[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Frequency vs temperature

Original poster: "Terry Fritz" <twftesla-at-qwest-dot-net>

Hi Jim,

At 05:37 PM 5/2/2002 -0700, you wrote:
>Looking over your extensive F, R, and Q data...
>How much would you expect the Fres to change over temperature...  You're 
>looking at tiny fractions of a % for frequency, and it seems to me that 
>dimensional changes in your coil and topload will be on that order...

A great thought!  Normally we would not be able to see frequency variation
due to thermal expansion but this things is really sensitive.  The data is
10,000 points long, 9 bits deep and the signal is averaged over 64 shots so
there is a lot of noise reduction.  The raw files are at:


The even files are the scope captures and the odd data is the temperature
probe signal.  A laptop computer picks the data off the scope and stores it
on the hot-streamer computer.  The Sun workstation grabs it and crunches
the data and stores back the nice HTML data file.  Paul's programs do all
the work :-)) 

>Resistance changes over temperature are going to be significant too.. and 
>that slightly changes the apparent resonant frequency (i.e. it's really not 
>=sqrt(1/LC) but a function of LC and R.. although.. your Q of 200 is high 
>enough that this effect might be small...
>A quick and dirty Wheeler calculation, while not accurate in absolute terms 
>to hundredths of a percent, could probably give you a sensitivity... that 
>is, L goes as some function (x/(x+1)?) of Length.... and so 
>forth...obviously N isn't changing...but length and diameter certainly are...

It is a white PVC tube wound with #24 copper wire.  There is no coating.

>Likewise, C of topload will also change as a function of temperature (CTE of 
>aluminum is what?  23 ppm/degree... so a 30C change will be a dimensional 
>change of about 700 ppm, and, to a first order, capacitance will change the 
>same (that is, about 0.07%), resulting in a  0.03% frequency change, which 
>is of the same magnitude as you are seeing...


>The elliptical F vs T curves seem to indicate a thermal/mechanical time lag 
>between parts of the system, that is, some things get bigger and smaller at 
>different rates than other things...
>This is pretty nifty, in that you can actually make reasonably precise 
>measurements (assuming that the error bars are reasonably sized!)

Paul has that stuff figured out so I will defer to him about error margin.
I try not to breath or go near it since it is sensitive to disturbances
from the experimenter ;-)

When rain collects on the roof of the garage, Q1 is obviously affected.
Apparently, loss in the wet roof sucks energy from the coil.  It is
interesting that q3 and q5 seem to be affected by the deeper dampness in
the woodwork or something.  Here are some pictures but they are a little
dated.  I am afraid to go near it to take new ones ;)