Differences between FFV on oscope and a dedicated spectrum analyzer?

Thread Starter

spinnaker

Joined Oct 29, 2009
7,835
I have an FFV option on my scope. I understand that I can do at least some rudimentary spectrum analysis with that feature. Can anyone tell me the differences between FFV on a scope and a dedicated spectrum analyzer?

I will have more FFV questions I will put in a new upcoming thread.
 

SLK001

Joined Nov 29, 2011
1,548
A FFT (or, what you call FFV) is based on the number and accuracy of the samples of the performing instrument. It is great for a quick and crude evaluation of your signal, but since it is dependent on the accuracy of the samples, it will miss most of the finer details.

A spectrum analyzer, OTOH, is a wideband receiver capable of viewing narrow chunks of that band accurately and in great detail. So, as spookman says, one is a toy and the other is a real instrument.
 

Wuerstchenhund

Joined Aug 31, 2017
189
I have an FFV option on my scope. I understand that I can do at least some rudimentary spectrum analysis with that feature. Can anyone tell me the differences between FFV on a scope and a dedicated spectrum analyzer?
Well, the biggest difference is that the scope is a time-domain instrument while a spectrum analyzer (SA) is a frequency domain instrument, and the design of both instruments is aimed at that specific purpose. Also, scopes show voltage vs time while SAs usually show power vs frequency. That has various consequences.

For example, the timebase of an SA is very stable while a scope's timebase generally isn't. This reflects on a scope's frequency accuracy in FFT mode.

A scope has 1M (low-end, 50ohms/1M switchable (mid-range/lower high-end) or 50ohms-only (upper high-end) inputs which go from zero volts to 5V (50ohms) while most SA's have 50ohms inputs that are DC intolerant (AC only, DC will destroy the input stage). So if your signal contains DC you need a DC block for your SA to avoid killing it. Better scopes also have system inputs for active/special purpose probes.

A scope's frequency range goes from DC to whatever it is specified for. Most SAs go from 9kHz or 10Mhz to the specified max (i.e. they leave the lowest frequencies out). That means for example to measure audio a normal SA is useless (there are special purpose audio SAs, though).

The ADC in most digital SAs is a 14bit or better type while most scopes only have an 8bit ADC. As a result, the dynamic range of the scope is much lower than those of an SA, which means with a strong input signal you won't see very smal frequency components. There are ways around that (i.e. oversampling aka HiRes mode, averaging) which can improve a scope's FFT resolution but it'll still be way off a proper SA.

Traditional SAs are swept SAs, they employ a heterodyne receiver which sweep over a certain frequency span. The SA front end is designed for high stability (low phase noise) and contains suppression of spurious signals that are caused by non-linearities in components which isn't fully avoidable. A scope front end consists of a few amplifiers, relay switches and maybe an DSP to correct linearity but that's it.

Because most SAs are swept SAs, they are limited in what frequency span it can capture at any given time (called Resolution Bandwidth or RBW). Older and entry-level SAs have RBWs up to 1Mhz or 5MHz, better traditional SAs go to 10MHz, and for modern Signal Analyzers/Realtime SAs it's 20Mhz and up to several hundred MHz. A scope however can, assuming there's sufficient sample memory, capture it's whole BW (0Hz to whatever MHz/GHz it's spec'd) in one go.

All that means that a DSO's FFT will not obtain the same RF performance as a SA, even a low-end one like the Rigol DSA800 or Siglent SSA3000X. Also, especially low end scopes perform pretty badly at FFT, with FFT sample sizes of 16kpts (Rigol DS1054z) or even less, all combined with a slow processing backend. However, if implemented well, FFT can be a great help to identify a frequency-related issue in situations a traditional SA would be misplaced.

I will have more FFV questions I will put in a new upcoming thread.
I'd suggest to use the much more common term FFT instead of FFV as the latter can (and I'm sure will) confusion.
 

Wuerstchenhund

Joined Aug 31, 2017
189
So, as spookman says, one is a toy and the other is a real instrument.
I disagree. FFT on a scope is not a toy, it's a versatile tool that never was intended to replace an SA and which is useful in certain circumstances. For example, good luck trying to capture all of a pulsed wideband signal with one of the cheaper or older SAs where the RBW ends at 1MHz or 5MHz. With a scope, no problem even for very extreme wideband signals.

Also, Nsaspook's EEVblog video is based on low-end scopes, i.e. scopes with limited FFT sample memory and comparatively slow processing. Of course FFT will be poor on those scopes. Let's try again on a Keysight DSO-S or LeCroy HDO9000 and the story is very different.

BTW: careful with those Tek MDOs. Despite Tek's colorful claims the "spectrum analyzer" in those scopes is nothing like a real SA (and especially the MDO3000's RF performance is piss-poor).
 
Top