PDM
Frank Valdes wrote on Sep 02, 1999
Q: I am attempting to use the phase-dipersion minimization tool in IRAF to
fit a collection of variable stars. I have 55 V band observations on four
nights, with about ten months separating the two pairs of nights. The
frequency and period spectra that I get from PDM have several minima and I
can either not get a satisfactory light curve or get degenerate solutions
which I am not experienced enough to discriminate without using a priori
assumptions about period-amplitude relationships.
The help file of PDM gives me the knowledge I need to use the tool, but
the subtleties are difficult. Is there any other documentation of PDM or
someone who can give me some tips on how to squeeze good light curves out
of it?
A: There is no other help documenation for PDM except the IRAF help page.
It contains a reference to the source of the algorithm which might
be useful to you.
About the only way to become proficient is to experiment. It may be
that you need to have some data which is less noisy to see how things
behave. You can make this with cl commands or a script. Below I
give a simple script that can make some sinusoidal test data.
cl> type mkpdm.cl
# Make noise. Use the scale factor to set the noise scale.
urand (200, 1, ndigits=4, seed=1, scale_factor=1., > "urand.dat")
# Add periodic term.
i = 0
list = "urand.dat"
while (fscan (list, y) != EOF) {
i = i + 1
x = 0.1 * i # Coefficient defines period/frequency
y = y + sin (x) + 2 # Add sinusoid. Coefficient is mean level
print (x, y, >> "pdm.dat")
}
list = ""
To create the data:
cl> cl < mkpdm.cl
To run it:
cl> pdm pdm.dat
You could try different levels of noise.
I hope this helps.
Cheers,
Frank Valdes
----- End Included Message -----
Frank Valdes wrote on Sep 03, 1999
Q: I am attempting to use the phase-dipersion minimization tool in IRAF to
fit a collection of variable stars. I have 55 V band observations on four
nights, with about ten months separating the two pairs of nights. The
frequency and period spectra that I get from PDM have several minima and I
can either not get a satisfactory light curve or get degenerate solutions
which I am not experienced enough to discriminate without using a priori
assumptions about period-amplitude relationships.
The help file of PDM gives me the knowledge I need to use the tool, but
the subtleties are difficult. Is there any other documentation of PDM or
someone who can give me some tips on how to squeeze good light curves out
of it?
A: There is no other help documenation for PDM except the IRAF help page.
It contains a reference to the source of the algorithm which might
be useful to you.
About the only way to become proficient is to experiment. It may be
that you need to have some data which is less noisy to see how things
behave. You can make this with cl commands or a script. Below I
give a simple script that can make some sinusoidal test data.
cl> type mkpdm.cl
# Make noise. Use the scale factor to set the noise scale.
urand (200, 1, ndigits=4, seed=1, scale_factor=1., > "urand.dat")
# Add periodic term.
i = 0
list = "urand.dat"
while (fscan (list, y) != EOF) {
i = i + 1
x = 0.1 * i # Coefficient defines period/frequency
y = y + sin (x) + 2 # Add sinusoid. Coefficient is mean level
print (x, y, >> "pdm.dat")
}
list = ""
To create the data:
cl> cl < mkpdm.cl
To run it:
cl> pdm pdm.dat
You could try different levels of noise.
I hope this helps.
Cheers,
Frank Valdes
----- End Included Message -----
Last post on Sep 03, 1999