USB network analyzers -- as bad as USB oscilloscopes, or not?

[hendorog]

Thats one advantage of a USB instrument - more of the instrument is software, and so can be upgraded.
Signalhound did this recently with a rewrite of their UI which was a massive improvement on the old UI.

Is that pre-or-post Shahriar's review?

The BB60C that Shahriar reviewed always had the 2nd gen software. I was referring to the SA44/SA124 models which originally came out with a fairly agricultural UI. Now the two share the same software and the API's have been brought into sync as well.

Just briefly, does anyone know what the Signal Hound actually sends over the USB link? I/Q samples, frequency/magnitude pairs, or something else? I'm trying to get a sense of what proportion of the cleverness is in driver/UI software vs actually on the device.

The Signalhounds have three modes - zero span, swept and real-time.
From the perspective of the API you can get IQ info when in zero span or swept mode, but only frequency/magnitude info in real-time mode. There was a discussion on their facebook page and it seems that IQ info isn't available for the realtime samples.
 
So I suspect that is how it comes from the device as otherwise they would just add that to the API.

[tggzzz]

It depends on what you mean by "convincingly". For the TDR example I can already clearly measure the lengths of 19 cm and 29cm open/short circuit stubs located 3.1m along cable. In theory 1.5GHz => "distance quantisation" of 3.3cm. I'm still improving my technique for measuring the impedance/magnitude of discontinuities.

As and when I'm satisfied with my techniques, I'll publish my code.

Cool! What software did you use to interface to the SDR? I.e., did you write your own code to talk to the USB directly, or did you use some software to take a frequency sweep or something for you? Did your code start with a list of frequency/magnitude pairs, or a pile of I/Q samples?

It is based on Kyle Keen’s program rtlpower http://kmkeen.com/rtl-power/ outputs frequency/magnitude pairs. I wrapped that with four Python scripts to do the processing I wanted.

Some preliminary results are at https://entertaininghacks.wordpress.com/2015/07/27/poor-mans-homebrew-tdr-with-4cm-resolution-part-1/ Please note the part 1 implies this is a start, and that I intend to explore improvements in subsequent parts.

Today I used the SDR to estimate the rise/falltime of a digital signal. It appears that 3 parallel 74LVC1G14 gates driving a 50ohm line have a rise/falltime of around 600ps.

[dr.diesel]

As someone who regularly uses instruments ranging from $50 to $500k

A preview of such a device would make a great video!   

[G0HZU]

Thats one advantage of a USB instrument - more of the instrument is software, and so can be upgraded.
Signalhound did this recently with a rewrite of their UI which was a massive improvement on the old UI.

Note that the SH44 concept is based on a frequency plan that is highly unconventional. It's a long time since I looked at the block diagram but this analyser relies on a very unconventional image detect/reject system that allows it to use very low first IF frequencies. This means they can design and build something very small with very low cost if they can manage to make this system workable.

I've not had one to play with but I'd expect the RF section to have some significant weaknesses and flaws. I suspect that this won't prevent it from being a very useful (and powerful) instrument but it doesn't possess the classic design integrity of a conventional analyser. However, I'd expect most users to be quite happy with it and they may well never use it in a way that exposes its unconventional design limitations.

[hendorog]

It's a long time since I looked at the block diagram but this analyser relies on a very unconventional image detect/reject system that allows it to use very low first IF frequencies. This means they can design and build something very small with very low cost if they can manage to make this system workable.

The detection scheme you are talking about is their method for doing two sweeps switching IF frequencies each time. Then they can use the differences to work out whats a real signal and what's not.
The downside is that hopping signals are hard to detect as the spur detection algorithm will hide them if they hop in between the two sweeps.

AFAIK That only applies to the swept mode, not the realtime mode.

[KE5FX]

In Tek's defense, I'd also point out that the RSA306's requirement for 40 MHz of realtime analysis bandwidth rules out some of the usual spur-avoidance hacks.  For instance, it's all well and good to plot the minimum response from N sweeps with N different IFs, as Signal Hound does, but that's not the least bit helpful if you need to demodulate the baseband signal or otherwise need to minimize dead time at (almost) all costs.

Not having examined it in detail, I'd guess that the worst spurious responses from the RSA306 will arise from integer-boundary spurs from the MAX2870 LO synth chip they use.  Assuming they can't vary the reference frequency dynamically, there is nothing they can do in real time to dodge those spurs in software.  When I need SFDR, nothing beats the old-school HP iron... but hey, I also can't get a 40 MHz realtime data stream out of my 8566B, so it's horses for courses.  The RSA306 and its competition from Signal Hound are really nifty gadgets.

[hendorog]

Not having examined it in detail, I'd guess that the worst spurious responses from the RSA306 will arise from integer-boundary spurs from the MAX2870 LO synth chip they use.  Assuming they can't vary the reference frequency dynamically, there is nothing they can do in real time to dodge those spurs in software. 

There is a magic switch for that exact issue in the MAX2870 according to the datasheet, perhaps turning that on was part of the fix:
If the device is in frac-N mode, it will remain in frac-N mode when fractional division value F = 0, which can result in unwanted spurs. To avoid this condition, the device can automatically switch to integer-N mode when F = 0 if the bit F01 = 1 (register 5, bit 24).

[Rupunzell]

Cannot agree with this practice of scaling testing and evaluation based on expected users and intended market. ALL instrumentation should be subjected to the same testing and evaluation rigor and standards even if these standardized test are extremely time consuming and proves the Instrument Under Test as poor and problem ridden or exceptional.

This line of thinking and methodology would be similar to alter the international standard for one volt to accommodate the instruments intended market.

It also tends to plays into potential user's belief of viewing (measurement) what they want to believe for a given instrument. While this serves to sell instrumentation, it can and will fool less astute users into believe their measurements are completely and absolutely accurate and honest under conditions that the user may not fully understand the possible sources of error and related problems. This specific item is most problematic for those with the least experience and greatest expectations.

Bottom line, one methodology for testing ALL spectrum analyzers regardless of cost, brand, type or else.

Anything less easily falls into deception and deceiving potential customers and their expectations.


Bernice

As someone who regularly uses instruments ranging from $50 to $500k on a daily basis, I can tell you that I have to tailor my reviews to the instrument's intended use and audience. This is inevitable. For example I put the Keysight S-Scope through much tougher tests than I did the USB-based spectrum analyzers (although I specifically showed and complained about the spurious response of the RSA). I specifically design tests to show the instrument's capabilities and limitations and that already takes more than one hour of video and more ten times longer to prepare. It is not matter of expertise necessarily, it is a matter of practicality and time.

[KE5FX]

Not having examined it in detail, I'd guess that the worst spurious responses from the RSA306 will arise from integer-boundary spurs from the MAX2870 LO synth chip they use.  Assuming they can't vary the reference frequency dynamically, there is nothing they can do in real time to dodge those spurs in software. 

There is a magic switch for that exact issue in the MAX2870 according to the datasheet, perhaps turning that on was part of the fix:
If the device is in frac-N mode, it will remain in frac-N mode when fractional division value F = 0, which can result in unwanted spurs. To avoid this condition, the device can automatically switch to integer-N mode when F = 0 if the bit F01 = 1 (register 5, bit 24).

The issue comes in when you're in fractional mode at a frequency close to an integer multiple but not quite coincident with it.  You get a spur at the integer multiple of the PFD comparison frequency.  The data sheet is (deliberately?) very vague about just how bad those spurs can be.  They can be very bad indeed.

Of course, if the wideband IF path is broader than strictly required to support the analysis bandwidth, that could give them a way out of that dilemma.  They may not even need to use those chips in frac-N mode, for that matter.

[hendorog]

The issue comes in when you're in fractional mode at a frequency close to an integer multiple but not quite coincident with it.  You get a spur at the integer multiple of the PFD comparison frequency.  The data sheet is (deliberately?) very vague about just how bad those spurs can be.  They can be very bad indeed.

Ah OK thanks I'll need to read up about it. The reason I'm so interested (sorry, I don't want to derail the thread) is that I've got a couple of MAX2870 eval boards here and I'd like to see the spurs. I haven't been able to make it happen yet, but I'm sure it will if I hit the correct scenario.

Edit: Spurs found

[Wuerstchenhund]

Based on the information available it wasn't really that hard to see where the flaws of the Tektronix device will be and to predict how it will perform. G0HZU also wasn't the only one who came up with that assessment.

To predict and find the spurs that I found required a degree of reverse engineering which I didn't see others doing.

When the Tek came out reviews seemed to pop up all over the places, and I'm pretty sure I read about similar problems in another. Although I didn't pay that much attention as I wasn't particularly interested in this Tek device.

To give an example see the image below. There is only supposed to be one test signal on the screen. But I was able to predict a particular centre frequency and span that displayed the same test signal several times at various levels. eg it displays a copy at -9dBc and another at -15dBc and another copy lower down still. There's only supposed to be one tone displayed!

Yeah, that looks pretty bad. Makes it pretty useless as a general purpose SA.

How was the software?

[pascal_sweden]

Did they sell a lot of that Tek device?

So they did never release a fix under all those years, and the actual fix still has to come?

How many years are we talking about, starting from the launch of the Tek device?

[hendorog]

Yeah, that looks pretty bad. Makes it pretty useless as a general purpose SA.

So they did never release a fix under all those years, and the actual fix still has to come?

You probably saw it by now, but w2aew posted a pic on the previous page showing a improvement with the update.

Shahriar's Signalhound BB60C review compared with the updated Tek at the time he compares the harmonics:
https://youtu.be/3txkDU-qy9k?t=1202

This is his review of the old one, at about the time he shows the harmonics:
https://youtu.be/GDcuRTOCj_s?t=960

[G0HZU]

Cannot agree with this practice of scaling testing and evaluation based on expected users and intended market. ALL instrumentation should be subjected to the same testing and evaluation rigor and standards even if these
standardized test are extremely time consuming and proves the Instrument Under Test as poor and problem ridden or exceptional.

This line of thinking and methodology would be similar to alter the international standard for one volt to accommodate the instruments intended market. It also tends to plays into potential user's belief of viewing (measurement) what they want to believe for a given instrument. While this serves to sell instrumentation, it can and will fool less astute users into believe their measurements are completely and absolutely accurate and honest under conditions that the user may not fully understand the possible sources of error and related problems. This specific item is most problematic for those with the least experience and greatest expectations.

Bottom line, one methodology for testing ALL spectrum analyzers regardless of cost, brand, type or else.Anything less easily falls into deception and deceiving potential customers and their expectations.

Bernice

Up until maybe 5 years ago I would completely agree with you and so would a lot of engineers. But I really do think that there has been a change away from the traditional (and expected) design integrity of a decent spectrum analyser. I think a significant chunk of today's users don't care much about spurious free dynamic range or many of the other benchmark tests associated with a spectrum analyser. i.e. they won't care (or even understand or even want to understand) the limitations of the RF converter in the instrument. They will simply want to buy it and use it and just accept what they see on the display unless there is an obvious anomaly.

For many users the features that are offered by the back end DSP and the amount of analysis bandwidth are the key features they want. If they can get these features cheaper by buying an instrument without the classic design
integrity expected in the RF converter of of a traditional analyser they will choose that option because they probably don't fully understand what tradeoffs they are actually making. In many cases their decision will prove to be OK because they may well never use the instrument in a way that shows up the design compromises of the cheaper instrument.

So I think it's a case of giving the customer what they want and being less open about the true limitations of the instrument. For example, the glossy datasheets for lower cost analysis instruments (from the major manufacturers) often concentrate on the back end analysis features but will deliberately be very vague about the specifications and capability of the RF converter section.

So you generally won't see the classic performance data that you normally expect to see from a classic HP/Agilent lab analyser. Less info here is 'more'

[hendorog]

So you generally won't see the classic performance data that you normally expect to see from a classic HP/Agilent lab analyser. Less info here is 'more'

Less price is more too Less weight, less age, less noise, less shipping cost, less repairs

For many users the features that are offered by the back end DSP and the amount of analysis bandwidth are the key features they want. If they can get these features cheaper by buying an instrument without the classic design
integrity expected in the RF converter of of a traditional analyser they will choose that option because they probably don't fully understand what tradeoffs they are actually making. In many cases their decision will prove to be OK because they may well never use the instrument in a way that shows up the design compromises of the cheaper instrument.

It's incorrect to assume people will all be tricked into buying a cheaper instrument. Many people - myself included - don't have the budget to afford the new prices commanded by the high end gear so look at ebay. Ebay is only useful for this if you happen to be close as long haul shipping these elderly boat anchors is expensive and risky.

Google is an easy way to find out things the manufacturer doesn't tell you before you buy. I bet a heap of people on this site found it because they were looking for an honest review of something or other.

Also I think there is a danger of lumping all of these USB devices into one bucket. In fact they are actually quite different.

The OP was looking at the SA44, which is an alternative to the traditional SA, just much cheaper.
The Tek and the BB60C are wideband and as you and w2aew mentioned, offer something that an old school SA can't do.

Signalhound are pretty upfront about what their devices can and can't do, and the company has a background of repairing HP spectrum analysers FWIW.

I have the SA124 and I think it's a fantastic tool when paired with the TG, and I haven't regretted buying it. I didn't look at the Rigol SA because it didn't offer the bandwidth. 12GHz vs 1.5Ghz was a bit of a no brainer for a bit more money. Since then Rigol have released higher bandwidth devices, but the prices are too high.

[G0HZU]

It's incorrect to assume people will all be tricked into buying a cheaper instrument. Many people - myself included - don't have the budget to afford the new prices commanded by the high end gear so look at ebay.

I didn't really mean 'all' people and I'm not sure 'tricked' is the right word to use to describe what I mean. I think there is an explosion of a new type of (casual?) user today who doesn't have the usual background that would normally lead them towards buying a spectrum analyser. I can remember back to the 1980s when only a select few individuals owned a commercial spectrum analyser and back then, not many people would actually want one anyway.

Yet today you can find people from all walks of life wanting to buy 'dongle' SDR type analysers or instruments like the Rigol 815 or the Signalhound.

Many of these people will not know about (or be able to give figures for) the more advanced figures of merit for a spectrum analyser or how to test for them and I'd argue that they probably aren't that interested in this info either. If they were interested then the various online blog reviews of spectrum analysers would be awash with critical questions about what key performance data was missing from the analyser review. I'm talking about classic, basic tests of a spectrum analyser.
 
I think the various manufacturers are aware of this and so a new breed of analyser is emerging that is geared towards this type of user. i.e. something cheap but flawed but the target user doesn't mind or care and probably doesn't want to fully understand the limitations of the instrument. They just want to play with the power of the back end processing and they aren't really that concerned about the design integrity of the RF converter ahead of it.

[tooki]

The general issue with using a computer is that the operating system could start to exhibit stability issues, when you install too many applications on it. [...]

OK, so I know this is basically off-topic, but that's just nonsense. If there are conflicting drivers, this can cause stability problems (though it's rare on modern OSes where drivers run in user space), but the number of applications installed is completely irrelevant to system stability.

[G0HZU]

How was the software?

Sorry, I missed that question.

I only had the RSA306 analyser for a few hours and my brief was to study the RF converter and report back. However, I obviously did have to use the PC GUI that came with the instrument.

In my opinion the actual spectrum display was quite pleasing and it really is nice to see a big colour screen showing the spectrum. But the user interface for the classic analyser controls was obviously written by a disinterested software engineer (AKA a 'typist'    )  because the control menus were very fiddly to use and placed far apart. The overall experience was a bit clunky and slightly irritating because it could easily have been a lot better. But I suppose this can easily be changed with an upgrade if they have another attempt with some input from a typical end user.

I didn't have time to play with data capture or data processing on the PC but I expect that this is one area where the instrument will shine because Tek already have a fair bit of experience in the RTSA market.

[ElectroIrradiator]

(...) I think there is an explosion of a new type of (casual?) user today who doesn't have the usual background that would normally lead them towards buying a spectrum analyser. I can remember back to the 1980s when only a select few individuals owned a commercial spectrum analyser and back then, not many people would actually want one anyway.

Yet today you can find people from all walks of life wanting to buy 'dongle' SDR type analysers or instruments like the Rigol 815 or the Signalhound.

Many of these people will not know about (or be able to give figures for) the more advanced figures of merit for a spectrum analyser or how to test for them and I'd argue that they probably aren't that interested in this info either. If they were interested then the various online blog reviews of spectrum analysers would be awash with critical questions about what key performance data was missing from the analyser review. I'm talking about classic, basic tests of a spectrum analyser.
 
I think the various manufacturers are aware of this and so a new breed of analyser is emerging that is geared towards this type of user. i.e. something cheap but flawed but the target user doesn't mind or care and probably doesn't want to fully understand the limitations of the instrument. They just want to play with the power of the back end processing and they aren't really that concerned about the design integrity of the RF converter ahead of it.

I would like to offer a different perspective, as I am personally quite seriously thinking about putting my own money down for one of these 'new breed' of spectrum analyzers. Specifically the one built into the Tektronix MDO3000 range of DSOs.

Some personal background may be in order. I will not claim to be the most exprienced RF/high frequency guy on the planet, nor the most sophisticated SA user there ever was. However I would like to believe my level of sophistication is somewhat beyond what you just described. I certainly do care about the detailed specifications of any instruments I own or may consider acquiring. If there are relevant techniques or specifications I happen not to know about in connection with them, then I'd certainly like to learn about them ASAP.

I do not believe the SA in the MDO3000 range is the best there ever was, not am I blind to its various limitations. Some of these are: SFDR, IMD, (apparent?) phase noise and lack of a TG. Sometimes you can work around these limitations when they get in your way, and sometimes you can't.

My philosophy here is that despite these limitations a SA is such an immensely useful tool, that I personally consider it to be a key part of the proverbial swiss army knife for the interested RF/high speed electronics experimenter.

Most of the time, when I use one of my instruments, I don't even scratch the possibilities nor push the limits of a given piece of gear. It is mostly bread-and-butter measurements, where I just use a high quality instrument because I happen to have it. If said piece of equipment magically lots the top of its high performance, it would still be quite useful for a lot of use cases, for much of the time I would spend with it.

The trouble I sense in for instance the online reviews are two fold.

Firstly, the super experienced people like Shahriar have a huge selection of gear to choose from. So he might not use a SA just to measure the power level of an oscillator, instead picking one of his dedicated high precision RF power meters for the job. But if you lack such an instrument, then a SA can of course easily do something as simple as this. Assuming, that is, you are happy about the accuracy of the measurement you get. For many types of experiments and personal educational tasks then yes, plus/minus a dB or so will do quite nicely. We are not all trying to develop the latest and greatest in satellite receiver technology, or something of similar sophistication.

On the other hand today only people like yourself, who have lots of experience with SAs and RF, also tend go out of your way to acquire one. This is a catch 22 for all the rest of the people out there. Today a SA is hard/expensive to get (new or possibly even just used), so few people without access to one know or realize just how useful they can be in many situations. So when you give them one to review, they frequently won't know how to demonstrate how useful it can be, despite any limitations.

A different way of putting this is thinking of DSOs versus analog scopes. Imagine you only have experience with fully analog scopes. Then suddenly somebody puts a Rigol DS1052E in your lap and asks for a review. I will submit that most people in that situation would not showcase the DSO to its full capacity. Yet despite this few people with experience with DSOs would argue that even a DS1052E - with all its limitations - doesn't make a lot of new measurements possible, compared to those available just with access to an analog scope.

If I do decide to buy one of the MDO3000 scopes, then I won't suddenly expect that I am, say, equipped to do proper EMC testing. Yet since I don't already have a SA in my home lab, then I'd *much* rather have a neat, new instrument like the MDO3000 than be completely without one.

The main reason why I don't already have one of the MDO3000 scopes is that I am not a famous blogger, who knows how to complain about the 'slow' interface. So I will unfortunately have to pay for it myself.

It is getting late, so I'd better stop here. Hope I am making at least a bit of sense.

[G0HZU]

The main reason why I don't already have one of the MDO3000 scopes is that I am not a famous blogger, who knows how to complain about the 'slow' interface. So I will unfortunately have to pay for it myself.

I'm not sure who you are referring to here but you did make me laugh out load 

It does seem odd that manufacturers give out $$$ test gear as gifts to bloggers like this. What is even stranger is that they don't seem to demo them properly. But I guess my version of a 'proper demo' isn't mainstream or box office. I guess I'm making the same point over and over...

I noticed that one part time blogger has been gifted a Tek scope/analyser that must have cost more than I paid for EVERY item of test gear I have ever bought over a 30-35 year timeframe. i.e. the entire contents of two rooms including my furniture.  If you include all their other freebies it's probably nearly double

[w2aew]

How was the software?

Sorry, I missed that question.

I only had the RSA306 analyser for a few hours and my brief was to study the RF converter and report back. However, I obviously did have to use the PC GUI that came with the instrument.

In my opinion the actual spectrum display was quite pleasing and it really is nice to see a big colour screen showing the spectrum. But the user interface for the classic analyser controls was obviously written by a disinterested software engineer (AKA a 'typist'    )  because the control menus were very fiddly to use and placed far apart. The overall experience was a bit clunky and slightly irritating because it could easily have been a lot better. But I suppose this can easily be changed with an upgrade if they have another attempt with some input from a typical end user.

I didn't have time to play with data capture or data processing on the PC but I expect that this is one area where the instrument will shine because Tek already have a fair bit of experience in the RTSA market.

FYI - the latest version of the UI (SignalVu-PC) has moved some controls and such around on the display - to group them together more logically and functionally.  So, rather than have Center Frequency, Reference Level, Span and RBW located far from each other - they are now grouped together in a ribbon along the bottom of the UI - making it easier to find and adjust.  They were placed on the bottom so that tablet users don't cover the display with their hand when they go to adjust the "big four".  I think you can see this on the screen shot that I posted earlier showing the SDFR improvement (under those conditions).

[KE5FX]

We looked at Copper Mountain last summer.     

https://www.coppermountaintech.com/

What'd you think?

[Rupunzell]

Bad drives out the good, pandering to what is know to be a larger market to increase profits.

There has and always will be a hobby electronics market. This is a very GOOD thing as it allows enthusiast access to tools (instrumentation and such) to further their passion, interest and knowledge of electronics overall.

What has happened with the boom of cyber based everything is the discounting and alteration of user expectations of tools and information presented and accessed by this generation of users. It remains best in many ways to be honest and straight forward about the limitations of low cost test gear, point out their inherent limitations and strengths rather than simply advertise-market on features that appear to be useful. It seems today, any techno widget that offers less than a complete set of every imaginable feature is not desirable. Couple that with the user believe that if they are not getting every possible feature, they are not getting their money's worth.

Readers, followers should be aware of the potential bias built into any published review, test or advertisement as those who do this work need funding to keep their lights on, question is where does the funding and access to items being reviewed come from?

Those involved with promoting and reviewing these USB based instrumentation should be honest and frank about their limitations and what they do best and educate their audience on why these limitations and strengths matter. This methodology is in the tradition of application notes and papers written to help their end users get the very most out of their offerings and further their education on each specific topic.

I'm not convinced that USB and PC based instrumentation is the solution to many instrumentation needs. This has happened due to the amount of computing power and capability for their cost. What is almost never discussed or mention is the fact that any switching device can and does generate RF interference both radiated and conducted. How much varies greatly dependent on the specific device in question. How do these factors affect measurement accuracy and how much error is a result of these innate problems. This does not mean USB and PC based instrumentation are all poor, it just means one needs to be fully aware of these potential problems and never take them for granted.

How many O'scope users have been fooled and spent countless hours trying to figure out a problem that never was as a result of high distortion probes, poor probing techniques and numerous related basic measurement skills errors?

The current generation of computer, keyboard, software centric users can be so easily lure into a sense of security by what has been presented to them on screen or by software, yet the reality of what is actually happening in a circuit or system can be completely different. The only real defense against this self deception is to fully understand what any given measurement should be rather than completely trusting any item of instrumentation or data presented by software via any display.

Software should be bug free as delivered. Updates should only happen when absolutely required for functionality and overly common practice of updating OS and software as a must is pure folly IMO.

What should be most valued in any spectrum analyzer is it's front end. All that follows is mostly dress up and could be added as required. This is the beauty of computer centric instrumentation, software can be modified as required. But users should be fully aware that actual performance and limitations are baked into the front end and no amount of software can make up for poor front end performance.

I'm guessing that most would consider me a bigoted cronemudgeon who has rather fixed expectations of what quality instrumentation must be and not willing to compromise.

The first spectrum analyzer owned was a hewlett packard 141T system with several plug ins, tracking generator and pre-selector during the 1980's. This was not an inexpensive item to have in the home lab, it was used a LOT and eventually replaced with newer SAs over time. I'm still of the opinion that ownership of a SA like the 141T or 140T remains quite valuable as a learning tool as full documentation is easily available with spares available as needed. What desist and discourages ownership of instrumentation like the 140 series spectrum analyzers is their size, bulk and lack of PC friendly and centric interface.



Bernice


[/quote]
Up until maybe 5 years ago I would completely agree with you and so would a lot of engineers. But I really do think that there has been a change away from the traditional (and expected) design integrity of a decent spectrum analyser. I think a significant chunk of today's users don't care much about spurious free dynamic range or many of the other benchmark tests associated with a spectrum analyser. i.e. they won't care (or even understand or even want to understand) the limitations of the RF converter in the instrument. They will simply want to buy it and use it and just accept what they see on the display unless there is an obvious anomaly.

For many users the features that are offered by the back end DSP and the amount of analysis bandwidth are the key features they want. If they can get these features cheaper by buying an instrument without the classic design
integrity expected in the RF converter of of a traditional analyser they will choose that option because they probably don't fully understand what tradeoffs they are actually making. In many cases their decision will prove to be OK because they may well never use the instrument in a way that shows up the design compromises of the cheaper instrument.

So I think it's a case of giving the customer what they want and being less open about the true limitations of the instrument. For example, the glossy datasheets for lower cost analysis instruments (from the major manufacturers) often concentrate on the back end analysis features but will deliberately be very vague about the specifications and capability of the RF converter section.

So you generally won't see the classic performance data that you normally expect to see from a classic HP/Agilent lab analyser. Less info here is 'more'
[/quote]

[rs20]

We looked at Copper Mountain last summer.     

https://www.coppermountaintech.com/

FWIW, those are slightly different beasts: VNAs rather than the SNAs, and accordingly $8k instead of $1k5.

[vk6zgo]

Even quite reasonable "standalone" SA's can fool you.
I've spoken before about the very nice IFR unit we had at my last work.

A PLL on the DUT had lost lock,& shifted from 433MHz to its default frequency of 470MHz.
(The fault was lost data due to a faulty plug on the ribbon cable from the PIC.)

The IFR showed this as a number of carriers distributed over that frequency range.
(Every time it sampled,the carrier was at a different frequency,so each sample was displayed as a discrete frequency.)

We spent ages trying to find the cause of all these spurii!!

A nasty,"cheapie" Atten analog device showed us the carrier as it shifted,& the error of our ways!