Practical Analyzer series

Latest count is around 50 million on food stamps... that's about 15% of the population :-( ...Jim Thompson

TUNER':

MHz

MHz

Hz

dBc

Yikes! That's horrid. This is how a good phase noise spec looks like:

Quote "On prototypes built for 1136 MHz and 1278 MHz, phase noise at

dBc/Hz

dBc/Hz

Yeah, that I could not use this thing for my work :-)

But as I said, for hobby or playing around it's probably quite good, considering the price tag.

It's far more than 15% because typically only one parent can file yet the whole family is on the dole. We can see an even worse trend at our church. The number of people lining up for our food closet has AFAIR more than doubled. These are folks from outside our congregation, people from the wider community. Many were middle class until a few years ago. So we regularly ask congregation members to pitch in more. Which is never a problem and so far our church always came through, but this does not bode well for our country.

TUNER':

MHz

MHz

Hz

dBc

dBc/Hz

dBc/Hz

Yup. My HP8568's phase noise is about -110 dBc at 1 kHz (500 MHz f_RF), which is very roughly 60 dB lower than yours.

Cheers

Phil Hobbs

More like 80dB lower :-)

On a sunny day (Sun, 24 Feb 2013 12:46:46 -0500) it happened Phil Hobbs wrote in :

TUNER':

max

dBc

dBc/Hz

dBc/Hz

Yes, that is much better indeed. What I was thinking is that if you can take say 2000 samples at a sample rate of

On a sunny day (Sun, 24 Feb 2013 09:29:41 -0800) it happened Joerg wrote in :

MHz

dBc

dBc/Hz

dBc/Hz

Hey, the -26 dBc/Hz is over the full width (1 kHz to 8 MHz), at 10kHz it is -80, and at 2 MHz -125. Where (in what frequency range) exactly is the noise and what form has it (gaussian?)? I think that is sort of important.

full speed on USB2

the size of a stamp.

Make it decode 'Clear QAM'.

An RF generator with a stepped sweep is just as usefull.

Maybe, but the economy went down the drain as well. So it is only logical more people are in need.

For tracking applications the only important things are how narrow your IF filter is and how quickly you want the plot. At least when measuring stable, linear networks, source PN is not critical.

In fact, spectrum analyzer users who have a max-hold button but no tracking generator sometimes use sweep generators or even noise diodes as the 'tracking' source. It's painfully slow -- brutally slow if you want 1 Hz resolution -- but it can be done.

Phase noise is normally treated as stationary for measurement purposes, for just that reason. The problem with high LO phase noise in an instrument like the Signal Hound lies in applications that have to run in real time. You can let your phase noise analyzer or your network analyzer run all night if necessary, not so much your QAM demodulator.

Still, the cheap LO synth chips are getting pretty good. The Signal Hound uses the ADF4350 IIRC, which is similar if not identical to what Tek uses in their gold-plated MDO4000 scopes. You can buy much worse spectrum analyzers than the Signal Hound, and spend a lot more money in the process.

-- john, KE5FX

's a

My guess is YTOs are dead within 5 years. The integrated VCOs are just getting too good.

Two things worth keeping in mind: 1) The mass-market Elonics chip is likely quite a bit noisier than the $10 part used in the Signal Hound (and MDO4000); and 2) a direct-conversion SDR with a baseband IF has an LO noise advantage over a traditional YIG-based spectrum analyzer that has to run its first LO at 2+ GHz when measuring at HF.

The noise advantage is significant for HF/VHF measurement but of course becomes less of a factor in the microwave range.

Actually the Elonics data is using a less-common metric for integrated jitter in that data sheet, namely SSB carrier/noise ratio as would be seen through a filter of the specified width. Usually you see the equivalent figures in ps or fs, and with different limits of integration (12 kHz - 20 MHz being common in telecom.)

The HP 8568A/B is quieter, but its comparable figure is probably more like -55 dBc (not dBc/Hz!):

I didn't have a plot out to 8 MHz handy, but mine shows -59.5 dBc between 1 kHz and 1 MHz. Most of the 8568's noise is going to be in the sub-100 kHz area anyway, so I wouldn't expect the 1 kHz-8 MHz performance to be very different.

For measurement of HF / VHF sources, the Signal Hound and HP 8568 will have fairly similar broadband PN floors. The flicker region, I'm not sure about.

-- john, KE5FX

That gets old really fast when you have to do wide sweeps at a few Hertz RBW.

Much of this was clearly avoidable. Some of the new laws were plain stupid. For example, the new medical device tax has killed tons of jobs out here. Predictably so. That's just one example out of many. It will have much more far-reaching consequences then most people see.

The economy is being held back *intentionally*. People are in *want*.

TUNER':

max

dBc

dBc/Hz

dBc/Hz

of 2,000,000/s (this device),

short time (1ms).

FFT,

like using a flash

the sidebands will come out OK.

absolute frequency,

fewer samples.

total spectrum,

noise.

Nope, because phase noise is not statistically stationary. Averaging it over long periods makes it worse.

Cheers

Phil Hobbs

MHz BW full speed on USB2

Nonsense. The measured response is convolved with the oscillator spectrum. A crappy oscillator means a crappy measurement.

Not so. Phase noise is Brownian in character at best. Google 'Leeson phase noise".

Sure. I can buy crap at any price I like. Or not. Fortunately there are still a lot of great boat anchors out there.

Cheers

Phil Hobbs

's a

TRIAL RF TUNER':

I just fired up my old HP 8555A/8552A a few days ago, pulled out of mothballs. Arcing by the CRT!! Arrgh!

(I don't use it very often these days.)

the absolute frequency,

nd fewer samples.

the total spectrum,

phase noise.

Not the case. See (e.g.) Rubiola.

In principle, if you had two Signal Hounds running on a common clock, you could use them to measure phase noise by plotting the average of the real portion of their cross spectrum.

-- john, KE5FX

TUNER':

max

dBc

dBc/Hz

dBc/Hz

Was it parked in a moist location? Chances are that there is a hefty accumulation of dust which became moist there. Or maybe some spider thought it was a nice shelter and now you've fried him.

nd 1 MHz BW full speed on USB2

he

ar

on

ere

Yep, I sure as hell didn't sell my HP 8568A when I bought my Signal Hound. :)

-- john, KE5FX