PSN-L Email List Message

Subject: Re: pictures help
From: ChrisAtUpw@.......
Date: Sun, 3 Apr 2005 23:25:05 EDT


In a message dated 03/04/2005, 1goss@........... writes:

> This is a screenshot of WinSDR It shows the 6 sec microseisms I think?
> https://home.comcast.net/~bryangoss/winsdr.jpg

Hi Bryan,

       The time scale given is four minutes. With a microseism period of ~6 
sec, you should be looking at ~40 cycles total. This could fit the faster 
vibrations or it could all be mostly noise. There is certainly a lot of longer 
period stuff in with it. 

> This is the boom pivot point a pipe union and razor blade
> https://home.comcast.net/~bryangoss/pipeunion.jpg  is the ball bearing 
> better than the razer blade?

       Razor blades do not tend to last very long. The idea seems to have 
arisen from the old system of knife edge bearings for lab balances. These used 60 
deg triangular agate bars resting on a flat agate plate. This only allowed 
about 200 gm load and the edge of the triangle had a tiny radius lapped onto it. 
A razor blade is about 15 deg and very sharp, so the edge load is huge and 
you are asking it to take > 2 kgm. Umm. Stainless ball bearings last for ~ever. 
You should be able to get >20 sec response without any problems

       You can buy stainless steel ball bearings ~3/16", which do not 
corrode. For a flat, you can use a tungsten carbide tool insert bit for a lathe. You 
can get flat triangular bits with about 0.3" sides, quite cheaply. If you 
epoxy the bearing into a V hole made with centre drill on the vertical support, 
turn the end of the boom rod flat and epoxy the tool bit onto this end, it 
should work very well. If you mount the ball in the end of the swing arm, you will 
have to do a major levelling operation every time you dismount the arm.....

> This is the damping
> https://home.comcast.net/~bryangoss/damping.jpg

       You shouldn't need oil AND magnetic damping! The magnet set-up you 
have should give both vertical and horizontal damping! I suggest that you remove 
the oil and then reduce the magnet separation till you get just sub-critical 
damping? If you can't get enough damping, use a 1/8" thick Cu plate. If you 
slowly deflect the arm 1/2" to one side and then release it, it should just go 
through zero by 1/16" or less and then fall back to zero. 
       There are several problems with oil damping. It is very temperature 
sensitive, relatively difficult to set up, it is messy due to oil creep and it 
collects bugs - giving 'bug-quakes'! It is also 'non linear' in that it 
selectively damps short period motion much more than long period motion. The 
professionals stopped using it over 50 years ago....
       I use a quad NdFeB bar magnet damping set-up. Two 1/4" thick mild 
steel plates 3.5" long by 2" wide are held apart by four 1/4" mild steel bolts and 
additional nuts at each corner. A NS pair of rectangular NdFeB magnets, 1" x 
1/2" x 1/4" are placed on the centre of the face of one plate with the long 
axis parallel to the 2". A pair of SN magnets are placed on the opposing face of 
the other plate and the separation is adjusted to about 1/8" to 3/16". The 
central field gradient is simply huge. A 1/16" or 1/8" thick plate of soft Al or 
Cu about 2" wide and 3" long is mounted on the arm. It is important to make 
the damping plate wider than the 1" of the magnets to avoid field edge effects. 
The main damping is over the central 1" NS magnet joint. If you mount a L 
shaped Cu blade horizontally, you can just slide the magnet array further over it 
to increase the damping. You also reduce the damping by increasing the 
central magnetic gap. The stray field is low with this design. I put gaffer tape on 
the magnet faces and then peel it off before I assemble the two plates, to 
remove any magnetic debris / whiskers. You always get some and it is very 
difficult to see...

> The bolt and 25mm wire
> https://home.comcast.net/~bryangoss/boltwire.jpg
> 
> The 5lb weight and coil
> https://home.comcast.net/~bryangoss/coil.jpg

       If you mount the sensor magnet on the arm, it will respond to changing 
magnetic fields, the earth's field, power line surges, fridge motors, even to 
passing cars and lorries. If you mount the coil on the arm, it will be ~free 
of these interfering signals. You may detect movements of a few millionth's of 
an inch!

       The effective length of the arm is a bit shorter than the distance 
between the centre of the mass and the knife edge. You want this to be ~30 to 40" 
to give a long period. You might mount the mass quite a bit further along the 
arm?

       It is usual to attach the top suspension wire to a ~3" high vertical 
extension plate above the screw arm. This gives a small vertical pendulum 
effect and prevents the arm from rotating when it senses movement. This is 
necessary if you use a ball or wire suspension.

> The overall I put the Seismograph in a clear box to prevent draft bit I 
> need to seal off the bottom this picture shows entire system and my messy 
> shop!!!!!
> 
>   https://home.comcast.net/~bryangoss/overall.jpg

       This looks OK, but you probably need to seal the edges of the case. 
You can buy 3" wide clear sticky tape for sealing greenhouse windows. This will 
take sunlight and it does not loose adhesion with time. Don't use cellotape, 
or ordinary pvc tape - they won't last long and you only want to do the job 
once!

    I would use two longitudinal L angles to support the whole base 
suspension, the sensor and the damping array plates. I use a single bolt at the centre 
of the mass end of the rig to alter the tilt angle. This makes adjustments a 
whole lot easier. You can set up each section / clearance in turn and the 
adjustments don't interact much. 

>   https://home.comcast.net/~bryangoss/setup.jpg
> 
   I would mount the computer and monitor further away from the seismometer. 
You will certainly see a signal when you turn the monitor on, due to the large 
magnetic pulse used to demagnetise the screen.

> Chris I am not sure how to do the FFT. I hope the screenshot of winsdr will 
> help, Thanks for the help If you all see something to improve I will try to 
> do it. 
> Thanks Very Much Bryan S Goss Corinth MS

       If you are using WinQuake, there is a tab labelled FFT on the top tool 
bar. You place the cursor where you want to start on the trace and click on. 
There should be a help file if you look for it.

       Hope that this is of some help! Are there two Corinths in MS?

       Regards,

       Chris Chapman
In a message dated 03/04/2005, 1gos=
s@........... writes:

This is a screenshot of WinSDR=20= It shows the 6 sec microseisms I think?
https://home.comc= ast.net/~bryangoss/winsdr.jpg


Hi Bryan,

       The time scale given is four minutes. W= ith a microseism period of ~6 sec, you should be looking at ~40 cycles total= .. This could fit the faster vibrations or it could all be mostly noise. Ther= e is certainly a lot of longer period stuff in with it.

This is the boom pivot point a=20= pipe union and razor blade
https://home.c= omcast.net/~bryangoss/pipeunion.jpg  is the ball bearing better tha= n the razer blade?


       Razor blades do not tend to last very=20= long. The idea seems to have arisen from the old system of knife edge bearin= gs for lab balances. These used 60 deg triangular agate bars resting on a fl= at agate plate. This only allowed about 200 gm load and the edge of the tria= ngle had a tiny radius lapped onto it. A razor blade is about 15 deg and ver= y sharp, so the edge load is huge and you are asking it to take > 2 kgm.=20= Umm. Stainless ball bearings last for ~ever. You should be able to get >2= 0 sec response without any problems

       You can buy stainless steel ball bearin= gs ~3/16", which do not corrode. For a flat, you can use a tungsten carbide=20= tool insert bit for a lathe. You can get flat triangular bits=20= with about 0.3" sides, quite cheaply. If you epoxy the bearing into a V hole= made with centre drill on the vertical support, turn the end of the boom ro= d flat and epoxy the tool bit onto this end, it should work very well. If yo= u mount the ball in the end of the swing arm, you will have to do a major le= velling operation every time you dismount the arm.....

This is the damping
https://home.com= cast.net/~bryangoss/damping.jpg


       You shouldn't need oil AND magnetic damping! The magnet set-up you have should give both verti= cal and horizontal damping! I suggest that you remove the oil and th= en reduce the magnet separation till you get just sub-critical damping? If y= ou can't get enough damping, use a 1/8" thick Cu plate. If you slowly deflec= t the arm 1/2" to one side and then release it, it should just go through ze= ro by 1/16" or less and then fall back to zero.
       There are several problems with oil dam= ping. It is very temperature sensitive, relatively difficult to set up, it i= s messy due to oil creep and it collects bugs - giving 'bug-quakes'! It is a= lso 'non linear' in that it selectively damps short period motion much more=20= than long period motion. The professionals stopped using it over 50 years ag= o....
       I use a quad NdFeB bar magnet damping s= et-up. Two 1/4" thick mild steel plates 3.5" long by 2" wide are held apart=20= by four 1/4" mild steel bolts and additional nuts at each corner. A NS pair=20= of rectangular NdFeB magnets, 1" x 1/2" x 1/4" are placed on the centre of t= he face of one plate with the long axis parallel to the 2". A pair of SN mag= nets are placed on the opposing face of the other plate and the separation i= s adjusted to about 1/8" to 3/16". The central field gradient is simply huge= .. A 1/16" or 1/8" thick plate of soft Al or Cu about 2" wide and 3" long is=20= mounted on the arm. It is important to make the damping plate wider than the= 1" of the magnets to avoid field edge effects. The main damping is over the= central 1" NS magnet joint. If you mount a L shaped Cu blade horizontally,=20= you can just slide the magnet array further over it to increase the damping.= You also reduce the damping by increasing the central magnetic gap. The str= ay field is low with this design. I put gaffer tape on the magnet faces and=20= then peel it off before I assemble the two plates, to remove any magnetic de= bris / whiskers. You always get some and it is very difficult to see...

The bolt and 25mm wire
https://home.comcast.net/~bryangoss/boltwire.jpg

The 5lb weight and coil
https://home.comcas= t.net/~bryangoss/coil.jpg


       If you mount the sensor magnet on the=20= arm, it will respond to changing magnetic fields, the earth's field, power l= ine surges, fridge motors, even to passing cars and lorries. If you mount th= e coil on the arm, it will be ~free of these interfering signals. You may de= tect movements of a few millionth's of an inch!

       The effective length of the arm is a bi= t shorter than the distance between the centre of the mass and the knife edg= e. You want this to be ~30 to 40" to give a long period. You might mount the= mass quite a bit further along the arm?

       It is usual to attach the top suspensio= n wire to a ~3" high vertical extension plate above the screw arm. This give= s a small vertical pendulum effect and prevents the arm from rotating when i= t senses movement. This is necessary if you use a ball or wire suspension.

The overall I put the Seismogr= aph in a clear box to prevent draft bit I need to seal off the bottom this p= icture shows entire system and my messy shop!!!!!

  https://h= ome.comcast.net/~bryangoss/overall.jpg


       This looks OK, but you probably need t= o seal the edges of the case. You can buy 3" wide clear sticky tape for seal= ing greenhouse windows. This will take sunlight and it does not loose adhesi= on with time. Don't use cellotape, or ordinary pvc tape - they won't last lo= ng and you only want to do the job once!

    I would use two longitudinal L angles to support the whol= e base suspension, the sensor and the damping array plates. I use a single b= olt at the centre of the mass end of the rig to alter the tilt angle. Thi= s makes adjustments a whole lot easier. You can set up each section / cl= earance in turn and the adjustments don't interact much.

  https://home.comcast.net/~bryangoss/setup.= jpg

   I would mount the computer and monitor further away from the s= eismometer. You will certainly see a signal when you turn the monitor on, du= e to the large magnetic pulse used to demagnetise the screen.

Chris I am not sure how to do t= he FFT. I hope the screenshot of winsdr will help, Thanks for the help If yo= u all see something to improve I will try to do it.
Thanks Very Much Bryan S Goss Corinth MS


       If you are using WinQuake, there is a t= ab labelled FFT on the top tool bar. You place the cursor where you want to=20= start on the trace and click on. There should be a help file if you look for= it.

       Hope that this is of some help! Are the= re two Corinths in MS?

       Regards,

       Chris Chapman

[ Top ] [ Back ] [ Home Page ]