开云体育

Hi, if I have an RF analyzer that has an analog bandwidth of, say, 10
MHz-4GHz, is it possible to add a front-end that will let it characterise a
device at audio rates, i.e. 10 Hz-100 kHz? What does this entail? Assume
the software (running on PC) and firmware can be modified. The device I am
specifically thinking of is the XTRX.

Thanks

The short answer is no. The long answer is that that you could conceivably do that with enough complexity, money, and time. You're better off getting equipment built for low frequency work - same principles, but different stuff, and generally simpler and cheaper.

The problem is that to RF/microwave stuff, audio may as well be DC - it's down in the mud, so to speak, and even though you can of course put low frequency signals on RF with various modulation techniques, it takes a lot to resolve what's going on. That's why your XTRX SDR thing for example, says "...down to 100 kHz with some degradation..." in discussions. There's not much point in using an RF system to analyze audio. It's for conveying it. There's no need to go through all the up and down conversion processes when you can just figure it out directly and more precisely in the baseband, especially with all the sophisticated analog and digital circuitry and DSP available.

Ed

Forgot to mention one thing - an old (1970s) example of doing something like this is in one of the HP141 modular RF/microwave spectrum analyzer systems. There's a plug-in (I forget which model - maybe 8553 or close) that goes from very low to 300 kHz or so. It upconverts the baseband to (as I recall) 50 MHz, which is one of the IFs of the system. The rest brings it back down to its 3 MHz IF for detection, logging, and display. The result is limited by noise and the stability of the LOs and the RBW of the crystal filters - about 1 kHz I think, which isn't all that great for audio.

This sort of thing is a way to get baseband up into the range of RF equipment. But again, why bother if you can figure it out directly?

And one last thing. If this application is for an audio network analyzer only, and not generally a spectrum analyzer, then no SDRs or SAs or RF stuff is needed at all - just generate test frequencies and measure the results, say with a low frequency generator and an AC voltmeter or detector. Maybe a PC sound card and software is enough. If you want something already built and ready to go, maybe an HP3561A or something similar and much newer and better would do.

Ed

If you are looking for an audio frequency network analyser, look at Bob Larkin's article: DSP-Based Vector Network Analyser for 10 Hz to 40 kHz in the May/June issue of QEX Magazine. You would have to assemble your own; however, boards and the software have been made available.

Bruce, KG6OJI

I've found the analog discovery most useful for low frequency work.
Not because it is especially great hardware, but it comes with
ready-made software and is relatively affordable.
Often the software that comes with this type of toy is poor and buggy,
but the network analyzer portion is well made in my opinion. It still
picks the spot frequency out of a very noisy switchmode supply with
great dynamic range so you can do control loop analysis. I wouldn't
want to use the scope part of the software for anything, it's crap.

<>

If you want a step up from a sound card with more frequency range and
at least rudimentary input attenuators this is a good choice.

ST

On Fri, May 4, 2018 at 4:21 PM, ebrucehunter via Groups.Io
<Brucekareen@...> wrote:

If you are looking for an audio frequency network analyser, look at Bob Larkin's article: DSP-Based Vector Network Analyser for 10 Hz to 40 kHz in the May/June issue of QEX Magazine. You would have to assemble your own; however, boards and the software have been made available.

Bruce, KG6OJI

The LF plugin is the 8556A. With an 8552A IF plugin, you can get down to 50Hz RBW. The superior 8552B can do 10Hz. The 1kHz you're thinking of is the minimum center frequency on the 8553 RF plugin.

The reason the 8556A only tunes up to 300kHz is that it's a fixed upconversion (allowing trims on the mixer to null out oscillator play-through), using the IF plugin's tuning range which can't exceed its input passband.

Dave Wise
________________________________________
From: [email protected] <[email protected]> on behalf of Ed Breya via Groups.Io <edbreya@...>
Sent: Thursday, May 03, 2018 10:10 PM
To: [email protected]
Subject: Re: [TekScopes] Using network analyzers below their stated bandwidth?

Forgot to mention one thing - an old (1970s) example of doing something like this is in one of the HP141 modular RF/microwave spectrum analyzer systems. There's a plug-in (I forget which model - maybe 8553 or close) that goes from very low to 300 kHz or so. It upconverts the baseband to (as I recall) 50 MHz, which is one of the IFs of the system. The rest brings it back down to its 3 MHz IF for detection, logging, and display. The result is limited by noise and the stability of the LOs and the RBW of the crystal filters - about 1 kHz I think, which isn't all that great for audio.

This sort of thing is a way to get baseband up into the range of RF equipment. But again, why bother if you can figure it out directly?

And one last thing. If this application is for an audio network analyzer only, and not generally a spectrum analyzer, then no SDRs or SAs or RF stuff is needed at all - just generate test frequencies and measure the results, say with a low frequency generator and an AC voltmeter or detector. Maybe a PC sound card and software is enough. If you want something already built and ready to go, maybe an HP3561A or something similar and much newer and better would do.

Ed

Dave Wise wrote:

"The LF plugin is the 8556A. With an 8552A IF plugin, you can get down to 50Hz RBW. The superior 8552B can do 10Hz. The 1kHz you're thinking of is the minimum center frequency on the 8553 RF plugin.
The reason the 8556A only tunes up to 300kHz is that it's a fixed upconversion (allowing trims on the mixer to null out oscillator play-through), using the IF plugin's tuning range which can't exceed its input passband."

Yup, that's the one - I could never remember which was which, except for the 8555A. I used to have a HP141T system and full set of plug-ins and external accessories way back. It was fun to fool around with, but always finicky, and I had to review how to set up and jiggle everything around whenever I used it. Once I graduated to the HP8566B, I decided to junk out the old system.

Besides saving all the RF and other goodies, I kept the front-end, tracking source, and upconverter guts from the 8556A LF unit, to do the same sort of thing for covering a lower frequency range with the 8566B, which doesn't play all that well below below 20 kHz or so. The plan was to convert the LF baseband input up to 50 MHz and view it on the 8666B. The LO would be derived from the SA's reference, so stability would be good. This is a very low priority project that will likely never be completed - an interesting possibility, but not essential, since I have an HP3561A that I hardly ever use anyway.

About the only rationale for maybe building it eventually is just "because," to round out the 8566B setup for the very bottom end coverage, including tracking. I converted the HP8443A to go with the 8566B and provide tracking from about 100 kHz to 250 MHz. Unfortunately, the original guts that I re-used are incapable of reaching much below 100 kHz, without serious modification - the output amplifier/leveler system cuts off, even though the mixer could reach DC if I wanted to build a different output system.

This low frequency cutoff issue is another example of why RF systems can't/shouldn't be used directly at LF/audio - it's usually unnecessary and impractical to go that far down.

Ed