Looking for a vector network analyzer for audio circuits

Hello everybody!

I am looking for a vector network analyzer that can measure audio signals (20kHz down to 20Hz).
I know that the "best" solution would be just using the soundcard from the computer and some software but I want a standalone device.

I already searched online but I only either found audio analyzers that would just analyze a incoming signal (no built in signal generator) or I would find vector network analyzers that worked from 9kHz to a few GHz.
But I want an actual VNA for audio frequencies.

Can anybody recommend me anything?
 
Thanks for the quick reply!
That looks just like what I need.
But I could not find a single vendor that actually specifies the price for one of those devices, which is not a good thing ...
I am willing to spend some money on this but not like 10k€ ... (does that make me sound naive?)
 
I know about "Hewlett-AgiSight", as I like to call them ;-)
But their (new) stuff is definetley outside of my current financial capabilities.

I guess I will just have to have some patience and hope that maybe some day some used equipment will be sold on something like ebay ...

I shouldn't be surprised that this kind of test equipment is so expensive.
A lot of time, money and effort goes into defeloping it and not that many people need it, so it is not really optimized or developed for cheap production.
Built to a spec, not to a price ...
 
  • Like
Reactions: rayma
Probably a Bode 100 is the best match. You do need a PC. They work from 1 Hz to 50 MHz. However they retail for over $6K and none show up on eBay. I'm not sure that you would get more with one that from an audio analyzer or soundcard + software. If REW or ARTA doesn't do what you want look into Virtin's: https://virtins.com/multi-instrument.html There is a learning curve but its very capable.
 
  • Like
Reactions: rayma
Last edited:
Another idea would be to build your own using an existing generator and scope or equivalent.

Look at Appendix 2:

Bateman Directional Coupler

But, I have to say that the SDR-Kits VNA's are really good, too. The nano VNA variants don't go down low enough in frequency for this application and would need that converter from QEX magazine.

How about this? Analog Discovery Network Analyzer

Even if you add a suitable computer so that the above solution is standalone, it's still pretty cheap compared to professional solutions. Professionals who make money from making measurements can and should pay for test equipment that is accurate and the calibration is traceable. Amateurs, even advanced ones, can come within a few percent of a professional's measurements these days by using the right test gear and a lot of common sense and understanding.

It all depends on what you want/need to do.
 
  • Like
Reactions: jasonRF
Thats usually called a dynamic signal analyzer or FFT analyzer.
One example of a dynamic signal analyzer that could be useful here is the HP 3562A and HP 3563A. The '63 adds a digital output but is otherwise identical to the '62. The main limiting factor with those is their 80 dB dynamic range.

The newer HP/Agilent/Keysight/whatever-name-this-week 35670A is nice too. Sadly it doesn't offer any significant improvement in dynamic range over the 3562/63.

Also Audio Precision of course.
Unfortunately AP isn't as useful as you'd think. It's not until fairly recently that the APx software has started to include the phase measurement in their frequency response plots. At least I think it now includes the phase measurement. I haven't actually verified this. I use the phase response a lot to determine pole/zero locations.

Also, to go beyond ~80 kHz you either need an APx555 ($30k+) or the fancy generator and fast FFT options ($$) on an APx525, SYS2700, etc.

Furthermore, the AP software barfs at frequencies below about 10 Hz. It's purely a software limitation and it's possible that there is a way around it. What seems to be going on is that the settling algorithm in the software gets confused by low-frequency signals so it bombs out with an "I can't find the signal" error. I must admit that I haven't tried real hard to work around this issue. It's pretty rare that I need to measure below 10 Hz anyway and I can easily pull out my HP 3563A for this.

For measurements above the 200 kHz limit of the AP I use an HP 3577A network analyzer.

Tom
 
There are a few HP3562s on ebay right now, and they are not too expensive.
I might go with one of those.
Need to save up a bit of money first ...

The Bode 100 is a very nice bit of test equipment!
I have used it a lot at work, can recommend it 100%.
But it is expensive ...
It used to cost only 4k€ last time I checked, but I think that was a year ago ...

Most relevant (right now) is just beeing able to quickly measure the frequency response of filters, amplifier circuits and things like that.
Until now I have been doing it "by hand" with my function generator and scope.
Yes, I could just write a simple bit of software that tells the function generator to do a sweep and then reads in the values from the scope.
But getting that to work is another can of worms and I want something that just works.
(The scope I have is an old DSO, a HP54645D)

Thanks for all the quck and helpful answers here!
 
Dunbar_Build_A_Vector_Network_Analyzer_1.jpg
 
  • Like
Reactions: Murray Allen
Hi Andre, what are your measurement tasks? You're telling you need a VNA (quite unuasual in the audio range) but probably you're interested in the standards like THD, noise, Bode plots etc. If so, I can highly recommend the good old R+S UPL. You'll find plenty of information in this forum. They used to be in the range of some XXk€ (a couple of decades back) but you'll find them today in really good shape for 2k€ - 3k€ on ebay.

https://www.ebay.de/itm/29577365273...JbbmqbNyt49A4z+91DObMJzq8x|tkp:Bk9SR5ycivH8Yw

If you need the famous B1 option (ultra low distortion generator) it will double the price. I've visited this shop in Duesseldorf/Germany where I've got mine. They have a great reputation and are trustworthy.
 
Using an APX555 here along with a HP3577A, 35665A, 3561A etc. Thought about the Bode 100, but lack of need/interest nixed that idea.
I built a Rogowski coil for wideband (or wide-ish band) signal injection and loop gain measurement. That worked very well with the 3577A.

So much snake oil in audio has discouraged further research into audio development.
No kidding. That's one of the reasons I'll stick with the DIYers. They tend to understand specs and performance measurements, and tend to appreciate the effort that went into designing something that works well.

Tom
 
Hi Andre, what are your measurement tasks? You're telling you need a VNA (quite unuasual in the audio range) but probably you're interested in the standards like THD, noise, Bode plots etc. If so, I can highly recommend the good old R+S UPL. You'll find plenty of information in this forum. They used to be in the range of some XXk€ (a couple of decades back) but you'll find them today in really good shape for 2k€ - 3k€ on ebay.

https://www.ebay.de/itm/295773652738?itmmeta=01HZKH2HR6MNJ0ZYF3YA0RRC0X&hash=item44dd7bc702:g😛k4AAOSw59NkkzZ6&itmprp=enc:AQAJAAAA4F7lzwYZ9O50ybmT+LcU7kW5s8gSSQfPOKaNAM2bqOznt/vwMhndR2vzNv+cxM2ptADr/kfPP49S3XUbT05q/Ssgxegj+YcfJMmLFqP9ZHgEEfCthshTBLzPp1kzj/Aoxc6s8NcmJvInmpBwQyPEYJuOYKxk/+jr/XJb1Ll+pFYdURnDhHDR9UvjCPhpeVb0NF8dIPgwR404x0tCk1Y6eYhto4fjPlzc9lkb3Xj5NVme7cU3m+SmmWGqaU+hKEBJ0FzlsrM9Llu8RGDZV4JbbmqbNyt49A4z+91DObMJzq8x|tkp:Bk9SR5ycivH8Yw

If you need the famous B1 option (ultra low distortion generator) it will double the price. I've visited this shop in Duesseldorf/Germany where I've got mine. They have a great reputation and are trustworthy.
Yes, I am currently mostly interested in Bode plots, THD and noise measurement would be a nice to have.
I am just trying to avoid getting something and then finding out a few years later that I need some fancy extra functionality.
Always get a bit more then what you need.

For my current needs (like you said, just the normal stuff for now) the instrument that you linked would be more then good enough.


That looks intersting.
I am pretty comfortable with SMD soldering, so I might get some PCBs made for that and try it out.
Open source hardware sure is an amazing thing.