In a message dated 12/03/2005, ke6pxp@....... writes:

At 07:20  PM 3/10/05 -0500, you wrote:
>George Bush wrote: 
>> I am  building a mercury-cup tiltmeter and it uses the INA121 opamp. I  
would
>> be interested in seeing your analysis of it's noise  also.

>I am interested in how the tiltmeter connects to the  amplifier.  What sort 
of signal >source impedance does this  represent?  Do you need the 
instrumentation amplifier
>differential  configuration?

It uses RF connected to the mercury to  excite antenna plates just above
two connected pools 24" apart. The antenna  signals are detected (rectified)
and fed into the INA121 in a differential  configuration. The DC impedance
is high (back impedance of the diodes), but  since they are charging .1 uf
caps, the AC impedance is low. The mechanical  design is old and originally
appeared in an early 1970's "Scientific  American" Amatuer Scientist article
"A Sensitive mercury tiltmeter that  serves as a seismometer". I still have
a copy of the pages (pp. 124-129),  but SA didn't print the date on the
pages. Recently Ron Newton modernized  the electronics and published an
article on the May 2003 "Nuts n Volts"  magazine, pp. 47-51, "Measure Earth
Tides with a Tiltometer"

George  Bush
Sea Ranch, CA, USA
38.73775N, 123.48882W
Hi George,
 
    The circuit in _http://www.nutsvolts.com/toc_Pages/may03toc.htm_ 
(http://www.nutsvolts.com/toc_Pages/may03toc.htm)  seems  to be flawed. It appears to 
depend on the reverse leakage current of the  rectifier diodes to provide the 
DC level stability and this is both variable and  highly temperature 
dependant. A resistor needs to be connected across both  capacitors to provide a fixed 
impedance to earth. You do need a true  differential opamp for this type of 
circuit to get the very high CMRR required. 
    The inherent weakness of this type of circuit is  that the properties of 
the tuned circuits and the diodes are temperature  sensitive and this limits 
both the accuracy and the drift. The temperature drift  on a Si diode is about 
2.5 milli V /C Deg. I note that there is no low pass  filter at the amplifier 
output! This will allow greatly increased output noise.  An INA118 amplifier 
could give better performance. This type of circuit was used  on seismometers 
for a short period during the 1970s before being replaced by fet  switch 
technology.
    Short length tiltmeters are also sensitive to  surface ground heave, wind 
noise and to movements caused by temperature and  rain / ground water level 
changes. 
    There is now absolutely NO necessity to go to such  high frequencies (4 
MHz), or to use resonant circuits and diodes, to generate  highly sensitive 
displacement signals using moving capacitor plates. This  can be done quite 
easily at about 20 KHz using fet switches. See "Bridge  Measures Small Capacitance" 
by Jeff Witt, Linear Technology Corp, in Electronic  Design, Nov. 4, 1996, 
pg. 110. This is also available on the Linear  Technology site in the second 
circuit compendium and uses the LTC1043 quad  switch / oscillator. 
    DO NOTE that liquid mercury has an appreciable  vapour pressure at room 
temperature and that the vapour is highly toxic.  It will also react strongly 
with many metals giving 'amalgams', which 'corrode'  the surface. Both Copper 
and Aluminum are highly susceptible. Steel and  glass may be used for the 
containment. Mercury is very dense and quite  expensive. 
    Details of more modern tiltmeters are given in  papers and references by 
Nicolas d'Oreye at the Walferdange site. Type  Walferdange tiltmeter or 
seismometer into _www.google.com_ (http://www.google.com)  See 
_http://www.ecgs.lu/index.php?option=displaypage&Itemid=51&op=page&SubMenu_ 
(http://www.ecgs.lu/index.php?option=displaypage&Itemid=51&op=page&SubMenu)  Water  with a layer of 
silicone oil on top has also been used very successfully in  tiltmeters. The 
silicone oil prevents evaporation and condensation problems with  the water. The 
high dielectric constant of water vs the low value for oil, is a  great help. 
    If you can find the full reference for the Sci Am  article, you ought to 
be able to get a photocopy through your local library,  although this may take 
a couple of weeks.
    
    Regards,
 
    Chris Chapman





In a message dated 12/03/2005, ke6pxp@....... writes:
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>At 07:20=20
  PM 3/10/05 -0500, you wrote:
>George Bush wrote: 
>> I=
 am=20
  building a mercury-cup tiltmeter and it uses the INA121 opamp. I=20
  would
>> be interested in seeing your analysis of it's noise=20
  also.

>I am interested in how the tiltmeter connects to the=20
  amplifier.  What sort of signal >source impedance does this=20
  represent?  Do you need the instrumentation amplifier
>differen=
tial=20
  configuration?

    It uses RF connected to the mercury to=
=20
  excite antenna plates just above
two connected pools 24" apart. The ant=
enna=20
  signals are detected (rectified)
and fed into the INA121 in a different=
ial=20
  configuration. The DC impedance
is high (back impedance of the diodes),=
 but=20
  since they are charging .1 uf
caps, the AC impedance is low. The mechan=
ical=20
  design is old and originally
appeared in an early 1970's "Scientific=20
  American" Amatuer Scientist article
"A Sensitive mercury tiltmeter that=
=20
  serves as a seismometer". I still have
a copy of the pages (pp. 124-129=
),=20
  but SA didn't print the date on the
pages. Recently Ron Newton moderniz=
ed=20
  the electronics and published an
article on the May 2003 "Nuts n Volts"=
=20
  magazine, pp. 47-51, "Measure Earth
Tides with a Tiltometer"

Geo=
rge=20
  Bush
Sea Ranch, CA, USA
38.73775N, 123.48882W
Hi George,
 
    The circuit in http://www.nutsvolt=
s.com/toc_Pages/may03toc.htm seems=20
to be flawed. It appears to depend on the reverse leakage current of th=
e=20
rectifier diodes to provide the DC level stability and this is both variable=
 and=20
highly temperature dependant. A resistor needs to be connected across b=
oth=20
capacitors to provide a fixed impedance to earth. You do need a true=20
differential opamp for this type of circuit to get the very high CMRR requir=
ed.=20

    The inherent weakness of this type of circuit i=
s=20
that the properties of the tuned circuits and the diodes are temperature=20
sensitive and this limits both the accuracy and the drift. The temperature d=
rift=20
on a Si diode is about 2.5 milli V /C Deg. I note that there is no low pass=20
filter at the amplifier output! This will allow greatly increased output noi=
se.=20
An INA118 amplifier could give better performance. This type of circuit was=20=
used=20
on seismometers for a short period during the 1970s before being replaced by=
 fet=20
switch technology.
    Short length tiltmeters are also sensitive to=20
surface ground heave, wind noise and to movements caused by temperature=
 and=20
rain / ground water level changes. 
    There is now absolutely NO necessity to go to s=
uch=20
high frequencies (4 MHz), or to use resonant circuits and diodes, to generat=
e=20
highly sensitive displacement signals using moving capacitor plates. Th=
is=20
can be done quite easily at about 20 KHz using fet switches. See "Bridg=
e=20
Measures Small Capacitance" by Jeff Witt, Linear Technology Corp, in Electro=
nic=20
Design, Nov. 4, 1996, pg. 110. This is also available on the Linea=
r=20
Technology site in the second circuit compendium and uses the LTC1043 q=
uad=20
switch / oscillator. 
    DO NOTE that liquid mercury has an appreciable=20
vapour pressure at room temperature and that the vapour is highly toxic.=20
It will also react strongly with many metals giving 'amalgams', which 'corro=
de'=20
the surface. Both Copper and Aluminum are highly susceptible. Steel a=
nd=20
glass may be used for the containment. Mercury is very dense and q=
uite=20
expensive. 

    Details of more modern tiltmeters are given&nbs=
p;in=20
papers and references by Nicolas d'Oreye at the Walferdange site. Type=20
Walferdange tiltmeter or seismometer into www.google.com See http://www.ecgs.lu/index.php?option=3Ddisplaypage&a=
mp;Itemid=3D51&op=3Dpage&SubMenu Water=20
with a layer of silicone oil on top has also been used very successfully in=20
tiltmeters. The silicone oil prevents evaporation and condensation problems=20=
with=20
the water. The high dielectric constant of water vs the low value for oil, i=
s a=20
great help. 
    If you can find the full reference for the Sci=20=
Am=20
article, you ought to be able to get a photocopy through your local library,=
=20
although this may take a couple of weeks.
    
    Regards,
 
    Chris Chapman