PSN-L Email List Message

Subject: RE: Instrumentaion Amps and transformers
From: "Blair lade" blairl@...............
Date: Tue, 5 Jul 2011 23:18:46 +0930


Here are some suggestions and ideas to try..

As  we know, Instrumentation Amplifiers are supposed to be good at reducing
common mode noise in signal / sensor systems.
So , lets look at what the symptoms are and what we can measure to try and
determine what is causing your LF 1.25 hz noise...

We have, and to a certain extent I'm  guessing or atleast suggesting how it
could or should be hooked up..

A sensor consisting of a coil(s) with a suspended magnet assembly that move
with respect to each other.

A metallic frame that holds some of the components that is or should be
grounded to the mains earth. (could be to the IA earth but not to both as
you might get circulating ground currents that will introduce noise..)

An instrumentation Amplifier with hi gain, the sensor outputs are connected
to the hi input impedance inputs of the IA and not across the gain setting

Some suitable input bias return resistors to the centre point of the IA
power supply which is well regulated.

A suitable dampening method, either resistive across the IA inputs or
mechanical like an oil dashpot, doesn't really matter but you must have
atleast one.
Cabling that is atleast 2 cores, screened, with the screen connected at the
IA end. The cabling to the seismo must not be run alongside any mains power
cables and should not be near any motors or transformers as all three
produce magnetic fields..

We see seismic signals and have some other LF signals that are present on
the output ..

The seismometer must not be located near any mains operated equipment ie..
transformers , mains cables carrying any large currents, motors, fridges ,
fluro lights, central heating systems, hot water systems etc 

We need to find out if the LF signal that we don't want is real, as in a
true seismic signal or is induced as in electrical or magnetic.

Things you can try.

Do you run a heater for the seismo, if so turn that off for testing, does
that make any difference?..

Replace the sensor coils with a resistor (or 2) that approximates the coils
DC resistance. Do this at the coil end.. 
Do we still have the 1.25Hz LF signal?  Hopefully not.. 

If No LF then the IA and the wiring are probably ok, any common mode signals
introduced into the wiring should be removed by the IA.. noting that with
screened cable, the only way LF signals can get into the wiring is by
magnetic fields, the screening should get rid of all electrical fields..and
with resistors on the ends, there are no ground loops and no magnetic fields
to worry about.

If you do still have the LF signal, then remove the wiring from the IA and
put the resistor (s) straight across the IA inputs...It all should be very
quiet and (LF) noise free.

If you have a large coil of wire like a solenoid, put that across the end of
the IA wiring (seismo end). Rotate the solenoid around, it should be fairly
quiet, it will be sensitive to magnetic fields and will also be quite
directional, you may be able to find all sorts of odd magnetic (60Hz )
fields, motors, transformers, HV and LV cables, central heating systems etc,
all produce magnetic field which could be picked up by the seimo..

If that is all quiet, connect the seismo up and clamp the magnet assembly
(or remove it). You have a similar situation to the solenoid, except that
there will be extra capacitance between the seismo coils and the metal
framework that is grounded (to remove or reduce electrostatic fields..)

If the LF signals are gone, but reappear when the magnet is freed or added,
then the signals may well be seismic in nature, mechanical vibrations from
motors, transformers and wind farms can travel considerable distances.
Motors can beat together or with transformers producing quite strong LF
vibrations that you cant hear. Out of balance fans are well known for LF

At the CTBT Seismic station in Scotland, they have an exclusion zone greater
than 40km in all directions in which wind farms are not allowed due to the
added noise from the turbine blades!

What is your sampling rate?

Do you have a low pass filter after your IA, if so what are its

Let me know your thoughts on all this

-----Original Message-----
From: psnlist-request@.............. [mailto:psnlist-request@...............
On Behalf Of gmvoeth
Sent: Monday, 4 July 2011 6:49 PM
To: psnlist@..............
Subject: Re: Instrumentaion Amps and transformers

Yes, lots of 60Hz common mode stuff
floating around here. I can measure
60Hz at like 23Vrms but there is little
current relating to it.
I think my DMM is like 20Meg or more.
If i try to measure the current I get
And static Electricity is bad also.

I have everything residing on aluminum
sheet metal which is tied to both
Circuit(chassis) and Power entrance (Earth)

I like the balanced nature of the Instrumentation
Amplifier especially since I use a split power supply.


----- Original Message ----- 
From: "Blair lade" 
Sent: Sunday, July 03, 2011 7:18 AM
Subject: RE: Instrumentaion Amps and transformers

> Hi Geoff, 
> The transformer idea could be made to work , however, you have to take it
> apart a bit (you need to change the magnetic assembly from a closed field
> (which is how a transformer works ) to an open field (which is how a
> solenoid works)... 
> If you pull the "Es" and "Is" apart and reassemble the transformer with
> the "Es" all pointing the same way it will sort of work.
> "C" core type transformers don't work well!
> hope this makes sense to everyone reading this.....
> Transformers are typically used with ac signals to change the level or
> isolation to the signal (ac voltage) which incidentally is why they are
> called transformers, and while seismic signals are an ac signal, the
> frequency of seismic signals is so low (typically 0.01 hz to 20hz) that
> transformer off the shelf just wouldn't have enough core in the right
> configuration to respond at these frequencies as a transformer.
> Just on the side....
> Typical mains (power) transformers in Aust are designed for 50Hz and have
> more core and inductance and are physically larger and heavier than the
> you have in the US which are designed for 60Hz, and likewise motors
> for Aust conditions have more core (the soft iron metal) than ones
> for the US.
> US transformers and motors don't perform well in Aust, they get hot and
> suffer from poor efficiency. Likewise , when Aust motors and transformers
> are used in the US at 60Hz they don't like it either..and it's got nothing
> to do with northern verse southern hemispheres and that water goes down
> sink in a different direction.....
> (Good) Audio transformers typically have enough core in them to work from
> or 30Hz upwards. They also use different core material.
> If you look at a low current relay coil with a pole piece, it has very
> amounts of wire and reasonable amounts of metal core (which is again
> different metal to that used in transformers).
> These could be used as a sensor for seismic but there are issues with
> linearity etc if you have large amounts of movement between the magnet and
> the core of the winding, again they are only able to respond to ac
> To test this, hook a coil up to a multimeter on dc volts and slowly move a
> magnet to and fro, you will see a varying dc signal, which is of course an
> ac signal.
> Now place the magnet on the end, once the meter settles down, you will see
> no voltage produced as the magnetic flux isn't changing. 
> It is a problem for all magnetic type sensors, poor dc (or low ac
> signal response.
> To a certain extent, the more core the better the low freq response, but
> there are practical limits.
> Getting back to Instrumentation Amps, why the decision to go down the IA
> path?
> Do you or do you think you have a common mode noise issue that has to be
> addressed?
> Blair
> -----Original Message-----
> From: psnlist-request@..............
> On Behalf Of gmvoeth
> Sent: Sunday, 3 July 2011 10:52 PM
> To: psnlist@..............
> Subject: Re: Damping CDR for HS10-1
> I tried to use a centertapped step up transformer
> but was unable to get that to work.
> It was like no signal at all.
> I guess you cant just grab a transformer off the shelf
> to make such a thing work.
> Also tried a centered tapped coil with a magnet
> which did work but too complex to make
> it is better to make two coils and use then
> together. or so it seems to me.
> regards,
> geoff
> ----- Original Message ----- 
> From: "Mark Robinson" 
> To: 
> Sent: Friday, July 01, 2011 7:16 AM
> Subject: Re: Damping CDR for HS10-1
>> Good thread.
>> I seem to remember Geoff's rig has a centre tap on the coil.
>> How should he connect that?
>> Mark
>> NZ
>> On 02/07/11 00:51, Blair lade wrote:
>>> Geoff,
>>>   Blair here in Aust,
>>> I do lots of (work) stuff with instrumentation amps, matching is
> certainly
>>> easier using some adjustable components...0.01% resistors are about $30
> each
>>> here!
>>> However, the increase in noise from using a 3 opamp IA config compared
> a
>>> traditional single input amp has to be weighed against the better common
>>> mode rejection with a true instrumentation amp.
>>> If you don't have to deal with Common Mode noise, then don't go down the
> IA
>>> path..
>>> There is much literature on the pluses and minuses of single opamp input
>>> verse 3 opamp IA input designs from a seismic point of view.
>>> With instrumentation amps (infact with all opamps), there has to be some
>>> bias current path from the inputs back to ground, usually provided by
>>> excitation circuitry (strain gauge stuff) but if you are just 'going in'
>>> with a floating signal from a coil, you will need to provide some bias
>>> resistance / reference to ground, probably a pair of 1meg resistors
>>> suffice , one from each input to ground (0 volts) depending on the IA
> input
>>> resistance, ofcourse, these could be your dampening resistors, the total
>>> value being split and the centre going to ground...
>>> I prefer the 1 megs as any mismatch in low resistance values here will
>>> affect your CMR a lot.
>>> So the dampening resistor is across the inputs with a couple of 1megs to
>>> ground.
>>> Need also to be careful to balance the capacitance of your input leads
>>> ground as well, but if you are using good quality 2 core screened
> microphone
>>> cable (with the screen tied to grounded only at the amplifier end) that
>>> should be fine for the frequencies of the signals you are looking at.
>>> The other end of the screen (seismo end) would normally be tied to the
>>> sensor's ground and should not be 'hard grounded'(eg. earth staked)
> unless
>>> you are expecting lots of trouble from lightning.
>>> I'd not bother with having a guard amp driving the screen for seismic
>>> signals.
>>> With careful design and correct choice of components, CMRRs of 150db at
> low
>>> frequencies with respect to output are achievable without trimming but
>>> you have that much common mode noise you should do something about it!
>>> The 1.25 hz periodic noise... what is your sampling rate and do you have
>>> some sort of low pass filter that lobs everything off before you get to
> the
>>> nyquist frequency (approx 1/2 your sampling rate)?
>>> Also note that induced magnetic fields at the coil end are not common
> mode..
>>> Blair
>>> -----Original Message-----
>>> From: psnlist-request@..............
> [mailto:psnlist-request@...............
>>> On Behalf Of Geoff
>>> Sent: Friday, 1 July 2011 3:59 AM
>>> To: psnlist@..............
>>> Subject: Re: Damping CDR for HS10-1
>>> I think I just wasted a lot of time
>>> trying to get rid of a signal which
>>> is real differential and not
>>> common mode,
>>> there seems to be a machine
>>> owned by a neighbor which is not
>>> always used.
>>> I am totally unable to rid the artifact of
>>> about 1.25 Hz which is periodical.
>>> Creating the multiple resistor pairs
>>> was a waste of time.
>>> A pot, most likely, is the beat way
>>> to balance the two against the ground.
>>> Like 100 Ohm or 10 Ohm 15 turn
>>> between two resistors matched already.
>>> 1 ohm is the best my DMM can do
>>> without help.
>>> I have trouble keeping my website
>>> to be real time. So its not
>>> 100% reliable.
>>> I think my website shows the
>>> instrumentation amp with the geophone
>>> on the negative leads of the instrumentation
>>> amp, will have to change that
>>> with a new schematic.
>>> Regards,
>>> geoff
>>> -----Original Message-----
>>> From: Geoffrey
>>> Sent: Saturday, June 25, 2011 8:39 PM
>>> To: psnlist@..............
>>> Subject: Re: Damping CDR for HS10-1
>>> Interesting Bob,
>>> But I'm using an instrumentation amplifier.
>>> In such an arrangement of three op amps
>>> you are using two positive inputs which means
>>> the input impedance is mega ohms to giga ohms.
>>> The only input is the the resistors which are
>>> split against ground. So in my case the you
>>> have verified my numbers to be basically correct.
>>> I have learned something new to myself in the past
>>> few days about this input.
>>> There seems to be common mode signals
>>> of an electrical nature coming in on the
>>> geophone input. The only way to balance out
>>> this unwanted signal has been to
>>> make several pairs of identical split resistors
>>> and see which pair will after installed eliminate the problem.
>>> It seems my test equipment can not resolve the measurements
>>> fine enough to properly match these two resistors.
>>> Therefore it is a matter of chance that the right
>>> combination can be achieved.
>>> I have never been able to do this balancing
>>> act with any configuration other than an instrumentation
>>> amplifier.
>>> It is my ignorance in combination with
>>> people who simply refuse to talk about this
>>> which has caused me years of headaches.
>>> In my case the Ge seems to reduce to
>>> (2.99 * 1302)/1742 or 2.234 v/(in/sec)
>>> But this is not how I handle this figure.
>>> I treat it as an overall loss of 20log(2.234/2.99) or -2.53dbv
>>> when calculating the final amplifier gain.
>>> Thanks for your feedback.
>>> Regards,
>>> geoff
>>> -----Original Message-----
>>> From: Bob McClure
>>> Sent: Saturday, June 25, 2011 6:11 PM
>>> To: psnlist@..............
>>> Subject: Re: Damping CDR for HS10-1
>>> For whatever it is worth, here is my computation of the shunt resistance
> to
>>> be applied to the HS-10 geophone to obtain a
>>> damping coefficient of 0.707. It confirms Geoff's latest results, but
> also
>>> allows for the loading provide by the amplifier itself.
>>> HS-10 properties
>>> Sensitivity, E = 2.99 V/ips = 117.7 volts per meter per second
>>> Natural Frequency = 1 Hz = 2*PI radians per second
>>> Natural damping = 0.031
>>> Inertial Mass = 33 oz = 0.936 kilogram
>>> Erhard Wielandt, in his chapter "Seismic Sensors and their Calibration"
> in
>>> the Manual of Observatory Practice
>>> presents a formula for electromagnetic damping.
>>> The formula is h = (E^2 / 2* M * wo * Rd) , where
>>>     E is the output in volt-seconds/meter,
>>>     h is the damping coefficient (0.5/Q),
>>>     M is the effective pendulum mass in kilograms,
>>>     wo is the natural frequency of the pendulum in radians/sec, and
>>>     Rd is the total shunt resistance.
>>> The recommended total damping is 0.707. Since the HS-10 has an open
> circuit
>>> damping of 0.031, we want the electromagnetic
>>> contribution to be 0.707 - 0.031 = 0.676.
>>> so,
>>> Rd = E^2 / (2*h*M*wo) = (117.7)^2 / (2 * 0.676 * 0.936 * 2 * PI) = 1742
> ohms
>>> Let us say the coil resistance is 440 ohms. The input resistance of the
>>> amplifier and its applied shunt resistor must then
>>> equal 1742 - 440 = 1302 ohms. The 1302 value is that of the external
> shunt
>>> resistor in parallel with the amplifier input
>>> resistance.
>>> Say the amplifier input resistance is 10K ohms.
>>> 1/Rext = 1/Rt - 1/Ramp
>>> 1/Rext = 1/1302 - 1/10000 = 0.000768 - 0.000100 =  0.000668
>>> Rext = 1497 ohms
>>> The applied load will reduce the sensitivity of the geophone. The output
>>> will be Rshunt/(Rcoil + Rshunt) times the open
>>> circuit value.
>> __________________________________________________________
>> Public Seismic Network Mailing List (PSNLIST)
>> To leave this list email PSNLIST-REQUEST@.............. with 
>> the body of the message (first line only): unsubscribe
>> See for more information.
> __________________________________________________________
> Public Seismic Network Mailing List (PSNLIST)
> To leave this list email PSNLIST-REQUEST@.............. with 
> the body of the message (first line only): unsubscribe
> See for more information.
> __________________________________________________________
> Public Seismic Network Mailing List (PSNLIST)
> To leave this list email PSNLIST-REQUEST@.............. with 
> the body of the message (first line only): unsubscribe
> See for more information.

Public Seismic Network Mailing List (PSNLIST)

To leave this list email PSNLIST-REQUEST@.............. with 
the body of the message (first line only): unsubscribe
See for more information.


Public Seismic Network Mailing List (PSNLIST)

To leave this list email PSNLIST-REQUEST@.............. with 
the body of the message (first line only): unsubscribe
See for more information.

[ Top ] [ Back ] [ Home Page ]