Spectrum Analyzer

[avrfreaks]

Hello I have some more questions again:

1. What is a Spectrum Analyzer and do I need one if I am purchasing a analog oscilloscope.

2. Are there advantages using a separate PC Spectrum Analyzer,

[cybergibbons]

A spectrum analyzer shows you a graph of magnitude/power against frequency.

I'd say they aren't very useful to an hobbyist - or at least if you are questioning if you need one, you probably don't. They are rather niche, and I would mainly say used for RF work. Anything that doesn't go up to RF frequencies is of limited use, I'd pretty much call it a toy.

Uses are:
1. Tuning the output power on a wifi router - you can observe when the output amp is pushed too far when harmonics are being produced.
2. Checking the output of a broadcast FM transmitter
3. Checking for causes of interference
4. Determining the modulation on a signal you have picked up
5. Detecting and analysing spurious harmonics

[jahonen]

I have found spectrum analyzer very useful for some specialized areas (niche, but quite recurring at my work), like checking/fixing switch-mode power supply outputs for spurious resonances, checking cleanliness of power supply rails, this could be done with a scope but SA has much better sensitivity and it gives clear view of the problem, checking phase noise (jitter) of an oscillator, debugging EMI/EMC issues, like checking RF leaks from low-speed-IO signals, measurement of common-mode RF-currents of IO-cables.

In generally, if some problem is defined in the frequency domain, then it is more suited to SA, otherwise it is scope stuff. It is nowadays common to have "poor man's spectrum analyzer" in scope with FFT, so you don't necessarily need a dedicated SA, unless you work with difficult signals.

Regards,
Janne

[saturation]

FYI the Rigol 1052E has such an option.

...

In generally, if some problem is defined in the frequency domain, then it is more suited to SA, otherwise it is scope stuff. It is nowadays common to have "poor man's spectrum analyzer" in scope with FFT, so you don't necessarily need a dedicated SA, unless you work with difficult signals.

Regards,
Janne

[alm]

But note that the 8 bits of vertical resolution of most scopes severely limit the dynamic range, so the emphasis is on 'poor man's'. The harmonics are often barely above the noise floor, at least on all DSO's I've seen.

[tecman]

Just an FYI, I have an HP audio spectrum analyzer that I am getting ready to sell.  It's an HP 3560A portable unit.  Good shape, just needs a new battery (5 x D NiCd assembly)

paul

[saturation]

Good to see here too:

https://www.eevblog.com/forum/index.php?topic=525.0

Alm's comment is vital.  But like all tools, know its limits to use it accurately.  For the Rigol its specs, on the back pages of the manual, is so far true.

But note that the 8 bits of vertical resolution of most scopes severely limit the dynamic range, so the emphasis is on 'poor man's'. The harmonics are often barely above the noise floor, at least on all DSO's I've seen.

[joelby]

I've found the Rigol's FFT to be pretty much useless! Assuming your signal and harmonics are large enough to show up at all, it's difficult to see what the amplitude and frequency of peaks are. For RF work, even at IF, it's a waste of time. I totally recommend investing in a proper spectrum analyser

[sigxcpu]

Did you use the cursors for amplitute and frequency? I'm at a very beginner level, but I find this very easy to be read.
Regarding the signal level, yes, you are right.

[alm]

What's the point of using a spectrum analyzer to look at the fundamental? Might as well use a frequency counter or power meter for that. You want to see the harmonics! So 400MHz is not enough by a long shot, something like 2GHz might be enough. But I think Joelby was talking about noise floor and (usually) clumsy controls.

[jahonen]

Harmonics measurements are useful with radio compliance testing. I believe most standards require measurement of ten first harmonics, so for 434 MHz carrier, one would need minimum upper spectrum analyzer frequency range of 4.34 GHz. For 2.4 GHz, it is 24 GHz, etc. But most often during development, one would like to see the close-by spectrum near the carrier, to check if there is a correct modulation. This is also often difficult with scope, since capturing the carrier would require high sample rate and getting enough samples to resolve the details near carrier would need a long record. Those requirements yield to inconveniently large number of samples, unless there is some form of downmixing.

Serious RF spectrum analyzers start with upper frequency range of 1-3 GHz. Unlike scope, they are basically a superheterodyne radio receiver with very wide input frequency range, adjustable attenuator, widely adjustable IF bandwidth and precision detector for determining the signal level. Most SA's don't work down to DC (at best down to few Hz or so, mine starts with 9 kHz), and even get damaged if fed with DC. Some contain demodulation features, so one could even listen a radio with it

Regards,
Janne