Can your typical electronics hobbyist benefit from a spectrum analyzer?

[Rene]

Hello,

I am considering buying a spectrum analyzer (probably a cheap used one) but I am wondering how much use will I be able to get out of it if I end up buying it. I know those things are expensive so I want to make sure that it is not going to just sit there in the off position.

As of today, my primary motivation for buying one of this things is for learning experiences. For example, I constantly see articles in books or online about passive filters and I would really (I mean really, really, really) like to build some filters myself and test their behavior in real life. Having a spectrum analyzer will allow me to compare the theory vs reality using different types of components, typologies, cable lengths, PCB vs breadboard etc (just for fun and learning). I am also thinking on getting a little more into RF and I think the spectrum analyzer could come in handy there (also mostly for learning).

Nevertheless, I have no illusions of one day acquiring enough knowledge to do electronics for a living, most likely this will remain a hobby and there is a big chance that my biggest accomplishment will end up being creating a blinking light circuit using a 555. However (and this is important), my goal is to create the best 555 blinking light circuit ever created, complete with impedance matched PCB traces and all. Not because the circuit needs to be that precised but because I believe you can learn a lot on how electronics work even from the most basic circuits as long as you push the circuit to the limits (at least I hope so and the reason why I may end up needing a spectrum analyzer).

So what do you guys think, can your typical electronics hobbyist benefit from a spectrum analyzer? Or is this a total overkill (the cost outweighs he learning benefits)?

Thanks.

[edavid]

1. My opinion is that when you need one, you will know.

2. A spectrum analyzer is not a very good tool for measuring filter response.   It is much better to use a sweep generator and an oscilloscope.

3. If you just want to learn about spectrum analyzers, try a sound card and software for audio, or a cheap SDR dongle for RF.  You will learn a lot that way, for a very low cost.

[Lightages]

What edavid said. 

[retrolefty]

Well it's my opinion that a spectrum analyzer was always the 'holy grail' for amature Ham radio operators/builders of past eras. Few actually ever got to obtain one especially new store bought, but today it is a much different time and is probably a more common appearance in a ham shack then ever before.

 One can do meaningful RF work without a SA, but it sure makes it easier, as a picture is always worth more then a list of numbers.

[rfeecs]

The best tool for filter measurement is a network analyzer. 

You might look at Analog Discovery.  It can be used as a low frequency network analyzer and spectrum analyzer:
https://analogdiscovery.com/

[schmitt trigger]

As a learning experience, the greatest benefit a Spectrum Analyzer provides, are the visual clues of looking at information in the frequency domain.

To me, my epiphany was when -back in the late 80s- I saw for the first time a modulated analog TV channel on a spectrum analyzer.

Being able to look at the main carrier and its subcarriers, the scanning-synchronized harmonics, the audio carrier's deviation, etc, etc, etc, was to say the least, an eye opener.

[tggzzz]

So what do you guys think, can your typical electronics hobbyist benefit from a spectrum analyzer? Or is this a total overkill (the cost outweighs he learning benefits)?

I like your attitude; "because I want to" is an unassailable reason But only you can assess the cost/benefit tradeoff.

An SA is a great general-purpose tool for assessing frequency-domain operation, but if you have a specific purpose in mind then there may be alternative options. For filter performance, a network analyser is ideal (essentially an SA plus a tracking generator). Other considerations are what other equipment you already have, and what what else you might want to spend your time (and money) on.

The key technical specs of an SA are its frequency range and its dynamic range. A key "weakness" of some SAs is that the maximum input power must never be exceeded, and that they cannot tolerate any DC voltage on the input.

As already mentioned, the Digilent Analog Discovery is a very good device within its limitations, and can be used as a network analyser plus other generally useful test equipment.

If you need higher frequencies, then (with limitations) you could use a cheap SDR dongle as an SA, and add a noise source as a poor-mans's tracking generator. See my .sig for more details.

Alternatively, with different limitations, a digitising scope can be used as a spectrum analyser.

My suspicion: your filters are audio or radio-ham frequencies; consider the Analog Discovery if you don't already have a scope and siignal generator.

OTOH, if you are concerned about PCB track impedance, then you will need much higher frequencies. A 1.5GHz SA is sufficient to resolve details ~4cm apart. Consider that a time domain reflectometer is a very good way of directly measuring a track's effect on a digital signal. And due to the duality between the time and frequency domains, it is possible to mathematically transform from one to the other.

[Inverted18650]

I posted a sweet Tek 7L12 inside the 7704 scope on eBay. Might be fun for you.

[Rene]

Thank you all for your responses.

One more quick question. Say that I was planning on getting something like the Rigol DSA815-TG (This is not the case but I just want to get a better feel for what it is being said here). Would that change anything? Would a piece of equipment of that caliber allow you to explore electronics in ways that the SDR dongle for RF, the Analog Discovery device or the oscilloscope function generator combination cant come close?

Sorry, I realize that this is probably a dumb question... just trying to get a better feel for what the alternatives offer by comparing them with what I hear is an excellent piece of equipment (probably not a relevant comparison but just curious).

Thanks.

[tggzzz]

Thank you all for your responses.

One more quick question. Say that I was planning on getting something like the Rigol DSA815-TG (This is not the case but I just want to get a better feel for what it is being said here). Would that change anything? Would a piece of equipment of that caliber allow you to explore electronics in ways that the SDR dongle for RF, the Analog Discovery device or the oscilloscope function generator combination cant come close?

Sorry, I realize that this is probably a dumb question... just trying to get a better feel for what the alternatives offer by comparing them with what I hear is an excellent piece of equipment (probably not a relevant comparison but just curious).

The AD is 0-10MHz. The Rigol is 9kHz-1500MHz. What frequencies interest you?
The AD is a based on a 14-bit ADC, which limits its dynamic range (~60dB at audio frequencies). The Rigol ought to have a wider dynamic range. What noise floor and distortion can you tolerate?
The AD is significantly cheaper than the Rigol, especially if you can get the educational discount.

My attitude is to figure out something I want to measure, then determine the specs required, then choose my equipment. I'll include the "foreseeable future requirements", but I'm wary of getting too many things that "might come in useful".

[mainman]

Rule one, never buy a tool you have no ESSENTIAL PROLONGED need for, a spectrum analyzer is a ungodly costly piece of equipment to just buy with no cost redeeming reasons. You will instantly regret buying one with no constant need for one.

That being said if you do buy, get a SA with at least a 3ghz band. Right now I am using the hobbyist RF Explorer but will be upgrading to a PSA2702

[David Hess]

2. A spectrum analyzer is not a very good tool for measuring filter response.   It is much better to use a sweep generator and an oscilloscope.

This would be my recommendation before buying a dedicated spectrum analyzer.

Better function generators provide either sweep capability or can be used in pairs to do sweeps and provide either a ramp output or sweep trigger which when combined with an oscilloscope can be used in place of a spectrum analyser and tracking generator for filter response analysis albeit only at low frequnencies.  And a good function generator is a universal instrument while being inexpensive.

A function generator and oscilloscope can also be used to make amplitude and phase measurements manually serving as a low cost low frequency vector network analyser.

[Electro Fan]

I can't cost justify any of my test gear in business terms (i.e. increased revenue or decreased costs sufficient to generate a ROI over some period of time) but after 4 oscilloscopes (which is 15 short of the 19 that is rumored to be par here), I "invested" in an older but still good (knock on wood) HP spectrum analyzer that works up to 6 GHz.  So far the main uses have been in the realms of ham radio, wifi, wireless key fobs, and some other radio signal testing/studying.  All of my gear really only serves one purpose, helping to measure and visualize circuits and signals in an attempt to learn what causes what.  For me the whole EE thing is a hobby that brings intrinsic enjoyment to the process of studying and learning.  It would be hard for lots of people to see how, for example, trying to generate and measure rise times in picoseconds could be worth anyone's time and effort much less any expense, but I find that the ability to hypothesize and test with some reasonable resolution and precision in the time domain and frequency domain with volts, amps, resistance/impedance, etc is fascinating in and of itself.  Some people would read for years about a distant mountain or city or other geography of interest and then save up to go visit.  With test equipment you can read about various circuit and signal properties and then "visit" them on your bench - and not only can you observe them but often you can manipulate the circuit and signal properties as you observe and measure them in a way that enables you to go from initially little understanding to gaining some control over what causes what to making those properties consistently reproducible to potentially mastering the properties.  Some people could spend travel $ and marvel at the site of the Grand Canyon (I can relate, been there, done that, and would recommend it) but the ability to hook up and test a wide variety of analog and digital devices pretty much whenever I have an idea and time and test them on the bench seems to be something I enjoy doing much more often than visiting the Grand Canyon.   

A spectrum analyzer isn't something I'd recommend for most enthusiasts before DMMs, power supplies, oscilloscopes, etc but when you are ready to explore RF it's a pretty cool piece of test equipment that can help you visualize, measure, and learn. 


This thing looks pretty interesting - it could be nice to for work away from the bench and it looks like it could save a lot of space on the bench:
http://www.tequipment.net/TTi/PSA2702/Spectrum-Analyzers/?gclid=CJHI6NG549UCFca2wAodiD0HTA

net, net:  if a SA is too hard on the budget there are probably other ways and other things to measure; if a SA fits the budget you might enjoy it; if it doesn't get enough use it can probably be resold.

[pelule]

I assume, you are not just interested in SAs for basic interest, but for "What an SA is useful for?"
Have a look to it more basically, so you may decide.
If you like to measure:
- DC or low frequency AC signals: use a DMM
- fast AC signal, data bit changes or busses: use an oscilloscope
- spectral portions of a signal, harmonics, EMI, demodulation of (digital) signals or transmission performance: use an SA

/PeLuLe

[jpb]

I can't cost justify any of my test gear in business terms (i.e. increased revenue or decreased costs sufficient to generate a ROI over some period of time) but after 4 oscilloscopes (which is 15 short of the 19 that is rumored to be par here), I "invested" in an older but still good (knock on wood) HP spectrum analyzer that works up to 6 GHz.  So far the main uses have been in the realms of ham radio, wifi, wireless key fobs, and some other radio signal testing/studying.  All of my gear really only serves one purpose, helping to measure and visualize circuits and signals in an attempt to learn what causes what.  For me the whole EE thing is a hobby that brings intrinsic enjoyment to the process of studying and learning.  It would be hard for lots of people to see how, for example, trying to generate and measure rise times in picoseconds could be worth anyone's time and effort much less any expense, but I find that the ability to hypothesize and test with some reasonable resolution and precision in the time domain and frequency domain with volts, amps, resistance/impedance, etc is fascinating in and of itself.  Some people would read for years about a distant mountain or city or other geography of interest and then save up to go visit.  With test equipment you can read about various circuit and signal properties and then "visit" them on your bench - and not only can you observe them but often you can manipulate the circuit and signal properties as you observe and measure them in a way that enables you to go from initially little understanding to gaining some control over what causes what to making those properties consistently reproducible to potentially mastering the properties.  Some people could spend travel $ and marvel at the site of the Grand Canyon (I can relate, been there, done that, and would recommend it) but the ability to hook up and test a wide variety of analog and digital devices pretty much whenever I have an idea and time and test them on the bench seems to be something I enjoy doing much more often than visiting the Grand Canyon.   

Well expressed.  Self-education and learning in a practical sense as well as being a worthwhile end in themselves may well pay dividends at some point. As someone who has worked as an engineer (albeit in research and development) I know that simply allowing accountants to rule in terms of immediate costs and returns can lead to much greater costs in the end.

It is, of course, sensible to resist the urge to get kit just because it is nice to have but on the other hand no particular measurement task is probably enough on its own to justify expensive purchases and these sometimes have to be made on the basis of future, as yet unknown, project work.

[IanMacdonald]

Well, when it comes to learning how waveshapes relate to spectra and so forth, yes you can benefit from seeing one in action. Whether it's worth buying one, doubtful though. You could spend the money more usefully.

"the best 555 blinking light circuit ever created"

Try making one that provides a flash of equal mark/space,  and immediately starts with a correct-length flash at powerup.   

[Syntax_Error]

I know this is not what the OP asked about, but I post it as a curiosity for them and maybe others who will search/read this thread in the future:

http://www.scottyspectrumanalyzer.us/

I would love to build and modify one of these, personally. Especially when considering the personal learning benefit of such a project.

[Zero999]

Well, when it comes to learning how waveshapes relate to spectra and so forth, yes you can benefit from seeing one in action. Whether it's worth buying one, doubtful though. You could spend the money more usefully.

"the best 555 blinking light circuit ever created"

Try making one that provides a flash of equal mark/space,  and immediately starts with a correct-length flash at powerup.   

That's not that hard to do.

[G0HZU]

Unless you have a definite requirement that means you have to be able wave a small/portable spectrum analyser around like a lollipop to make very crude measurements, then I'd advise you to run a mile from that TTi PSA2702 analyser with it's awful specifications. Don't look back... don't ask why... just keep running with your cash still in your pocket

[mainman]

Unless you have a definite requirement that means you have to be able wave a small/portable spectrum analyser around like a lollipop to make very crude measurements, then I'd advise you to run a mile from that TTi PSA2702 analyser with it's awful specifications. Don't look back... don't ask why... just keep running with your cash still in your pocket

I am going to ask why.

[bd139]

TTi did a 1GHz unit that you plugged into your scope that wasn't totally shit, the TSA1000. You can pick them up for virtually nothing. I got outbid on one at £40.

I really would like an SA but I'm happy with DG1022Z -> DS1054Z sweeps for my current use cases. These are good to 25MHz. RC low-pass here. Sweep is log, frequency is marked at the expected 3dB point. yellow = in. blue = out. dark blue = marker. win:



Also working on an SA using the DS1022Z as the sweep source. Plan is to cover 1-30MHz so mix incoming and sweep gen to a higher frequency (50-80MHz), filter, downconvert, log amp (this bit works already). This is all pie in the sky at the moment and I haven't even thought about harmonics and sidebands really.

As for the original question, an SA is really only good if you're building transceivers as a hobbyist as you will want to see harmonic byproducts and you can just filter them out with a low pass anyway and rely on faith. Perhaps there is some value in band scanning for amateur radio, SSB, carrier monitoring etc. I can't really think of many more use cases that are high enough frequency that can't be served by the crappy FFT in a cheap DSO or via some other means that are far cheaper. The DS1054Z FFT is borderline useless above about 10MHz even with frigging source and sample rates for ref. If you have the money to burn, there is an education there though

Edit: before I forget, if you're going to get into RF, buy a copy of "Experimental Methods in RF Design". The book will walk you through the concepts and measurements and give you some hints at building your own SA, filters etc. You need to know these fundamentals before the application of an SA will make much sense.

[G0HZU]

I am going to ask why.

Here's a few reasons and they are in no particular order...

It only has a choice of 3 RBW settings and the smallest RBW is 15kHz.

The analyser only works down to 1MHz and the smallest span size is 270kHz.

The analyser has only two choices for setting the reference level at either 0dBm or -20dBm.

The smallest signal it can detect is -95dBm.

The IP3 SFDR is only about 63dB with a 15kHz RBW.

The sweep rates are very slow.

The display only has 271 points on the frequency axis.

The phase noise is poor

To quote Father Ted...

"Run Dougal, run quite fast... quite fast!" 

[Howardlong]

I am going to ask why.

Here's a few reasons and they are in no particular order...

It only has a choice of 3 RBW settings and the smallest RBW is 15kHz.

The analyser only works down to 1MHz and the smallest span size is 270kHz.

The analyser has only two choices for setting the reference level at either 0dBm or -20dBm.

The smallest signal it can detect is -95dBm.

The IP3 SFDR is only about 63dB with a 15kHz RBW.

The sweep rates are very slow.

The display only has 271 points on the frequency axis.

The phase noise is poor

To quote Father Ted...

"Run Dougal, run quite fast... quite fast!" 

I bought the previous version quite some years ago, the PSA2701T, and by far the biggest drawback are the limited RBW settings which are the same three (15k, 280k and 1MHz) in the newer 2702 version. Its only use case for me is out in the field, where I supplement it with an SDR for closer in analysis. In the lab I use proper bench units.

The battery in my PSA2701T is useless, and you have to be really careful about using an external supply with it as most wall warts spew out allsorts of RF nasties which the device doesn't filter out at all well.

[RoGeorge]

You don't need a spectrum analyzer.

Since you are talking about filters and breadboards, I assume your filters will be in audio frequencies (breadboards can not be used for radio frequency). A spectrum analyzer is dedicated for radio frequencies, and most of them can not work with audio signals.

For audio, a PC is enough and a program is all you need, and it is more precise.
http://www.radio.imradioha.org/pc_based_test_gear.htm

[Mr. Scram]

Rule one, never buy a tool you have no ESSENTIAL PROLONGED need for, a spectrum analyzer is a ungodly costly piece of equipment to just buy with no cost redeeming reasons. You will instantly regret buying one with no constant need for one.

That being said if you do buy, get a SA with at least a 3ghz band. Right now I am using the hobbyist RF Explorer but will be upgrading to a PSA2702

I have never regretted buying a good tool, even if it used sparingly. The pain of spending the money goes away, but the tool remains. A good tool is a good tool.