Spectrum Analyzer, which one to get?

[Bicurico]

May I add the following:

The Spectrum Analyser can only handle a certain signal power level. If it is too high, you risk damaging the input stage. This is written next to the input connector! i.e. "MAX +10dBm"

The attenuator allows to lower the signal power level, so that the remaining signal is "safe".

As ntnico said, if you attenuate the signal too much, you won't distinguish the signal from the noise. Hence why you need a configurable attenuation.

If you use the internal attenuator, the spectrum analyzer will automatically do the math plus the attenuation has been calibrated (it is never perfectly flat across the frequency range).

If you use an external attenuator, then you need to subtract its attenuation from the displayed values, as the spectrum analyser obviously just measures the input signal, not knowing what you have done with it outside the spectrum analyser.

For example, if you are dealing with antennas and/or transmitters, it is easy to exceed the maximum input rating. Doing so will quickly break the spectrum analyser. What I do is to connect a switchable external attenuator, that has a range up to -60dB (if I am not mistaken). I start with the biggest attenuation and lower it, always keeping an eye if the input signal is under the spectrum analyser's rating (written on the Input connector). Ideally, the signal does not need any external attenuation - but I like to be on the safe side.

In order to get you as paranoid as myself: if you are using i.e. a cheap Baofeng radio transmitting 5W and hold the antenna close enough to the spectrum analyser's input connector, you may break it, as 5W are roughtly 37dBm!

Spectrum analysers are sensitive equipments.

Why don't you consider buying yourself a cheap "SMA Simple Spectrum Analyser" device for aruond 50 Eurio from Chinese eBay? It can do 35MHz up to 4GHz and you can run it with my software (http://vma-satellite.blogspot.pt/2016/12/vma-simple-spectrum-analyser-download.html). This gives you a cheap introduction to spectrum analysis with little risks.

Also, do read the manuals of spectrum analysers! You will learn a lot without spending any money.

Regards,
Vitor

[rf-loop]

It is confusing alright ! 
I was playing with an Anritsu MS2721B Spectrum Analyzer at work and it has an Internal Attenuation Level control.
I fed it a signal of -25db.
Then I changed the Int Att Level to 10db, 20db and then to 0db but it did not affect the graph or apparent signal strength...It remained at -25db

If it do not work like this, it need repair or set for recycling of electronic waste!

It is designed to show signal level in input connector! Your signal is -25dBm so what ever internal attenuator you set it need show your signal level. Of course it remain -25dB on the screen becvause your signal IS -25dBm.  If you add EXTERNAL attenuator, say example 10dB, then your signal is after attenuator (in SA input connector) -35dBm and your SA show this level, because tjhis is signal level in SA input connector.   If you change SA internal attenuator, your signal level in input connector do not change and it display it just right.
 

All SA's what I know over tens of years, and I know many, are designed for display correct signal level what is connected to SA input connector independent of what is internal attenuator level as long as signal is still well over base noise level..

[TurboTom]

The internal attenuator of an SA is less useful for input protection but rather to keep the signal level entering the active circuitry of the SA fronend within the linear range. Too strong an input signal produces harmonics and intermodulation products long before dangerously high levels are approached. The Siglent SSA specifies a damage threshold of +30dBm (IIRC) but will provide half-way decent input characteristics at 0dB attenuation at signal levels lower than -10dBm. These signal levels refer to the strongest signal present at the input, not just what´s visible within the selected frequency/span. So, to identify the strongest individual signal (if performing measurements of unknown signals), it´s always a good idea to start with full span / wide bandwidth, the select input attenuation to keep the strongest peaks attenuated to -10dBm at least (i.e. if you see peaks up to +10dBm, you should select an internal attenuation of at least 20dB) and only then narrow down frequency/span/RBW parameters. If you just "play along" and keep input levels within the SAFE range, you can also decrease attenuation and observe how too strong signals at the active input circuitry will affect the spectrum. Eventually, the SSA will start "buzzing" at you to notify that the input signal is too strong. But this all happens long before dangerously high levels are reached.

Cheers,
Thomas

[Bicurico]

 @TurboTom:

Thank you for your excellent explanation - this is why I love this forum. Again I learned and consolidated knowledge.

@amdnra:

Not breaking the spectrum analyser is something you need to consider.

This is a famous Youtube video:https://www.youtube.com/watch?v=E6P-R0C9ro0&t=261s

Regards,
Vitor

[rf-loop]

The internal attenuator of an SA is less useful for input protection but rather to keep the signal level entering the active circuitry of the SA fronend within the linear range. Too strong an input signal produces harmonics and intermodulation products long before dangerously high levels are approached. The Siglent SSA specifies a damage threshold of +30dBm (IIRC) but will provide half-way decent input characteristics at 0dB attenuation at signal levels lower than -10dBm. These signal levels refer to the strongest signal present at the input, not just what´s visible within the selected frequency/span. So, to identify the strongest individual signal (if performing measurements of unknown signals), it´s always a good idea to start with full span / wide bandwidth, the select input attenuation to keep the strongest peaks attenuated to -10dBm at least (i.e. if you see peaks up to +10dBm, you should select an internal attenuation of at least 20dB) and only then narrow down frequency/span/RBW parameters. If you just "play along" and keep input levels within the SAFE range, you can also decrease attenuation and observe how too strong signals at the active input circuitry will affect the spectrum. Eventually, the SSA will start "buzzing" at you to notify that the input signal is too strong. But this all happens long before dangerously high levels are reached.

Cheers,
Thomas

Optimal "mixer level" (level after internal step attenuator) for normal use is -20dBm in SSA3kX. (depending some other settings and needs for measurements quality (including unwanted productions by mixer etc)  also bit higher can use)

I do not know what is 1dB compression level in Siglent.

Because sometimes it is boring to use Keysight material, here from Anritsu
It is good that user know this kind of things least roughly for his analyzer.

With signal under test try set SA input attenuator so that not exeed 1dB compression level.

[amdnra]

rf-loop: So if the internal attenuator does not change the level or anything else on the display, then what's the point of having it? TurboTom says it is there to adjust the signal so that it works within the linear region. Ok fine, but this implies that the displayed signal moves either up or down, and not stay at -25db as my previous example.Please someone address this Specific point

[TurboTom]

It´s been mentioned more than once in this thread: The SA displays signal level at the RF input N connector, not what it "sees" at the input of the active frontend. The attenuator is physically placed in front of the active circuitry and since the SA "knows" its settings, it shifts the measurements up and down on the screen accordingly (it does much more than this to compensate for internal gain deviations from the ideal flat line -- that´s why every digital SA has extensive internal calibration tables that are unique to every machine).

Cheers,
Tom

[amdnra]

it shifts the measurements up and down on the screen accordingly

Exactly! Then the damn thing does not stay at -25db!!
Answer this question please: If the signal does not change one bit from the aforementioned -25db with either the Attenuator ON or OFF then of what use it relly is and how can you tell that the Internal Attenuator is on? Does anybody understand the question??

[TurboTom]

I guess most members here understand your question and also see the point where you find it difficult to understand what the SA is doing and why it has to have the attenuator and comensates its setting on the display.

First of all, consider the SA as an instrument to show you the spectral RF power at its input connector over a quite wide dynamic range and also over a decent frequency span. It´s a black box, what´s going on internally is basically irrelevant.

The parameters that you can adjust on the SA are there to either "zoom" into the signal at the frequency range that you´re particularly interested in (center frequency, span, bandwidth) or to compensate for the instrument´s shortcomings dynamics-wise. And that´s what the preamplifier or the attenuator are for.

The "real" dynamic range of an SA´s signal processing section may be, let´s say 80dB (-100dBm to -20dBm, depends on the selected bandwidth and of course on the amount of effort put into the circuitry). Yet, you want to be able to measure in the range of maybe -120dBm to +20dBm, which is a span of 140dBm. It becomes immediately obvious that with the available signal processing hardware, this isn´t directly possible and definitely not so in a single setting. If you attenuate the input signal so the strongest components of +20dBm won´t overdrive the SA´s front end, you´ve got to attenuate the signal by 40dB. Yet, now the lowest signals that won´t vanish in the SA´s noise floor will have to be stronger than -60dBm. So as long as there are signal components as strong as +20dBm, you SA cennot resolve any components lower than -60dBm. If oyu made sure there are no powerful signal components at the SA´s input, you can reduce attenuation and maybe even activate the 20dB preamplifier, effectively lowering the noise floor fo the SA to -120dBm. Yet this means, signals as low as -40dBm may already overdrive the input of the SA and cause harmonics or intermodulation.

Consider the attenuator a parameter of the vertical position of the "window of interest" that you can slide over the "real" spectrum at the SA´s input socket. Thsi should make clear why the reported signal strength must not change when you change the SA´s internal attenuator setting.

Hope this helps to clarify the situation.
Cheers,
Thomas

[amdnra]

Thomas: I appreciate the fact that you are taking the time to explain things... but it is still not all clear because I see the tendency not to answer the questions posted but instead people go on with their own explanations and tangents. To be clear I'm not being ungrateful, quite the contrary but one wishes to have one's questions answered as posted, to clear some sticking points.

Now let me ask you a SA question: If you have a signal sitting at, say, -20db and the User adds an 'Attenuation Level' of -25db
then the signal on the screen should shift and show the signal sitting at -45db. Or would it? From previous replies I gather it would still show -25db. 
In other words, the addition of discrete internal attenuation by the User does not show up as having any kind of effect (since the signal STILL shows -25db). Then what's the point if nothing changes on the display!! (I did this with an Anritsu model MS2721B).
I'm not concerned here about when the SA does its own special switching for Internal Attenuation that is automatically chosen by the unit itself; I'm talking about the User programmable input from the keypad of the instrument to set the Internal Attenuator to some value.
Please advise.Thanks

[nctnico]

Now let me ask you a SA question: If you have a signal sitting at, say, -20db and the User adds an 'Attenuation Level' of -25db
then the signal on the screen should shift and show the signal sitting at -45db. Or would it? From previous replies I gather it would still show -25db. 
In other words, the addition of discrete internal attenuation by the User does not show up as having any kind of effect (since the signal STILL shows -25db). Then what's the point if nothing changes on the display!! (I did this with an Anritsu model MS2721B).

The point is to stay in the measurement region with the least distortion and thus the best measurement results.

[amdnra]

and how's that going to be possible when the display does NOT change one bit from before and after the attenuation? remember, per you folks it is still pegged at -25db

[CJay]

In simple terms the SA nulls out the -25dB and displays the signal levels as it is presented to the attenuator, not the front end of the SA.

As others have already explained, that can be useful if the input signal is going to overload the SA and cause the front end to create distortion.

If you stick an external 25dB attenuator on the SA and don't tell the SA it's present then the indicated level will correctly drop 25dB.

[amdnra]

If you stick an external 25dB attenuator on the SA and don't tell the SA it's present then the indicated level will correctly drop 25dB.

Alrighty...CJay: then What happens when you have a -25db signal present and then you stick a 25db Internal Attenuator on the SA and the SA knows it's there, Pray tell? According to the group in the know here, nothing Happens!! It stays at -25db!! It changes nothing

[idpromnut]

Now let me ask you a SA question: If you have a signal sitting at, say, -20db and the User adds an 'Attenuation Level' of -25db
then the signal on the screen should shift and show the signal sitting at -45db. Or would it? From previous replies I gather it would still show -25db.

<snip>

On my older HP that has an internal attenuator, it is a bit clearer what is going on here. On my SA, the attenuator knob also has the "reference level" setting, and both change when you change the attenuator setting. So if the reference level is say 10dB (that is what the strength of a signal that touches the top of the screen, where the reference level is, would be), and the attenuator is set to 0dB, then if you feed in a -25dB signal, you should see the top of your signal reach the -25dB mark on the screen, -35dB from the reference (this is important!).

Now, you decide to change the attenuator setting to -20dB. This will /at the same time/ change the reference level (i.e. what the top line of the screen means) to -10dB. Now that same -25dB signal being fed in is into the SA is at the same relative "place" on the screen, but the scale is different, it is now at -45dB. But it is /still/ -35dB from the reference level, so as people have been saying, your signal strength has not changed, how ever the reference level combined with the attenuator setting have now brought that signal lower in an absolute sense, but on the SA screen, it should still be the "same" relative to the reference level (which is at the top of the screen). 

In other words, the addition of discrete internal attenuation by the User does not show up as having any kind of effect (since the signal STILL shows -25db). Then what's the point if nothing changes on the display!! (I did this with an Anritsu model MS2721B).
I'm not concerned here about when the SA does its own special switching for Internal Attenuation that is automatically chosen by the unit itself; I'm talking about the User programmable input from the keypad of the instrument to set the Internal Attenuator to some value.
Please advise.Thanks

Again, the internal attenuator allows the user to bring the input signal down in absolute strength to a point where the instrument has better performance and you can get more detail about the characteristics of that signal. You are right however, that from a relative display perspective, there should be no change (i.e. the signal will "look" the same on the screen) but the measurement scales will change and different performance characteristics of the SA will come into play.

Whether the instrument automatically or the user manually adjusts the internal attenuator, the above still applies. If you wanted to further protect the instrument from a particularly high input signal (i.e. a +100dB signal where your SA has a max of +30dB input), then you would need to use an external attenuator. You might at the same time use the internal attenuator to adjust what level of signal is applied to the SA front end to gain more accurate measurements (so for example, a 100dB of external attenuation, and then 20-30dB of internal attenuation to bring the signal down to the optimal strength for analysis).

Cheers,
id

[nctnico]

and how's that going to be possible when the display does NOT change one bit from before and after the attenuation? remember, per you folks it is still pegged at -25db

It is about the signal which goes into the spectrum analyser's first mixer. That will be attenuated! It is the same as using a 1:100 probe on a modern oscilloscope. If you tell the oscilloscope you use a 1:100 probe it will multiply all numbers on the display by 100.

[amdnra]

To member id: Read reply number 13 by myself and reply number 14 by DaJMasta . Comment on whether reply 14 makes sense

[idpromnut]

To member id: Read reply number 13 by myself and reply number 14 by DaJMasta . Comment on whether reply 14 makes sense

I will need to recheck my setup at home; I'm pretty sure that the relative signal position on the screen will not change, but that the vertical scale will change (on my SA). The signal reaching the SA after the internal attenuator will be reduced in strength and this is reflected in the change of vertical scale.

[CJay]

If you stick an external 25dB attenuator on the SA and don't tell the SA it's present then the indicated level will correctly drop 25dB.

Alrighty...CJay: then What happens when you have a -25db signal present and then you stick a 25db Internal Attenuator on the SA and the SA knows it's there, Pray tell? According to the group in the know here, nothing Happens!! It stays at -25db!! It changes nothing

And the group here is correct, the signal will still display as -25dB even though the signal presented at its input is -50dB (which is a nonsense because dB has to be relative to something, be it uV, watts, milliwats etc.)

The important bit is 'it changes nothing on the SA display' because the SA is adjusting the scale to take that attenuator into account by adding +25dB to the displayed figures.

This is really very simple and I'm begining to wonder why or even if you're having a hard time understanding it.

[amdnra]

Aha! Eureka!!  nctnico takes the cake when bringing up the scope probe analogy. That's all it took. Finally got it.

 

   This is really very simple and I'm begining to wonder why or even if you're having a hard time understanding it

I have never used a Spectrum Analyzer before (that's the reason I'm here at the Forums asking questions. Because I didn't know. Duh ) And Yes, it is very simple when somebody explains it intelligently and provides a common and good example.
and About the last part of your quote, Strange you included a stupid implication after your previous decent posts. There's no need to post anything if all you're going to offer is some veiled insult.
Thanks to all that tried to help.

[CJay]

Aha! Eureka!!  nctnico takes the cake when bringing up the scope probe analogy. That's all it took. Finally got it.

 

   This is really very simple and I'm begining to wonder why or even if you're having a hard time understanding it

I have never used a Spectrum Analyzer before (that's the reason I'm here at the Forums asking questions. Because I didn't know. Duh ) And Yes, it is very simple when somebody explains it intelligently and provides a common and good example.
and About the last part of your quote, Strange you included a stupid implication after your previous decent posts. There's no need to post anything if all you're going to offer is some veiled insult.
Thanks to all that tried to help.

Simply put, it was beginning to feel like you were trolling, there were numerous explanations that if read and thought about gave the correct reasons and explanations of the behaviour you would see.

[tautech]

All make very good points and offer sound advice of SA usage. While I'm no expert and only have a year or two with SSA3000X, hard wired connection into the RFin port should be first attempted with extreme care. Answers above indicate the reasons why.
Initial use can wisely be restricted to weak coupled inputs until confidence and familiarity with the UI.

Yesterday while exploring Peak from a recent discussion on the forum, I had local spurious emissions via a coax sniffer loop on the display. Settings and attenuation were such to have peaks well within the dynamic range displayed but occasionally I noticed something at ~900 MHz. Cell phone I suspected as it was in my shirt pocket.
I shifted the coax loop and as it came close to me the cell handshake with the cell tower and immediately the SA warned of excess input......loop and phone had 300mm separation. 

These SA things are super sensitive........use with great thought and care.

[nctnico]

For some tests you may want an RF limiter. A couple of months ago I spoke to someone who blew up the input of a SA using a LISN without limiter while measuring a DUT which produced a lot of output on the LISN.

When not in use I have a 50 Ohm terminator on the input of my SA.

[amdnra]

All good points. Thanks guys 

[usagi]

you may also want to protect your SA with a DC blocker if you are sending in completely unknown signal source.