Hi Brett
I have been running three sensors with the smt8 type=A0triple feedback. The=
 ~30" vertical with woofer coil and LVDT seems to work very well. I oscilla=
te the LVDT with a 6500hz phase shift oscillator. I have a 10 hz single pol=
e passive low pass filter after the maxim dg419 demodulator. I then run it =
thru a gain of 5 and into the triple feedback. The ~12" leaf spring vertica=
l uses the same 6500 hz oscillator with a VRDT=A0 and STM=A0home built =A0f=
eedback coil. I have been getting a large very low frequency oscillation ar=
ound 0.005 hz. Looking at the winquake helicorder it doesn't last all day (=
maybe 2-10 hours at a time) . I have been trying to see if it's environment=
al but it doesn't appear to occur at the same time periods each day.Could t=
his oscillation have anything to do with the feedback? For others using the=
 VRDT is degaussing the center element very important and does this have to=
 be done more than once? I did make a circuit diagram using easytrax
 which I can email you. More questions later.
Regards
Barry
=A0
=A0
--- On Thu, 9/18/08, Brett Nordgren  wrote:

From: Brett Nordgren 
Subject: Re: VRDT Experience
To: psn-l@..............
Date: Thursday, September 18, 2008, 4:43 PM

Chuck,

The biggest drawback that I can see is that they don't seem to do too well=
=20
with higher drive frequencies.  I usually see them run at 5-6kHz or so,=20
while some other sensor types are happy at much higher frequencies.  After=
=20
you demodulate the signal, it contains a large second harmonic component=20
along with the higher even harmonics, and you need to filter it to recover=
=20
the baseband seismic signal.  But in a feedback seismograph, that filter=20
can introduce phase shifts which tend to make the loop oscillate.

If you move your carrier up to 20kHz, you can design a filter which cuts=20
off at a higher frequency and which, as a result, doesn't add as much phase

shift in the 10-30 Hz region where loop oscillations are an issue.  The=20
STS-2 uses a 20kHz carrier and a 3rd order Bessel filter designed to cut=20
off at 1600 Hz.  That filter adds less than two degrees of phase at=20
30Hz--practically nothing, but it reduces the 40kHz 2nd harmonic to below=
=20
0.018% of its starting value.

With a 5kHz carrier it becomes harder to fit the filter between the high=20
frequency corner at 10-30Hz (gain crossover) and the 10khz second=20
harmonic.  So I am not that fond of magnetic transducers.

If you do want to go that route and use 5kHz, a 3rd order filter that cuts=
=20
off somewhere around 400Hz should be about optimum.  It would add about 5=
=20
degrees at 20Hz, not too bad, and it should nicely filter the harmonics.

Does anyone know of a nice 400Hz 3rd order, low noise, Bessel filter design=
=20
to use here?

Regards,
Brett



Hi Brett
I have been running three sensors with the smt8 type triple feedback. The ~30" vertical with woofer coil and LVDT seems to work very well. I oscillate the LVDT with a 6500hz phase shift oscillator. I have a 10 hz single pole passive low pass filter after the maxim dg419 demodulator. I then run it thru a gain of 5 and into the triple feedback. The ~12" leaf spring vertical uses the same 6500 hz oscillator with a VRDT  and STM home built  feedback coil. I have been getting a large very low frequency oscillation around 0.005 hz. Looking at the winquake helicorder it doesn't last all day (maybe 2-10 hours at a time) . I have been trying to see if it's environmental but it doesn't appear to occur at the same time periods each day.Could this oscillation have anything to do with the feedback? For others using the VRDT is degaussing the center element very important and does this have to be done more than once? I did make a circuit
 diagram using easytrax which I can email you. More questions later.
Regards
Barry
 
 
--- On Thu, 9/18/08, Brett Nordgren <brett3nt@.............> wrote:

From: Brett Nordgren <brett3nt@.............>
Subject: Re: VRDT Experience
To: psn-l@..............
Date: Thursday, September 18, 2008, 4:43 PM

Chuck,

The biggest drawback that I can see is that they don't seem to do too well 
with higher drive frequencies.  I usually see them run at 5-6kHz or so, 
while some other sensor types are happy at much higher frequencies.  After 
you demodulate the signal, it contains a large second harmonic component 
along with the higher even harmonics, and you need to filter it to recover 
the baseband seismic signal.  But in a feedback seismograph, that filter 
can introduce phase shifts which tend to make the loop oscillate.

If you move your carrier up to 20kHz, you can design a filter which cuts 
off at a higher frequency and which, as a result, doesn't add as much phase

shift in the 10-30 Hz region where loop oscillations are an issue.  The 
STS-2 uses a 20kHz carrier and a 3rd order Bessel filter designed to cut 
off at 1600 Hz.  That filter adds less than two degrees of phase at 
30Hz--practically nothing, but it reduces the 40kHz 2nd harmonic to below 
0.018% of its starting value.

With a 5kHz carrier it becomes harder to fit the filter between the high 
frequency corner at 10-30Hz (gain crossover) and the 10khz second 
harmonic.  So I am not that fond of magnetic transducers.

If you do want to go that route and use 5kHz, a 3rd order filter that cuts 
off somewhere around 400Hz should be about optimum.  It would add about 5 
degrees at 20Hz, not too bad, and it should nicely filter the harmonics.

Does anyone know of a nice 400Hz 3rd order, low noise, Bessel filter design 
to use here?

Regards,
Brett




[ Top ] 
[ Back ] 
[ Home Page ]