Author Topic: What VNA to get? (Read 14297 times)

LukeW · « **on:** May 11, 2018, 10:40:33 am »

Looking at purchasing a VNA for work.

What models should I look at? Any recommendations?

Let's say $25k-30k AUD.

Going up to 5.8 GHz would be nice, but I don't think it needs to go higher than that.

Application contexts would include PCB-level RF design, and design of antenna-matching LC networks between the IC and the antenna, at frequencies that would typically be ~2.4GHz as well as ~400MHz and ~900MHz. Antenna design, antenna measurements and calculation of VSWR, and validation of PCB microstrip/CPWG widths and characteristic impedance would be relevant uses.

Spectrum analysis, and analysis of transmitted RF power, and basic preliminary in-house checks for RF harmonics and regulatory compliance would be a useful use-case, too.

Mechatrommer · « **Reply #1 on:** May 11, 2018, 10:54:14 am »

Rohde & Schwarz ZNLE6 2 Port 6GHz Vector Network Analyzer $17,720.00

dmills · « **Reply #2 on:** May 11, 2018, 11:08:03 am »

If your working band only goes to ~2.5Ghz, you will find that you don't need to spend anything like that to get perfectly good results.

I know the 8753D or even a 8753C (Which I own) is old but they still produce perfectly good results about as well as anything more modern, and I would expect to buy four with change for the cal kit(s) for that money. The real expense here is in the calibration kits and the test port extensions.

If you wanted something a little classier, then a PNA should be doable, has the advantage of direct support for TRL calibration which makes the cal kit pain a bit less tedious.

If you have money to burn then go for the PNA-X or such, drooool.

Most VNAs do not make very good spectrum analysers, I would probably make that a separate box, apart from anything else, you want to see harmonics, which means that if your working band is 2.5GHz you want to see out to 20GHz or so, a 20GHz SA is **MUCH** cheaper then a 20GHz VNA.

Have a look at folk like Copper Mountain, particularly if you are pushed for lab space, and R&S are always a good option but IMHO there is a reason the HP stuff is so popular.

Regards, Dan.

LukeW · « **Reply #3 on:** May 11, 2018, 11:26:12 am »

Hmmm, if I have left over budget in my proposal maybe suggestions for a spectrum analyser would be nice too

hagster · « **Reply #4 on:** May 11, 2018, 12:53:41 pm »

My recommendation is to go for a Copper Mountain Compact VNA like the s5085 or cheaper s5065. They are really fast and accurate instruments and you don't end up paying a fortune for all the additional 'options'. If you have money left over you could get a E-Cal kit too.

For designing matching networks give AnTune a look. It automatically calculates optimal matching networks using real component libraries in real time(takes a live S11 feed from the VNA) http://www.antune.net/index.html

As others have said you are best to go for a separate Spectrum Analyser. I have played with a Signal Hound one recently and it seemed pretty good.

If you really need an all-in-one VNA/SA the Keysight Fieldfox might be worth a look.

LukeW · « **Reply #5 on:** May 11, 2018, 01:35:03 pm »

The Copper Mountain Technologies S5085 looks nice, but I'm generally a bit skeptical about all PC-based instruments.

I've never used a PC-based VNA before, but all PC based scopes and stuff generally seem a bit crap, and just not as good as the "real thing".

I'd like to do the job properly, with a real good tool, and if I need to spend a few extra bucks that's OK.

Would also like recommendations for SAs (like ~$3000 I suppose?) if anybody has a favorite they want to suggest.

dmills · « **Reply #6 on:** May 11, 2018, 03:32:05 pm »

I second the slight twitchyness about PC based instruments, USB based things doubly so.
It can work, but I am not sure I like random USB cables with what may or may not be adequate isolation in my RF lab.

Copper mountain does however have a decent rep, and if you are happy with that operating style and the box does what you want, then why not?

As to spectrum analysers, the usual suspects really, Keysight, R&S, Anritsu all got decent gear with slightly different operating styles, all of it is fine.

One figure of merit sometimes ignored (especially by the PC based instrument vendors) is sweep rate, and in the case of FFT IF based designs the maximum bandwidth for which POI is 1, especially when working with digital modes, being able to set peak hold and capture the third harmonic of an entire channel with high POI is a very useful party trick.

Having a quasi peak detector and a 9KHz RBW option is useful if you ever want to do precompliance testing.

Regards, Dan.

hagster · « **Reply #7 on:** May 11, 2018, 04:08:34 pm »

Quote from: LukeW on May 11, 2018, 01:35:03 pm

The Copper Mountain Technologies S5085 looks nice, but I'm generally a bit skeptical about all PC-based instruments.

I've never used a PC-based VNA before, but all PC based scopes and stuff generally seem a bit crap, and just not as good as the "real thing".

I'd like to do the job properly, with a real good tool, and if I need to spend a few extra bucks that's OK.

Would also like recommendations for SAs (like ~$3000 I suppose?) if anybody has a favorite they want to suggest.

I have used some crappy PC based scopes too. CMT is in a different league.

You can download their software for free. It has a demo mode which I can vouch for being near identical to the real thing.

Most of these companies will lend you a demo unit to evaluate if you ask nicely btw.

w2aew · « **Reply #8 on:** May 11, 2018, 07:48:18 pm »

The 6GHz Tektronix TTR506A is about $12k - it also runs a PC-based user interface similar to the Copper Mountain units.
https://www.tek.com/vna/ttr500

Software available for free to try as well.

There is a LE option (slightly lower performance) 6GHz unit that lists for just under $6K US. See the bottom of the above webpage.

KE5FX · « **Reply #9 on:** May 11, 2018, 08:27:16 pm »

Quote from: LukeW on May 11, 2018, 01:35:03 pm

The Copper Mountain Technologies S5085 looks nice, but I'm generally a bit skeptical about all PC-based instruments.

If you buy a new benchtop VNA today, you're going to get a PC-based instrument whether you want one or not. It would be crazy to build one any other way.

So it's just a question of whether it uses your PC or theirs.... and how much you pay for it.

hagster · « **Reply #10 on:** May 11, 2018, 09:31:51 pm »

We had a 40GHz HP VNA that ran WinNT on a really old embedded PC. While you can connect it to the network. It was basically a huge honey trap of unpatched vulns. We could have upgraded it to WinXP, but the involved replacing the motherboard and the front panel.

dmills · « **Reply #11 on:** May 11, 2018, 11:23:10 pm »

True that everything is a PC these days, been that way for a while, but you live and die on the user interface, and I rather like having buttons and encoders even if the underlying compute platform is a PC.

Regards, Dan.

KE5FX · « **Reply #12 on:** May 12, 2018, 12:22:45 am »

Interestingly, musicians have already solved this problem by using MIDI to separate their control surfaces from their instruments. Someday we may see gadgets like this:

... for PC-based network analyzers, spectrum analyzers, oscilloscopes, anything bigger than a DVM.

G0HZU · « **Reply #13 on:** May 12, 2018, 12:50:22 am »

I guess the used choice in this price range would be something from the Agilent E5071x range. Typically covering 9/300kHz to 8.5GHz. These big old VNAs have a fairly good user interface (apart from the cal kit setup) and are set up to be easy to use remotely, especially in an ATE or manual test environment. The RF performance is generally very good and the measurement speed is quite good. But they are getting quite old now and still command high used prices although the prices vary a lot from seller to seller. We have several of them at work and they have been quite reliable over many years. I've only seen one fail and go back for repair and I'm not aware of any others that have failed like this. Note: the display on the B version is only VGA but it is still useable. I think the C version offers 1024x768. It is possible to get higher res modes on the B version with an external monitor but this is fiddly to set up.

I would naturally shy away from any RF test gear made by Tektronix for various reasons. But if you are buying a new piece of gear from them I'd check out the long term cost of maintaining it under a service contract. Check out what happens if/when you leave the $$$ contract. How much will it cost you for Tek to repair your faulty VNA in 5 years' time? I also watched Tek's own youtube videos of the TTR500 series VNA and I'm not impressed.

The Copper Mountain VNAs look to be a class above but I've not played with one so I can't be certain. I was offered a demo of any CMT model I chose via a UK rep (think it was Melcom) but this will have to wait for a less busy time at work.

The little Pico VNA is another unknown but seems remarkably cheap for a 2 port VNA. It is very basic with a very dodgy looking GUI but it might be good enough for your needs although the limited range for the RF source power would be a deal breaker for me.

LukeW · « **Reply #14 on:** May 12, 2018, 10:06:22 am »

A self-contained system which does not need to connect to a separate PC would be better, because (among other reasons) I use a Mac and Linux machines, and usually avoid Windows unless it really can't be helped or can't be run inside VirtualBox.

tszaboo · « **Reply #15 on:** May 12, 2018, 10:31:47 am »

Quote from: hagster on May 11, 2018, 09:31:51 pm

We had a 40GHz HP VNA that ran WinNT on a really old embedded PC. While you can connect it to the network. It was basically a huge honey trap of unpatched vulns. We could have upgraded it to WinXP, but the involved replacing the motherboard and the front panel.

So, connect it to a firewall? For about 100 EUR there are mini PCs that have 2 or more Ethernet ports, and any semi-competent network administrator can setup something like that.

Mechatrommer · « **Reply #16 on:** May 12, 2018, 10:53:08 am »

or not connect it at all. i dont get a reason why we have to when we can sit there and do the job...

dmills · « **Reply #17 on:** May 12, 2018, 11:49:04 am »

Some form of connection to a PC is extremely useful, be it Ethernet, GPIB, USB or RS232 so that you can use tools like softplot and for automated measurements.

But yea, working assumption is that the OS is full of holes so treat appropriately.

Regards, Dan.

hagster · « **Reply #18 on:** May 12, 2018, 12:07:57 pm »

The network connection is useful when you need some kind of automation or remote control. I often have a need to control a signal generator form a different room for instance.

Also if you need to do documentation its really handy to have screnshots of data files ready to embed in a document. Yes USB and floppy disks work, but its clunky.

Portability is a big thing for me these days too. Being able to move the instrument to the DUT rather than visa versa is awesome.

Omicron · « **Reply #19 on:** May 12, 2018, 01:51:32 pm »

Quote from: LukeW on May 12, 2018, 10:06:22 am

A self-contained system which does not need to connect to a separate PC would be better, because (among other reasons) I use a Mac and Linux machines, and usually avoid Windows unless it really can't be helped or can't be run inside VirtualBox.

I'd cast another vote for the R&S ZNLE6. I've played around with one and the UI is way better than the TTR506A which I found confusing. You can use Windows Remote Desktop to control it from a Mac using the official RDP client for OSX from Microsoft, I've done so and the UI adapts itself to that mode and is not limited to the resolution of the built in lcd so you can run it full screen on whatever monitor you have hooked up to your Mac. It also automatically adds the buttons that are normally on the front panel of the device to the GUI when running via RDP. I was really impressed on how polished it all is. Definitely ask R&S for a demo before you decide.

G0HZU · « **Reply #20 on:** May 12, 2018, 02:41:31 pm »

I looked over the specs for the ZNLE6 and the source power can only be varied over about -10 to +2dBm typical. This alone would be a dealbreaker for me. I'd want to be able to vary the source power well below -30dBm. The other specs don't look that great but it is a low cost entry level VNA so there will be compromises in terms of price vs performance.

I'd suggest it is best suited for general purpose passive work where accuracy isn't critical. Where I work, few people would want to use this instrument unless it was for casual research work and all the other VNAs were in use already. So we would be unlikely to purchase one if we had one on trial.

hagster · « **Reply #21 on:** May 12, 2018, 03:38:35 pm »

So what are some of the features that people value in a VNA?

I find TDR and Time Gating to be handy.

PnpNpn · « **Reply #22 on:** May 12, 2018, 09:26:15 pm »

+5 für den deutschen Tiger, No need for maintenance

G0HZU · « **Reply #23 on:** May 14, 2018, 10:53:14 pm »

Quote

Application contexts would include PCB-level RF design

I don't think the ZNLE-6 VNA qualifies for this if you want to do any basic amplifier design or verification. It appears to be crippled by its limitations wrt the source power and the signal handling of the receiver port.

If you set the source as low as it can go (-10dBm) and them put a 15dB gain LNA inline for test it will overload the VNA receiver as the receiver max level appears to be 0dBm. So it is not really suitable for full s2p tests on a basic LNA that other VNAs can do easily. It really does look to be optimised (crippled?) for passive testing. If you want a decent/versatile lab VNA for active and passive work I'd run a mile from the ZNLE-6 based on the datasheet specs alone.

It's a shame that gear like this doesn't get sent out for independent (critical) testing because otherwise you are at the mercy of the marketing dept of each manufacturer unless you get one on trial for your own evaluation.

radar_macgyver · « **Reply #24 on:** May 14, 2018, 11:45:56 pm »

It's not a bench instrument, but consider the Keysight Fieldfox. It can be upgraded from VNA to VNA+SA using a license key. The portability comes in handy when doing antenna tests and such. The one thing I don't like about it is the relatively low-res display.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: What VNA to get? (Read 14297 times)

Share me