PSN-L Email List Message

Subject: Re: Seismograph noise problem
From: ChrisAtUpw@.......
Date: Mon, 11 Feb 2008 18:53:34 EST


In a message dated 11/02/2008, lconklin@............ writes:

Regarding corrosion under solder joints, I haven't really checked  
thoroughly for that yet, but the board looks very clean.
Hi Larry, 
 
    I had a whole Sony TV with crevice  corrosion. The joints looked perfect, 
but you could peel them off a black  corrosion coat on the circuit strips. I 
had to unsolder, clean and resolder  every joint on the **** board.

Re. your  and Roger's comments about the cover, It will take me a little 
time to  build a new one. 
    Putting the cover on the floor isolates the  seismometer from pneumatic 
effects. I use 2" Celotex, but it isn't cheap. Bubble  wrap can also be very 
useful in reducing temperature changes and  drafts ....

One  other thing that I'd be interested in hearing your thoughts on is 
the fact  that the "noise" isn't really as random as I'd expect from 
something like  bad solder joints and such.  If you look at the data I 
posted,  especially from the LF channel, there definitely seems to be a 
dominant  component with a period of around 20 seconds or so.
    I had noted that. It looked as if you were  experiencing greatly 
increased gain, rather than just random noise. This could  be resistors, solder 
joints, diodes or the opamp.
    Do measure the DC levels on TP. Does the Red/Green  LED ever light up?
    The NE5534 does take about 0.5 micro A to drive it  - quite thirsty...
    Try tapping the components?
    Unsoldering C49 would isolate everything downstream  of the first two 
opamps.
    Could there be any strong radio signals at 2 / 4 /  8 / 12 MHz? You do 
have two radio receivers on the input... Electricity Utility  time switches work 
off radio signals on the power lines.
    Try connecting an audio amplifier to the circuit  before the integrator 
and listening with headphones? I have solved some noise  problems this way. I 
was getting quite large random pulses which seemed to be  real, not 
instrumental. When I listened, I heard a heavy lorry approaching a  sunken drain grating 
on the corner of the local main road....

I've got  too many irons in the fire right now to devote full time to 
debugging this,  but I plan to revisit all of your suggestions.  Aside with messing 
with  the cover a little, the other very easy test will be 
to disconnect the  oscillator to see what happens.  I still have an old test 
data file from  the last time I tried it, and it doesn't look much different 
from what I saw  then with everything hooked up, or now.
    I would expect there to be a simple fault which is  sensitive to humidity 
- since heating the board reduced the signal. Good  Luck!
 
    Regards,
 
    Chris Chapman







   





In a message dated 11/02/2008, lconklin@............ writes:
<= FONT=20 style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000=20 size=3D2>Regarding corrosion under solder joints, I haven't really checked= =20
thoroughly for that yet, but the board looks very clean.
Hi Larry,
 
    I had a whole Sony TV with crevice=20 corrosion. The joints looked perfect, but you could peel them off a bla= ck=20 corrosion coat on the circuit strips. I had to unsolder, clean and reso= lder=20 every joint on the **** board.
<= FONT=20 style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size= =3D2>Re. your=20 and Roger's comments about the cover, It will take me a little
time to= =20 build a new one. 
    Putting the cover on the floor isolates the=20 seismometer from pneumatic effects. I use 2" Celotex, but it isn't cheap. Bu= bble=20 wrap can also be very useful in reducing temperature changes and=20 drafts ....
<= FONT=20 style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size= =3D2>One=20 other thing that I'd be interested in hearing your thoughts on is
the=20= fact=20 that the "noise" isn't really as random as I'd expect from
something l= ike=20 bad solder joints and such.  If you look at the data I
posted,=20 especially from the LF channel, there definitely seems to be a
dominan= t=20 component with a period of around 20 seconds or so.
    I had noted that. It looked as if you were=20 experiencing greatly increased gain, rather than just random noise. This cou= ld=20 be resistors, solder joints, diodes or the opamp.
    Do measure the DC levels on TP. Does the Red/Gr= een=20 LED ever light up?
    The NE5534 does take about 0.5 micro A to drive= it=20 - quite thirsty...
    Try tapping the components?
    Unsoldering C49 would isolate everything downst= ream=20 of the first two opamps.
    Could there be any strong radio signals at 2 /=20= 4 /=20 8 / 12 MHz? You do have two radio receivers on the input... Electricity Util= ity=20 time switches work off radio signals on the power lines.
    Try connecting an audio amplifier to the circui= t=20 before the integrator and listening with headphones? I have solved some nois= e=20 problems this way. I was getting quite large random pulses which seemed to b= e=20 real, not instrumental. When I listened, I heard a heavy lorry approaching a= =20 sunken drain grating on the corner of the local main road....
<= FONT=20 style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size= =3D2>I've got=20 too many irons in the fire right now to devote full time to debugging this= ,=20 but I plan to revisit all of your suggestions.  Aside with messing wi= th=20 the cover a little, the other very easy test will be
to disconnect the= =20 oscillator to see what happens.  I still have an old test data file f= rom=20 the last time I tried it, and it doesn't look much different from what I s= aw=20 then with everything hooked up, or now.
    I would expect there to be a simple fault which= is=20 sensitive to humidity - since heating the board reduced the signal. Good=20 Luck!
 
    Regards,
 
    Chris Chapman

 

[ Top ] [ Back ] [ Home Page ]