Good focussing with exceptional seeing: fainter stars

I had good star images based on good seeing of 1.8 arc seconds on Sunday 30th at 3.00 am. These exceptional seeing conditions can result in seeing more fainter stars. This is rather like been on a mountain at a height of 7000 feet. This is well above the average Irish seeing conditions which would be about 3 arc seconds. Long exposure imaging of 360 seconds showed fainter stars as a result of excellent focussing and exceptional seeing conditions.

In order to benefit of getting down to faint magnitudes, assuming good collimation of the optics, there is focus and seeing. This combination can result in an extra 0.5 magnitude. I used a combination of FocusMax.and MaximDL in processing the star images. This was used 5 times and took nearly an hour of processing. Clearly the accuracy of the focus estimate can be improved by repeating the measurement in order to average out the effects of seeing fluctuations, but it is time consuming. Focus routines which are independent of changes in the intrinsic seeing are preferable. From these routines I got the average of the five. Normally I would have processed once. But wanted to see how best I could get the focus.
However much larger errors can occur in determining the correct focus position in the first instance. Rapid changes in the PSF (point spread function) width, due to seeing variations, limit the accuracy with which the optimum focus position can be estimated. The accuracy of this method is limited by changes in the intrinsic seeing during the focus run. Substantial seeing fluctuations may be expected on short timescales.

I now use this routine every 30 minutes, which takes about a minute. As the temperature of the night drops, so the focus position changes. If you are doing long runs of imaging example of about 5 hours, the Optec TCF focuser has this advantage.

Eamonn A

Hi Eamonn,

Very interesting report. This is something I have given some thought to while imaging. All the software I use for focussing have one thing in common, they use the Full width half modulation FWHM method to determine the preciseness of the focus. As you mention the seeing conditions often mean that the values given over short period of time for the same focus position can vary greatly. Even guiding/tracking issues or wind can throw these values out by a long way. I normally choose a star which is only bright enough to give a resonable image over about 10 seconds or so. The result of this is that the seeing tends to averagre over that time.

I wondered about the possibility of writing some software which would take a series of these images analyse them individually and return the average FWHM, point spread value and/or peak intensity over the series. Of course this would increase the time it takes to focus but anything that results in better focus will result in better imagery whether and make for far more usefulscientific data or even just prettier pictures.

OTOH it should be a simple matter to use this data as part of an autofocus routine which could iterate to the best possible focus.

Your statement of an extra half magnitude is interesting. It may not sound much but from a scientific point of view can be the difference in getting astrometric data on a faint object or not. I dont have data to hand but i'm sure I could do some quick math which would demonstrate that this extra capacity would have a more significant impact on the s/n ratio for a stated image or set of images.

Fascinating stuff!

Eamonn/Dave

I would have thought that the seeing has nothing to do with focus? Unless of course you focus when the seeing is particularly poor - then you don;t really know for sure if you are focused or not...

Bus assuming you are at correct focus and the seeing is variable, at some point you will be in focus.

However, if you are not at focus and the seeing varies, then even if the seeing becomes crystal clear, you will still not be focused...

Am I on the right track?

Not having the advantage of a temperature compensated auto focuser (yet ) but having researched what's out there, it seems that major image capture software packages have routines for focusing. I think there may also be some standalone packages.

AFAIK, they allow you to train the software (especially if you have a temperature compensating focuser) to achieve focus automatically (e.g for unattende/robotic imaging).

I don't know if the packages allow an averaging over a series of imges though. But I would have thought that you may not want this anyhow. If seeing is variable, you'll just get an average focus position which may or may not be the actual correct focus...

I have heard that some folks actually de-fous slightly to minimise the effects of a night of poor/variable seeing.

As Dave says, it sure is another area of imaging that is fraught with difficulty and complexity....

Cheers

Dave