Time series photometry problem

[james_screech]

I take time series photometry readings mainly of eclipsing binaries using ASI1600MM-C camera. Initially I used a ED70 refractor without any problems, but more recently I've upgraded to a 200mm Newtonian, this gives me a much greater dynamic range, all well and good. However a couple of time using it I've needed to perform a meridian flip during an observing run and I've noticed a discontinuity in the resulting photmetry graph at the flip. My flats look OK, though no symmetrical. Due I suspect to the  focuser  intruding into the light path and possibly the colimation not being perfect (its the best I can get it). But then that's the whole point of flats to remove these issues. 

As the nights are getting longer now, a meridian flip is becoming more likely in an observing run, are there any suggestions on what could me wrong, and how I can correct it?

Thank you
James 

Edited August 25, 2019 by james_screech

[andrew s]

Did you use a calibration star or stars ? What does it/they do on the flip?

Maybe there is some movement in the Newt on the flip causing a shift the flats don't pick up. 

Regards Andrew 

[james_screech]

4 hours ago, andrew s said:

Did you use a calibration star or stars ? What does it/they do on the flip?

Maybe there is some movement in the Newt on the flip causing a shift the flats don't pick up. 

Regards Andrew 

I use MuniWin software and had a check star selected, using MuniWin's find variables function, it looks like all stars have a similar problem.

I suspect there may be something moving in the imaging train, I just don't know what, as everything appears secure. All screws are tight etc.

Regards

James 

[andrew s]

3 minutes ago, james_screech said:

I use MuniWin software and had a check star selected, using MuniWin's find variables function, it looks like all stars have a similar problem.

I suspect there may be something moving in the imaging train, I just don't know what, as everything appears secure. All screws are tight etc.

Regards

James 

Could just be a shift in the collimation especially if the secondary is undersized.

Can you check it either side of the mount?

Regards Andrew 

[james_screech]

22 minutes ago, andrew s said:

Could just be a shift in the collimation especially if the secondary is undersized.

Can you check it either side of the mount?

Regards Andrew 

As the system is not permanently setup I'm not sure how to do that. I usually check the collimation with a laser collimator during the day in doors, with the OTA not even on the mount but just standing it on the floor.

[andrew s]

45 minutes ago, james_screech said:

As the system is not permanently setup I'm not sure how to do that. I usually check the collimation with a laser collimator during the day in doors, with the OTA not even on the mount but just standing it on the floor.

Just collimate it laying down then turn it "head to tail" i.e long ways and see if it's still collimator.

Regards Andrew 

 

[james_screech]

Collimation didn't appear to change with the the orientation of the OTA. What I did notice when removing the coma corrector / camera was that there was a tiny amount of play (perhaps 1/4mm) between the coma corrector and the eyepiece holder. I've now hopeful fixed that and will see if that was the problem next time.

Thanks for the suggestion Andrew.

Regards
James

Edited August 26, 2019 by james_screech

[Whirlwind]

On an aside how did you get the magnitudes out of Muniwin.  I thought it just gave you relative magntiudes?  I've never been able to see a function where you can astrometrically link your stars?

[james_screech]

21 hours ago, Whirlwind said:

On an aside how did you get the magnitudes out of Muniwin.  I thought it just gave you relative magntiudes?  I've never been able to see a function where you can astrometrically link your stars?

It doesn't astrometrically link stars, it asks for the actual magnitude of the comparison star(s) and calculates it from that.

Regards
James

[Whirlwind]

Sorry, I'm still unclear on how this is done.  I can only see the differential magnitudes as an option. I guess this is selective blindness, where do you get to enter the magnitudes specifically?  I'm just wondering how this is done.  In particular do you state this in two separate files before and after the flip?

[james_screech]

15 hours ago, Whirlwind said:

Sorry, I'm still unclear on how this is done.  I can only see the differential magnitudes as an option. I guess this is selective blindness, where do you get to enter the magnitudes specifically?  I'm just wondering how this is done.  In particular do you state this in two separate files before and after the flip?

When you save the photomtric data (in my vase BAAVSS format) the software asks for various information including the magnitude(s) of the comparison stars.

[Whirlwind]

Ah OK, so if you aren't saving the files in a certain format you might never get to see that option until you've entered your reference details. 

So to return your issue when you create your light curve do you determine the differential magnitudes separately pre- and post- flip or all at the same time.  It's actually not that uncommon where a star has not been relocated back onto the same location to get some discontinuity.  When you stated that every star has the same discontinuity is it the same difference in differential magnitude in each object?

[james_screech]

11 hours ago, Whirlwind said:

Ah OK, so if you aren't saving the files in a certain format you might never get to see that option until you've entered your reference details. 

So to return your issue when you create your light curve do you determine the differential magnitudes separately pre- and post- flip or all at the same time.  It's actually not that uncommon where a star has not been relocated back onto the same location to get some discontinuity.  When you stated that every star has the same discontinuity is it the same difference in differential magnitude in each object?

Everything is done at the same time, the discontinuity isn't exactly the same for all stars. I think as andrew s suggested there was some movement in the optical train. 

[Whirlwind]

That could probably be tested as you should see less difference at larger apertures and if you pick different comparison stars.  To get such a discontinuity you need something to change the relative flux between the objects so that the apertures are no longer suitable to capture equivalent flux between pre and post flip (perhaps because one star is now no longer circular).  If you expand the aperture used then this effect should reduce because you start capturing the flux that has been spread out because of 'some' movement.  If it is, say, due to tilt then selecting comparison stars closer to the target object should also reduce this effect as all these stars will have similar optical defects.  

If none of these resolve the issue then it may not be due to the optical train and perhaps to do with the camera.  If you don't place the star on exactly the same position then if there are any discontinuities on the camera then these would result in an absolute difference in your photometry.  This is why for scientific work most people recommend higher grade cameras (although that's no protection against later discontinuities arising).  It's very common even in professional set ups to see discontinuities where there is movement of the object onto another areas of the CCD and using differential photometry.

[coatesg]

If the plot is relative magnitudes between the reference and the eclipsing binary, then it shouldn't show a discotinuity like this around a flip, even with focuser light path intrusion.

I suspect, like Whirlwind says, there's some mirror flop/loss of focus/collimation that's adversely affecting the profile of the star. If the aperture settings are too tight for one or the other, then this might cause this kind of issue. It doesn't really matter about focus being absolutely bang on for photometry, as long as you have the aperture settings dialled in correctly in the photometry software. If you have some raw data, then I'd be happy to take a look and plot it out.

[james_screech]

23 hours ago, coatesg said:

If the plot is relative magnitudes between the reference and the eclipsing binary, then it shouldn't show a discotinuity like this around a flip, even with focuser light path intrusion.

I suspect, like Whirlwind says, there's some mirror flop/loss of focus/collimation that's adversely affecting the profile of the star. If the aperture settings are too tight for one or the other, then this might cause this kind of issue. It doesn't really matter about focus being absolutely bang on for photometry, as long as you have the aperture settings dialled in correctly in the photometry software. If you have some raw data, then I'd be happy to take a look and plot it out.

Thanks for the offer, the problem would be that the images files are about 40GB. I have wondered if it might be the EL panel I used for flats, so last night I switched back to an old LED panel. I haven't processed those yet, but hope for an improvement.

[james_screech]

I've just analysed my images from last night and they still have the problem, so the issue isn't to do with the flats panel.

If I can't find what is changing soon, I'll go back to using my ED70 for photometry, possibly on my SE4 mount in EQ mode so I won't even need a meridian flip. A nice small setup that is quicker to setup and much lighter to move around. But just lacks the dynamic range in images that I get with the SW200PDS scope.

[coatesg]

On 07/09/2019 at 10:03, james_screech said:

Thanks for the offer, the problem would be that the images files are about 40GB.

Just choose a few from either side of the discontinuity 😉 I'm sure it's not the flats, my guess is something mechanical impacting the stellar profile, bit needs image inspection to be sure.

Edited September 9, 2019 by coatesg

[james_screech]

16 hours ago, coatesg said:

Just choose a few from either side of the discontinuity 😉 I'm sure it's not the flats, my guess is something mechanical impacting the stellar profile, bit needs image inspection to be sure.

OK, attached are a few files either side of the flip and the master flat. I know guiding isn't good on some of them, but that shouldn't effect photometry, am finder guider is pushing guiding a bit when the main scope has a focal length of 900mm.

flip.zip

Edited September 8, 2019 by james_screech

[coatesg]

OK - well I get the same discontinuity:

 but I get it for both the variable, and also doing a check against a few of the comparison stars in the field against each other - but the discontinuity goes the other way in the stars I chose here (and it varies in magnitude according to the reference stars used):

 

 

There's something odd here in that for the variable, using one set of comparison stars, the discontinuity goes one way, whereas you've got it going the other way. I can't quite work out why this would be the case, but there's definitely something odd. Could you try doing single star comparisons against individual reference stars in the field perhaps (from the AAVSO sequences) - that might give a clue as to the scale and direction of the discontinuity when using different comparison stars over the image? Otherwise, I'm a bit stumped...

 

 

Edited September 9, 2019 by coatesg

[andrew s]

Bit of a punt but did you do a 180 rotation of the flat for the set taken after the meridian flip. If the flat has a gradient it might explain things if you did not de-rotate the images before processing.

Regards Andrew 

[sdufoer]

Yes, I had this issue and headache too.  It's your focal plane that is slightly shifting, giving wrong readings for ref/ck stars when the image is turned upside down.  I suspect your imaging train is too heavy and your focuser + tube bends.

In my case it was the primary of the ODK16 that was shifting.

CCDInspector should give you more insight.  Look for tilt in image.

PS: try LesvePhotometry for automated image processing for variables.  It's a huge time saver.

[james_screech]

I also had the discontinuity in different direction with different stars in the field, as said, I suspect movement of something just not sure what as everything is done up tight, but it probably doesn't need much movement to cause the problem.

[JeremyS]

Hello @james_screech, as others have suggested, either a non-uniform flat-field (the images gets rotated 180 deg on flipping, which means the stars in the image fall on a different part of the chip) or movement in the imaging train.

Some discussion on AAVSO here and in their exoplanet manual (search the doc for "merdian")

[james_screech]

6 minutes ago, JeremyS said:

Hello @james_screech, as others have suggested, either a non-uniform flat-field (the images gets rotated 180 deg on flipping, which means the stars in the image fall on a different part of the chip) or movement in the imaging train.

Some discussion on AAVSO here and in their exoplanet manual (search the doc for "merdian")

The images don't need rotating as CMuniWin matches stars without it. So the only thing must be something is moving. As I never had this problem with my refractor I think I will just use the newtonian for visual work and go back to the refractor for photometry.