[time-nuts] Wavecrest noise floor
SAIDJACK at aol.com
SAIDJACK at aol.com
Tue May 1 05:03:55 EDT 2007
In a message dated 4/24/2007 11:49:41 Pacific Daylight Time,
henk at deriesp.demon.nl writes:
>Hi Said,
>Thanks for the picture.
>I have inserted a 100MHz lowpass between the PRS10 and the Wavecrest
>tot remove the spurious signals.
Hi Henk,
slightly off-topic - I got my Wavecrest DTS-2075C working now :) Turns out
it was my old ISA GPIB card, Visi just didn't like the Agilent Driver. I got a
PCI NI-488, and it's working great now :) Thanks for the help everyone!
I attached a plot of the noise-floor of the Wavecrest, it's quite amazing.
To do the measurement, I feed a 10MHz CMOS square wave from my Fury GPSDO into
a MiniCircuits resistive divider, and then have one short cable to Chan1 and
one long cable to Chan2. I measure the time interval between both channels.
Since it's the same signal, chan2 just being delayed, I get to read the noise
of the instrument. The cable delay is about 7.142ns.
The plot is slowed down to 1 update per second, 25 measurements per second.
The unit can do significantly more (up to 40Ks/s of course). Magenta is max,
red min, and blue average reading of the set of 25 measurements per second.
Increasing to 100 sample bins per second doesn't really reduce the noise much.
Pretty impressive, the overall noise is 28ps peak to peak over 300s
measurement intervall, the result of averaging the 25 sample-bins every second is
about 7ps peak to peak per second.
That's pretty much in line with the 2.7ps RMS jitter that I measure on a
very good oscillator with this unit...
bye,
Said
************************************** See what's free at http://www.aol.com.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DTS2075_NoiseFloor.jpg
Type: image/jpeg
Size: 80535 bytes
Desc: not available
Url : http://www.febo.com/pipermail/time-nuts/attachments/20070501/c6a27974/attachment-0001.jpg
More information about the time-nuts
mailing list