Tiers of gear

[LogicalDave]

I think many EEVblog readers in the RF topic area would appreciate a task-based guide to selecting an SA and a VNA.  There are a bewildering wealth of choices these days, but affordable (I know that's a subjective term) equipment seems to fall into four broad tiers:

• entry-level/hobby gear like TinySA Ultra and NanoVNA SAA-2N
• budget bench gear from Rigol, Siglent, Owon, etc.
• USB gear from Libre, Copper Mountain, etc.
• used (eBay) pro gear from HP/Agilent/Keysight, R&S, Anritsu, etc.

Because the bottom end capabilities have steadily improved (even the entry-level devices are no longer toys), it has become harder to know when you need to step up from one tier to the next, often at substantial cost. I think many readers would benefit from examples or a task-oriented guide to what you can, or more importantly cannot do with each tier of equipment.  If you're designing a custom 868MHz PIFA, do you really need a $5K (used) Keysight VNA?  How about sweeping filters? characterizing components?  Optimizing antennae and feed lines?  Tuning diplexers?

Assume the reader understands what SAs and VNAs are used for, what a TG is, what the frequency range and power levels are for their application (if not, there are loads of other posts covering those).  The question I'm asking is: for what common circumstances is the greater dynamic range of a LibreVNA *required* over a NanoVNA (or SVA10xx)?  When does the lower phase noise of a R&S SA really matter over an SSA30xx? What *tasks* really force you to step up a higher tier of equipment for a given type of functionality (SA or SA+TG or VNA)?  I understand that there are a lot of scenarios, but thanks in advance for any examples that could help folks understand which tier of gear they need.

[coppercone2]

ooohh  A measuring contest

the best ones are damn big and have veiny cables
or are the best ones the ones with capable operators?!

[EggertEnjoyer123]

Some of the disadvantages of hobby level gear I can name:
1. It's easy to blow them up. The LiteVNA and TinySA can only take +10dBm before blowing up. They also can't take a lot of DC voltage. (To be fair, it's easy to blow up the HP spectrum analyzers too in 0dB attenuation mode, and they also do not like DC voltage). The advantage the HP spectrum analyzer has is that it has an attenuator, so you could put in +30dBm as long as you have everything set correctly.
2. The bandwidth is usually limited (~6GHz).
3. Low dynamic range. If I'm testing amplifiers on the LiteVNA I usually add a 10dB attenuator on the input and a 20dB attenuator on the output because I don't want to saturate the amplifier or blow up the VNA (see point 1). That means I lose 30dB dynamic range with S21 and 20dB dynamic range with S11, which is quite a lot considering that the dynamic range is usually around 50dB at the higher frequencies.
4. Can't set output power easily (have to add attenuators).
5. Can't do all 4 S parameter measurements. If I want to check the stability of an amplifier, I need to physically reverse the connections and then use a Python script to plot the stability factor.
6. More variability. I tried comparing the LiteVNA performance to a professional VNA, and the LiteVNA would deviate by around 1dB with no averaging (see picture where I measured a filter). If you average 25 samples, you get fairly close to the professional VNA. Averaging 25 samples takes forever though. Of course if I just want to check whether my amplifier is matched I don't really care.

I have a LiteVNA and a HP 22GHz spectrum analyzer that I bought for $325 (not too far off from the $100 that you'd need for a tinySA)

[Bicurico]

I would say that it all comes down to your particular requirements and budget.

Examples:

1) Your budget has no limits: you buy a R&S spectrum analyser and nobody will ever point the finger at you for having purchase the wrong equipment
2) You have a very small budget, need to visualize the spectrum from 1GHz to 2GHz: you buy a TinySA Ultra
3) You need to carry out pre-complicance EMS tests, have a limited budget and need to document measurements: perhaps a Siglent class device will fit well

A - A hobby user, student or hacker with limited budget might find the TinySA Ultra a perfect entry level equipment.
B - A more serious user or professional user will go for Siglent or Rigol.
C - A professional user with special requirements (calibration, certification, precision) will go for Keysight, R&S and likes.

The second hand market for HP, Keysight, Agilent, etc., test equipment is normally for those that know how to repair and maintain such devices. These are C-type (see above) users, that get a kick of privately owning equipment they have used on their job 10-30 years ago. I would say that it is more a TEA issue than a reasonable purchase.

Hope this helps, this is just my subjective opinion.

[LogicalDave]

Thank you @EggertEnjoyer123 and @Bicurico for sharing your thoughts.  I agree with @Bicurico's categorization of users (A,B,C) and which tier they should generally buy.  Nevertheless, I think it would be useful for many users to understand specifically what they need each tier for given the big cost steps.

Consider tier 1: I've owned several generations of TinySA; initially, they were just a curiosity; the idea that you could look at RF spectrum with a $50 device that fit in your shirt pocket was nuts.  However, their limitations precluded most practical applications (at least above HF/VHF).  However, each generation improved (RBW, amplitude/freq accuracy, etc.) and despite the many well documented limitations, I now often grab my TinySA Ultra FIRST when I want to see if an IoT device is transmitting at roughly the right level or frequency or is frequency hopping properly; I often don't care whether the signal is +20 or +22dBm, I just want to know it's in the general vicinity.  Similarly, when I want to quickly check the resonant frequency of an antenna, I often grab the NanoVNA SAA-2N out of the same desk drawer.  Although I haven't used it for this purpose, from what I can see, the NanoVNA is quite capable of sweeping filters, tuning antennae, characterizing components, and many of the most common VNA applications through 3GHz *as long as high accuracy and big dynamic range isn't necessary* (and it often isn't).

My tier 2 gear comes out when I am spending significant time building/measuring/debugging something and must make precise amplitude/frequency measurements below 3GHz.  This is mainly for accuracy and convenience. In the past, I mainly used cellular field service gear from eBay (MT8222a, E7495, etc.), but now I grab my SVA1032x; its capabilities are similar: tracking generator, good amplitude and frequency accuracy (and the ability to slave it to a lab standard), extensive suite of automated measurements, but it is new, calibrated, has a bigger screen, modern UI, easier PC interfacing, etc. all of which make it quite useful for most of my practical RF development and testing.  If you use an SA or VNA often or for work and your time is valuable, tier 2 is an easy decision, but for hobby/home use, I suspect that the large majority of practical use-cases for both SA and VNA can now be handled using the combination of TinySA4 and NanoVNA at a much lower price.  If you can think of important practical tasks below 3GHz that can be done with tier 2 gear but not with the Tier1 SA+VNA combo, I'm interested in them...that's really what this post is about.

I don't have any tier 3 gear and am interested in where others see it fitting in.  Exactly what do you do with it that could not be done with a NanoVNA or SVA?  When should a user step up from tier 2 to tier 3?

These days, I only grab my (old, big, heavy) tier 4 gear when I really need to.  I mainly use it to prepare for compliance testing when I need to accurately measure (or suppress) spurious/harmonics in protected bands up to 10GHz.  I may occasionally use it when I need greater accuracy than tier 2 gear provides, but that is rare.  So I agree with @Bicurico about who buys tier 4 gear (user type C), but disagree that tier 4 is mainly a TEA issue; it still seems to be a necessity for testing that must be done before bringing products to the certification lab; I've never been able to get by without a good SA with 10-22GHz BW for this reason.

So @coppercone2, I am not interested in a measuring contest...there are many of those already.  The question I'm asking is: what practical tasks can you actually DO with each tier of gear that you can't do with the lower tiers?  I've offered a few examples but am interested in more and believe others likely are too.

[coppercone2]

the question is so broad and the answer is so highly dependent on skill/knowledge and opinions of quality/confidence that this is probobly not going to get good answers. its going to be 'well if you just did that 1 step then' or 'if you can accept ....' and 'we just don't need that because.... *something about confidence that may or may not be misplaced*' or 'well that is absolutely improper'

I just expect a fight to break out with people trying to boast about financial efficiency. You can't even come to a conclusion on how cheap you can go with a passive component like a resistor. Some people care about thin leads and some don't ? . or wire insulation quality requirements.

like you can start a 30 page flame war trying to push teflon wiring

 It seems like just a way to try to greenlight some thing lol

broad question on judgement for complicated science topic relating to the economy.

some people think its silly to have any tiers and that its either excellence is a habit or that everything is the same. you can argue going cheap for teir 1 means you can miss some important exploration. like Columbus cheaping out on the compass and not finding america (or whoever, vikings). or that its always a waste of money so long your skilled enough. its personality, possibly personality that you HAVE to have to fit in within a particular organization. the risk is that you get it completely wrong and form an incorrect opinion on the viability of something or nother.

[LogicalDave]

@coppercone2, if you are a forum moderator and believe this topic is inappropriate or unlikely to be lead to productive discussion, then I will, by all means, withdraw it.  If you are not, but think the question is too broad, perhaps you can suggest a proper narrowing. If you think proper answers depend on the skill/knowledge of the reader, then perhaps you can pick one skill level and answer in ways useful to those readers. I have no interest in a flame war with you or anyone else, I'd just like to have a productive discussion of the topic.

If the topic was poorly stated or insufficiently clear; I'll try to restate and summarize it: there is a lot of new and interesting SA and VNA gear spread across a much broader price range than has historically been the case.  There are many threads covering tear-downs, specifications, and recommendations, but very few that discuss specific applications and why a particular tier of gear would be appropriate for a particular application.  I offered a concrete example of why one might need a tier 4 SA with greater bandwidth than tiers 1 or 2 typically provide.  I expect that there are similarly common use cases that demand, for example, the larger dynamic range of a tier 3 or 4 VNA over tier 1 or 2 and I'm interested in what they are.  I'm looking for answers like: I use a LibreVNA to do X because it would either take me hours to do the same thing with a NanoVNA or SVA1032x or I couldn't do it at all due to insufficient dynamic range or precision or lack of some key feature.

[joeqsmith]

For the VNA,

What frequency range, resolution, power range do you need?   My latest VNA for example is limited to 300kHz.   Personally, I would have liked to support lower frequencies for PDN and other measurements.   High side is now limited to 9GHz.  TDR resolution suffers.  I couldn't use it for my waveguide experiments.   This one has a resolution of 1Hz and provides support for a reference clock.   

What sweep type (CW, linear, log ...), number of data points, and how fast do you want it to sweep?   

Type of interface connectors, number of ports, bias-Ts are needed?

Dynamic rage? 

Accuracy? 

E-Cal, unknown Thru, TRL.....? 

Software?  That would be a lengthy topic.   I ended up buying a used VNA rather than a new Pico Technology because of their software.   

I've never tried to blowup any of my equipment and can't suggest that one would be more robust than another.   This seems more like a user problem.

If you don't know what you need, my advice is don't buy anything.  Or start with one of those $50 NanoVNAs and learn some of the basics.  Worse that happens is you damage it and are out $50, or you come to conclusion this is complex topic and you are in way over your head and the VNA gets tossed into the junk drawer.   

I also suggest finding some books on the subject.   

Sure the NanoVNA misses a lot of features.  Then you have that squarewave drive you need to deal with.   I made this video showing a typical problem you may run into.  The real problem is the user may not have the skills to know there is even an issue with stability.

[joeqsmith]

Recently a member posted about their new V2Plus4 Pro.  I had them look at the noise floor and compared it with some of my low cost VNAs.  Spending more doesn't automatically mean you will achieve higher performance. 

https://www.eevblog.com/forum/rf-microwave/nanovna-custom-software/msg5039302/#msg5039302

[xmo]

With respect to one question asked by LogicalDave: "When does the lower phase noise of a R&S SA really matter over an SSA30xx?",  I am sure others can post many professional level measurement situations where low phase noise matters, but here's just one.

The TIA provides industry guidance through many standards and bulletins including documents containing recommended methods for performance modeling of land mobile communications systems utilizing a variety of modulation standards (e.g. Analog, DMR, Tetra, P25, etc.).

Contained therein are descriptions of the modulation formats including SPD graphs and ACPR data.  The basis for these is data obtained by using a spectrum analyzer.  The TIA provides the software tools and instructions necessary for anyone to perform these measurements individually simply for the educational value or to compare the performance of a format not included by the TIA - e.g. D-Star.

Not as a measurement 'contest' but more of a challenge, I have posted this same set of facts on various forums - as yet - zero takers willing to step up with their own instruments and skills.

There are several complexities, including the need for low phase noise and the ability to configure the instrument correctly to acquire a data fie with the necessary format.

The instructions state that data points are collected for +/- 50 KHz (100 KHz Span) in a bin size of 31.25 HZ for a total of 3201 points. The spreadsheet tool expects that resolution.  Siglent? Rigol?

To show how the phase noise issue affects the results, examine the attached comparison which shows the SPD of a P25 C4FM signal with an ACPR highlighted.  The image on the left is directly out of the TIA standard document.  The one on the right is one I created using the TIA tools and a data file I created using an Agilent CXA spectrum analyzer.

Clearly, even this $20,000 instrument falls far short of accurately depicting the true performance of the modulation format as captured from the UUT.

To match the TIA results requires the extremely low phase noise of a high tier instrument such as an Agilent (Keysight) PSA or PXA.

[Wallace Gasiewicz]

Quote from: Bicurico on September 10, 2023, 07:44:25 am>

I would say that it all comes down to your particular requirements and budget.

Examples:

1) Your budget has no limits: you buy a R&S spectrum analyser and nobody will ever point the finger at you for having purchase the wrong equipment
2) You have a very small budget, need to visualize the spectrum from 1GHz to 2GHz: you buy a TinySA Ultra
3) You need to carry out pre-complicance EMS tests, have a limited budget and need to document measurements: perhaps a Siglent class device will fit well

A - A hobby user, student or hacker with limited budget might find the TinySA Ultra a perfect entry level equipment.
B - A more serious user or professional user will go for Siglent or Rigol.
C - A professional user with special requirements (calibration, certification, precision) will go for Keysight, R&S and likes.

The second hand market for HP, Keysight, Agilent, etc., test equipment is normally for those that know how to repair and maintain such devices. These are C-type (see above) users, that get a kick of privately owning equipment they have used on their job 10-30 years ago. I would say that it is more a TEA issue than a reasonable purchase.

Hope this helps, this is just my subjective opinion.

I think that the older gear like HP definitely requires repair abilities as Bicurio states.I have a lot of it and it is all old and it requires occasional repair, so if a person wants something that is going to work, that person needs to consider the considerable time and effort that goes into the maintenance of these old machines. Many of us who have these old things spend a lot of time on the forums getting help with repair. Just look around the Forums.
It is at least a good part a TEA issue. Especially for people like me who are relatively new to these things.On the other hand, there are a lot of resources and help available for repair.
Also with the big old stuff you need a room that is big enough to keep the equipment and a lot of AC power. It also keeps the room quite warm and if you have a bunch of it the temp of the room can be a problem for the equipment . Especially if you are concerned about temp coefficients when evaluating a device.

On the lighter side, I used to have a rule that I could not buy it if I could not carry it out to my car, On this note, anyone want a Marconi 2380 Spectrum Analyzer??? HA!!

Just some thoughts.

[jwet]

Good thread, I've been thinking about this lately as I get older and am thinking about thinning out my "Type 4", lab.  Its taken me a lifetime to accumulate all this stuff, most is stuff I've used at work at some point and have fond memories and "always wanted".  Most required some repair or at least TLC and alignment.  I'm in my 60's and  live in a big empty nest house with plenty of space but if I ever moved, I'd have to make some tough decisions.

I think Logical Dave asks the right question- What practical use are these tiers as far as functionality?

I can and do make do with Type 1 through 3 equipment often.  I have a vacation house and have a mini lab that is all sub type 4 equipment.  The nano VNA is a really special duck- it has so much functionality.  I can do 95% of what I need to design and service gear with Type 1 - 3 type equipment.  Type 4 generally brings quantitively valid results, dynamic range and repeatability.  This is epecially true for a VNA- dynamic range means being able to see more subtle effects.  I also just like working with one or two generation old pro gear- I feel like I can do anything.

[G0HZU]

I generally only have Tier 4 gear here. Nothing from Tier 2 from Rigol, Siglent, Owon or Hantek etc. I also have an Analog Discovery 2 and I guess that might be classed as USB gear from Tier 3.

The only Tier 1 gear I have is an original Hugen nanovnaH and this is a very impressive device for the price, especially when used below about 50MHz. I do intend to buy a Tiny SA Ultra at some point.

I think the Tier 4 gear could become Tier 4 and Tier 5. This would reflect the difference between mid range pro test gear (from HPAK and R&S and Anritsu etc) and the really exotic and incredibly expensive high end pro test gear (also from HPAK and R&S and Anritsu etc).

I've got a few items here that would probably qualify as Tier 5 that each would have cost >$50k when new. Some would have cost over $100k new maybe 10-15 years ago but can be purchased for a fraction of this cost today.

[G0HZU]

Some examples of mid range pro RF test gear would be
Marconi 2024 sig gen
HPAK ESG(D) sig gens
HPAK EXA and MXA spectrum analysers
R&S FSP spectrum analyser
R&S ZVB VNA
HPAK ENA network analyser


Some examples of $$$ high end pro RF test gear would be:
HPAK PSG sig gen
HPAK PSA and PXA/UXA spectrum analysers
HPAK PNA-X network analyser
Tek 5000 and 6000 series RTSA analysers.
R&S FSV and FSW spectrum analyser