Re: V455 And and CD Ind...

From: Joe Patterson <*jop_at_astro.columbia.edu*> Date: Wed, 19 Sep 2007 09:30:43 -0400 (EDT) · jop_at_astro.columbia.edu

Hi Arto,

Well, I've been studying the rapidly sampled data we have so far, and the 
fast periods seem to have not yet reappeared.  At first glance, anyway. 
So no information lost, I imagine.  But even a little bump from 17 s (even 
to 15 s) would be prudent.  Or even to 21 s; that sort of hides the faster 
signal, but keeps the 68 s in very plain view.

Some people tend to feel that one must resolve periodic signals with miles 
to spare: e.g. 7-10 bins across the signal.  But actually, if the signal 
is truly stable (this one is, at least at quiescence), BARELY resolving it 
is OK - you pay a little price in amplitude, but you might need to pay it 
for decent signal-to-noise.  Also, most people haven't studied their 
time-keeping, shutter times, etc. enough to be really confident about very
short integrations.  So when in doubt, go a little longer - but avoid 
those unlucky numbers of 17 and 34!

joe

On Wed, 19 Sep 2007, Arto Oksanen wrote:

> Thanks for the input, Joe!
>
> Guess who has been observing with 10 s exposures plus 7 s dead time
> most of the outburst!  And I have been happy for such a short cycle
> time. :-/
>
> I could do 13 s cycles by using a smaller subframe and observing
> unfiltered (whenever weather improves).
>
> br,
>
> arto
>
>
> 2007/9/19, Joe Patterson <jop@astro.columbia.edu>:
>>
>> Dear ,
>>
>> Pardon me for the brevity.  The start of the school year has set me
>> back!  Still, I have been analyzing all the data coming in... and here's
>> my report/suggestions.
>>
>> V455 And = HS2331, as you all know, is having an outburst for the ages.
>>   The coverage is almost as good as that of WZ Sge in 2001, and vastly
>> better than any other dwarf nova in history.  It's great to have bright
>> eruptions from stars transiting near local midnight, and in this case
>> with many periodic clocks to study the effects of outburst.
>>
>> A few of you are obtaining multicolor data; except possibly for the very
>> beginning of outburst, there appear to be no significant color effects
>> in the star.  It's basically a big blue light bulb, and the main reasons
>> for observing it filtered are these:
>>
>> (1) Archival.  With V data, it's possible to say meaningfully "how
>>      bright it is".
>>
>> (2) (Differential) extinction.  Through a very wide or no filter, a CV's
>>
>>      extinction always exceeds that of the comparison star, introducing
>>      unwanted waves in long light curves.
>>
>> (2) is an item of some consequence, and worth doing if you can afford
>> the roughly 6x loss of photons (in which case V is probably the best
>> choice).  But I tend to be skeptical about (1).  CVs are forever
>> variable anyway, and there is likely no significance to their exact
>> brightness.  *Someone* needs to do V measurements; and everyone - even
>> unfiltered diehards - needs to make an estimate of the brightness level.
>> But the sensitivity of time series to periodic signals is greatly
>> improved by a healthy count rate.  So for our purposes, unfiltered
>> generally wins.  Exceptions are a very bright star (like V455 And the
>> last 2 weeks!) and an interest in timescales natural to extinction (say
>> 6-24 hours).
>>
>> There's another very specialized reason that applies to this star.  It
>> has strong 68 and 34 second signals at quiescence.  Their origin is
>> still unclear, but we FERVENTLY would like to measure how they are
>> affected by the outburst.  So we need quite short integrations.  To
>> resolve both, you need a cycle time (integration + dead time) of <17 s,
>> or let's say 13 s to be more comfortable.  Can you do that?  I think for
>> most of you the answer is no; if that's true, try for ~20 s, which will
>> fail to resolve the harmonic properly, but do a decent job on the
>> fundamental.  Avoid, however, cycle times very close to 17 s and 34 s;
>> when the period is an exact multiple of the sampling time, terrible
>> things happen!  (Like trying to measure the Earth's rotation period with
>> brightness measures at noon and midnight every day!)
>>
>> For the moment, V455 And is certainly the best northern object.  To say
>> the least.
>>
>> In the south, it's time to quit on V1432 Aql.  We've covered 1.2 cycles
>> of its 50-day supercycle, done very well (primarily due to Berto), and
>> the data clarify the ephemerides and timescale for synchronization.
>> It's also time to quit on EC2117 ("Ind").  This is a great eclipsing
>> binary, with hardly anything yet published on it; we have excellent
>> data, and one or more of you southerners might want to write something
>> up - I'll be happy to send you all the data and a preliminary analysis.
>>
>> Apart from its faintness, CD Ind is a great, great target this month and
>> next.  This too is an "asynchronous polar", and our data (mainly Bob
>> Rea's) show many effects cycling on the six-day beat between Porb and
>> Prot.  They also specify the long-term ephemeris, which has never yet
>> been published.  This star may well be the most revealing of this small
>> class of CVs (only five known) - revealing of the magnetic torques
>> present in this "asynchronous" geometry.  Try hard for CD Ind!
>>
>> Not so brief... but in a few days, I hope to make up more detailed
>> target suggestions.
>>
>> Please keep up the cba-chat, it's working very well I think!
>>
>> joe
>>
>>
>
>
> -- 
> Arto Oksanen
> arto.oksanen@jklsirius.fi
>
>
>