[time-nuts] Lucent KS-24361, HP/Symmetricom Z3809A, Z3810A, Z3811A, Z3812A GPSDO system

Rex rexa at sonic.net
Sat Nov 1 00:32:50 EDT 2014

Here's another reference on driving 10-ish MHz square wave outputs via 
digital chips.

A few years ago I hacked my HP Z3816 to covert its 4 - 19.6608 MHz 
square wave outputs to be 4 more 10 MHz outputs. In the process I 
reverse engineered some of what was there. I found each of these outputs 
came from one 74ACT040 inverter chip per output connector with several 
gates in parallel through 100 ohm resistors to give low impedance drive. 
Maybe all the parallel gates are overkill for most needs, but anyway, in 
the process I drew a schematic of the arrangement that was found there.

You can find the schematic picture, labeled "One of the 19.6608 MHz 
Outputs", near the middle of this page for the whole hacking project:

Bob, I might add parenthetically, that while your responses always seem 
accurate and informative, many times they are presented in such a 
sketchy bullet-point way that only those who already understand what you 
are describing can accurately follow what you are trying to share. Maybe 
it is just my less-than-expert point of view, but I think a lot of your 
posts would benefit if you could give a bit more detail or maybe a link 
to some kind of example or explanation. I appreciate all you offer, it 
takes time to read and reply, but I think often you are preaching to the 
choir when a little more detail could reach the whole congregation. 
Change or not, please keep posting. Even the cryptic stuff contains 
meaning, perhaps a spur to dig deeper.

Bruce, when posting here, used to baffle me too, but he often shared 
links to papers or schematics to aid in following the details of what he 
was describing.

On 10/31/2014 5:02 PM, Bob Camp wrote:
> Hi
> A  $0.15 each dual / quad / hex / octal buffer IC’s will get you > 15 dbm per pair of gates. For under $10 in active parts you can have 30 or 40 outputs. I suspect that if you look inside the 3812 that’s exactly how they are generating the 10 MHz you are looking at.
> Bob
>> On Oct 31, 2014, at 7:01 PM, Graham / KE9H <ke9h.graham at gmail.com> wrote:
>> Bill:
>> On cable TV systems, 50 MHz to 500 (or higher) are the forward channel.
>> (Head-end to client.)
>> Below 30 MHz is the reverse channel, for data going from the client to the
>> cable company.
>> The band 30 to 50 is a cross over zone for the band splitting filters.
>> It is designed to not amplify the forward direction below 50 MHz.
>> --- Graham
>> ==
>> On Fri, Oct 31, 2014 at 4:07 PM, Bill Riches <bill.riches at verizon.net>
>> wrote:
>>> Hi Bob,
>>> I connected 10 MHZ test jack output to a 15 db el-cheepo CATV amp and the
>>> output of that to a 4 way splitter.  Splitter outputs went to 3336-3586 and
>>> counter.  All seem to like the ref signal.  Output of the amp takes makes
>>> the semi-square wave into a sick saw tooth.  Amp is only rated from 50 to
>>> 500 mhz so strange things are happening with 10 MHZ input.  Other CATV amps
>>> do have better low fx response - will play with that later.
>>> I have 10 mhz pulse from Lucent into trigger input of 465 scope and
>>> Thunderbolt gps output into vertical input of scope.  Time base is set for
>>> .01 Usec per div. I notice that trace moves right to left then left to
>>> right about every 5 min or so.  Moves about 3 div before changing
>>> directions.  Why?  Is the Lucent still making a list?  It has only been on
>>> for a few hours.  It takes about 10 min for GPS to go out from a cold
>>> start. (My Thunderbolt and RB do not change direction when using one as
>>> trigger and the other for vert input to scope)
>>> I ordered a USB to RS422 converter cable - will be here next week.  What
>>> program is sort of working?  Using Windoze 7 64bit and have an old XP
>>> machine available.
>>> Sure do appreciate all the info from our time nuts gurus!
>>> 73,
>>> Bill, WA2DVU
>>> Cape May, NJ

More information about the time-nuts mailing list