VNA Recommendations

[garboui]

For work I am about to jump in to the purchase of a VNA. This is mostly to verify matching of on board chip style antennas to an RFIC (2.4GHz). As these devices are generally expensive new I am approved to buy in the used/ebay market. Right now it seems that the selection is split betwen teh small handheld/benchtop Anritsu units and the large rack mount units from Advantest and HP. For similar price and frequency, these rackmount units tend to be at least twice the age sporting floppy drives.

Is there any inherent advantage/disadvantage to going Anritsu vs old rack mount? I do like the idea of being able to fit on my desk.

[D3f1ant]

Me too  Checking/tuning pcb chip antennas (verify reference designs). So expensive  Its not really core business so want something that will the job done for cheap. Whats wrong with a PocketVNA or miniVNA Tiny for this kind of work?

[rx8pilot]

I have been wanting one of these just for the learning opportunity. Right now I have no real need and don't even fully understand what they are able to measure, why and how. Wanted to join the thread to see where it goes - I don't really have any technical contribution.

[Muxr]

I have been wanting one of these just for the learning opportunity. Right now I have no real need and don't even fully understand what they are able to measure, why and how. Wanted to join the thread to see where it goes - I don't really have any technical contribution.

Yeah same. Probably not relevant to the OP as he's looking for a commercial unit for work.. but I've been toying with the idea of building one like this dude did just recently: http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html

[nctnico]

For work I am about to jump in to the purchase of a VNA. This is mostly to verify matching of on board chip style antennas to an RFIC (2.4GHz). As these devices are generally expensive new I am approved to buy in the used/ebay market. Right now it seems that the selection is split betwen teh small handheld/benchtop Anritsu units and the large rack mount units from Advantest and HP. For similar price and frequency, these rackmount units tend to be at least twice the age sporting floppy drives.

Is there any inherent advantage/disadvantage to going Anritsu vs old rack mount? I do like the idea of being able to fit on my desk.

I think you just have to compare the specs (noise, dynamic range) but for antenna matching you probably don't need a large dynamic range or extremely low noise. I'm wondering at which Anritsu units you are looking though.

[rx8pilot]

I just took a look at the PocketVNA - do you suppose that would accomplish your seemingly simple needs? Looks like you may see around 50+db dynamic range at 2.4Ghz.

Do you have a budget or any technical targets you are hoping to hit? Are you trying to buy bigger than today's requirement so you have some future proofing? Do you have a lot of room on the target bench for one of the older and giant instruments?

[TheSteve]

If you have a tighter budget I'd aim for an older HP - 8714 of some sort perhaps. You'd get 100 dB of dynamic range at least.

[rx8pilot]

Even those ancient boxes are fairly expensive. Nuts. Any idea, out of pure curiosity, did those cost new. They look like mid-80's or so.

Sent from my horrible mobile....

[KJDS]

HP8713 were about £20k new in the late 90s

An HP8510C was about £160k in the late 80s

[TheSteve]

Even those ancient boxes are fairly expensive. Nuts. Any idea, out of pure curiosity, did those cost new. They look like mid-80's or so.

Sent from my horrible mobile....

Well you have to be patient and watch ebay to get a good deal. I have seen working 8714C's go for 1K or less several times over the past year.
There is an 8753C with a test set(of some sort) for 1500 right now.

[D3f1ant]

Not that excited by really old gear, seems risky if dont know anything about this stuff and a few grand is a lot to spend on something that could die any day now.

I'll beat the MegiQ shill, that always posts in these threads all over the net, and suggest http://www.megiq.com. More expensive than a pocketvna or minivna...and cheaper than a coppermountain or Anritsu. From a learning perspective that sandbox kit looks like something.

Wish there was more independent hardware reviews somewhere.

[rfbroadband]

since you said it is for work I will recommend

R&S ZNB, officially called medium class but has almost high end class performance.

Very fast , very good dynamic range, very quiet !

I love the unit.

https://www.rohde-schwarz.com/us/product/znb-productstartpage_63493-11648.html

[rx8pilot]

Any idea what the price range is on those for 6Ghz or whatever closest option level to that may be?

[hendorog]

Not that excited by really old gear, seems risky if dont know anything about this stuff and a few grand is a lot to spend on something that could die any day now.

I'll beat the MegiQ shill, that always posts in these threads all over the net, and suggest http://www.megiq.com. More expensive than a pocketvna or minivna...and cheaper than a coppermountain or Anritsu. From a learning perspective that sandbox kit looks like something.

Wish there was more independent hardware reviews somewhere.

Their products all seem expensive. Dont go to low frequency either. Based on the frequency it just looks like a synth chip combined with a adl5380 demodulator.

[D3f1ant]

They dirt cheap compared to a R&S ZNB that start at >US$32000 
Surely somebody here owns one of these budget friendly modern usb type things. The quality of the software is going to be important. The MiniVNA seems to get fairly positive feedback elsewhere.

[hendorog]

Heh, yep very true. But its not comparable with that,  just seems too high compared with the likes of the minivna and vwna. The open source one mentioned sounds interesting though.

[Howardlong]

For the OP I've had an 8753a with a T/R test set for about 15+ years. It works well except for the CRT which in recent years started acting erratically. I replaced it with an LCD screen from Newscope about 18 months ago and it's behaved beautifully ever since. The electronics in these workhorses is superb. My only fault is that the internal battery/supercap failed before I even took delivery of it, and I've yet tomfigure out how to fix it. It just means I need to recal every time I switch it on, which is no big deal in my use cases as pretty much every time I switch it on it's for a different frequency range anyway.

I also have an Anritsu Sitemaster which is one port only (I think some are two port, not sure), so these are fine for antennas, but not for measuring filters etc. The one I have is again quite old, I bought it second hand about 12 years ago, it is a VNA but the built in display only gives you scalar return loss. You need to use the long-in-the-tooth software and a serial cable to see the Smith charts. It's battery powered so you can use it in the field, but the batterries, ISTR they're NiCds, or nickel metal hydride, are the weak link. Next time I get around to replacing them I'll put in eneloops.

On the bench, the 8753a is what gets used, out in the field it's the Anritsu.

Keep in mind that as well as needing a "proper" cal kit, you'll need to make up and validate your own test fixtures and cal kits if you're measuring on board integrated antennas. Plus, in the design itself, placing optional DNP micro RF connectors on top of in line zero ohm links (or ceramic caps) at one or two points will be a big help. Not a bit deal, but it's worth mentioning. I use these http://www.mouser.co.uk/ProductDetail/Murata-Electronics/MM8030-2610RJ3/?qs=dnvuLCnOKpt0o1NKcCkttA%3D%3D

[KJDS]

I've got an HP8753A with a test set and an R&S ZVRE that's good to 4GHz, both of which I've had for a while and now I'm back to real work I rarely use either.

Anyone interested in buying either drop me a PM

[joeqsmith]

I have been wanting one of these just for the learning opportunity. Right now I have no real need and don't even fully understand what they are able to measure, why and how. Wanted to join the thread to see where it goes - I don't really have any technical contribution.

Yeah same. Probably not relevant to the OP as he's looking for a commercial unit for work.. but I've been toying with the idea of building one like this dude did just recently: http://hforsten.com/cheap-homemade-30-mhz-6-ghz-vector-network-analyzer.html

I read through his blog.  Pretty neat!   

I have a very old 8754A that I would NEVER recommend it for commercial use.   That's my advice to the OP.   
One day I may get something better for home use but for now the old relic fits my needs.

[bson]

I use these http://www.mouser.co.uk/ProductDetail/Murata-Electronics/MM8030-2610RJ3/?qs=dnvuLCnOKpt0o1NKcCkttA%3D%3D

Interesting!  What do you use for cabling?  Murata seems to have an SMA-to-SWG(2610) probe for these ($80 on Mouser).  This looks like exactly what I've been looking for to step up from Hirose U.FL and pigtails.  Any experience with it?

[Howardlong]

I use these http://www.mouser.co.uk/ProductDetail/Murata-Electronics/MM8030-2610RJ3/?qs=dnvuLCnOKpt0o1NKcCkttA%3D%3D

Interesting!  What do you use for cabling?  Murata seems to have an SMA-to-SWG(2610) probe for these ($80 on Mouser).  This looks like exactly what I've been looking for to step up from Hirose U.FL and pigtails.  Any experience with it?

Not cheap, but then this is for R&D, not production. http://www.mouser.co.uk/ProductDetail/Murata-Electronics/MXHQ87WA3000/?qs=sGAEpiMZZMuur7c7MCGU2AqtzKLnOBpO

You need to be a bit dainty with them, and make sure that whatever you're testing will work logistically as they're about 16mm high but they work.

[Howardlong]

Pics of the comparative sizes

[D3f1ant]

I'm going to try a pocketvna. If its sh!t its not so expensive that I'll lose sleep over it. Will post a teardown if anybody is interested.

[garboui]

I just took a look at the PocketVNA - do you suppose that would accomplish your seemingly simple needs? Looks like you may see around 50+db dynamic range at 2.4Ghz.

Do you have a budget or any technical targets you are hoping to hit? Are you trying to buy bigger than today's requirement so you have some future proofing? Do you have a lot of room on the target bench for one of the older and giant instruments?

I want to stay away from a purely USB based device. A nice machine UI is what I am after. This equipment may also at some point in its life become attached to an ATE so having standard SCPI/ VISA interface (Ethernet preferred but not required) is also important.

Not that excited by really old gear, seems risky if dont know anything about this stuff and a few grand is a lot to spend on something that could die any day now.

I'll beat the MegiQ shill, that always posts in these threads all over the net, and suggest http://www.megiq.com. More expensive than a pocketvna or minivna...and cheaper than a coppermountain or Anritsu. From a learning perspective that sandbox kit looks like something.

Wish there was more independent hardware reviews somewhere.

This! Also being for work if the ole' faithful dies because its old and needs repair and then the repairs take weeks because of parts scarcity this will be no bueno! A couple grand for something more current is easily justified in this case.

In the past I have had the (dis)pleasure of using a AEA VIA Echo. I commend its features to price ratio though its interface if abysmal. To get useful, it really needs the PC software talking to its (closed) usb interface. This interface also was super buggy and slow!

As far as buying for something taht will leave room to grow. We likely wont be seeing anything past 2.4g in my current market segment. Though if the mater is 4-6GHz then this gives a bit of headroom. Having capability for 5GHz wi-fi work could still be useful in the future.

[chscholz]

It's really hard to beat the performance of a ZNB!

We also have a value version of the ZNB, the ZND.
 
ZND comes only as 2-ports version, only up to 4.5 GHz or 8.5 GHz (no 20 GHz or 40 GHz version like you can get for the ZNB). Lowest frequency is 100 kHz (as opposed to 9 kHz), lower output power, lower dynamic range, slightly slower sweep speed, only 5,001 points per sweep (as opposed to 100,001 for the ZNB)
But the ZND has the same GUI, physically the same touch screen, knobs and buttons, same fan, same overarching design with two receivers (reference and measurement) per port. Low-cost VNAs usually have a shared reference receiver, you'll never get "benchtop performance" with these instruments.

ZND is aggressively priced, even less expensive than the portable ZVL, but still more expensive than a USB VNA  

since you said it is for work I will recommend

R&S ZNB, officially called medium class but has almost high end class performance.

Very fast , very good dynamic range, very quiet !

I love the unit.

https://www.rohde-schwarz.com/us/product/znb-productstartpage_63493-11648.html

[hendorog]

ZND is aggressively priced, even less expensive than the portable ZVL, but still more expensive than a USB VNA  

Could you share some idea of the prices?

[loxodes]

ZND is aggressively priced, even less expensive than the portable ZVL, but still more expensive than a USB VNA  

Could you share some idea of the prices?

It looks like the 8.5 GHz ZND goes for about 20k USD.. (https://www.microlease.com/us/products/rohde-schwarz/rf-network-analyzers/znd8?basemodelid=111969&bestbuyonly=1)

[tatus1969]

The MiniVNA seems to get fairly positive feedback elsewhere.

I am using it and am not very happy with it. IMO it has two big problems:
- the entire's signal path frequency response is not very flat, they heavily rely on compensation there. You can even see sawtooth residues from that
- the signal gen's output impedance over frequency is also not very flat. As this cannot be compensated, this is my main concern.
Anyhow, for that low price a great tool.

[Performa01]

...
- the signal gen's output impedance over frequency is also not very flat. As this cannot be compensated, this is my main concern.
Anyhow, for that low price a great tool.

The vast majority of these instruments generally don’t have a perfectly matched input/output impedance over the entire frequency range.

The usual cure is to introduce some fixed inline attenuator with 6 or 10dB, which improves the return loss by 12 or 20dB respectively. Of course, this is at the cost of input sensitivity / output power, but that shouldn’t be a big concern for most applications.

[tatus1969]

thanks for the tip, will consider that next time running into that problem. Doesn't help for reflection measurements with the miniVNA's built in directional coupler though. Just want to say that this one is good for hobbyists but I would not recommend using it in any more professional tasks.

[hendorog]

ZND is aggressively priced, even less expensive than the portable ZVL, but still more expensive than a USB VNA  

Could you share some idea of the prices?

It looks like the 8.5 GHz ZND goes for about 20k USD.. (https://www.microlease.com/us/products/rohde-schwarz/rf-network-analyzers/znd8?basemodelid=111969&bestbuyonly=1)

Thanks for that. By the way, how are you getting on with your VNA project?

[loxodes]

Thanks for that. By the way, how are you getting on with your VNA project?

(for context, hendorog is talking about my open hardware VNA: http://github.com/loxodes/vna)

The VNA works from 500 MHz to 3.5 GHz with about +/- .5 dB of accuracy, 40 dB or so of dynamic range, and a dozen points per second. I'm waiting on the next revision of some boards to push the maximum frequency up to 6 GHz and improve the accuracy and dynamic range.

The current topology relies on a broadband I/Q demodulator with an output that goes down to DC (an ADL5380), and I don't know of any of those above 6 GHz. I haven't decided how I'm going to push it past that, but I'll probably end up ditching the I/Q downconverer and doing that in software like Henrik Forstén's VNA. I'm playing with the AD9864 to handle the sampling and downconversion with a BeagleBone PRU for high speed IO. The LTC5548 mixer/doubler could enable operation up to 14 GHz off my existing synthesizer, I'm waiting on a test board for that.

I wouldn't recommend that anyone reproduce what I'm doing yet, Henrik Forstén's VNA is a better option for a DIY VNA at this point. When I'm done I hope to have something an order of magnitude slower, less accurate, and cheaper than a commercial VNA.

[hendorog]

The VNA works from 500 MHz to 3.5 GHz with about +/- .5 dB of accuracy, 40 dB or so of dynamic range, and a dozen points per second. I'm waiting on the next revision of some boards to push the maximum frequency up to 6 GHz and improve the accuracy and dynamic range.

The current topology relies on a broadband I/Q demodulator with an output that goes down to DC (an ADL5380), and I don't know of any of those above 6 GHz. I haven't decided how I'm going to push it past that, but I'll probably end up ditching the I/Q downconverer and doing that in software like Henrik Forstén's VNA. I'm playing with the AD9864 to handle the sampling and downconversion with a BeagleBone PRU for high speed IO. The LTC5548 mixer/doubler could enable operation up to 14 GHz off my existing synthesizer, I'm waiting on a test board for that.

I wouldn't recommend that anyone reproduce what I'm doing yet, Henrik Forstén's VNA is a better option for a DIY VNA at this point. When I'm done I hope to have something an order of magnitude slower, less accurate, and cheaper than a commercial VNA.

Thanks, and congrats on the good results.

I made something similar using an ADL5380 and a MAX2871 PLL. It worked but the accuracy wasn't great.
The plan is to add a second receiver for a reference, a second PLL and move to mixers instead of the ADL5380.

The 2nd PLL will be used to shift the phase - ala N2PK VNA - which removes the need for IQ demod, permits harmonic cancellation and DC offset cancellation.