Vector Signal Analyzer vs. Spectrum Analyzer

[mk...@interquest.com]

We are considering purchase of a Agilent E4406A Vector Signal
Analyzer.

http://cp.literature.agilent.com/litweb/pdf/5968-7618E.pdf

Can anyone please advise what is the difference, in terms of
operational capabilities, between this and a spectrum analyzer of
similar specified bandwidth?

IOW can it be used as a conventional spectrum analyzer, and, if so,
are there any limitations.

And yes, we do have a learning curve ahead.

Martin King

[John Miles, KE5FX]

On Tuesday, February 18, 2014 1:40:26 PM UTC-8, mk...@interquest.com wrote:
> We are considering purchase of a Agilent E4406A Vector Signal

> IOW can it be used as a conventional spectrum analyzer, and, if so,
> are there any limitations.
>

Yes and yes. They are neat boxes, especially if you can find one for less than $1000-$1500. I wouldn't pay much more than that for one, though. Agilent was careful not to allow them to compete with the "real" E4400 SAs. The maximum span width is limited to a few MHz (10?) and there are a lot of blank menu entries and nonfunctional buttons.

On the upside, they have a lot of digital RBW settings, going down to 0.1 Hz. I use mine when I want to look at port isolation or interconnect leakage down to -120 dB and below. That's about the only time I turn it on. I normally prefer my old 8566 and 8568 boat anchors because of the operational limitations, subpar user interface (IMHO) and long boot time imposed by the E4406A.

A few of my quick-and-dirty GPIB tools (http://www.ke5fx.com/gpib/readme.htm) can be used with it, including the phase noise and spectrum recording apps.

-- john, KE5FX

[miso]

Looking at the datasheet, this isn't that much better than a spectrum
analyzer. If it was a VNA, I'd be a lot more interested. But it does have a
power meter built into it.

The noise spec is a bit high, but they spec it at a 1kHz BW. Most spec at
the 100Hz BW.

The Agilent box can input IQ. I presume it can display a constellation, but
I haven't read the manual.

A spectrum analyzer can be just that, i.e. a spectrum analyzer, or it could
incorporate a power meter. Better yet, it could have a tracking generator to
sweep filters.

If you want to test antennas, you need a dedicated device like an Anritsu
Site Master, or a network analyzer with impeance bridge. I suppose if you
want to characterize any port, you would need a VNA. Products like a Site
Master are very noisy, since they are intended to be used for cable testing,
impedance sweeps, etc., i.e. using the built in source. A VNA is generally
as low in noise as a spectrum analyzer, but more expensive.

It might make more sense to indicate what you want to accomplish.

[mk...@interquest.com]

On Tue, 18 Feb 2014 20:23:17 -0800, miso <mi...@sushi.com> wrote:


>Looking at the datasheet, this isn't that much better than a spectrum
>analyzer. If it was a VNA, I'd be a lot more interested. But it does have a
>power meter built into it.
>

Can it be used as an "ordinary" spectrum analyzer within the frequency
range specified, or is there some compromise in features and
performance because it is a "vector" signal analyzer?

Reading between the lines, you seem to suggest it can do what a
spectrum analyzer does and more. Is this correct?

We don't want to buy one and then find out we need to buy a separate
spectrum analyzer ... that's the general meaning of my question.

Thanks,

Martin King

[Robert Lacoste]

wrote in message news:ebk7g9tm9eduhi48d...@4ax.com...

> We are considering purchase of a Agilent E4406A Vector Signal Analyzer.

> Can anyone please advise what is the difference, in terms of
> operational capabilities, between this and a spectrum analyzer of
> similar specified bandwidth?
> IOW can it be used as a conventional spectrum analyzer, and, if so,
> are there any limitations.

Hello Martin,

As John said, its first limitation as compared to a classical spectrum
analyzer is its maximum span : you can analyze no more than 10MHz of
spectrum at a time. The second disavantage is that it is a quite complex
equipment, therefore you will need to spend a little more time to use it
efficiently (and if not you could get wrong measurements, for example with
pulsed signals, as it is FFT-based and not sweeping). Lastly it covers from
7 to 314MHz and from 329MHz to 4GHz, so check that you have no interest in
the intermediate small band...

Now the advantages : as a vectorial signal analyzer such a product can do
things that a classical SA can't. In particular it can demodulate signals
and display I/Q signals, constellations, etc. However for anything else than
basic applications you would need to have the associated software options
installed on your 4406A : GSM/EDGE/CDMA analysis, etc.

Have fun with it, nice equipment and available at reasonnable costs as
plenty of telecom manufacturer bought it to test their transmitters...

Robert
www.alciom.com

[mk...@interquest.com]

Thanks for the helpful replies. i didn't realize the span was limited
to 10MHz. That would not suit our purpose.

Martin King

[miso]

I would track down the operating manual if you really want to know how it
works.

Again, do you need to sweep parts? Not every spectrum analyzer has a
tracking generator.

Again, do you need complex impedance?

[miso]

Robert Lacoste wrote:

> wrote in message news:ebk7g9tm9eduhi48d...@4ax.com...
>> We are considering purchase of a Agilent E4406A Vector Signal Analyzer.
>> Can anyone please advise what is the difference, in terms of
>> operational capabilities, between this and a spectrum analyzer of
>> similar specified bandwidth?
>> IOW can it be used as a conventional spectrum analyzer, and, if so,
>> are there any limitations.
>
> Hello Martin,
>
> As John said, its first limitation as compared to a classical spectrum
> analyzer is its maximum span : you can analyze no more than 10MHz of
> spectrum at a time.

> (snip)
> Robert
> www.alciom.com
I'd say 10MHz at a time would be annoying.

[josephkk]

In many cases this gets down to what the instrument class does. A VNA
will measure two port parameters of most any arbitrary DUT, over a defined
frequency range in fixed steps (linear and maybe logarithmic). An SA will
measure the general amplitude of (quasi-) steady signals over a defined
range with a defined sweep speed and receiver bandwidth. The sweep speed
and RBW interact.

That said what things do you wish your instrument to do?

?-)

[Simon S Aysdie]

I think people may be confusing max instantaneous bandwidth with sweeps. You need to check if it sweeps. I'd be surprised if it didn't but I suppose it is possible (older VSAs did not include SA operations, but they usually do these days).

For example, look at the image under
NI 5665 14 GHz VSA Analysis Time, Preselector Enabled
http://sine.ni.com/ds/app/doc/p/id/ds-321/lang/en

That instrument has a 50 MHz max instantaneous bandwidth, but there you see a sweep of essentially 14 GHz.

Spectrum/signal analyzers have always had limited instantaneous bandwidth, of course. If the displayed spectrum is beyond the max instantaneous bandwidth, then it is swept, whether the sweep is analog or stitched digital.

[G0HZU]

A late reply, but the E4406A doesn't have the (downconverter) frequency plan of a conventional spectrum analyser. It downconverts to a first IF of 321.4MHz with no preselection so there is no image rejection and also a hole in its tuning range at 321.4MHz. It will also suffer from other significant spurious terms. eg subharmonics of 321.4MHz will have limited rejection. You are supposed to use it for vector signal analysis under controlled conditions rather than as a spectrum analyser.

[Simon S Aysdie]

On Friday, March 27, 2015 at 8:13:37 PM UTC-7, G0HZU wrote:
> A late reply, but the E4406A doesn't have the (downconverter) frequency plan of a conventional spectrum analyser. It downconverts to a first IF of 321.4MHz with no preselection so there is no image rejection and also a hole in its tuning range at 321.4MHz. It will also suffer from other significant spurious terms. eg subharmonics of 321.4MHz will have limited rejection. You are supposed to use it for vector signal analysis under controlled conditions rather than as a spectrum analyser.

That is the case with the NI 5663 too, if I remember the model number correctly. It is also true of some Aeroflex models. However, I will point out that some old SA's---that did not have "signal analysis" in any sense that we mean it now---also did not have image rejection. I think I used to test with a "Cushman" that had the problem, but my memory is suspect these days. (It was at a place I did not work at for very long, in the '80's.)

The 5663 is really intended for signal analysis in a "controlled environment" only. I mean, one can do spectrum analysis with it, but has to be aware of the limitations.