[time-nuts] Using commercial video amplifier for 10MHz clock distribution.

[Charles Steinmetz]

Luciano wrote:

>     The Extron is a very nice solution, certainly of modest price, but we must consider   *   *   *   the low separation between the output channels.

Well, there is a distinct difference between DAs that use a single 
amplifier (discrete or integrated) to drive each output channel (as your 
design does), and DAs that drive multiple outputs from each amplifier 
using build-out resistors.  In the case of the Extron integrated circuit 
DAs using CLS409 opamps, the isolation is 30dB between outputs fed from 
one amplifier, and >90dB between outputs fed from different amplifiers 
(see my document detailing recommended modifications to the Extron DAs, 
downloadable from):

<http://www.ko4bb.com/getsimple/index.php?id=download&file=02_GPS_Timing/Extron_ADA_6_modifications_for_use_as_10MHz_distribution_amp_STEINMETZ.pdf>

So, simply by using only some of the outputs of an Extron IC-based DA, 
one gets isolation roughly equal to your design.  Of course, one could 
easily build a one-output-per-amplifier DA using the Extron circuit 
(with LMH6702 or other suitable opamps) to have a DA with ~100 dB of 
isolation between all outputs (the exact isolation will be dependent 
upon the layout, just as it is with your design).

One advantage of the IC design is considerably lower distortion -- the 
harmonics are all below -65dBc at 10MHz, +16dBm output, compared to 
-45dBc at +13dBm for your design (per your table).  Further, the 
distortion products of the IC design are essentially all odd-order 
harmonics, while your design shows predominant even-order harmonics.

Odd-order harmonics are symmetrical, so they do not generate additional 
amplitude-to-phase modulation when the output feeds a squarer/zero-cross 
detector.  Even-order harmonics, conversely, *DO* generate additional 
amplitude-to-phase modulation when the output feeds a squarer/zero-cross 
detector. This is very important because the vast majority of 
measurement equipment likely to be fed by a DA/iso amp feeds its inputs 
directly to a ZCD.  See reference in next paragraph.

We have, in the past, disagreed about the importance of low distortion 
in DA/iso amplifiers, but I have satisfied myself that a timing DA/iso 
amp should have, at a bare minimum, -65dBc harmonic suppression.  For 
just one reason, see the NIST publication, "The Effect of Harmonic 
Distortion on Phase errors in Frequency Distribution and Synthesis" 
(Walls and Ascarrunz), available at:

<http://www.ko4bb.com/manuals/73.134.121.2/Harmonic_distortion_and_phase_errors_in_frequency_distribution_and_synthesis_NIST_1995.pdf>

Best regards,

Charles