Low cost PC-based 3 GHz two port VNA with USB interface

I'm looking for a low cost PC-based 3 to 5GHz 2 port VNA (vector network analyzer), preferably with a USB interface and good user interface software for the PC. Any suggestions will be gratefully received.

**broken link removed**

I have so lusted after one of those Copper Mountain VNAs!
Although they're a good deal compared to a *real* VNA for professional use, they are by no means "low-cost" in the amateur/hobbyist sense of the word.

Their Australian distributor recently quoted me (A$ ~= US$):
~$28,000 for the 804/1 model,
~$16,000 for the 304/1 model, and
~$4,000 for the 1300/1 model.

(+ ~$700 for the calibration kit). Ouch.

If you can get by with scalar measurements, the Signal Hound + tracking generator (**broken link removed**) is worth a look for (a far more reasonable) $1500 total cost.

Thanks BigBoss and thylacine. Much more affordable than Anritsu etc. Unfortunately I have to have a VNA, so scalar is out. There is also this (with RS232 port):

**broken link removed**

I believe these sell for around GBP 6000 which is a high price for something that doesn't seem nearly as good as the Planar. Then there is this:

http://miniradiosolutions.com/extender

which, together with their miniVNA-pro gives 1.5GHz (so it's no use to me). Very affordable (and hobbyist) - and I'd be happy with it if they would make a 3GHz version! So it looks like the Planar for me.

thylacine1975 said:

I have so lusted after one of those Copper Mountain VNAs!
Although they're a good deal compared to a *real* VNA for professional use, they are by no means "low-cost" in the amateur/hobbyist sense of the word.

Their Australian distributor recently quoted me (A$ ~= US$):
~$28,000 for the 804/1 model,
~$16,000 for the 304/1 model, and
~$4,000 for the 1300/1 model.

(+ ~$700 for the calibration kit). Ouch.

If you can get by with scalar measurements, the Signal Hound + tracking generator (**broken link removed**) is worth a look for (a far more reasonable) $1500 total cost.

Click to expand...

Whoa..
They are crazy if the prices are around 28K$..
I can buy a brand new Agilent VNA-maybe higher performance refurbished one.

It is a true shame that some small company has not come up with affordable spectrum analyzers and network analyzers by now. There should be a two port DC-10 Ghz vector network analyzer out there for maybe $4000 USD. There is not much inside of these things--it could be done on one PC board.

I have a project, for instance, where I need to measure amplitude and phase of S21 thru a media, 1 to 2.5 GHz, and I think I can do that for around $120 USD in good quantity. It is not going to be quite as accurate as an HP network analyzer, but certainly not $50K!!!! All you need are 2 phase locked sources, some printed couplers, mixer chips, and a DSP processor chip.

biff44 said:

It is a true shame that some small company has not come up with affordable spectrum analyzers and network analyzers by now. There should be a two port DC-10 Ghz vector network analyzer out there for maybe $4000 USD. There is not much inside of these things--it could be done on one PC board.

I have a project, for instance, where I need to measure amplitude and phase of S21 thru a media, 1 to 2.5 GHz, and I think I can do that for around $120 USD in good quantity. It is not going to be quite as accurate as an HP network analyzer, but certainly not $50K!!!! All you need are 2 phase locked sources, some printed couplers, mixer chips, and a DSP processor chip.

Click to expand...

Yup. There's a big gap in the "low cost" VNA market. All the components are out there, someone just needs to put them together. BigBoss has changed my mind about getting the Planar - for AUD 28k there's this brand new Agilent: http://www.home.agilent.com/en/pd-1...k-analyzer?nid=-536902639.912674&cc=AU&lc=eng

biff44 said:

There should be a two port DC-10 Ghz vector network analyzer out there for maybe $4000 USD. There is not much inside of these things--it could be done on one PC board.

Click to expand...

If you look at the "real" VNA, there is quite a lot inside. That's really different from the DIY analyzer boards - in design and in performance.

thylacine1975 said:

I have so lusted after one of those Copper Mountain VNAs!
Although they're a good deal compared to a *real* VNA for professional use, they are by no means "low-cost" in the amateur/hobbyist sense of the word.

Their Australian distributor recently quoted me (A$ ~= US$):
~$28,000 for the 804/1 model,
~$16,000 for the 304/1 model, and
~$4,000 for the 1300/1 model.

(+ ~$700 for the calibration kit). Ouch.

Click to expand...

Those VNA prices are really excessive, though IF the calibration kit is well done, $700 is not that much. But that is a big IF.

For 3 GHz, a used HP 8753A, B or C would be a far better choice. To get to 6 GHz, you need an 8753B or later, with the 6 GHz option - but there are plenty of them on eBay.

I paid $2700 for an 8753A (0.3 MHz to 3 GHz) with an S-parameter test set. I don't think that was particularly cheap. If you hunt around, you will find them for less.

I also bought a used HP 8720D (50 MHz to 20 GHz) for $16000, which has the TDR option. That was a pretty good deal. I never realized how complicated the TDR was - I assumed it was "just" an Inverse Fourier Transform (IFT) of the frequency domain data, but there is a lot more to it than "just" an IFT. See:

https://www.amazon.co.uk/Handbook-Microwave-Component-Measurements-Techniques/dp/1119979552

For some strange reason, that book is only available for pre-order on Amazon.com, as Amazon.com claims it has not been published, despite giving a publication date in September of this year. But I've had my copy a few weeks, bought on Amazon UK. It can also be bought directly from Wiley, the publisher. (You can read 85% of it online for free at Amazon UK if you log in. But it is well worth buying if you are going to seriously use a VNA)

I know I'd rather have an HP than a Copper Mountain. A used HP wont depreciate much - it has done most of its depreciation. Buy an old HP today, and you are likely to get 90% of that back in a year. With a bit of luck, you might make a profit. I bet you would have a hard time getting 25% of the cost of a $28000 Copper Mountain VNA tomorrow, if you bought it new today.

Also worth noting is the Agilent VNA forums are very good, and many Agilent staff are paid to read them and answer questions on them. I'm sure someone like Joel Dunsmore, the author of that book, knows more about VNAs than all the staff at Copper Mountain put together. He will answer questions on the Agilent forums, even if it is about old kit.

Just a warning. If someone does go and buy one of the higher frequency VNAs with 3.5 mm connectors, budget another $2000 for used test cables. If you want it to actually work beyond 9 GHz, budget another $4000 for a used 3.5 mm calibration kit and another $5000 for a used N calibration kit.

I'm in the position of having a 20 GHz VNA, but can't really afford the cost of the 18 GHz (N) or 26.5 GHz (3.5 mm) calibration kits. I'm making do with cal kits to 6 GHz (N) and 9 GHz (3.5 mm). I really need a cal kit in N to 9 GHz and might buy one of them new, as I can't seem to find a used one. For 18 GHz, the cost is very high.

thylacine1975 said:

I have so lusted after one of those Copper Mountain VNAs!
Although they're a good deal compared to a *real* VNA for professional use, they are by no means "low-cost" in the amateur/hobbyist sense of the word.

Their Australian distributor recently quoted me (A$ ~= US$):
~$28,000 for the 804/1 model,
~$16,000 for the 304/1 model, and
~$4,000 for the 1300/1 model.

(+ ~$700 for the calibration kit). Ouch.

If you can get by with scalar measurements, the Signal Hound + tracking generator (**broken link removed**) is worth a look for (a far more reasonable) $1500 total cost.

Click to expand...

Gentlemen,

My name is Alex Goloschokin, and I am the Managing Director of Copper Mountain Technologies in Indianapolis. I would like to thank you all for taking interest in our VNAs, it is certainly greatly appreciated! The discussion on this board has been brought to my attention, and with your permission I would like to take a few minutes to clarify a couple of points:

1. Please let me assure you that our analyzers are, indeed, *real* VNAs for professional use. Our Data Sheets can be downloaded from the product pages on our website www.coppermountaintech.com, and I would like to invite you to compare our specs to those of the competing Agilent units (or we can e-mail the comparison to you upon request). You can also download and explore our Demo Software from the same pages. Here is what JFW Industries' Chief Engineer Brett Chermansky wrote to us: "I have finished my testing and I was happy with the results. The 8GHz model was nearly identical or better to the 8.5GHz ENA Agilent unit." JFW have since bought our analyzer and are very happy with it. Here are some other satisfied customers of ours: Canadian Department of Defense, US AirForce Institute of Technology, Harvard University (Massachusetts General Hospital), Johns Hopkins Advanced Physics Lab, Motorola Solutions, Finisar Corporation, Networks International Corporation, EMR, Utrecht Hospital (Netherlands), Embedded Solutions (Austria), Advanced Cyclotron (Canada), and many others.

2. Pricing - apparently there was a bit of miscommunication, and our distributor was quoting a three-year service plan together with the equipment, including shipping the analyzers to and from the US for annual verification (calibration) three times. We've since identified a lab in Australia that can perform annual cal, so I've asked our distributor to re-quote just the equipment. I think you will be pleasantly surprised with the prices, especially considering that our analyzers include Time Domain, Fixture Simulation, Gating, and Frequency Offset as standard features. The Agilent unit for $28K mentioned above is a 3 GHz unit, and the price does not include some of the options that are standard in our models. Our 3 GHz models is priced under $12K, and the 8 GHz model is priced under $20K, so considering that our customers find our analyzers' performance equal to or better than comparable Agilent models, I believe our pricing is very reasonable.

Once again, I sincerely apologize for the miscommunication on the pricing that has occurred in Australia. If you have any questions or need any additional information, please feel free to contact me at alex.g@coppermountaintech.com or +1-317-222-5400. With more than 1,700 units sold throughout the world, we are definitely looking forward to serving the Australian RF and Microwave community!

If one wants a VNA, there are several options.

1. Agilent, Anritsu, Rhode and Schwarz bench instruments.
2. Portable instruments such as the Agilent FieldFox. http://www.home.agilent.com/upload/cmc_upload/All/clp_ID.2219184_eng.html?cmpid=zzfindfieldfox
3. Copper Mountain
4. Agilent CertiPrime - fully reconditioned by Agilent. See **broken link removed**
5. Used HP/Agilent
6. Low end hobbiest/semi-pro VNAs.
7. Design and build your own - no easy task, but several people have done it.

Given enough money, an Agilent benchtop unit would be my choice. If I had less money, then I'd look at an Agilent CertiPrime unit, or used Agilent/HP unit. Given even less money, a unit such as those from SDR kits would be my only option. Given the prices of the Copper Mountain units, they don't attract me.

Although older HP units on the used market are not offically supported by Agilent, in practice there is a lot of support available for them.

1. Full service manuals available often as a free download, or for a modest price if you want paper.
2. A lot of support from Agilent forums
3. Plenty of people know how to use them.
4. Tons of application notes.

I recently purchased an Agilent FieldFox N9923A - this is a 6 GHz portable VNA. I was not over impressed with this, as it had what I considered an unacceptable number of firmware bugs. Support from Agilent was first-class. Eventually I decdided I did not want to keep the unit, so Agilent agreed to a refund. They picked up the shipping costs both ways too. UPS sent me an email, and all I had to do was to pick a data/time convenient to me, arrange the pickup via UPS, and Agilent pay UPS directly. Despite my concerns about this particular model, I'd still buy Agilent again.

I'm sure I'm missing something, and expect CopperMountTech can clarify this, but the only calibration kit I can find on the Copper Mountain web site is a 1.5 GHz N calibration kit. I looked at the data on that, and see this seems to follow the Agilent method, of giving the characteristics of the opens in terms of an offset delay and a frequency dependant fringing capacitance. I was puzzled why the female open offset delay is -13.68 ps. Certainly on the Agilent ones, the distance of the open to the reference plane is always positive and since the female on an N protudes beyond the reference plane, it seems odd to have a negative delay.

Dave

Generally speaking, Copper Mountain Technologies (CMT) does not produce calibration kits. Making a precision calibration kit is not a trivial task, and there are several manufacturers in the world who excel at that, Rosenberger definitely being the top provider (and the most expensive one), Maury Microwave (who actually produce cal. kits for Agilent), Spinner, and some others. CMT VNAs can be calibrated with any SOLT or SOLR cal. kit available on the market today (and the 804/1 model also supports TRL calibration) – we have a library of most commonly used kits built-in into our software, and if the kit of choice is not in the library, it can be added by specifying its parameters from its Data Sheet or by importing its parameters file (if such is provided by the kit manufacturer).
Our N1.1 kit is the only kit we produce ourselves, and it provides a lower-cost alternative for the use with our TR1300/1 model (300 kHz – 1.3 GHz). The TR1300/1 is priced under $3K, and if a user does not already own a calibration kit, it is highly unlikely that he/she would spend $5K on a 6 or 9 GHz Rosenberger or Maury kit when purchasing a $3K 1.3 GHz VNA. So the N1.1, priced at around $600, combined with the $3K TR1300/1 is some sort of an “economy pack.” The N1.1 is a good enough kit for its frequency range but one still can achieve higher accuracy error correction using higher precision (and higher cost) calibration kits.
The CMT N1.1 Female Open Component is specified to have a -13.68 ps time delay due to its simplified design. Normally, a Female Open calibration component would have a protruding center conductor with a length of 13.68 ps (or more if the measure is offset). The Female Open of N1.1 does not have a center conductor, so it is specified as -13.68 ps (which is the distance from the port plane to the reference plane), which allows us to make precise measurements of the absolute values of the phase.

CopperMountTech said:

Normally, a Female Open calibration component would have a protruding center conductor with a length of 13.68 ps (or more if the measure is offset). The Female Open of N1.1 does not have a center conductor, so it is specified as -13.68 ps (which is the distance from the port plane to the reference plane), which allows us to make precise measurements of the absolute values of the phase.

Click to expand...

When you say "female open", do you mean a device which looks somewhat like a female N connector, or one which connects to a female test port, and so looks like a male N connector? That might seem a silly question to some, but when attaching an "open" calibration device with a female connector to a male test port, one has to select "Open (M)" in the menus on HP/Agilent equipment. The sex is defined as that on the test port, not on the calibration device you are attaching.

I've seen people select "Open (F)" or "Short (F)" in HP menus when the calibration device in their hand is a female N. This is wrong, and they should select "Open (M)" or "Short (M)", as they are connecting to a male test port. To me at least, this is counter-intuitive.

Just to make matters worst, "Open -F-" has the dead opposite meaning to "Open (F)".

If you buy the "Qucikcal" option or the Agilent N9923A Fieldfox, which works to 6 GHz, you can calibrate just using an open-connector. There is no need for any cal kit. So when doing a calibration, it asks you to select the DUT connector type and sex, then just leave whatever you connect to the DUT open. That works to at least 6 GHz, though it is not quite as accurate as a mechanical calibration kit. I assume this Quckcal works the same way on the higher frequency portable VNAs, but I'm not sure of that. Some of the portable units go to over 20 GHz.

The method the Quickcal uses is proprietry Agilent technology, and so the method is not published. But I'm quite impressed with it.

I can undertand someone not waning to pay $3k for a VNA and then $5k or a cal kit. That said, the Agilent 85032F (9 GHz N), is about $2200, so less costly than the VNA.

There is at this minute someone on eBay trying to sell the used obsolete 85032B (6 GHz N) cal kit, at a little over $3000, when the Agilent replacement is rated to 9 GHz and costs less than he wants for a used one

**broken link removed**

Admittidly his kit does have the set of N to APC-7 adapters, but it still seems excessive when a new current model can be bought for less than an used obsolete one.

Dave.

If you are simply working for personal use you could use this VNA for about £550 with a cal-kit: **broken link removed**

Cheers,

Joel Richard
ca.linkedin.com/in/joelrichard/

I recently tried copper mountain's two port and one port VNAs, and I liked them both.

RFdesignHQ said:

If you are simply working for personal use you could use this VNA for about £550 with a cal-kit: **broken link removed**

Cheers,

Joel Richard
ca.linkedin.com/in/joelrichard/

Click to expand...

The original poster was looking for 2-5 GHz - what is well beyond the capabilites of a VWNA.

Although the VWNA is cliamed to have redunced performance between 500 MHZ and 1.3 GHz, I have seen some results from a version 2 of that, and I would have said it was unusable.

Dave

drkirkby said:

The original poster was looking for 2-5 GHz - what is well beyond the capabilites of a VWNA.

Dave

Click to expand...

True, I guess when it comes to VNAs nothing is free. I can't help but wonder about increasing the frequency range with external mixers.

The lowest cost VNA that goes up to 4GHz I am aware of is the MegiQ VNA-0440
At some 4000usd it gives you a full bi-directional dual port VNA with USB interface and software for the user interface.
You can visit their site www.megiq.com

There are other nice VNA's mainly for amateur use, but some are serious instruments. However I have not found one that goes far higher than 1GHz.

For anyone interested, there is the 2port VNA called "miniVNA Tiny" that works from 1-3000MHz for $580 US. Site: http://miniradiosolutions.com/minivna-tiny

It can also interface with your android device.