hey everyone in
order to make better
coatings in my sputtered chamber I need
to accurately control what gases are
being added in real time to the chamber
so I started looking into these mass
flow controllers and I thought I'd talk
about what I learned at first I thought
I would be able to accurately control
gas flow with a manual flow meter also
called a rota meter and these are pretty
cool because they are completely
unpowered and so there's really no
chance that the electronics can die and
they're typically much less expensive
than a true mass flow controller so my
plan was to just turn the needle valve
and adjust the flow to what I needed and
then call it quits but this actually
doesn't work because what these flow
meters measure is the volume flow rate
and if the pressure changes like it will
if the pressure regulator is not perfect
or if the temperature varies a little
bit then the actual amount of gas going
through the flow meter will change a
little bit or the reading will change it
becomes inaccurate so what these meters
are actually measuring is the volume
flow rate they don't actually know how
dense the gas is and if the pressure
goes up or the temperature changes then
the density will change and a different
amount of gas will flow through the
meter for a given flow rate I eventually
found a pair of true mass flow
controllers on ebay for quite a bit less
than what they usually sell for
surprisingly even on eBay these things
typically go for a few hundred dollars
each and in this case I got these two
for substantially less the flow
controller is an active device and it
consists of a sensor and some analog
amplification circuitry and a valve to
regulate the flow of gas going through
the whole thing so the setpoint is
determined by an external voltage and
then that's fed into this analog
circuitry so we basically have a closed
loop servo control system so the way it
actually measures gas flow is by
applying a little bit of heat to this
sensing tube here and normally if no gas
were flowing through the sensor to the
thermal profile we look something
like this so that the tube is hot here
and it's not as hot at the ends because
it's attached to the rest of the system
so there would be this nice smooth
Gaussian profile distribution however if
there's gas flowing through this little
sensing tube the profile will look
something more like this it will be
shifted over and it will be spread out
it'll be a little bit more lopsided so
basically we've got to round coils
around the sensing tube and if there's
gas flowing through there the further
down part of the tube will be hotter in
other words this coil will be hotter
because the temperature profile has
shifted over and we can sense that
difference with the analog circuitry so
we know that there's gas flowing through
there what's interesting is that if the
gas pressure changes so let's say
suddenly our input pressure doubles or
something like that then there's just
more molecules in here to soak up the
heat as they're traveling through the
tube and we will get a reading like well
we can detect when the pressure changes
similarly if the input temperature
changes we don't really care too much
because we're mostly just interested in
the shape of this profile so even if the
overall temperature goes up it doesn't
matter because really all we're making
is a differential measurement here one
consideration is that we want the gas
flowing through the small tube to be
laminar flow so that turbulence doesn't
cause our readings to be weird so what
we do is we put a restriction in the
main flow path that has almost well has
an own resistance so that when we split
the flow into this little metered flow
and the main flow that ratio holds
constant for the operating range of the
valve so here's a valve where I've taken
the cover off this is the cover here
this little chamber here presumably has
the metering tube in it and then here's
the control valve and then most of the
gas goes straight through through that
restriction in the bottom there these
valves need to be calibrated for the
specific gas that you're going to be
measuring so it's no fair to calibrate
it on oxygen and then switch to argon
it will not it will still compensate for
pressure differences but the overall
reading will be off because argon and
oxygen have different thermal properties
since what we're measuring here is the
ability for the gas to carry heat from
one heater to the next if the gases
thermal properties change then that is
going to change the reading this unit is
calibrated for a hundred standard cubic
centimeters per minute of argon and this
valve that I have hooked up is 10 cubic
centimeters of oxygen per minute in
order to do a rough sort of sandy check
calibration on these devices I hope this
one up and set it to mid scale and I put
the tube in a beaker of water so you can
see how much gas is visually coming out
it's set to mid scale on the full range
is 10 standard cubic centimeters a
minute so in theory this is what five
standard cubic centimeters of oxygen
looks like I also have a rota meter
inline going into the flow controller
just to get a reading a comparison
reading and the numbers appear to be
agreeing pretty well the rota meter is
coming up with about 3.6 or 3.7 standard
cubic centimeters of air however the
rota meter is calibrated for one
atmosphere and currently I have about 12
psi or almost two atmospheres going into
the flow meter this actually makes sense
since earlier I said that pressure will
cause a pressure change will cause the
rota meter to be incorrect and in this
case a higher pressure means a lower
volume flow so the rota meter will read
low when the pressure is above its
intended calibration point and as it
turns out the difference or the the
correction factor is the square root the
ratio of the pressure differences so in
this case we've got about square root
two since we're running at about 12 psi
that's almost 2 atmospheres absolute and
that's about 1.4 square root of 2 and so
if we multiply 1.4 by the 3 point 6 or 7
we get up to about 5 and that makes
sense since I I set the flow meter to
the flow controller to mid scale here's
a closer look at the flow controller
itself inside there there's actually an
LM 324 quad op-amp there
and it's got a bunch of internal pots to
set presumably slope and gain internal
gain stuff there's only one pot exposed
to the outside there's a hole in the
outside cover for calibration so
apparently once all these other ones are
set in fact this one even has it do not
adjust sticker on it then the meter is
good for anything and to adjust for
different gas types and different
maximum flow rates I think they just
changed this one pop here and change an
aperture like a pinhole or something
inside to basically set the overall flow
rate annoyingly the meter requires plus
and minus 15 volts to operate and then
accept a 0 to 5 volt input control
signal to tell it how much flow you want
it has these two to-220 transistors to
control power going into the valve and
it also outputs a 0 to 5 volt signal to
tell you what it thinks the flow rate is
so you can actually tune the or at least
you can be aware if the thing is not
servo incorrectly if your application
requires rapid flow changes you can see
how the actual flow rate is not is or is
not meeting your requested flow rate
okay well I'm going to get these install
on my vacuum chamber and hopefully be
sputtering pretty soon ok see you next
time bye
s
Posts
No comments:
Post a Comment