[DSP-10] 1296 RSU -- more lessons learned

[Bob Larkin]

Hi Courtney,

I have been away for a while, and came home to your interesting discussion 
of the 1296 RSU experiments.  I was not tuned into Paul Wade's project, and 
so went to 
http://www.w1ghz.org/MBT/1296MHz_Transverter-Right_Side_Up.pdf  to catch up.

I have a couple of thoughts.  The most obvious is the possibility of a 
simple Yagi made from hobby store parts to pick up a bunch of dB on the 
antenna end.  6 or 7 elements is not too sharp in azimuth, but works on the 
cheap part of the gain curve.

On the box mounting, you ask really interesting questions!!  In a previous 
life, I dealt with the idea of RF stuff on boards, that end up on standoff 
posts.  The problem of stray coupling between such boards is 
surprisingly  difficult to suppress.  The source of all this is this thing 
we call "ground."  Each of the boards connector spots is at its own version 
of ground.  When there are more than two of these connector spots, there 
will be mutual coupling on the shields of the connecting cables, hooked to 
some other "ground."

The simplest example does not occur on the 1296 RSU, but is an amplifier 
board  on standoffs with cables connected at the input and output ends of 
the amplifier.  Normally these cables will go from the board to a coaxial 
connector on the ends of a metal box (that holds the board).  The input RF 
current flows to the board on the center conductor of the input 
cable.  Most of the input current returns on the shield of  the input 
cable.  But not all.  A small amount will return on the output cable 
shield, which is sort of in parallel with the input cable shield.   And the 
same story for the output currents.  This provides coupling between the 
input and the output.  If the amplifier gain is 10 dB, this is a non-issue, 
usually.  If the gain is 30 dB, it might change the response of the 
amplifier, putting in ripples, and so forth.  If the gain is 50 dB, it 
might oscillate.

If this was a  1296 RSU, these same effects can cause any of the signals, 
such as  the LO, to appear on the wrong side of a filter,  or amplifier, etc.

Solutions are several.  Putting each board in a separate milled box with 
the ground only cut out below the boards where necessary, and each 
connector mounted to a wall next to the board will do it.  That, of course, 
misses the whole point that we want to build on the kitchen table.

Even putting the LO into a separate box, with the output connector attached 
solidly to the metal wall can be worth the effort.  Likewise, maybe the 
1296 transverter board could have the transmit and receive connectors done 
this way.  The I-F cable is usually less of a problem, as is any lower 
frequency, since the mounting post inductance, along with the box  becomes 
a short across the cable shields.

But this separate box stuff is still unnecessary for something like the 
1296 RSU.  One large Bud or LMB type of chassis could be used along with a 
couple of tricks.  First, mount board so that the 1296 RF connectors are at 
the wall, and the wall is screwed to the connectors.  Next,  all the inside 
cables can have some  form of ferrite beads placed on them that will 
increase the mutual inductance of the cable shields.  This can be multiple 
small beads, like was in the DSP-10, or bigger torroids  with multiple 
turns around them.  Increasing this inductance decreases the current 
through the mutual inductance, and can be surprisingly effective.  Beads on 
the power wires are cheap insurance, as well, as they also provide common 
paths for the signals.  Finally, try to route wires and cables as close to 
the metal walls as possible.  This reduces the tendency for coupling loops 
to be formed by cable shields and wires.

For accurate layout, somebody probably has thoughts to offer, but I like 
making large marks with a Sharpie pen (blue would be traditional), and 
laying out the holes by scribing on the pen ink areas.  Then use a enter 
punch, like the General "automatic" punches ( 
http://www.generaltools.com/Products/Automatic-Center-Punch__77.aspx ), to 
make a starting point for the drill bit.  General also has good rules and 
scribes.

Keep the story coming, Courtney.  It will be fun to hear of the pile of QSO's!

73,  Bob   W7PUA

At 03:29 PM 9/19/2009 -0700, Courtney Duncan wrote:
>I monitored all week without hearing anything, or seeing anything that
>turned out to be real in the spectrum display.

....snip....

>Anybody have any
>hints about how to:
>
>- choose a box for something like this
>- position boards in it
>- wiring harness
>- place connectors
>- ?
>
>so that NOTHING interferes mechanically with anything else, or creates
>an electrical hazard?
>
>73 and thanks,  Courtney, n5bf/6
>
>
>"It's science. They don't know anything; they just make a lot of
>educated guesses."  -- Viannah Duncan