Flat is the goal
The perfect parabolic figure on a mirror as imaged by an interferogram and analyzed by Open Fringe would be a perfectly flat surface. This image shows the progress we have made in previous sessions that has focused on the central hill – now it’s a central ripple. We are leaving that mountain now to focus on the hilly outlands that comprise our astigmatic figure.
To orient yourself in comparison to previous images, you need to stand on your head. Or turn the page upside down. We made a change to the interferometer that results in the image being captured directly through it, instead of reflected off of the beamsplitter surface, which introduces a reflection. The effect in our igram images is that top is at the bottom and there is no left to right reversal. We think this will be less confusing, but time will tell!
In addition to the instrument changes, Steve and Mark have found that they are able to better stabilize the fringes on the camera focal plane by taking a feed of the image in the viewfinder and sending it to a computer monitor. The bigger image makes it easier to see when all is in focus and the fringes are sharper across the face of the mirror. In this run, we got the best quality igrams ever, as judged by the smaller number of unwrap errors.
A brief explanation is in order. The Open Fringe software implements algorithms that trace the fringes from edge to edge. The software has been designed so that it identifies errors in the fringe tracing to give the operator some sense of the quality of the igram – after all, you will be making some possibly momentous decisions about removing glass based on this analysis. The 40-inch F/3.6 mirror we are making presents particular problems because there are so many fringes. Since each fringe represents the difference of a wavelength of light in the height of the surface above some well-chosen plane, and because an F/3.6 mirror is a bowl instead of an F/8 plate, if you take my meaning, there are many more wavelengths of light between the center and the edge on such a fast mirror. More fringes puts them closer together and harder to separate on the camera’s image surface, hence more unwrap errors.
We’re striving for unwrap errors that number less than 1,000. With these latest igrams, we reached 2,500, down from 9,000 in earlier sessions. Once Open Fringe does all of this algorithmic tracing, it then averages and visualizes the surface. In order to make it easier for the eye to detect the errors in the figure that this work is all about, it removes the parabola mathematically. If the surface were perfect, this “artificial null” would make the analyzed image perfectly flat – a plane.
On Monday, Larry, Steve and Mark used the 12 inch tool to work for 30 minutes each on the outer zone high spots. We are pursuing a strategy of local correction instead of wholesale re-spherizing and re-figuring. In the Open Fringe analysis images we are moving toward flat.
Another figure of merit we are interested in improving is the astigmatism that is apparent from the igrams we take in each session (we used eight on this one). This session produced a landmark in the analysis you are looking at because this is the first time that all of the individual igrams have been even visible on this graph. You can see them clustered in the upper right and lower left quadrants. Our next goal will be to move them to the center and to make the number of wavelengths of light distant from that center smaller and smaller. That will tell us that the mirror is flat in all directions from the center – no astigmatism. And that, my children, will tell us it’s time to aluminize.