In a message dated 29/08/05, davefnelson@....... writes:

>  I initially planned to use a capacitative sensor taking advantage of the 
> high dielectric constant of water. I am very familiar with precision 
> capacitative bridge circuits such as those used in the MKS capacitance manometer and 
> other high sensitivity pressure sensors.  
>  
>  The problem is that these sensors must work in a gas not a liquid. I have 
> also built capacitative sensors for use in water. The water must be very pure 
> or the conductivity becomes a contaminating factor in the bridge causing a 
> phase shift.

       Can you operate an MKS sensor in oil? It would have a different 
sensitivity. 

    You need to boil distilled water to get the air out and then cool it in a 
sealed container. If you leave pure water open to air, it absorbs CO2 (as 
well as O2 and N2), which gives you very dilute carbonic acid. You fill the 
tiltmeter / seismometer etc and then either cover the ends with oil or seal in a 
Nitrogen atmosphere. 

 The water/oil interface seems like a good configuration but it has problems 
if you 
> want to have some portability. You have to fill the thing where you want to 
> use it. 

       You can get vacuum valves which fit water pipe sizes. These have a 
thick rubber diaphragm compression closure and O ring clamp fittings for the 
pipes. If you fit one in the centre of the apparatus, you can close off the middle 
and prevent flow to either end for handling.
 
 The coaxial configuration I currently have in my sensor was originally 
intended to be 
> capacitative but the conductivity of even distilled water was an issue 
> particularly if you add something to keep the water from "sticking" to the 
> electrode. The use of detergents to eliminate some of the surface tension effects 
> also increased the water conductivity. I finally said if you can't beat it -- 
> exploit it so, I made based my system on variable contact area in a 
> conductive fluid in a vertical coaxial configuration. The water is mixed with 
> dishwasher sheeting agent like made by "Cascade" in a 100 to 1 ratio. This gives the 
> right conductivity and makes the meniscus ride up and down on the brass pipe 
> electrode surfaces without sticking. The meniscus stays at a fixed contact 
> angle. I have also used much lower concentrations of various detergents and 
> very small amount of salt to add the ions needed for conductivity. I built a 
> test cell to evaluate different mixtures. 

       Another capacitance method is to coat the metal surfaces with marine 
varnish to prevent conduction. It can be dusted with Aluminium or Titanium 
Oxide while wet.  Alcohol is quite effective at reducing surface tension.
       In the traditional tiltmeter applications you have a larger diameter 
pot at either end of the horizontal tube, which increases the period, a 
horizontal disk electrode just below the water surface and a horizontal disk 
electrode either in free gas or submerged in oil just above the water surface. You may 
use 10 to 50 pF total. Due to the high dielectric constant of water, this 
arrangement is not too sensitive to changes in conduction. If you use oil with a 
dielectric constant of 2, you can allow a greater physical separation of the 
water surface from the bottom of the top electrode for a comparable capacity.
       Have a look at 
http://ntweb.mcn.org/gbush/Tiltmeter/tiltmeter%20Main.html
       Another method is to use the capacitance as part of a high stability 
RC oscillator and measure the change in frequency at both ends. This can give a 
linear position readout. Some frequency meters have a RS232 digital output 
which can be hooked up to your computer.
       I would advise against adding salt or any other halide to water to 
increase the conductivity. They are much too corrosive. Have you tried 
bicarbonate of soda? 
                              
> The transducer is of course the core of the instrument -- I have tried a 
> lot of things and do not claim to have found the optimum but it does work 
> pretty well and is easy to deal with. I can pick the whole instrument up, not 
> worry about mixing different fluids move it without any concern about handling it 
> an any special way. It make some time to restabilize but in a day or so it 
> will be back to normal operation. Bubbles are a big problem on initial fill 
> and following rough handling but they go away. 

       You can buy a water jet pump to provide a near vacuum for the initial 
fill. Even pumping out by connecting the suction end of a small paint spray 
compressor will considerably reduce the volume of any trapped air. 

> Regarding pressure sensors and water based seismometers. I tried lots of 
> configurations with some limited success but the usual problem was noise from 
> atmospheric pressure changes. Even if the system is completely closed the 
> tubing or pipe will be compressed by the pressure change. When you are working 
> with pressure changes of 10e-6 torr in a 1000 torr atmosphere the noise can be 
> overwhelming no matter how you try it isolate it. I gave it up after trying 
> many different configurations. For those interested in infrasound some of the 
> 
           
       I am puzzled as to why you are seeing ANY signals due to atmospheric 
pressure / noise changes.

       The more usual arrangement is to provide a central inverted U tube 
half filled with oil and a differential pressure sensor at the top. The oil 
floats on the water and the sensor is only in contact with oil. If you have sealed 
end pots on the ends of the two arms but connect them together with a plastic 
tube, while the gas pressure may change with time and temperature, it is the 
same for both arms and the differential sensor will not see it. It is an 
advantage to use silicone fluid, since it is inert. You can use 'transformer' oil. 
It comes in several viscosities. Lubricating oil may have a lot of quite nasty 
additives and is best avoided. You can use a large piezo sounder disk as the 
sensor if you are only interested in periods up to about 5 to 10 seconds. These 
also make quite good infrasound sensors, but they need a very stable 
temperature environment.  
       Another type of diaphragm sensor may be made by sticking discs of thin 
Permalloy or similar onto the centre of a circular stretched mylar membrane, 
with two coils in ferrite transformer cups on either side. The coils are 
connected up as a LR relaxation oscillator and you detect the Mark/Space variation 
as the discs move toward or away from either coil. These are used to measure 
very small air pressure changes as a glider gains or looses height. 

       Regards,

       Chris Chapman
In a message=20=
dated 29/08/05, davefnelson@....... writes:



 I initially planned to use=
 a capacitative sensor taking advantage of the high dielectric constant of w=
ater. I am very familiar with precision capacitative bridge circuits such as=
 those used in the MKS capacitance manometer and other high sensitivity pres=
sure sensors.  

=20

 The problem is that these sensors must work in a gas not a liquid. I ha=
ve also built capacitative sensors for use in water. The water must be very=20=
pure or the conductivity becomes a contaminating factor in the bridge causin=
g a phase shift.




       Can you operate an MKS sensor in oi=
l? It would have a different sensitivity.=20



    You need to boil distilled water to get the air out a=
nd then cool it in a sealed container. If you leave pure water open to air,=20=
it absorbs CO2 (as well as O2 and N2), which gives you very dilute carbonic=20=
acid. You fill the tiltmeter / seismometer etc and then either cover the end=
s with oil or seal in a Nitrogen atmosphere.=20



 The water/oil interface seems like a good configuration but it has p=
roblems if you 

want to have some portab=
ility. You have to fill the thing where you want to use it. 



       You can get vacuum valves which=20=
fit water pipe sizes. These have a thick rubber diaphragm compression closur=
e and O ring clamp fittings for the pipes. If you fit one in the centre of t=
he apparatus, you can close off the middle and prevent flow to either end fo=
r handling.

=20

 The coaxial configuration I currently have in my sensor was originally=20=
intended to be 

capacitative but the con=
ductivity of even distilled water was an issue particularly if you add somet=
hing to keep the water from "sticking" to the electrode. The use of detergen=
ts to eliminate some of the surface tension effects also increased the water=
 conductivity. I finally said if you can't beat it -- exploit it so, I made=20=
based my system on variable contact area in a conductive fluid in a vertical=
 coaxial configuration. The water is mixed with dishwasher sheeting agent li=
ke made by "Cascade" in a 100 to 1 ratio. This gives the right conductivity=20=
and makes the meniscus ride up and down on the brass pipe electrode surfaces=
 without sticking. The meniscus stays at a fixed contact angle. I have also=20=
used much lower concentrations of various detergents and very small amount o=
f salt to add the ions needed for conductivity. I built a test cell to evalu=
ate different mixtures. 




       Another capacitance method is to=
 coat the metal surfaces with marine varnish to prevent conduction. It can b=
e dusted with Aluminium or Titanium Oxide while wet.  Alcohol is quite=20=
effective at reducing surface tension.

       In the traditional tiltmeter applic=
ations you have a larger diameter pot at either end of the horizontal tube,=20=
which increases the period, a horizontal disk electrode just below the water=
 surface and a horizontal disk electrode either in free gas or submerged in=20=
oil just above the water surface. You may use 10 to 50 pF total. Due to the=20=
high dielectric constant of water, this arrangement is not too sensitive to=20=
changes in conduction. If you use oil with a dielectric constant of 2, you c=
an allow a greater physical separation of the water surface from the bottom=20=
of the top electrode for a comparable capacity.

       Have a look at http://ntweb.mcn.org=
/gbush/Tiltmeter/tiltmeter%20Main.html

       Another method is to use the capaci=
tance as part of a high stability RC oscillator and measure the change in fr=
equency at both ends. This can give a linear position readout. Some frequenc=
y meters have a RS232 digital output which can be hooked up to your computer=
.

       I would advise against adding salt=20=
or any other halide to water to increase the conductivity. They are much too=
 corrosive. Have you tried bicarbonate of soda?=20

             =
;            &nb=
sp;    

The transducer is of cou=
rse the core of the instrument -- I have tried a lot of things and do not cl=
aim to have found the optimum but it does work pretty well and is easy to de=
al with. I can pick the whole instrument up, not worry about mixing differen=
t fluids move it without any concern about handling it an any special way. I=
t make some time to restabilize but in a day or so it will be back to normal=
 operation. Bubbles are a big problem on initial fill and following rough ha=
ndling but they go away. 




       You can buy a water jet pump to pro=
vide a near vacuum for the initial fill. Even pumping out by connecting the=20=
suction end of a small paint spray compressor will considerably reduce the v=
olume of any trapped air.=20



Regarding pressure sensors=20=
and water based seismometers. I tried lots of configurations with some limit=
ed success but the usual problem was noise from atmospheric pressure changes=
.. Even if the system is completely closed the tubing or pipe will be compres=
sed by the pressure change. When you are working with pressure changes of 10=
e-6 torr in a 1000 torr atmosphere the noise can be overwhelming no matter h=
ow you try it isolate it. I gave it up after trying many different configura=
tions. For those interested in infrasound some of the sensors I investigated=
 would be ideal.

           

       I am puzzled as to why you are seei=
ng ANY signals due to atmospheric pressure / noise changes.



       The more usual arrangement is to pr=
ovide a central inverted U tube half filled with oil and a differential p=
ressure sensor at the top. The oil floats on the water and the sensor is=
 only in contact with oil. If you have sealed end pots on the ends of the tw=
o arms but connect them together with a plastic tube, while the gas pressure=
 may change with time and temperature, it is the same for both arms and the=20=
differential sensor will not see it. It is an advantage to use silicone flui=
d, since it is inert. You can use 'transformer' oil. It comes in several vis=
cosities. Lubricating oil may have a lot of quite nasty additives and is bes=
t avoided. You can use a large piezo sounder disk as the sensor if you are o=
nly interested in periods up to about 5 to 10 seconds. These also make quite=
 good infrasound sensors, but they need a very stable temperature environmen=
t.  

       Another type of diaphragm sensor ma=
y be made by sticking discs of thin Permalloy or similar onto the centre of=20=
a circular stretched mylar membrane, with two coils in ferrite transformer c=
ups on either side. The coils are connected up as a LR relaxation oscillator=
 and you detect the Mark/Space variation as the discs move toward or away fr=
om either coil. These are used to measure very small air pressure changes as=
 a glider gains or looses height.=20



       Regards,



       Chris Chapman