[amsat-bb] QO-100 bare minimum: GPS referenced LNB + IC-402 (read: no dish, 70cm IF!)

Fri Mar 1 09:28:50 UTC 2019

On 1.3.2019 6:19, Dave Hartzell wrote:
> Very cool!  
> 
> Can you elaborate a bit more on the signal chain that you have built?

Hi Dave,

a very short description: the lnb is modified for external frequency
referenencing and at the radio end I have a simple high pass filter
for the IF going to the IC-402 (or other), a low pass filter for DC
supply to the LNB and a simple LC series circuit to inject the reference
frequency (26,5 - 28 MHz is OK for reliable locking) up the IF cable.

More detail:

I am using the Octagon OSLO Single LNB OPTIMA which is a PLL LNB with
an original reference crystal of 27 MHz (poor quality, no temperature
control or even compensation). I have made an extremely simple mod to
this by removing the crystal and feeding an external reference
frequency. Most hams seek dual or more output models or drill an
extra hole for connector in order to apply this reference, but after
some experimentation I finally found a stable and reliable way of
extracting the reference frequency from the "L band IF" connector:

http://fletcher.fi/27MHz_BPF_Octagon_Oslo_sized.jpg

The series LC circuit within the LNB MUST be right next to the L band
connector for stability. Also care should be taken in avoiding any
shorts (Polyimide/ Kapton tape etc.). Also, the cover and all screws
MUST be torqued down as the PCB is in compression contact, sandwiched
between the bottom and cover, hence needing excellent 360 degree
microwave grounding all around.

With some reliability, but still risky, you can identify this particular
model of LNB by the "3D"-text on the box. Beware, there is zillions of
versions on the huge market of PLL LNB's and not all are that easy
to modify or perform well at 10.4 GHz.

Here is a quick comparison test I made, using an external, radiative
noise source, flatness maybe +- 1dB or so:

http://fletcher.fi/modified_lo_octagon_oslo_single_lnb_optima_oh2aue_27_02_2019_sized_annotated.png

So it is quite amazing the 144 MHz IF is also otherwise fine, but the
PLL VCO is REALLY at it's limit of reliable locking (see the other video
on my Youtube channel "oh2aue". At least it just about locked at a few
degrees C below zero ;-) And obviously the gain is also way down
too, maybe 10 dB from my 200 MHz highpass filter in the triplexer.
However, the noise figure has not yet appreciable deteriorated.

Some similar single ouput models others have experimented with are:

"Universal Single-LNB GOLDEN-MEDIA GM-201, 0,1 dB "
"Golden Interstar Universal Single LNB GI-201 Platinum X"
"Goobay 67321"

Typically I use around 40 dB of attenuation in the path in order to
set the receiver AGC knee at a suitable point for pleasant copy. My
attenuators have a T choke affair to bypass the LNB DC supply. There
are also cheap commercial attenuators available, but the largest
values are usually only 12 dB. Then again, homebrewing a 40 dB
attenuator good to beyond 1 GHz is not very trivial either.

There are many models out there that have differing external appearance,
but contain the same electronics and vice versa: there are also models
that have the same external appearance but different electronics. To
confuse things further, the text "PLL LNB" does not imply that that the
electronics are PLL based. And vice versa, no mention of "PLL" does not
mean the LNB is a DRO LO type, rather it can well be synthesized and
crystal controlled.

In the video I am using the Leo Bodnar Dual Output GPS Disciplined
Frequency Reference. Obviously the supplied reference signal has to
be very stable and low phase noise as in e.g. the Octagon LNB above,
the fractional synthesis effectively multiplies by 361.1111111...
multiplying any residual phase noise and jitter by the same order.

At the receiver/radio end I have a very simple triplexing circuit
for LNB DC bias (and 22 kHz/DiSEqC as the case may be), highpassing
the IF from about 200 MHz up plus the 27 MHz series circuit. A lowpass
is also recommended if your reference signal is not pure sine. I only
have sketchbook drawings, but here is an estetically pleasing photo of
OH3FWG's own implementation of such:

http://oh3tr.fi/~oh3fwg/p4a/triplexer.jpg

For my personal experimentation I have also put together a "laboratory
grade" receive system (flat group delay, flat response, impedance
matching for each stage, etc.). This receiver frequency selectively
mixes down the spectra of both transponders to HF (10 - 21MHz) and
there is a video of this receiver feeding my Barlow Wadley XCR-30,
merely because I can :-)

http://fletcher.fi/kiwisdr_eshail2_receiver_29_12_2018.jpg