Author Topic: What will a "proper" VNA do that an SA and tracking gen won't?  (Read 4849 times)

0 Members and 1 Guest are viewing this topic.

Offline Chris Wilson

  • Super Contributor
  • ***
  • Posts: 1029
  • Country: gb
  • Race car engineer, dog lover, hoarder.
What can a proper VNA do that a spectrum analyser, tracking generator and directional coupler will not be able to do, assuming both have the same frequency ranges? Thanks.
Best regards,

                 Chris Wilson.
 

Offline JackOfVA

  • Supporter
  • ****
  • Posts: 350
  • Country: us
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #1 on: June 02, 2015, 11:17:57 am »
Measure phase, so that you can determine the complex impedance of the D.U.T.

 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 4210
  • Country: ca
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #2 on: June 02, 2015, 05:02:16 pm »
Also SA will typically not correct for systematic errors, i.e. does not remove effect of connections, input impedance of SA and output impedance of TG on measurements.
Facebook-free life and Rigol-free shack.
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 9765
  • Country: my
  • reassessing directives...
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #3 on: June 02, 2015, 05:10:33 pm »
Also SA will typically not correct for systematic errors, i.e. does not remove effect of connections, input impedance of SA and output impedance of TG on measurements.
dave showed how to correct that using rigol dsa815 SA feature in video #343.
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline hagster

  • Frequent Contributor
  • **
  • Posts: 360
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #4 on: June 02, 2015, 05:58:29 pm »
Lots of things.

Primaraly it measures both phase and amplitude rather than just amplitude. Good ones also do time domain measurements such as group delay, TDR and gating.

The other key thing is that they measure reflection (S11) as well as transmission (S12).

You can do some stuff with a spectrum analyser, tracking generator and a directional coupler, but it's not coherent (ie doesn't measure phase).
 

Offline hagster

  • Frequent Contributor
  • **
  • Posts: 360
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #5 on: June 02, 2015, 06:04:33 pm »
That said a VNA is not much use for non linear devices that generate harmonics etc. Unless you get one of the just announced Agilent VNAs with a simultaneous SA mode that is.
 

Offline KJDS

  • Super Contributor
  • ***
  • Posts: 2442
  • Country: gb
    • my website holding page
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #6 on: June 02, 2015, 06:16:23 pm »
Also SA will typically not correct for systematic errors, i.e. does not remove effect of connections, input impedance of SA and output impedance of TG on measurements.
dave showed how to correct that using rigol dsa815 SA feature in video #343.

Whilst that will sort of correct, you won't correct for source match errors and you won't correct for load match errors. Having said that, I've done a lot of work when I was young just using a spec ana with a tracking gen, but a VNA makes life so much easier. If you want to improve return loss, then the spec ana and a bridge will tell you the magnitude, but with a VNA you will know the phase as well and that is necessary if you wish to calculate a circuit to correct the return loss.

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #7 on: June 02, 2015, 07:16:14 pm »
Quote
You can do some stuff with a spectrum analyser, tracking generator and a directional coupler, but it's not coherent (ie doesn't measure phase).

There are some (usually old and expensive) spectrum analysers that can do this.

My old Advantest 4172 spectrum analyser has a built in tracking generator but it is also fitted with a factory option for measuring impedance. This requires an external bridge but in this mode it can do some things a VNA can do. Eg measure impedance.

But a decent modern VNA will outclass it for speed and accuracy. Also a decent 2 port VNA can correct for source and load mismatch uncertainty when SOLT calibrated and it's possible to measure something like an unmatched xtal/bandpass filter (or an unmatched amplifier) just once and export an s2p model to a PC and then do all the small signal matching on an RF simulator.

Then build it for real and check it on the VNA :)
« Last Edit: June 02, 2015, 07:20:31 pm by G0HZU »
 

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #8 on: June 02, 2015, 07:59:38 pm »
Quote
Primaraly it measures both phase and amplitude rather than just amplitude. Good ones also do time domain measurements such as group delay, TDR and gating.

Group delay is just rate of change of phase wrt frequency so I'd expect the vast majority of VNAs to be able to display group delay.

For this reason my old Advantest spectrum analyser can also measure group delay using its internal tracking generator as one of its party tricks. I think this is a standard feature rather than an option. The impedance measuring option needs the external bridge.
 

Offline hagster

  • Frequent Contributor
  • **
  • Posts: 360
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #9 on: June 02, 2015, 08:07:24 pm »
Group delay is just rate of change of phase wrt frequency so I'd expect the vast majority of VNAs to be able to display group delay.

For this reason my old Advantest spectrum analyser can also measure group delay using its internal tracking generator as one of its party tricks. I think this is a standard feature rather than an option. The impedance measuring option needs the external bridge.

I haven't come across one of these 'vector' spectrum analysers before. They sound interesting. I guess if you had one of these you may not need a VNA.
 

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #10 on: June 02, 2015, 08:19:30 pm »
Group delay is just rate of change of phase wrt frequency so I'd expect the vast majority of VNAs to be able to display group delay.

For this reason my old Advantest spectrum analyser can also measure group delay using its internal tracking generator as one of its party tricks. I think this is a standard feature rather than an option. The impedance measuring option needs the external bridge.

I haven't come across one of these 'vector' spectrum analysers before. They sound interesting. I guess if you had one of these you may not need a VNA.

Advantest/Takeda Riken caused quite a stir in 1981 when the 4172 was released. HP brought out the HP4396A 'combined' Spectrum/VNA about 10-12 years later.

But I wouldn't get too excited about the TR4172 in VNA mode. It 'can' measure impedance and display a smith chart and also measure phase and group delay etc but the performance isn't very good by today's standards and I wouldn't recommend purchase of one if you wanted it for serious VNA work. The TR4172 is also huge and very heavy!

It has a very long and fiddly (and very strange!)  calibration procedure. The HP4396A is going to be much better as a VNA but they still fetch a high price today despite boing over 20yrs old. For the same cost I would much rather have an early 8753 VNA and a HP8568B analyser if space permits.

 

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #11 on: June 02, 2015, 08:20:09 pm »
Note: You can also measure and model components right up into the GHz region with a decent VNA. Eg measure the behaviour of SMD resistors, caps, inductors, transistors etc etc up to many GHz and export the data to a PC as a small signal model that you can use in an RF simulator.

These s1p or s2p or sNp models of the real component will improve the accuracy of the simulation compared to crude lumped models based on guesswork.

« Last Edit: June 02, 2015, 08:21:50 pm by G0HZU »
 

Offline KJDS

  • Super Contributor
  • ***
  • Posts: 2442
  • Country: gb
    • my website holding page
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #12 on: June 02, 2015, 09:12:15 pm »
Note: You can also measure and model components right up into the GHz region with a decent VNA. Eg measure the behaviour of SMD resistors, caps, inductors, transistors etc etc up to many GHz and export the data to a PC as a small signal model that you can use in an RF simulator.

These s1p or s2p or sNp models of the real component will improve the accuracy of the simulation compared to crude lumped models based on guesswork.

Whilst that is true, it needs a great deal of care to ensure that the measurements are accurate. It's very easy to have a minor issue with the cal for SMD measurements and you're worse off than educated guesswork

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #13 on: June 02, 2015, 09:34:36 pm »
Note: You can also measure and model components right up into the GHz region with a decent VNA. Eg measure the behaviour of SMD resistors, caps, inductors, transistors etc etc up to many GHz and export the data to a PC as a small signal model that you can use in an RF simulator.

These s1p or s2p or sNp models of the real component will improve the accuracy of the simulation compared to crude lumped models based on guesswork.

Whilst that is true, it needs a great deal of care to ensure that the measurements are accurate. It's very easy to have a minor issue with the cal for SMD measurements and you're worse off than educated guesswork

Agreed. I use dedicated test fixtures and a custom user calibration that lets me measure something like an 0805 or 0603 component quite accurately. eg a typical (extracted) model will give the expected package capacitance and inductance even for a tiny component like this.

In case anyone thinks this is a straightforward task, it does take a lot of experience to be able to set up a system that can do this accurately. It takes more than just a decent VNA and a $$$ calibration kit. One of my passions is accurate modelling and simulation and I've been doing stuff like this with a VNA for about 25 years :)

You can also download free S1p and S2p models from some manufacturers who measure and model the components for you or you can pay for SMD component models from 3rd party companies like Modelithics. They independently model the components from various manufacturers and offer them for sale.
« Last Edit: June 02, 2015, 09:38:51 pm by G0HZU »
 

Offline rastro

  • Frequent Contributor
  • **
  • Posts: 380
  • Country: 00
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #14 on: June 02, 2015, 10:45:19 pm »
There's several places on eBay that sells PCB's and/or kits that allow you to choose various MMIC's to roll your own RF amplifier.  However if you are building your own you would want some way to measure/verify power and spectrum on the input/output.  Ideally a spectrum analyzer - at minimum a power meter or RF detector. 

This is one example from the RF Basic Store - Designer Kit for RF MMIC Amplifier with SOT-89 Package:
http://www.ebay.com/itm/Designer-Kit-for-RF-MMIC-Amplifier-with-SOT-89-Package/400822346962?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D1%26asc%3D20131003132420%26meid%3Df915ddc99aff440a8306bce5b2e62327%26pid%3D100005%26rk%3D2%26rkt%3D6%26sd%3D400822343856&rt=nc

 

Offline chscholz

  • Regular Contributor
  • *
  • Posts: 85
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #15 on: June 05, 2015, 04:13:43 pm »
Hi Chris,

conceptually a VNA is really nothing else but a generator, directional element(s) (directional couple, bridge, etc.) and spectrum analyzer, a switch if you need to do forward and backward measurements, plus some software tying the whole thing together.
Error correction? Conceptually you can write that all yourself (your choice of 7-term error model, 12 term error model, or anything in between). No magic there, just a bunch of linear algebra and some smart ways of arranging unknowns or assuming certain properties of your DUT (such as reciprocity for certain calibration techniques).

But in praxis, it is super difficult to build a VNA that has a large frequency range, high dynamic range, low measurement uncertainty, low noise floor and fast measurement speed, all at the same time. Also, modern VNAs have two receivers per port (one reference receiver, one measurement receiver). To replicate that you would need two SA's per port. Many modern VNAs have multiple internal sources, too for nonlinear measurements (mixer measurements, intermodulation measurements, etc.) High-end VNAs also have phase coherent sources (e.g. for true differential measurements). In short, things tend to be come rather complicated pretty quickly.

But again, conceptually, and with a lot of work, you can do everything a VNA can do with signal generator(s)/spectrum analyzer(s)/directional elements(s) and maybe a few switches.

Just last month I put together a little marketing piece that talks a bit about VNA basics.
http://www.techonline.com/electrical-engineers/education-training/courses/4439323/VNA-Basics

Chris


What can a proper VNA do that a spectrum analyser, tracking generator and directional coupler will not be able to do, assuming both have the same frequency ranges? Thanks.
 

Offline rfeecs

  • Frequent Contributor
  • **
  • Posts: 693
  • Country: us
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #16 on: June 05, 2015, 09:02:05 pm »


But again, conceptually, and with a lot of work, you can do everything a VNA can do with signal generator(s)/spectrum analyzer(s)/directional elements(s) and maybe a few switches.



As has been noted, a conventional spectrum analyzer does not measure phase.  That is why it is called a Vector Network Analyzer.

Also already noted, but I will repeat it:  without measuring phase, you cannot do full error correction, and you cannot measure impedance.
 

Offline f5r5e5d

  • Frequent Contributor
  • **
  • Posts: 349
Re: What will a "proper" VNA do that an SA and tracking gen won't?
« Reply #17 on: June 05, 2015, 09:24:45 pm »
Quote
Just last month I put together a little marketing piece that talks a bit about VNA basics.
http://www.techonline.com/electrical-engineers/education-training/courses/4439323/VNA-Basics

Chris

any link that doesn't require a capsule bio of information to register?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf