►
From YouTube: ORI FPGA Standup 24 May 2022
Description
Open source RF model for advanced digital waveforms, cool music, and more.
Want to join? https://openresearch.institute and click Getting Started
A
A
What
we
do
is
talk
about
what
we've
done
over
the
past
little
bit
and
what
we
are
going
to
be
doing
over
the
next
little
bit.
What
we
pick
for
the
little
bit
as
a
week
if
we
have
any
roadblocks
that
are
in
our
way
or
if
we
need
any
resources
in
order
to
do
the
work,
and
we
have
a
lot
of
different
things
going
on
represented
by
this
group.
A
So
I
will
start
the
past
week
has
been
chaos
and
lots
of
different
things.
The
the
things
most
appropriate
are
some
some
efforts
in
getting
additional
human
resources
for
fpga
design.
That's
that's
been
successful.
I'm
looking
for
names
to
show
up
today,
but
I
don't
see
any
yet,
but
we
will
have
some
fpga
and
asic
designers
that
will
be
joining
to.
B
A
Out
getting
the
encoder
to
work
for
the
demo
for
august
for
defcon,
so
getting
the
encoder
that
swattos
and
and
everest
have
worked
on,
and
other
people
have
worked
on
too
to
getting
that
working
of
the
air
and
verified
and
validated
so
so
progress
there.
Looking
for
more
help,
so
you'll
see
it
on
slack
and
in
email.
A
A
So
the
usual
complaints
are
that
I
don't
have
enough
time
and
just
like
everybody
else,
and
the
roadblocks
are
non-technical.
Really
the
hard
part
is
never
the
technical
part
for
what
we
do
all
right.
So
next
on
the
list
is
onshul
and
then,
when
you're
done
with
your
report,
pass
it
on
to
the
next
person
that
you
see
and
so
on.
B
Sure
so
for
me,
the
progress
is
that
now
I
can
boot
my
c706.
So
I
resolve
the
kernel
issue,
which
I
described
last
time,
so
I
can
boot
it
now.
I
am
stuck
how
to
test
the
complete
dvds
too.
Maybe
it's
because
of
my
lack
of
experience
there,
so
I
have
set
up
a
time
tomorrow
with
everest
and
we
should
be
able
to
find
a
way
forward
and
I
should
be
able
to
test
it.
I
have
everything
on
the
board
bit
stream.
Is
there
kernel?
B
Is
there
everything
is
there,
but
I
just
need
to
test
it.
So
that's
me.
Apart
from
that,
I
had
a
couple
of
meetings
with
art
and
bob
to
understand
their
software
module
and
how
I
can
proceed
with
that
there.
The
roadblock
is
that
I
need
a
transmitter
device
which
art
is
preparing
for
me
and
he
will
ship
it
over
maybe
next
week
so
yeah.
B
C
B
B
B
B
All
mentioned:
I'm
james
kilojoul
at
seven
kilo
delta
echo
reporting
in
from
labs
out
in
some
more
conditional
replicas
there's
not
really
much
that
we've
been
encountering
lately.
No
new
strange
requests.
I'm
sorry,
I
don't
know
of
any
monika
and
we
haven't
gotten
any
requests
from
that
individual
here.
So
things
have
just
been
doing
as
they
do
here
and
not
much
else
to
report.
C
Hey
how's,
it
going
okay,
so
I'm
still,
you
know
trying
to
figure
out
where,
where
I
can
fit
in,
but
so
I
I
think
I
I
discussed
with
a
couple
people
last
week.
One
of
the
things
that
I
could
do
is
so
on
the
res
for
the
receiver.
We
need
a
method
of
being
able
to
get
simple
timing
and
then
also
carrier
lock.
C
What's
in
chat,
anyways
thanks
michelle
anyways,
so
I
kind
of
dusted
off
some
code
that
I
had
and-
and
I
added
like
the
different
constellations
for
dvds2
I'd
like
to
share
my
screen
at
this
point
and
then
just
show
you
what
I
have
and
sometimes
that
that
can
help
and
see
how
you
guys
can
use
it.
This
is
mainly
for
receiver
stuff,
but
in.
C
I
have
to
warn
you:
this
is
not
a
finished
product
and
the
other
thing
I
have
to
warn
you.
It's
it's
combining
two
different
visualization
techniques.
I
was
having
problems
with
plotly
and
trying
to
sometimes
some
of
these
things
are
just
so
complicated
they
put
too
much.
You
know
the
the
works
into
the
thing
anyways,
okay,
so
part
of
it's
going
to
be
displayed
here
in
this
browser.
C
Hopefully
here's
a
code,
I'm
just
going
to
run
it
and
then
we'll
talk
to
it
and
then
we
can
kind
of
explore
such
sections
in
it.
So
just
remember
the
audience
here,
just
it's
to
help
guide
the
implementer
on
what
potentially
they
could
do
for
to
achieve,
receive
timing,
simple
timing
and
and
do
whatever
scenarios.
So
it's
this
is
kind
of
like
an
rf
model
and
in
fact
I
have
various
parameters
in
here.
This
is
a
single
file
here
I
I've
renamed
it
dvs2
constellation
timing
and
sims.
C
C
I
create
a
class
called
md,
which
is
for
modulation,
d
mod
whatever,
and
here
you
put
in
whatever
constellation
you
have,
and
then
you
put
your
code
rate,
the
only
thing
that
this
does
with
the
code
rate,
it
doesn't
do
any
encoding
or
anything
I
mean
any
fec
associated
with
it,
but
on
the
16
and
32
aps-k
modulations,
the
code
rate
will
adjust
the
amplitude,
and
this
does
that
anyways.
So
one
of
the
things
I
do,
I
generate
an
iq
list
for
the
constellation.
C
Some
of
these
figures
that
you're
seeing
at
the
bottom
can
be
enabled
or
disabled
depending
a
lot
of
it
is
just
kind
of
for
debugging
of
like
let's
say
you
go,
adjust
something
and
you
want
to
figure
out
what's
going
on.
So
that's
what
this
is
for.
Okay,
so
it's
it's
kind
of
done.
This
first
thing
here
right
now
is
generating
some
other
stuff
up
here.
Let's,
let's
look
at
these
diagrams
here.
C
First
of
all
I
put
in,
we
could
run
this
again
in
a
second
I
put
on
qpsk.
This
looks
like
a
psk
constellation
just
in
case
anybody's
wondering
the
the
next
thing
here.
Figure
two
is
our
one
of
the
things
we
do.
What
this
thing
does,
maybe
I
should
kind
of
walk
through.
It
is
it'll,
generate
your
filters
and
it
generates
two
one
for
essentially
the
transmitter
and
one
for
the
timing.
C
C
I
forgot
what
the
term
is:
no,
no,
no
polyphase,
no
anyways.
He
he
uses
as
a
template
to
generate
the
polyphase.
You
know
so
so
I
I
generate
essentially
the
template
kind
of
the
the
reference
filter
and
then
then
I
generate
the
poly
phases
and
I
kind
of
overlay
them.
So
this
here
is
actual
points
and
this
one
here
the
green
is
actual
reference
and
then
inside
of
that,
you
see
all
the
different
phases
overlaid
on
top
of
each
other.
So
it's
just
kind
of
a
interesting
way
to
visualize.
C
C
What
it
looks
like
and
maybe
should
have
been
dots,
but
it's
connecting
the
knots
so
just
say
that
so
the
reference
shaping
filter
versus
one
of
the
phases.
This
is
phase
zero.
What
it
looks
like
on
one
of
the
multi-rate
filters
and
then
on
the
receive
side
we
we
need
to
do
timing
associated
with
that
and
one
of
the
things
that
we
do.
We
over
sample
this
quite
a
bit
and
generate
a
multi-rate
filter.
So
this
is
kind
of
the
over-sampled
kernel
and
then
this
is
the.
C
Oh
all
right
all
windows
at
its
best,
all
the
kernels
overlaid
on
top
of
each
other.
So
if
we
zoom
in
on
this,
we
can
see
that
it's
a
fairly
high
resolution
right
so
just
kind
of
the
idea
is
that
you
want
to
get
a
fairly
good
representation
of
of
the
data
on
the
receiver
because
you're
once
this
comes
once,
the
data
comes
in
you're
down
you're
down
sampling
it
to
essentially
two
samples
per
symbol
with
this
technique.
C
This
is
all
this
has
gotten
from
or
modeled
off
of,
one
of
fred
harris's,
I
suppose
either
lectures
or
one
of
his
he
he
kind
of
sent
me
over
some
matlab
code.
So
this
is
a
highly
derivative
version
of
that.
I
suppose
you
could
say
so
after
that,
one
of
the
things
to
get
phase
lock
the
phase
lock
loop
to
work
and
also
the.
C
C
C
C
C
C
So
all
the
data
and
everything's
in
there
the
html
files
are
a
little
bit
on
the
large
side
for
an
html
page.
But
the
idea
is
that
you
can
it's
very
interactive.
You
can
go
in
there
and-
and
you
know,
look
at
data
points.
So
let's
say
if
we
want
to
say:
okay,
let's
zoom
in
onto
this
thing
here
and
see,
what's
going
on
at
the
very
beginning,
and
let's
see
here,
this
is
going
to
work.
First
of
all,
maybe
not.
C
Okay,
well
maybe
too
many
data
points
right
here:
anyways,
okay,
so
here's
just
the
data
there's
4
000
symbols
going
through
and
as
part
of
how
the
multi-rate
filters
work
to
recover
your
symbol.
Timing,
one
of
the
things
you
do
is
you
generate
this?
This
instance
or
implementation
has
32
phases
right,
so
I
take
each
one
as
the
data
goes
through
it
and
you
can
kind
of
look
at
the
data
as
it
goes
through
the
phases
and
you'll
see.
C
C
There's
a
rollover
crossing
boundary,
then
there's
a
loop
response
and
then
it
settles
here
right
then
this
is
the
the
derivative
filter
on
you
know
what
what
it
thinks.
The
values
are.
Let's
see,
oh
so
here
it
decided
to
zoom
in
this.
Is
you
know?
So
you
have
these
little
tools
here.
Zoom
pan,
whatever
lasso
auto
scale,
is
all
in
html
land.
So
it's
kind
of
interesting
to
have
the
plotlist
stuff
represented
and
then
not
only
that
you
can.
Let's
say
we
zoom
in
on
a
section
here.
C
You
can
look
at
these
values
pretty
easily
and
then
there's
also
added
information
that
you
can
display
on
these
little
tag
values.
You
know
the
tag
that
sticks
off
the
side
in
machine
learning.
You
know
there
could
be
a
lot
of
you.
You
plot
it
with
respect
to
something,
but
then
you
can
have
all
the
extra
information
on
there
here.
It
just
gives
you
your
your
you
know.
C
Basically,
x
y
coordinates,
so
that's
one
advantage
of
using
poly
okay,
so
here's
the
loop
response,
you
see
it,
it
has
its
response
and
it
settles
out.
So
this
is
pretty
good
actually
and
then
so
I
plotted
a
pre,
p,
pre
and
post
it's
just
pll
here,
but
I
think
this
is
simple
timing,
I
believe
maybe
not
and
then
here's
our
constellation
that
that's
getting
recovered
here
for
the
qpsk.
C
Timing,
so,
along
with
simple
timing,
you
also
need
to
do
carrier
lock
where
you
know,
there's
symbol,
timing
and
then
there's
yeah.
What
do
you
call
it?
Just
timing?
I
I
it's
mainly
curious
all
right,
yeah,
the
facelocks
look
here
I
mean
these
are
pretty
small
numbers,
but
you
see
it
just
doesn't
settle.
You
know
and
then
here's
your
error
here
so
yeah.
A
C
And
then
here's
here's,
here's
the
post
pll
on
that
which,
for
some
reason
it
looks
like
it's
being
plotted
over
here
or
over
here
as
well.
I
don't
know
why
that
is,
and
then
here's
a
constellation
now
because
we're
down
sampling
to
two
samples
per
symbol
to
do
all
of
this-
and
I
don't
know
if
this
is
a
garner
or
early
late,
whatever
it's
one
of
fred
harris's
things.
C
So
you
have
your
your
even
samples
or
your
odd
samples,
depending
on
you
know
how
it
starts
starts
off.
One
of
those
even
samples
are
going
to
be
your
actual
symbol
data
and
the
other
one
is
going
to
be
the
intermediate
transitions.
You
know
between
symbol,
qpsk
here.
Obviously
you
have
these.
These
odd
ones
here
are
basically
the
ones
that
don't
change.
You
know
that
stay
at
this
coordinate
point,
and
these
here
are
when
it
transitions
from
this
guy.
B
C
Obviously,
you
know
kind
of
going
through
the
center
points.
So
you
see
these
are
the
odd
ones.
So
at
some
point,
you'll
have
to
the
design
will
have
to
make
a
determination.
Do
I
use
the
even
ones
or
the
odd
ones?
Okay,
so
that's
that
let
me
go
ahead
and
clear
this
you
know.
Obviously,
qpsk
is
not
exactly
damn
it.
Sorry
keep
that
out.
C
C
Okay,
it's
done
so
for
for
each.
You
know,
modulation!
There's
a
these
are
dictionaries
in
python.
If
you
guys
don't
know
so
the
first
and
each
one
is
the
key
value
is
just
the
index
for
the
modulation
number.
You
know
there's
four
in
qpsk
and
then
each
entry
in
the
dictionary,
the
first
entry
is
just
kind
of
the
the
amp
the
ring
number.
So
there's
for
these
guys,
there's
r0
one
two
there's
three
rings:
basically,
and
because
it's
zero
based,
if
we
look
on
in
the
well,
let
me
bring
up
the.
C
You
know
if
we
look
into
the
etsy
org
doc,
you
know
you,
you
have
your
constellation
and
then
you
got
your
code
rigs
and
your.
So
this
table
is
essentially
in
the
in
the
code
that
I
showed
you
I'll
show
you
where,
and
so
you
have
your
different
constellation
points
and
your
gammas
here,
which
determines
what
your
your
ratio
between
r2
and
r1
and
r3
and
r1,
and
in
and
like
in
the
case
of
32,
aps
k.
C
C
A
C
C
That
happens
here
and
these
guys
here
these
are
dictionaries
as
well
and
then
there's
the
the
class
itself.
The
instance
of
the
class
has
essentially
the
modulation
default.
Modulation
is
qpsk,
but
you
just
set
that
whenever
code
rate
that's
one
of
those
numbers
down
there
and
then
some
other
stuff.
I
forgot
what
this
is,
but
if
I
wanted
to
save
something,
so
I
generate
a
magnitude
list
and
then
I
generate
the
constellation
and
then
in
the
generate
file.
C
C
Generate
data
so
in
generate
data.
Basically
I
go
in
there
based
on
the
constellation.
It
tells
you
how
many
data
points
there
are,
and
then
I
generate
the
this
thing,
so
I'm
just
basically
generating
random
data
and
then
between,
let's
say,
zero
and
how
many
points
are
in
the
constellation,
and
I
return
them.
That's
what
you're
seeing
okay
simulation
has
been
finished.
So,
let's,
let's
look
at
those
things
there.
C
C
This
is
the
kernel
again
and
we
we've
been
through
all
these
filters
here,
but
they're
the
same
filter
just
you
know
these
are
there's
a
derivative
filter
with
all
the
different
phases
in
there.
Okay
and
then
here
we
look
at
the
data.
We
can
look
at
the
data
here.
This
is
the
data
prior
to
filtering
and.
C
Yeah,
so
one
thing
that
plotley
has
is
is
a
problem
with
when
there's
too
many
data
points
in
a
plot.
It
takes
too
long
for
it
to
do
anything,
and
you
we
see
that
right
here,
it's
just
not
very
reactive,
anyways,
it's
normally
you'd
be
able
to
zoom.
This
in
matplotlib
doesn't
have
this
issue.
That's
one
reason
why
I'm
going
back
to
map
okay,
so
here's
a
filter
again.
B
C
C
This
is
the
recovered
constellation.
So
one
thing
that
I
like
to
see
is
these:
these
sampling
points
to
be
a
little
bit
better.
This
is
you
know
the
simulation
without
any
noise
added
to
it,
so
that
this
concerns
me
a
little
bit.
These
should
be
kind
of
you
know
almost
as
good
as
these
guys
here
and
these
guys,
you
don't
see
any
noise
on
this
at
all.
The
variation
here,
I
believe,
is
due
to
the
phase
arm.
You
know
you
gotta.
Imagine
it's
it's
toggling
between
a
phase.
C
C
I
suppose
you
know
you're
going
between
plus
one
minus
one
and
kind
of
averaging
out
and
and
if
you
remember
you're
on,
let's
pick
one
of
these
things
here,
imagine
you're
on
one
of
these
phases
and
you're
selecting
between
a
couple
of
them
right,
so
you're
you're
toggling
between
a
couple
of
these
things,
so
that
the
gain
remember
how
these
filters
work.
You
you
have
everything
in
the
thing.
It's
all
added
together,
you,
you
convolve
it
with
a
data
point
and
a
couple.
A
C
B
B
C
C
C
C
So
remember
this
is
just
nor,
if
you
let
this
run
right
without
the,
if
you
add
in
an
impairment
in
there,
like
a
frequency,
offset
these
right
now,
this
simulation
is
not
run
with
it.
These
points
are
going
to
be
spinning
right,
you're,
going
to
get
a
circle
here,
you'll
get
a
ring
here
and
you
get
a
ring
here
right.
C
Hopefully,
it's
useful
for
somebody
to
use
the
idea
is
that
once
you
get
carrier
lock,
these
symbols
here
go
into
the
decoder.
You
know
you
you've,
you've,
just
you've
achieved
simple
timing
and
you've
de-rotated
the
constellation
and
locked
it
in
place.
That's
what
this
would
represent
and
then
you
put
that
into
the
decoder
and
hopefully
with
any
kind
of
luck.
You
get
your
data
back
or
video,
whatever
you've
transmitted
anyways.
Does
that
make
sense?
Does
anybody
have
any
questions
comments,
questions
suggestions,
I
open
the
to
the
you
guys.
A
C
B
C
B
C
B
B
A
Wow
well,
that
was
plenty
thank
you
so
much
and
we'll
we'll
get
it
into
the
repo
and
talk
about
it
with
the
team,
on
slack
and
in
our
weekly
report,
all
right.
Anybody
else
with
work
that
they've
done
over
the
past
week
and
work
that
they
have
scheduled
over
the
next
week
and
roadblocks
and
resources
needed
speak
up.
D
The
the
thing
that
I
was
mentioning
yes
last
week,
which
was
spectrum
analyzer
stack
and
just
basically
using
it,
hopefully
for
seeing
the
spectrum
and
measuring
levels,
do
does
the
lab?
Do
the
labs
already
at
ori?
Have
their
own
spectrum
analyzer
kits
that
you're
interested
to
do
that
or
that
will
be
later
in
the
development
process.
D
Oh
okay,
all
right,
I'm
actually
asking
because
I
I'm
just
for
my
own
use
case,
I'm
making
some
test
boards.
I've
got
designs
in
progress,
but
just
just
basically
try
just
generating
a
carrier
and
just
doing
on
off
keying
seeing
them.
So
I
thought
I'd
just
sync
with
your
future
requirements.
Just
in
case
you
wanted
to
get
a
another.
You
know
a
report.
A
Oh
yeah,
no,
no,
sorry
not
to
cut
off
paul
or
anything,
but
the
higher
higher
than
six
gig,
which
is
where
our
spectrum
analyzers
top
out
at
would
be
welcome.
So
whatever
you
got
we'll
use
we're
looking
at
well,
we
have
a
dual
band:
feed
for
10
gigahertz,
slash,
24
gigahertz,
which
we're
sending
out
to
people
this
week
and
we're
looking
at
47
gigahertz.
So
you
know
they
will
need
measurement
at
the
higher
frequencies.
So
anything
that
you
have
that's
a
resource
is
welcome.
D
Okay,
I've
sent
you
an
email,
it's
in
your
mailbox
michelle,
but
generally
the
the
up
to
22
gigahertz,
I'm
already
set.
So
the
question
is:
what
levels?
What
kind
of
things
would
you
like
to
see
on
analyzer
I
mean
we
could
we
can
reduce
the
power
until
we
can't
see
it,
I'm
just
curious
what
would
be
the
kind
of
test
case
test
test
requirements.
B
A
Sort
of
I
mean
the
the
requirements
for
any
particular
mission
are,
are
going
to
be
obvious
to
us.
So,
for
example,
requirements
to
go
on
the
upcoming
sounding
rocket
missions
are
obvious
and
there
are
gates
that
we
have
to
pass
at
these
various
things
and
then
to
actually
be
on
a
mission.
There's
there's
going
to
be
rubrics
and
credentials
and
qualifications
and
test
schemes.
But
as
far
as
our
own
internal
I
mean
because
that's
what
you're
at
right
now
is
like.
A
A
Yeah,
so
I
mean
I'm
just
throwing
just
thrown
out
there
because,
like
a
tenth
of
a
db,
would
be
for
for
a
normal
like
a
typical
benchtop,
but
what
you're
talking
about
with
the
equipment
that
you
have
is
a
little
bit
better
than
that.
So
I
mean
it's
just
simply
a
question
of
communicating.
Here's
the
here's.
What
I
think
the
calibration
level
is,
here's
the
last
time
it
was
calibrated
just
to
be
transparent,
and
then
I
think
that
that's
good
enough
and
that's
super
helpful.
D
I
think
I
can
create
the
mask.
I
mean
I
can
report
the
side.
Lobes
I
mean
from
from
a
center
frequency
minus
and
a
center
frequency
plus.
I
can
write
automation,
script
using
these
bench
measurements
and
actually
sweep
and
give
a
test
device
under
test,
basically
a
report
that
it's
generating
such
and
such
rf
at
such
and
such
points
away
from
center
frequency.
Oh.