cheapish spectrum analyzer

these people use to have 12G unit, I don't know what happen to that one however, they do have a USB 1hz..4.4Ghz model for under $1k

In the audio range, there are many tools That run on a pc and use the audio card For input.

Mark

R&S handheld ones look pretty nice, not sure I'd call them "cheapish" though.

--sp

Best regards, Spehro Pefhany

Re-reading that, it looks like you might be happy enough with audio almost range- we have a little portable box about the size of a breadbox that has a PCI backplane and an NI 4-channel 24-bit ADC board as well as some other stuff. It runs XP and Labview (not great) but usually we just transfer the data to another computer and use Matlab or Scilab.

Not as convenient as a purpose-built instrument, but pretty good up to tens of kHz (max sample rate is 100kHz IIRC). The very-low-frequency performance sucks though- 1/f corner is around 1Hz and drift is tens of uV/K (set at +/-10V input scale).

Of course if you could use a sound card it would almost be free.

--sp

Best regards, Spehro Pefhany

I was just trying to measure the -3 dB point of a bandlimited noise generator, expected to be around 10 KHz. An oscilloscope FFT is, well, too noisy to resolve the corner frequency. I'd need something that can sweep really slow, with a low resolution bandwidth.

--

John Larkin Highland Technology, Inc picosecond timing precision measurement

jlarkin att highlandtechnology dott com

If you don't have to do it very often just start averaging the fft power spectra. In my limited experience 100 or 1000 should be plenty to give a decent shape. Just let it average until it's good enough. Just don't average the time domains :-).

----- Regards, Carl Ijames carl.ijames aat deletethis verizon dott net

Got any Agilent DSO/MSO6000 models around? They released a firmware patch a while back that enables FFTs on their entire acquisition record, or at le ast a much larger chunk of it. It could probably resolve down to at least a kHz or so at baseband.

I have plenty of real (no pun intended) spectrum analyzers around here, and I still find myself using my MSO6054A's FFT function all the time. It's t he only scope I've ever used whose FFT doesn't utterly suck. There are int erleaving spurs, of course, but they don't really get in the way...

-- john, KE5FX

Sounds like you need a "dynamic signal analyser".

I have an R9211C it is big and clunky and slow but very useful and I have grown to love the thing. It goes down to millihertz.

I keep thinking about making a modern version but who has the time.

In my home lab I have a Picoscope 20Msa 12bit scope I use for pre compliance conducted emission measurements. It can update the up to 30MHz band many times per second, so time relared effects can be investigated

I just ordered the Picoscope 5444B, which has 16bit 60Msa, so that is even better. It has AWG generator so you can do frequency sweeps of filters etc also. Costs less than 1800 usd

Cheers

Klaus

Do you need the 3 GHz? I'd really like you to buy the Rigol DSA815-TG, 9kHz - 1.5 GHz.

Then you can report on it's performance. :^)

George H.

Everything you ever did or did not want to know about the Rigol, with some biased and unbiased reviews:

Steve

They are really low priced in general. Many years ago I bought a Agilent funktion generator. The price of that instrument could have payed for a complete Rigol instrument park.....

Cheers

Klaus

As we've discussed here before, the SDR-type spectrum analyzers have about 3 orders of magnitude worse close-in phase noise than a real YIG-tuned boat anchor. If you have the space, you can get an HP 8566B in good condition for a grand or so.

Cheers

Phil Hobbs

The Rigols look nice. Their scopes are excellent.

But what's with 9 KHz? Seems like most "RF" spectrum analyzers quit below 9 KHz. They are obviously giving up market by cutting off the audio range.

Say Phil, (or anyone) is the HP8568B basically the same instrument? (Just a bit less on the upper range... 1.5GHz.)

They look cheaper on ebay. (Ughh, deleting my long link ~$1k +shipping)

George H.

We've got an SRS box that does mHz to 100k Hz, so the Rigol would be fine.

For the low freq stuff, if you don't need great accuracy (~10% or so) you could make a variable HP filter or maybe BP, and do the numbers.

(Time or money...)

How about a lockin as spectrum analyzer (single point at a time)... I'd have to get some math help to get the band width right.

George H.

I wonder why. Seems to me like superheting down a slice of spectrum, digitizing it, and doing an FFT for the fine detail, would work fine. The only requirement would be low phase noise on the fixed-frequency LO, which would seem, to me, to be easier than getting low phase noise on a swept LO.

I'm a big fan of the HP 35665A (I paid $300 for mine about 5 years ago). You can get an EPROM with all the options turned on, which is nice. The HP 3562A is another nice one.

Cheers

Phil Hobbs

A lockin needs a phase reference! With a fixed-frequency reference, it becomes a mixer. Its AC output could be lowpass filtered and slowly measured to estimate the noise density near the ref frequency.

I do have a really old HP selective voltmeter box. I wonder if it still works.

Yes, the 8568B is basically the low-frequency half of the 8566B. (I have one of those too.)

One nice thing about the 8568 is that it'll stand some DC on its input, which the 8566B won't.

Cheers

Phil Hobbs