►
Description
AMSAT-DL Online Satellite Symposium on 26th September 2020
The ADALM Pluto as part of the AMSAT-DL LEILA QO-100 control station in Bochum by Mario DL5MLO.
A
The
topic
of
mario
of
lawrence
delgalima
5
mike
lima
oscar,
is
the
adult
pluto
based
ground
station
in
bochum.
So
the
control
station
in
bochum
is
getting
converted
by
mario
to
pluto.
Mario
is
a
long-term
member
of
amsterdl,
he's
an
expert
in
digital
signal
processing,
processing
for
dsps
fpgas
and
he's
also
very
interested
in
deep
space.
A
A
So
he
set
up
the
the
system
and
the
receiver,
and
this
project
is
still
ongoing
in
both.
So
we
are
receiving
still
the
sickness
from
the
nasa
stereo,
satellite
and
stream
them
to
nasa
in
the
us,
mario
implemented
the
beacons
and
the
lila
functions
for
the
control
station
of
q100
and
with
the
expansion
of
the
narrowband
transponder
to
500
kilohertz.
He
changed
the
hardware
and
associated
software
from
red
pitaya
to
adam
pluto,
and
that's
what
his
talk
will
focus
on
today.
A
B
B
Many
of
the
people
that
are
regular
attendance
of
the
symposium
in
bochum
will
probably
have
seen
me
before
presenting
one
or
the
other
topic
here
with
amsat
others.
Now
here
on,
the
satellite
and
worldwide
live
stream
will
probably
see
me,
maybe
the
first
time
I'm
not
very
much
active
on
the
satellite
I'm
listening
from
time
to
time,
but
it's
always
a
matter
of
let's
say
time
constraints,
especially
most
of
the
time
I'm
playing
with
the
hardware
and
getting
like
lila
improved
and
so
on.
B
B
B
It's
a
it's
a
feature
that
has
been
flown
on
well
ao40.
Of
course,
and
it's
a
it's
got
several
functions
and
the
functions
here
is
mainly
to
help
the
users
to
operate
the
satellite
in
various
ways
and
these
various
ways
are
like
finding
it
with
the
beacon.
This
was
not
a
function
back
then,
but
here
it's
mainly
get
the
power
levels
right.
B
Lila
2,
unfortunately,
could
not
be
implemented
on
the
satellite
itself.
Of
course,
we
would
have
loved
to
have
it
directly
at
the
satellite.
This
would
have
made
some
things
easier.
On
the
other
hand,
the
change
that
we
just
did
like
changing
the
hardware
completely
would
have
been
completely
out
of
the
question
where
lila
directly
on
the
spacecraft.
B
Stuff
that
is
installed
in
qatar
and
in
bochum.
We
mostly
run
the
amsterdam
installation
in
bochum
just
for
practical
reasons,
but
we
would
have
the
other
backups.
Unfortunately,
these
backups
are
still
in
the
hardware
that
you
see
here.
This
is
the
original
leila
two
hardware
and
the
objective
of
leila
ii
is
like
to
generate
the
beacons
to
mark
the
satellite
position
so
that
you
can
point
at
it.
B
What
you,
mostly
when
you
mostly
interact
with
lila,
is
the
warning
tone
that
you
hear
if
you
are
transmitting
too
loud.
So
if
you
hear
that
tone,
please
just
reduce
your
power
a
little
bit
and
then
lila
will
be
quiet
and
will
not
interfere
with
your
qsos.
So
that's
just
a
general
public
servants.
Public
service
announcement
here
leila
was
planned
and
built
for
the
old
band
plan.
B
Things
changed
fast,
so
maybe
some
people
remember,
maybe
others.
Don't
we
had
this
band
plan
here
with
a
250
kilohertz
wide
transponder.
This
was
basically
the
plan
from
the
beginning.
I
shouldn't
say
of
time,
but
from
the
beginning
of
the
project
that
we
would
have
a
250
kilohertz
transponder.
I
was
told
that
yeah.
Well,
maybe
the
filters
are
not
entirely
flat,
so
maybe
a
little
bit
wider.
B
So
when
I
plant
lila,
I
plant
for
350
kilo
cycles
with
a
which
is
a
digital
system,
which
I,
which
you
could
then
run
at
384
kilohertz
sample
rate,
and
we
used
low-pass
filters
in
the
in
the
iq
modulator
that
had
a
cutoff
rate
of
one
megahertz
and
that's
something
for
this.
You'll
have
to
you'll
have
to
see
how
the
lidar
two
hardware
is
structured
for
purposes
of
not
having
to
build
everything
from
scratch.
B
For
example,
the
iaf,
where
we
interface
with
the
scc
station
in
qatar,
is
in
the
middle
of
l
band.
In
the
in
the
bohan
setup,
we
used
the
traditional
70
centimeter
uplink
chain
to
transmit,
and
so
so
everything
was
slightly
different.
But
so
we
had
this
iq
modem
and
the
design
there
basically
was
more
or
less
fixed
for
this
lower-
quite
quite
quite
low
bandwidth
with
some
reserves.
But
we
really
did
not
know
or
did
not
anticipate,
that
the
transponder
would
be
technically
that
wide.
B
Now,
if
you
look
at
this
hardware,
you
see
the
rapid
tire
on
how
the
repetition
is
built
in
where's,
my
somebody
moved
the
mouse
and
now
okay.
So
here's
here's.
What
you
have
inside
of
the
red
pitaya
and
the
red
pitaya
technically,
is
a
xilinx,
a
sync
fpga
system,
with
lots
of
peripherals
around
it,
mainly
the
ad
converters,
to
get
the
analog
to
the
digital
domain
and
the
dac
converters
to
to
go
from
the
digital
to
the
analog
domain.
B
B
B
We
could
have
done
it
in
the
beginning,
but
I
would
I
had
preferred
to
have
the
cpu
idle
by
at
least
50
just
in
case,
and
then
that
resulted
in
me
having
to
say:
okay,
the
we
have
to
move
functions
into
the
fpga,
so
it
was
back
then
already
on
the
old
lila
that
I
did
an
fpga
design
where
I
had
many
of
the
of
the
logic
moved
into
the
fpga.
It
took
me
quite
some
time
it
was
my
first
fpga
design,
I
have
to
admit,
but
yeah.
B
B
All
nice
or
fine
until
I
think
it
was
peter,
came
up
with.
We
would
like
to
change
the
band
plan
to
a
new
plan
and
utilize
all
the
transponder
bandwidth,
that
is
there.
We've
got
indications
from
the
frequency
coordination
stuff
that
this
can
then
be
done,
so
they
had
to
re-notify
and
whatnot
recoordinate
the
frequencies,
because
everything
was
planned
for
the
250
transponder,
oh
yeah,
and
then
the
question
was
now:
can
we
do
it
with
lila?
B
So
first
we
discussed
how
the
new
transponder
vent
plan
should
look
like.
There
are
some,
let's
say,
peculiarities
that
people
may
maybe
do
not
realize.
First
of
all,
we
decided
that
we
should
have
three
beacons
in
the
beginning.
We
had
two
beacons.
We
need
one
on
the
left
or
the
lower
end
and
one
of
the
upper
end,
which
is
obvious
because
we
want
to
show
where
is
the
beginning,
and
where
is
the
end
of
the
plan?
B
Now
for
leila
to
accommodate
this
band
plan
means
I
need
to
generate
the
signals
now
on
a
five
kilohertz,
500
kilohertz
bandwidth,
and
it
also
means
that
I
need
to
monitor
the
500
kilohertz
bandwidth.
All
of
this
happened
on
and,
of
course,
in
quite
short
notice,
as
it
always
happens.
Oh
I've
got
this
idea.
We
could
do
that.
B
How
fast
can
we
do
this,
and
it
turns
out
that
the
hardware
design
of
lila
put
a
put
a
stop
to
that,
mainly
because
we
have
the
id
low
passes,
which
would
have
to
be
extended,
because,
even
though
I
said
that
they
are
one
megahertz
wide,
that
was
only
half
of
the
story.
Unfortunately,
inside
this
half
megahertz
we
had
some-
let's
say
local
qrm,
due
to
a
voltage
converter,
so
that
has
to
be
out
of
the
passband.
B
So
I
could
not
find
the
two
times
500
kilohertz
bandwidth,
that
we
needed
to
receive
and
to
transmit
the
beacons,
and
that's
why
we
said:
okay
we'd
have
to
change
that
filter
which
would
have
been
okay.
But
then
it's
modifying
some
30
components,
analog
components
on
the
pcb
board,
and
nobody
really
wanted
to
do
that
and
then
okay,
we
could
build
or
produce
new
lylas
and
then,
when
we
computed
the
cost
about
that,
we
found
out
that
maybe
it's
cheaper
to
just
use
a
pluto.
B
We
had
a
spare
already
on
hand
and
hat
is
already
installed
in
bohum
for
testing
purposes.
It's
cheap.
It
matches
the
specifications
in
terms
of
frequencies
in
terms
of
bandwidth
and
so
on.
It's
supported
by
a
common
sdr
driver,
so
porting
the
leila
software
to
the
pluto
was
actually
achieved
in
less
than
two
weeks,
but
it
was
two
weeks
in
real
time,
but
I
of
course
only
was
able
to
work
on
it
sometimes
in
the
evening.
B
When
you
look
at
pluto,
then
you
know
that
pluto
has
some
issues.
They
are
very
well
known,
very
well
documented.
By
now.
The
most
important
one
is
that
the
original
built-in
crystal
is
basically
unusable
for
narrowband
work.
It
drifts
or
it
makes
your
transmit
frequency
drift
by
about
a
year
in
this,
in
this
case
about
500
hertz
per
minute.
This
is
slightly
temperature
and
slightly
device
dependent.
B
Now
the
pluto
internally
works
with
a
40
megahertz
reference
oscillator
and
we
did
not
have
a
or
we
we
first
tried.
Okay,
let's
create
the
40
megahertz
out
in
various
different
ways.
Mostly,
we
had
the
idea
of
using
a
cheap
pll.
There
was
like
small
it
ic80
something
these
are
basically
six
pin
chips.
You
connect
power,
you
connect
the
10,
megahertz
and
you're
supposed
to
get
40
megahertz
out.
B
B
But
finally,
I
found
a
way
basically
originally
by
patching
the
pluto
image
directly
to
set
it
to
10
megahertz,
and
we
got
this
to
to
work
and
to
lock
at
10
megahertz
and
the
performance
is
okay,
although
I
must
admit
that
the
phase
noise,
if
you
give
it
a
40,
megahertz
reference
signal,
you
get
still
better
face
noise,
so
we
are
still
trying
to
get
a
cheap,
reasonable.
Let's
say
that
not
cheap!
B
B
So
this
gives
you
a
square
wave
instead
of
a
sine
wave
for
an
lo,
and
that
means
that
you
have
harmonics
of
the
pluto
that
are
suppressed
only
by
some
10
db
for
the
harmonic
and
so
on.
So
that
is
not
really
acceptable,
but
then
again,
if
you
pass
this
through
your
amplifier
and
your
amplifier
filters,
the
higher
frequencies
anyway,
it
may
or
may
not
be
such
such
a
problem.
B
It's
just
a
small
hint.
Please
check
your
output
signal
also
for
harmonics.
If
you
are
capable
of
doing
that.
The
pluto
is
not
that
nice
there
and
that's
also
important
if
you
use
the
pluto
on
different
frequencies
on
different
lows,
if
you
use
it,
for
example,
to
transmit
on
70
centimeters
the
sdr
sam
or
the
sdrs
that
are
based
on
iq
sampling,
they
have
a
an
issue.
That's
called
a
dc
dc,
offset
problem
or
lo
leakage.
That
means
you
have
a
center
frequency,
and
here
you
have
a
small
leak
of
rf
energy.
B
B
If
you
look,
if
you
look
closely
by
the
way
you
see
that
the
beacons
they
don't
have
the
same
level,
the
upper
beacon
is
slightly
transmitting
slightly
more
power
such
that
at
the
transponder.
The
the
downlink
level
of
the
beacons
again
is
the
same,
but
anyway
here
is
here
is
this:
here
is
a
slight
spur.
B
B
B
B
You
find
the
analog
devices,
the
ad9363
radio,
trans
transmitter
and
in
the
middle
you
find
designs.
Xc7
set
10
fpga
again,
which
is
interestingly,
very
much
the
same
as
it
was
on
the
red
pitaya
and
therefore
on
the
original
lila.
So
we
are
basically
dealing
with
old
and
known
friends,
and
that
makes
some
of
the
things
a
little
bit
easier,
especially
what
it
now
makes
possible
is
makes
possible
to
do
some
comparisons
between
the
pluto
and
the
original
lila.
B
If
you,
if
you
look
at
the
fine
print,
even
for
the
analog
devices,
radio
chipset,
you
will
find
that
you
actually
have
a
bandwidth
of
at
least
20.
Megahertz
is
what
they
guarantee
for
the
cheap
version.
The
interesting
question:
how
much
of
of
the
wider
bandwidth
you
could
possibly
use
when
overclocking
it,
because
these
six
gigahertz
they
were
also
not
part
of
the
of
the
spec,
but
anyway,
that
is
way
more
than
we
have
at
the
transponders.
B
The
problem
here
is:
we
cannot
get
this
spectrum
processed
out
or
past
pass
through
out
of
the
pluto
through
the
usb
port,
and
the
solution
to
this
is
obvious.
The
data
processing
should
happen
directly
on
the
pluto,
and
I
believe
that
is
what
analog
devices
what
and
you
have
to
say
also
the
other
people
that
contributed
to
the
pluto
project
like
silence.
B
So
if
you
look,
if
you
look
at
the
processing
chain
and
the
relative
bandwidths,
which
is
what
I
did
here,
if
you
look
at
the
original
spectrum
that
you
have,
which
is
very
wide
you
can
you
have,
you
have
a
filter
and
you
have
a
passband
which
will
take
a
part
of
the
spectrum
and
then,
if
you
pre-process
it
and
even
demodulate
it
or
in
the
transmit
chain,
it
would
be
the
modulator
directly
on
the
pluto.
Then
your
usb
bandwidth
is
ample
to
get
it
to
the
pc
or
anything
else.
B
So
that
is
the
basic
idea.
If
you
look
into
the
pluto
fpga,
which
I
guess
a
few
people
did
so
far,
you
find
other
interesting
things
which
I
have
not
done
or
looked
at
further.
For
example,
they
have
hardware
preparation
here
for
a
gps,
one
pps
signal.
So
I
believe
that
at
some
point
you
can
just
connect
a
one
of
these
cheap
gps
receivers
that
you
get
on
on
ebay,
with
the
1pps
connectors
to
the
available
pins,
and
you
can
run
a
gps
new,
more
or
less
directly
inside
the
pluto.
B
B
B
I
understand
it
partially,
but
not
fully
why
this
is
the
case,
because
the
pluto
has,
in
my
opinion,
much
less
processing
than
what
delilah
had
delilah
has
a
has
a
complete
fft
core
in
the
fpga,
for
example,
to
analyze
the
spectrum
which
the
pluto
does
not
have,
and
the
uplink
and
or
filters
for
for
up
and
down
conversion
and
stuff,
like
that,
I
also
have
in
in
delilah.
So
I
believe
that
the
fpga
utilization
in
the
pluto
also
can
be
optimized
significantly
for
amateur
radio
purposes.
B
B
B
B
The
reason
why
analog
disables,
the
second
core,
which
then
you
sometimes
enable
again,
may
not
only
be
a
limitation
of
the
fpga,
but
also
a
limitation
of
the
available
power.
Remember
in
a
usb
2
device.
You
only
have
500
milliamps
at
five
volts,
which
makes
two
and
a
half
watts
of
total
power,
so
for
all
of
the
the
people
that
run
pluto's,
it
may
be
very
wise
to
think
about
externally
powering
that
pluto
and
also
maybe
adding
a
heatsink
to
the
fpga.
B
If
you
want
to
experiment
a
little
bit
more
with
pluto
integrated
signal
processing,
if
it's
just
going
to
pass
the
data
through,
it's
it's,
it's
simple
and
there
is
nothing
going
on
on
the
pluto.
But
if
you
do
experiments
or
like
we
will
run
the
lila
on
it
on
it,
at
least
in
part,
then
it
is
important
to
have
additional
cooling
and
additional
power
supply
for
lila.
B
So,
to
recap:
this
talk
leila
now
runs
on
the
pluto
implementing
time
to
get
this
running
on,
the
pluto
was
really
a
fraction
of
what
was
spent
compared
to
the
fpga
design.
The
fpga
was
mostly
on
on
or
off
about
one
year
until
I
had
everything
working
back
then
right
now
we
are
punting
the
issue.
It's
just
the
quick
thing
put
the
pluto
plug
it
in
on
a
pc
and
run
the
software.
The
total
hardware
cost
certainly
was
not
higher
this.
The
development
cost
certainly
was
much
lower.
B
So
if
you
can
do
this,
I
think
software-defined
radio
is
still
better
than
fpga
defined
radio,
because
it's
just
that's
it,
but
in
any
case
the
fpga
has
lots
of
potential
lots
of
ways
to
experiment
with
and
if
once
once
you
are
proficient
with
that,
there's
probably
much
more
magic
that
you
can
do
now.
What
comes
next,
it
has
been
mentioned
again
and
and
again
in
the
past.
B
We
did
this
with
an
effort
to
go
on
quickly
and
with
the
idea
to
be
able
to
return
a
month
later
to
install
the
second
backup
system
and
everything
so
that
we
can
have
a
like
proper
system,
but
then
came
corona
and
everything
was
swatted
here
so
right
now
we
do
have
the
lilac
there,
but
we
have
not
yet
touched
it.
We
had
plans
of
meeting
in
pohum
yesterday
to
do
this,
but
again
due
to
corona.
B
This
did
not
happen,
so
we
will
have
to
see
when
we
go
to
boho
next
time
to
get
the
system
up
and
get
get
a
backup
installed,
and
until
that
time
I
don't
want
to
really
touch
lila.
We
have
some
new
plans
for
lila
to
improve
things,
mainly
with
regards
to
the
beacon
the
beacon
used
to
was
quite
boring
in
the
beginning,
and
then,
when
we
switched
to
pluto,
then
even
the
voltages
and
so
on
that
we
had
on
the
old
lila
were
not
available
anymore.
B
So
there
is
not
even
any
good
telemetry
on
the
beacon
at
this
time,
but
we
hope
that
this
will
change
in
the
near
future.
It's
just
a
matter.
We
don't
want
to
mess
with
lila
completely
remotely
if
we
don't
have
a
real
hot
backup
next
to
prevent
any
outages,
so
please
bear
with
us
for
a
little
longer.
B
A
A
B
The
pluto
firmware
utilization
that
I
have
shown
here
is
from
I
just
compiled
or
well
synthesized
the
original
pluto
design.
It
does
not
have
any
specific
additions
there
I
have
seen
in.
I
think
it
was
one
or
two
weeks
ago
on
on
twitter.
There
was
some
discussion
about
this
fpga
usage,
especially
when
evarista.
B
Complained
about
that,
this
utilization
is
now
so
high
since
in
since
analog
added
the
revision
c
processing,
but
then
again
I
I
this
is
part
of
what
what
I
did
not
understand
the
let's.
Let's
go
one
step
back,
maybe
some
of
the
people
here
on
this
feed.
They
do
not
know
that
analog
apparently
is
working
on
a
new
revision
for
the
pluto.
B
It
has
been
discussed
for
quite
some
time,
but
if
you
look
at
the
fpga
code,
you
see
that
they
now
made
some
specific
changes
and
the
specific
changes
that
they
made
is.
The
pluto
has
more
than
one
rf
interface.
They
have
a
complete
second
set
of
channels
in
there
and
the
request
was
that
these
be
pinned
out
right
now.
They
are
grounded,
unfortunately,
grounded
directly
at
the
chip.
B
So
it's
also
not
easy
easily
possible
to
just
scratch
off
a
little
a
little
bit
of
copper
and
contact
some
wire
really
unfortunate,
but
yeah,
that's
it
and
so
the
the
new
revision.
Apparently
I've
never
seen
it.
I
don't.
I
don't
know
anything.
I
have
no
inside
information
on
that
plans
to
make
these
additional
rf
ports
accessible.
B
So
the
firmware
now
has
the
support
for
this
additional
rf
channel.
However,
in
the
fpga
there
is
basically
a
sample
rate
conversion
for
this
second
channel,
and
this
is
not
there
for
the
second
channel,
so
the
amount
of
resources,
especially
dsp
resources
that
I
could
find
that
are
spent
on
the
new
on
the
new
design
for
the
revision
c.
They
are
not.
They
are
not
that
much
and
these
90
percent.
I
really
there's
lots
of
that
that
goes
into
this
decimation
chain.
B
So
I'm
I'm
really
wondering
if
maybe
re-implementing
that
this
that
this
decimation
chain
will
free
up
lots
of
resources.
I
I
have
started
to
look
into
this.
That's
probably
one
of
the
the
next
things
that
I
look
and
yeah
we
can.
We
can
just
discuss
that
offline
or
yeah.
Just
just
just
drop
me
an
email,
let's
say
any
more
questions.