Comments needed on this circuit

[gmv]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello,

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

Every time the Air conditioner comes on a tremendous
electrical disturbance is generated in the input
of this circuit causing the class "A" baseline to follow
a nonsymmetrical cycle of waveform maybe 30 to 60 seconds
in period.

I am seeking recommendations to stop or significantly
reduce this electrical disturbance. The offending device
is actually a heat pump and the signature of the disturbance
is quite different depending upon whether the device
is heating or cooling.

I am unable to relocate the sensor so other solutions are necessary.

Any help here is appreciated.

gmv

 

[Andrew Holme]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello,

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

Every time the Air conditioner comes on a tremendous
electrical disturbance is generated in the input
of this circuit causing the class "A" baseline to follow
a nonsymmetrical cycle of waveform maybe 30 to 60 seconds
in period.

I am seeking recommendations to stop or significantly
reduce this electrical disturbance. The offending device
is actually a heat pump and the signature of the disturbance
is quite different depending upon whether the device
is heating or cooling.

I am unable to relocate the sensor so other solutions are necessary.

Any help here is appreciated.

gmv

Why does the shield stop 2 feet from the sensor?

How about putting some/all of the amplification at the sensor end?

 

[John Woodgate]

I am seeking recommendations to stop or significantly reduce this
electrical disturbance. The offending device is actually a heat pump and
the signature of the disturbance is quite different depending upon
whether the device is heating or cooling.

The unbalanced input stage is wide open to every disturbance under the
sun. You need a balanced input stage and a twisted-pair shielded cable
to your sensor. With that huge gain in the first stage, I suspect the
whole design, frankly.

 

[gmv]

You are probably right.
I am trying to go as cheap as absolutely possible.
Is it possible to use a diff amp on the input if there
is no center tap from the sensor ?

Surprisingly this circuit produces fairly good results
so long as there is no electrical disturbances in the vicinity.
I have found using electrolytic condensors works quite well
so long as you do not use them in a feedback type filter.

 

[John Woodgate]

I am trying to go as cheap as absolutely possible. Is it possible to use
a diff amp on the input if there is no center tap from the sensor ?

Centre-tapping is not necessary unless your source and amplifier input
both have very high common-mode impedance. In that case, putting a pair
of resistors (preferably closely matched in value) in series across the
sensor and grounding their junction may help.

 

[gmv]

Why does the shield stop 2 feet from the sensor?

How about putting some/all of the amplification at the sensor end?

The geophone came with out a shielded cable it only has a
single pair going to the coil/magnet assy inside.
Or so it seems. I was trying to get away as cheap as possible
and could not find the fancy shielded wire you guys speak of at the
local radio shack store. Yes I think a preamp at the sensor might be the way to go.
Or possibly an instrumentation amplifier but it all increases the cost.
Is it possible to create a three wire system from the two wire
sensor so I can use a Diff amp as the input ?

 

[Larry Brasfield]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello, Hi.

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

Every time the Air conditioner comes on a tremendous
electrical disturbance is generated in the input
of this circuit causing the class "A" baseline to follow
a nonsymmetrical cycle of waveform maybe 30 to 60 seconds
in period.

I am seeking recommendations to stop or significantly
reduce this electrical disturbance. The offending device
is actually a heat pump and the signature of the disturbance
is quite different depending upon whether the device
is heating or cooling.

I am unable to relocate the sensor so other solutions are necessary.

Any help here is appreciated.

The first three steps should be to convert your single-ended
preamplifier into a differential configuration, extend the shield
to the rest of the cable and around the geophone, and tie the
shield to the same ground the input diff-amp uses as reference.

At the same time, I would redistribute the gain so that, rather
than x 2947 followed by x 1, you would have x 54 and x 54.5.
This would prevent overload of the input stage on signals that
are going to be rejected by subsequent filtering.

The circuit shown below has approximately the same input
impedance, gain, and frequency response as your original
two stages with the above suggestions incorporated.


||
.------||----.
| || |
| 12n |
___ | ___ |
.----o---|___|--o-o----|___|---o
| | | |
| | 1.07k | 57.6k |
| | | |\| |
| .-. '-|-\ | || |\
| | | | >---------o-----||----o---------------|+\
| | | 511 .-|+/ || | | >---.
| '-' | |/| .-. .---|-/ |
| | | 47u | | | |/ |
| | | | |348k .-. |
| | ___ | ___ '-' | | |
| .--o---|___|--o--o---|___|---. | | | 348k |
| | | | | '-' |
| | 1.07k | 57.6k | | ___ | ___ |
| | | | o---|___|---o---|___|--'
| | | || | |
| | .---| '-----||----o === 2.00k 107k
| '--| | || | GND
'----| | 12n |
'---| ===
GND
(created by AACircuit v1.28.4 beta 13/12/04 www.tech-chat.de)

 

[Rich Grise]

The geophone came with out a shielded cable it only has a
single pair going to the coil/magnet assy inside.
Or so it seems. I was trying to get away as cheap as possible
and could not find the fancy shielded wire you guys speak of at the
local radio shack store. Yes I think a preamp at the sensor might be the way to go.
Or possibly an instrumentation amplifier but it all increases the cost.
Is it possible to create a three wire system from the two wire
sensor so I can use a Diff amp as the input ?

You pretty much already have that - just extend the shield, and use it as
your ground connection. Depending on the frequencies and stuff, you could
send power down the other two in "common-mode", with the signal
differential.

Of course, it'd be simpler if you had, say, 4 conductors + shield -
I got some of this at Home Depot once.

I also suspect power supply droop - is it possible to put the sensor on a
different mains circuit, or maybe put in a line stabilizer of some kind,
or even a UPS?

Good Luck!
Rich

 

[Nico Coesel]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello,

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

The first stage should be a differential input. This will eliminate
most of your problems. Try to connect the shield from the cable to
ground of your circuit where the power supply is connected to your
circuit.
I also noticed you are doing a lot of low pass filtering. You could
look into a 'Sallen-key' filter ('Google' for it). This is an easy
filter to construct with an opamp and will probably save you several
opamps.

 

[Rich Grise]

The circuit shown below has approximately the same input
impedance, gain, and frequency response as your original
two stages with the above suggestions incorporated.

I'd break this between the two opamps, and use a differential output amp
for U1, and put that part of the circuit at the sensor.

||
.------||---.
| || |
| 12n |
___ | ___ |
.----o---|___|--o-o----|___|--o
| | | |
| | 1.07k | 57.6k |
| | | |\| | |\

| .-. '-|-\---------o-----------||-----o---------------|+\

| | | | > cable |
| | | 511 .-|+/---------------------||---- | ----------.---|-/
| | | | '-' | |/|

At this point I gave up on trying to do ASCII art. Hopefully, it gets
the point across.

.-. | |/

| | | 47u | | |
| | | | |348k .-. |
| | ___ | ___ '-' | | |
| .--o---|___|--o--o---|___|--. | | | 348k |
| | | | | '-' |
| | 1.07k | 57.6k | | ___ | ___ |
| | | | o---|___|---o---|___|--'
| | | || | |
| | .---| '-----||---o === 2.00k 107k
| '--| | || | GND
'----| | 12n |
'---| ===
GND
(created by AACircuit v1.28.4 beta 13/12/04 www.tech-chat.de)

(royally munged by Rich Grise)

Cheers!
Rich

 

[Larry Brasfield]

I'd break this between the two opamps, and use a differential output amp
for U1, and put that part of the circuit at the sensor.

I agree that such a solution would provide lower
susceptibility to noise pickup. However, given that
the OP has gotten useful results already, and that
line-frequency noise is not particularly significant in
seismic signals, I figured the simpler solution would
be to revise the existing circuit. It will likely work
well enough to keep the noise from taking the input
stage into overload, so whatever measures were
good enough before to take out the 60 Hz and low
harmonics should be more than good enough after
applying those changes.

Another consideration: This is obviously a one-off,
amateur solution. The concept of good-enough applies
more than usual in such situations.

 

[Jamie]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello,

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

Every time the Air conditioner comes on a tremendous
electrical disturbance is generated in the input
of this circuit causing the class "A" baseline to follow
a nonsymmetrical cycle of waveform maybe 30 to 60 seconds
in period.

I am seeking recommendations to stop or significantly
reduce this electrical disturbance. The offending device
is actually a heat pump and the signature of the disturbance
is quite different depending upon whether the device
is heating or cooling.

I am unable to relocate the sensor so other solutions are necessary.

Any help here is appreciated.

gmv

its obviously a Hi-Z problem due to the nature of your app.
Microphone wire is crap most of the time from what i have
seen.
you could try some tightly shielded RG174..

in any case, after looking at your circuit i think moving it
all to the location of the sensor and having a low Z output.
or at least move one stage to give you a unity gain with
low Z, make sure you use Large by pass caps in the sensor
end for the +/- lines to keep the EMF out of that.
your basicly picking up the EMF and your cable is the secondary.

 

[Nico Coesel]

its obviously a Hi-Z problem due to the nature of your app.
Microphone wire is crap most of the time from what i have
seen.
you could try some tightly shielded RG174..

And lose the twisted pair? Bad idea. With a differential input
amplifier and shielded microphone wire, you can have several hundred
feet of wire between the microphone and the amplifier without any
problems. I see no problem in the setup of the OP.

 

[Fred Bloggs]

http://files.photojerk.com/gmvoeth/diagram1.gif


Hello,

The circuit above has been used by me for the last year to
receive regional and teleseismic signals but I have been
plagued by a problem that requires me to seek help.

The lead in cable is a shielded, twisted pair,
microphone cable. Its shield stops maybe 2 feet from the sensor.
The sensor cable itself has no shield. Its entire length
is approximately 50 feet and parallels a major airconditioner
power supply line about 4 or 5 feet away for maybe 30 of the
50 feet.

Every time the Air conditioner comes on a tremendous
electrical disturbance is generated in the input
of this circuit causing the class "A" baseline to follow
a nonsymmetrical cycle of waveform maybe 30 to 60 seconds
in period.

I am seeking recommendations to stop or significantly
reduce this electrical disturbance. The offending device
is actually a heat pump and the signature of the disturbance
is quite different depending upon whether the device
is heating or cooling.

Is that 432 ohm at the sensor something you added or is it internal to
the geophone? Your input lines are not "balanced", a terminology which
simply means each input line from the sensor has identical impedance to
your circuit common over all frequencies in this case. Your interference
problem runs deeper than just electrical. From your description, it is
clear that the heat pump assembly is transmitting shock and/or acoustic
wave components into the ground that are within the frequency range of
the geophone.

 

[Ken Smith]

line-frequency noise is not particularly significant in
seismic signals,

This is not true in the general case. I assume that the OP is working at
very long ranges where the frequency content is all very low frequencies.

 

[Fred Bloggs]

http://files.photojerk.com/gmvoeth/circuita.gif


http://files.photojerk.com/gmvoeth/circuitb.gif


Thanks to input from you guys I tried an instrumentation amplifier
on my input and that seems to have eliminated the 60HZ noise I had
in my first stage.

circuita.gif is the noisy devil.
circuitb.gif is the one I am now using.

gmv









this file is now circuita.gif

That's a lie if I ever heard one.

 

[neon]

the sensor is your souce therefore the 2 wires are for signal and ground the shield must be ted at the source to sensor ground and nowhere else ever.