VNA-nuts ? Show your VNA

[Po6ept]

I do not get notifications for my own topic ??

I don't believe that starting a topic changes notification settings.  Under your profile settings, you can select whether or not to enable notifications.  Go to Profile, Modify Profile, Notifications

[The Electrician]

The cal-kit is all important, and there's no reason not to buy one of these--good up to 6 GHz!

http://www.ebay.com/itm/RF-Network-Analyzer-Calibration-Kit-2-Kits-included-Plug-and-Jack-/370981777421?hash=item56603cb00d

[branadic]

I've one of those VNWA 2.6 at home and developed those 36MHz TCXO upgrade kits for the 2.x series a few while ago, as a direct replacement for the internal crystal.
By the end of the last year we ordered the latest model for work.

[G0HZU]

I suppose I should also mention I also have an Advantest TR4172 1800MHz spectrum analyser + TGen  that has an impedance measuring option fitted.

You are supposed to buy an external RLB from Wiltron with this option but I don't have the RLB. But I do have several home made bridges that seem to work well.

However I rarely use this feature on the analyser as the VNA calibration procedure is quite lengthy and very strange! Also, the HP8714B is so much nicer to use as a VNA.

But the TR4172 is probably the 'strangest' VNA on here so far. I'll post up some pics the next time I have the RLB handy and I can spare the time to set up the calibration on it

OMG I just remembered... I've actually got TWO of these TR4172s now. I recently bought a non runner for spares (now fixed) for £250 a few months ago. So I am in the VNA qty3 club as well 

Note: If you google TR4172 and look at the images you can see one in VNA mode with the smith chart display turned on.
It's one of the ugliest analysers on the planet but it was one of the very best (and one of the most expensive) spectrum analysers in the world in the 1980s.

[chscholz]

I'm a bit VNA addicted.

Me too. I work for R&S America and specialize in VNAs; I am a marketing guy.

[marshallh]

Can anyone recommend a book/website detailing VNA usages and what the numbers mean?

[PA4TIM]

http://www.microwaves101.com/encyclopedia/couplers.cfm Website with lots of info

The agilent impedance measuremet handbook : https://www.home.agilent.com/agilent/editorial.jspx?cc=GB&lc=eng&ckey=2162694&nid=-34051.536879632&id=2162694

One off the two vna "bibles"
http://www.books.rohde-schwarz.com/ the second book on this page from M. Hiebel

And the one from J. Dunsmore, the vna desinger from Agilent
http://books.google.nl/books/about/Handbook_of_Microwave_Component_Measurem.html?id=GEnBhrvPFQwC&redir_esc=y

@chscholtz:

I am a marketing guy

OK, because it is VNA marketing we forgive you 
That must be a great job. To be surrounded by those wonderfull instruments . I have never used a modern professional VNA, so I know those only from books, but i like the R&S VNA's very much. (but they are waaaaay out of my hobby budget)
When I bought my Hameg the guy who demonstrated the scope was from the service department and we had a nice chat. It turned he knew my website because he used to work at Fluke working on the calibrators I have. And now he did support on the R&S VNA's and offered me to analyse my calkit. To bad I had a computer crash and lost his email adres (and can not remember his name)

[PA4TIM]

Tonight a friend came by and he gave me 3 sealed bags from Huber-Suhner. Inside 3 sucoflex 104E cables with N connectors. They have serial numbers and a test certificate. They are very beautifull. I did some tests and they perform very well. Even better as my amphenol sma cables (the vnwa has sma so I had to use SMA to BNC and BNC to N to mount them and because I have no N cal kit I used N to GR-874 and calibrated the set with a GR calkit and results were above expectation. .

[KJDS]

I'm still looking for a proper set of Gore cables for mine. They also make the Agilent ones. So beautiful to use. Very flexible but with no spring to them and also no phase change with flex. They'll also last forever.

[marshallh]

Thanks PA4TIM for those links.

[TerraHertz]

Yes, thanks PA4TIM. And everyone for this thread. Lots of interesting reading.

Quite a while ago I too decided I wanted to learn about VNAs, or rather high frequency RF design in general.

I've been gradually building up a pile of RF test gear, starting with the really early basic stuff like slotted lines, through phase/gain meters, and early VNAs. All very old stuff, on the principle that the service manuals are available, the operating principles are discussed in practical terms in the manuals and app notes, it's all relatively cheap (ie I can afford it), and who cares if the accuracy is not so great any more.
I'm not afraid of fixing stuff where necessary. Also, I just like old test gear. Love the way it's built, and the awesome manuals.

Unfortunately the 'too many time demands' syndrome has defeated me so far, and all the gear has just been sitting idle while other life issues get sorted out. Or not.

Anyway, here's some of it. Top down:
 HP 8405A   vector voltmeter,
 HP 8418A   aux display holder, with 8414A polar display (also have the 8413A  Phase-gain meter)
 HP 8410B   Mainframe, with 8412B   Scope (rectilinear)
 HP 8411A   Harmonic freq converter
 HP 436A    Power meter (also have some assorted sense heads)
 HP 8750A  Storage normalizer (not sure I have the right interface card and cables for this and the 8410B)
 HP 8743A  Trans/Refl TS  (2-12.4GHz)
 (and a bunch of other stuff elsewhere)

I didn't know there were small USB-based VNAs these days. Interesting.

[marshallh]

I've been gradually building up a pile of RF test gear, starting with the really early basic stuff like slotted lines, through phase/gain meters, and early VNAs.

Mind if I move in? 

[TerraHertz]

I've been gradually building up a pile of RF test gear, starting with the really early basic stuff like slotted lines, through phase/gain meters, and early VNAs.

Mind if I move in? 

Heh, usually it's Chinese and Russian ladies wanting to marry an Aussie for citizenship.
Even if you were cute enough, there are other... difficulties.

Hey, does anyone have a list of the best HP App Notes on this topic?
I have a few, like 117-1 Microwave Network Analyzer Applications, and the HP soft-cover book Microwave Theory and Measurements.
If only I had time to read them! (Motivation is also a factor, sigh.)

After investigation I'm kicking myself, in that the HP 8410B apparently isn't easy to convert to computer controlled operation, while the 8410C is.

[PA4TIM]

That is a nice stack of HP gear. (I have all my calibrators in a 19" rack)
I have the 8407 with 8412 and the analog gain/phase meter. But not the 8414 polar display.
I stil use the 8407 but not much, because that little German wonderbox can do the same and much more. One cool thing I did with the HP is using it as a RF-IV analyser.  I made a S21 setup but used a Tek CT1 inline current probe for the reference channel and a T to measure the voltage over the DUT with the test channel. That worked very well but involved some math to translate it to impedance .


HP has some nice appnotes made in the 60-70's when they made the first vna's. I learned my first vna steps from those. The reason I made my serie of tutorial is because it is very hard to find information for starting vna-nuts. The books I mentioned are great but a bit overwelming for starters.

I found the smith chard picture for the advantest vna. http://www.technologymuseum.eu/phpwebgallery/picture.php?/196/category/60&pwg_id=ab6a46a1fbd069bc1c507aa3f4de2311
I do not think it is an ugly vna, I like the looks of crt's with readout.

[chscholz]

OK, because it is VNA marketing we forgive you 
That must be a great job. To be surrounded by those wonderfull instruments . I have never used a modern professional VNA, so I know those only from books, but i like the R&S VNA's very much. (but they are waaaaay out of my hobby budget)

Best job I ever had (not necessarily the easiest, I guess that is sort of related). It's amazing what difference it makes to work for a privately held company.

[ivaylo]

It's amazing what difference it makes to work for a privately held company.

Is it worse or better? Also are you comparing with publicly traded companies or with government organizations?

[larry42]

Its a pity the SDR-kits website is such an uncomprehensible pile of words. I tried to look for the kit, a simple explaination on how to use it or specs but they don't seem to exist. If it can be tricked into going over 3GHz at reasonable sweep speeds I might consider buying one. I just have to get my head around to read those Smith charts.

The website is indeed pretty confused and messy ("Documents" links to some random data sheets ,but not to the VNA) then the "webshop", sometimes referred to the the "e-shop" on the site will bring you to buy the thing.

But the Help file (which is just a PDF version of the Windows Help) is filled with references to the Version 3, the 3E, expansion boards etc. Page two then shows a board with a serial connector and a USB interface board. Talks about LPT interfaces (parallel).

I've just gone back to the webshite now - Apart from a couple of sets with cal kits, you can buy the VNWA3 or the 3E, which is referred to as a 3+ with expansion board. Finally on page 11 of the help file:
"VNWA3E was formerly named VNWA3+ but both are the same product. The DG8SAQ VNWA3E is the VNWA3 fitted with the VNWA3 Expansion Board which provides additional functionality as follows:..."

I wish they just had a line on the webshop description that said: "1kHz-500MHz (1300MHz with reduced performance) 2-port VNA with integrated Transmission/Reflection test set"


Anyway after I figured out which one to buy, I received it and promptly had lots of issues getting the thing to work at all. Admittedly with a Win7 VM.  The problem is that the Setup routine was very un-polished, with a confusing (IMO) method of calling the items e.g. there is a Window area called "Auxiliary Audio Settings"
with text that says "Aux. Audio measure THRU", but then a pulldown for "Aux. reference = right channel")
The software has some quirky UI, that is definitely not according to any UI guidelines - something like a utility you might get from a motherboard maker.

I was rather chagrined when I initially received the item after the poor purchasing experience due to the confusing website. Which was then promptly followed by the audio setup that had to be done each time I plug it in, a VMware issue, but still the setup was just strange - requiring me to follow the precise procedures in the manual each time, because the names and placement of the controls (choice between left/right, main/aux and thru/reflect). Afterwards I was rather vocal on the Yahoo site, complaining about the interface* and that for 500 quid it should be a bit more polished, I ended up giving some constructive criticism and Tom Baier (DG8SAQ) incorporated my suggestion of an auto-setup for the mapping of thru, reflect, left, right, main and aux. Therefore for most people they just have to disconnect cables and press "auto setup".

*I have used VNAs, programmed them and written cal routines for them since 1998, so I consider myself fairly au-fait with the concepts. Moreoever, I had lent the VNWA to a university professor, who is also a practical engineer running a successful radar company, and he had problems following the setup as well.

So the conclusion is that yes, it's pretty good for what it does. I don't consider it a wonder box. I really wish that the interface was open source, because there are a number of dialog boxes I want to change, the calibration is not what I'm used to (HP VNA background), but probably no worse than the R&S analyzers I have fought with. The thing is not very good as a spectrum analyzer, it's very rudimentary for that. Tom Baier is responsive and helpful for those that are experiencing problems.

It's certainly lighter than any other VNA!

I'm in the process of making an S-parameter set for it, and maybe swapping out the mixers, to give better performance at the higher frequencies.

I've not been able to do any real comparisons to Agilent VNAs. Much of the accuracy comes down to the cal kit, but as most people don't have 3 grand to spend on a real cal kit, I wouldn't worry too much about it. Above 500MHz the lack of DR will play a role. I assume there will be a few spurs as well due to the use of the DDS aliases.

[G0HZU]

I also had a look at the VNWA website a few days ago and wasn't too impressed...

The pdf help file has some good and bad points. But I find the user interface to be poorly designed and visually unappealing. It's easy for me to sit here in a chair and criticise but I can't understand why the interface is so poor when you have more display and control versatility/area on  PC screen compared to a typical HP/Agilent VNA.

I really wish that the interface was open source, because there are a number of dialog boxes I want to change

Maybe it could be rewritten to allow a degree of external control. eg the ability to use a custom 'skin' and also allow the user to change labels and menu names to something more intuitive.

[KJDS]

...., but probably no worse than the R&S analyzers I have fought with. ....

Does anyone like the R&S user interface? I quite like the ZVF analyzers capabilities but using it drives me round the twist.

[G0HZU]

LOL!
At work we have various R&S VNAs from the very strange 20GHz ZVM to the cute little two and 4 port jobbies. eg the ZVB range.

I like the RF performance but hate the user interface on all of them. Plus the ZVM isn't easy to use via GPIB when importing data into Genesys. It doesn't export the corrected data and often goes a bit loopy afterwards.

The little ZVB VNAs are very portable and pretty but I feel like I've captured a VNA from another planet when I try to use one as the user interface seems to have been deliberately designed to be 'different' to Agilent.

[G0HZU]

Much of the accuracy comes down to the cal kit, but as most people don't have 3 grand to spend on a real cal kit, I wouldn't worry too much about it.

What do people use for a VNA cal kit here? At work I have access to the usual HP/Agilent cal and ecal kits which is useful.
However, for casual research work at home I often use a homemade SMA based system that works very well with my HP8714B. It was extremely cheap to make and I don't have to worry about losing parts of it or it wearing out as it is so cheap to make a replacement.

I have access to the Agilent kits at work if I need to make critical measurements at home but I get pretty good results with the homemade kit.

There does seem to be a recent influx of cheap(ish) VNA cal kits appearing on ebay etc and I wonder if anyone has tried them?

[KJDS]

Until recently I was using a home made short and open and a Narda load.

I've now picked up a couple of HP cal kits, one I'll keep and the other is for sale.

[PA4TIM]

For sma I have a Rosenberger and amphenol calkit, for bnc I have 3 homemade kits. I also have a HP and narda load but I get better results (in a t-check) with my homemade ones. For N I have only very nice cables and a load at this moment. But my most used kit is a GR-874 calkit. I made the open and short but found a brand new 50 Ohm load. I use that because I have a whole bunch of GR stuff.

For those not familiar with that, here are some examples. I also have trompones, an  adjustable stub, cables etc and a collection of bridges and generators.


About the vnwa, it is not a >15K R&S or agilent that need calkits more expensive as the vnwa and cables you do not dare to touch without gloves ;-)
But I call it a wonderbox because fot that money I can not get a better one and it performs very well (in my opinion, I have measured the strangest things with it from fF and mOhm upto things like impedance of mud)
It can be used as counter, can be used as signal generator, the TDR is neet, it has a LCR meter so for that money I will not complain. (I do not needs all those functions because I have like 5 counters, 10 LCR meters, a bunch of generators, timestandards, etc but if you have not, I think its handy)
Indeed Jan is not a webside builder, he is a HAM and a retired engineer. Tom is a professor on a university in Germany. SDR-kits is  more or less a hobby and they had not expected the VNWA to become so populair. But they mean well, ar far from commercial  and are very helpfull. If you find a bug or have a good idea Tom solves it very fast or adds it and I think some professional companys can learn from that. So for me it's great but if I was used to digital R&S or Agilent I maybe had an other opinion. Concidering the price (the cheapest R&S or Agilent is far outside my range )  I am afraid I will never have that exerprience. I envy you guys who are allowed to play for a job with those great toys for boys ;-)

So now I know who to thank for the auto setup. That is a very good feature. I never had problems on my XP computerr but then came that bloody W7 and it drove me nuts. If all worked and I used it the next day W7 replaced drivers and this other stupid things but because I use it for years it takes me less as a minute to get it working again but the autoset option is very nice because W8 (i use now)  is even worse in this, (for me as Linux user)

I have no experience with other user interfaces so I do not know better and I use it for something like 4 years now. For me it is good. It can be easy customized with other background, colors, size of markers, grids, custom traces . The trace can be saved to files and used in things like qucs or other software (I have no experience with that but I know people who use that a lot. The help file has lots of info but I agree it can be confusing because it is grown a bit out of it's shoes. It is becoming a bit better over time. The first versions where just dumps from the helpfile and there was no index, pagenumbers and subjects were spread all over the file. That is the downside from amateurs making stuff, but I am sure if they were commercial we had to pay for things like updates , would have to live with bugs and the thing would cost over 1000 euro and still sell because the first better alternatrive would still be unobtainium, it probably had a more impressive looking cabinet, nice manual etc. But thanks to this I got VNA's as a nice hobby and learned a lot and some functions in it are made on my request, I do not think R&S or Aghilent would do that.

So that's why I like it so much and I tend to be a bit over enthousistic, but, hey, I'm just a hobby network analist (and I know, we are a rare unusual breed)
 

[larry42]

Yep I will agree that once one acutally understand the website and the product, both Jan and Tom are nice and helpful individuals.

I feel that the VNWA could actually be a much better product with a few improvements. Improved website, the help file edited by someone skilled in writing user manuals (right now too many words and German-> English grammar construction, which is completely understandable, just a slight nuisance when trying to understand the thing quickly), and some of the diaog boxes and cal stuff could be improved.

Further afield - electrically I would like it if a few changes where made - I wonder if the newer 1 and 2.5GHz DDS's could be integrated (along with a better mixer) - but I fear that would add at least 100 USD to the BOM.

About to make my own cal kit soon. And have to remind Poul on the importance of cables in the cal process - he seems to neglect this in all his otherwise useful documents.

[G0HZU]

About to make my own cal kit soon.

Here's a quick comparison I did today where I compared the same RF choke (on a test fixture) on my HP8714B with one of the Hp8753ES analysers at work. The HP8753ES is cal'd with a HP 85033E 3.5mm calkit  and my HP8714B is calibrated with my homemade calkit.

The test was only carried out up to 3GHz because that's the limit of my HP8714B but you can see how well my calkit agrees.

To get this level of accuracy I had to Dremel the homemade cal kit to fine tune it and also I run it with a 'user' cal file. But I use this cal kit all the time and it cost me nothing to make because I made it from salvaged 18GHz SMA parts

To build it from new parts would cost about £20.

Obviously, this isn't exactly a conclusive test but you only need to add a 0.2ps delay error in one set of results to see the results diverge.

The plot below shows the real and imaginary components of the complex impedance of the SMD choke. The results agree so well that the 2 pairs of traces are sat on top of each other

The second plot is of a 1K 0805 SMD resistor in a test fixture measured on both analysers.