[time-nuts] What equipment can I economically use to measure jitter and noise sidebands?

[Magnus Danielson]

From: Peter Vince <pvince at theiet.org>
Subject: Re: [time-nuts] What equipment can I economically use to measure jitter and noise sidebands?
Date: Sun, 15 Jul 2007 02:12:35 -0700 (PDT)
Message-ID: <600116.75422.qm at web26605.mail.ukl.yahoo.com>

> ); SAEximRunCond expanded to false
> Errors-To: time-nuts-bounces+magnus=rubidium.dyndns.org+magnus=rubidium.dyndns.org at febo.com
> 
> Thanks for your reply, John.
> 
> >Too many dimensions to that question. :)
> >What carrier frequencies...
> 
> 3.57 MHz, 4.43 MHz, 5.0 MHz, 10.0 MHz, 135 MHz, and maybe 742 MHz.  Yes,
> you guessed, those used in television.

You would probably like 6.75 MHz, 13.5 MHz, 27 MHz, 270 MHz and if god is good
also 1.4835 GHz and 1.485 GHz.

There's 180/360 MHz and 540 MHz but we don't see much of those links and
2.97 GHz is comming up.

Actually those are half and full frequencies of the bit clocks for 270 Mb/s,
360 Mb/s, 1.4835 Gb/s, 1.485 Gb/s and 2.97 Gb/s links.

Yes, you guess it, I fiddle around with television stuff too. :-)

> >If your work is limited to DC-HF measurement then you certainly can't
> >beat the Wavecrest units people have been talking about...
> 
> >I don't know anything about measurements on modulated signals; most of
> >the time these measurements are made on CW carriers or perhaps on pulsed
> >ones (see HP's AN 386).  Maybe the Wavecrest software helps with this?
> >Since the BER of a digitally-modulated signal depends on carrier phase
> >noise and phase hits among other things, I would guess that running your
> >link's existing BER diagnostics under controlled conditions would be the
> >easiest way to observe noise.
> 
> The 135 MHz signal has simple cycle-by-cycle NRZ on-off modulation. 
> While we have a BER test, we don't have a way of measuring the jitter
> other than eyeballing it!

What's your actual signals? It sounds like you should look at a Tek VM700A
(with the SDI option) or VM700T. It does jitter and wander measurements on your
SDI signal. It supports your analog signals but since I rarely sees them in
the wild I haven't checked what I can measure. For SDI, VM700 is THE choice.
The VM700 has a flaw in its design, but it can be overcome with not that much
trickery. I think I was down to two chips or so...

Wavecrest's need a litte help to trigger up on a known point in the stream.
So some form of signal synchronous to it needs to be generated, and only then
the software can do anything. 

Another option is to use a good scope. I've measured jitter down to a few ps
RMS that way. The scope solution assumes non-modulated signal or that trigger
signal.

Cheers,
Magnus