quasi ideal IF gaussian filter - as in most HP analyzers (different story for my beloved Tektronix ones) - is inversely proportional to the minus 3dB IF resolution bandwidth.
To make a long story short, if the -3dB bandwidth of the IF gaussian filter is 1 Hz it will take about 1 second to get the reading of a single spectrum frequency; hence if you needed to analyze 10 Khz @ 1 Hz resolution it will take 10000 seconds (more because of the video filter) for a single sweep!

For sake of comparison with the above case, by using a modern scope (or the A/D card of a PC) you only needed to acquire about 1 second of signal time history to get the 1 Hz (FFT bin) frequency resolution (I am provisionally neglecting the effect of FFT windowing because it doesn't change so much the math). As the input bandwidth of modern scopes largely exceeds so much the conventional 20 Khz audio bandwidth you can for example limit your A/D sampling frequency to 100 Khz and your maximum frequency resolution is only limited by the maximum available record length (you need 100 thousand samples for 1 second FFT time history @ 100 Khz sampling frequency). As at such 'low sampling frequency' the number of effective bits of 'vertical resolution' of the time history will be higher than 12, that is, more than 80 dB of on-screen 'intermodulation free' dynamic range are now common in even PC based FFT audio analyzers. If the audio signal were really stationary it would be even possible to digitally average an ensemble of ten 1 second time histories and getting about 10 dB of premium additional dynamic range!

In the case of the HP analyzer of the video I seriously doubt that the on screen intermodulation free dynamic range is greater than about 70 dB.....

In addition, what is evident in your video is that the signal is postulated to be stationary, so that the analog sweeped spectrum analyzer had enough time to scan the horizontal span and build up a screenshot.

A FFT analyzer basically works in single shot and it works fine even with time varying signals so that you might measure TID as well as THD.

This post is made possible by the generous support of people like you and our sponsors:

Follow Ups Full Thread

Topic - This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - Michael Samra 23:35:38 05/13/16 (8)

RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - Triode_Kingdom 09:31:53 05/14/16 (6)

one of the best instruments for audio spectrum work is the analog Stanford Research 760. It will trounce eithe - Michael Samra 20:47:35 05/14/16 (0)

RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - Michael Samra 12:27:41 05/14/16 (4)

RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - Triode_Kingdom 13:42:13 05/14/16 (3)

Well be that as it may,I need a new RF signal generator as my old - Michael Samra 20:20:15 05/14/16 (2)

RE: Well be that as it may,I need a new RF signal generator as my old - Triode_Kingdom 09:47:32 05/15/16 (1)

I hear you.I hate the digital stepped output control because - Michael Samra 12:06:34 05/15/16 (0)

RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - 6AS7_6SN7 03:25:05 05/14/16 (0)

Follow Ups

RE: This is interesting where Dave is comparing an old HP dedicated Spectrum analyzer - 6AS7_6SN7 03:25:05 05/14/16 (0)

Post a Message!