Suggestion for cheap network analyzer / impedance analyzer

Klaus-

I wonder if the MFJ-225 Graphic Analyzer would do what you want? It is meant to be used for antenna analysis, but can also do some network analysis. It lists at $400. Amateur Electronic Supply sells it for $370.

From the MFJ description, "Two Analyzers in One Out in the field, MFJ-225 is a compact completely self-contained handheld analyzer. On the bench it becomes a full-fledged two-port (S21) desktop machine when teamed up with your PC. Using powerful IG-miniVNA freeware, you`ll run detailed data analysis and then print out stunning color-graphic plots to document your work!"

The built-in display of an SWR plot does not impress me, but the computer software should do better.

Fred

100Hz may be tough but if 1kHz works you could try an SRD kit:

But it doesn't have a reference input which I usually need.

Then there is the Signalhound series where I have both the track gen and analyzer. Supposedly one can do scalar network analysis with it but I've never done that:

They cost a little more in Europe than here and the SA44/TG44 set that I have does cover down to 100Hz and lower:

It doesn't get much smaller than that.

--
Regards, Joerg 

http://www.analogconsultants.com/

Unfortunately it won't go below 1.5MHz.

--
Regards, Joerg 

http://www.analogconsultants.com/

The kit seems to cost as much as the signal hound you suggested. The hound really looks nice, and the price is good

I did program a generator and scope combination to act as a network analyzer with very good results, but it's slow and my lack of spare time means the GUI is not user friendly and has few functions

Cheers

Klaus

I like it. It's not very good for pulsed stuff, digital TV signals and such. But I did EMC pre-compliance scans with it and the lab plots came in almost verbatim the same. The Signalhound is also somewhat programmable so it probably has cabapilities I'll never even scratch the surface of because I am not exprienced in programming.

You can email Signalhound about a VNA application, they are very responsive. The problem with a scope is the paltry 8-bit converters up front and no filters so it's going to be noisy. Impossible to get 80dB.

Oh, and you should see my latest "GUI". Push pressure valve with left knee, then when a pulse comes by immediately press CTRL-PrntScrn for documentation. It's one of those days where I have to wear safety googles, don a hard hat and ban the dogs from my lab in case something goes kablouie and flies across the room :-)

--
Regards, Joerg 

http://www.analogconsultants.com/

100Hz may be tough but if 1kHz works you could try an SRD kit:

But it doesn't have a reference input which I usually need.

Then there is the Signalhound series where I have both the track gen and analyzer. Supposedly one can do scalar network analysis with it but I've never done that:

They cost a little more in Europe than here and the SA44/TG44 set that I have does cover down to 100Hz and lower:

Try this puppy;

Isolated generator and other nice features. Cost = BODE 100/2

Cheers; Harry

Over $2k though. Always keep in mind the AD converter granularity. With the usual 8-bitters there is no chance to achieve what Klaus wants to do. Gussied up oscilloscopes are generally no the answer in this scenario. Maybe SDRs are.

--
Regards, Joerg 

http://www.analogconsultants.com/

Over $2k though. Always keep in mind the AD converter granularity. With the usual 8-bitters there is no chance to achieve what Klaus wants to do. Gussied up oscilloscopes are generally no the answer in this scenario. Maybe SDRs are.

--
Regards, Joerg 

Always keep in mine that if you don't read the data sheet you may quote  
erroneously data. 
I read 14 bit ADC and 100MSPS. 
What do you read? 

Harry

d e I

The scope I have been using is a Picoscope with 12bit ADC. If used with run ning average filter (if the frequency is low enough to do that), the resolu tion is boosted to 14-16 bit.

That combined with standard o-scope analog front end (20V/div down to 5mV/d iv), function generator with similar output range levels (max 10V down to 1

0mV into 50 ohm) and syncronized averaging with swept single frequency sine waveform, yields dynamic range over the entire frequency range of excess o f 100dB (only problem it is slow)

Regards

Klaus

Almost all of them do this "resolution boost". But it's not real unles you have ideal signals all the time with no pulsating stuff in there. It looks good on paper but in the end one needs a real hi-res ADC and those are normally only found in gear such as SDR. They can do this because they down-converter. For example, the last stage of my Signalhound has a bandwidth of only 250kHz for this reason, so they can use a hi-res ADC.

In medical we call this instantaneous dynamic range. If you want 80dB for that you realistically need a 16-bit converter with 14 bits or so of ENOB.

As long as you have no pulsating stuff in there.

--
Regards, Joerg 

http://www.analogconsultants.com/

Your response: "Always keep in mine that if you don't read the data sheet you may quote erroneously data. I read 14 bit ADC and 100MSPS. What do you read?"

I read in there on page 140, quote "Moving average filter with 40n ?

1.28us time constant. Resolution improvement to 12 or 14 bits. 20 MHz anti-alias filter".

To me that sounds like marketing speak for "Well, we don't have a hi-res converter in there but we can average the signal". Everybody can average but that does not work in all real-life situations. Where did you see the 14-bit converter listed?

BTW, please adjust your newsreader. The quoting is wrong and it's not good to respond underneath siglines. Makes a mess. I had to hand-copy your response back.

--
Regards, Joerg 

http://www.analogconsultants.com/

To be fair, it's going to be better than the data from your average 8-bit scope: in that case, the LSBs will be much noisier on a (presumed otherwise stable) signal, and hence the quantization noise will be not noisy but stable, and not averaged out.

The equivalent bandwidth for averaging is the DFT of whatever's in the buffer (i.e., the data points, and the number of data). With strong attenuation for signals out of band (too fast to be sampled, or too slow to look like anything beyond DC offset), or uncorrelated (averaged out).

In the old days where "average" meant exactly the 320 samples x 256 levels visible on screen, it didn't win you much if any accuracy, but with larger buffers today (assuming they actually save the LSBs from the averaging process), and ADCs suited to the noise level, it's a clear win.

The resulting equivalent bandwidth is not fixed, but spans from the equivalent sample rate to Fs / (buffer length) or thereabouts.

The 'real' (outside world to 'what you see') bandwidth is, I think, Fs / (N averages), so for example, it drops when Fs(equiv) is higher but Fs is lower in equivalent time sampling, if the scope uses that mode.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com

You mentioned a 'kit', how about a 'design'? I've designed a Network Analyzer that will do out to around 2GHz at a "onesy's" parts cost in the range of $250. It has some 'super limitations', but at that cost, who cares? The design can be adapted to include spectrum analyzer. [120-140 dB ranges]

I'm in the middle of an instrumentation development Project that requires the Network Analyzer part and need to construct, etc. ...so you want to work in collaboration, I'm open. Not a kit, but you'll be able to do all your own calibration and servicing. ;)

It sounds interesting. Does it go all the way from 100Hz and up?

Collaboration could be fun, I often wonder why more DIYs don't go together to make really nice projects, but I guess most person have their own opinion on what they want and are not open to compromise.

Just like government projects, specifying all kinds of requirements instead of just using a proven project than can handle 95% of the requirements.

I am open to a collaboration, but I cannot be sure to offer much time, I have two jobs right now and 2 kids, a dog, a rabbit, a hamster, etc ;-)

Regards

Klaus

th

he

V

cy

he

dB

es

l

r to make really nice projects, but I guess most person have their own opin ion on what they want and are not open to compromise.

ad of just using a proven project than can handle 95% of the requirements.

have two jobs right now and 2 kids, a dog, a rabbit, a hamster, etc ;-)

a bit expensive but a start:

3/

-Lasse

Really nice, thanks for the link

Could be better with more ADC Bits, but could be ok :-)

Regards

Klaus

Capability is down to 10Hz, ...I think. Didn't look as closely there since once into the high end tend to look only at high end.

DIY's NOT sharing? I've always envisioned the image of a student who, while participating in a study group, 'hides' his papers away from fellow students for fear of having something 'copied'. For me, not sharing comes from a fear of embarrassing myself by doing osmething completely stupid.

This project is like making an instrument 'capable' of doing what you want, a solution looking for a problem, you just have to 'live with its quirks' is the cost.

Hmmmm, time may be a limiter. If you've ever written code, you know what I mean. Writing code can quickly give you that vacant stare off into space with the attributes of someone from "The Walking Dead" [the vacant stare, stumbling walk, 'food' streaked down off your face], make you a stranger in your own home, perceived enemy to your pets, ...and, in extreme DIY projects, create a really, really cranky spouse.

Other than that, DIY projects can be fun.

Does your email address accept any type of attachment? for example gmail accounts simply delete any email with a .zip attachment and tells no one. Not even quarrantine the email. so end up renaming as .piz then goes right through

I'm a software engineer and writing code holds no horrors for me, and would be interested in such a collaboration.

But how would such a project differ essentailly from Scotty Sprowls' modular analyser?

GREAT!

Glad did not offend you, no offense intended, I was describing MY personal affliction when writing code.

Don't know, might be the same! need EXACT URL to look, couldn't find anything.