I wrote the article in rich text and then converted it to a pdf. I'll have a bash at writing it in html, but I'm really slow with html, I find it so confusing, it makes my brain ache. Get it to look right in Safari & its not OK in Opera, &c, &c. But I'll have a go later today.

The notion that the overlapping nth order image spectra make the spikes appear almost as bright as the zero order image occurred to me. I was about to make one after the webmaster in our local astro. soc. brought it to my attention. I have the foam board, the craft knife, and the steel straight edge, and I'm a trained draughtsman. No problem in drawing one and making it. But it was a beautiful sunny day, so I spent it on my sun lounger instead, pondering its design function.

The first amateur astronomer to make use of a coarse diffraction grating to my knowledge was Colin Pither, a member of the Webb Society who invented the diffraction grating micrometer for measuring double stars, back in the 1960's. There is a relationship between the bar/slit spacing and the useful angular separation range of measurable pairs. (Its described in Bob Argyle's Vol.1 Webb Society Double Star handbook).

That's when it occurred to me that all that was required were the three elements of the mask. So why had Pavel turned it into a coarse diffraction grating? Because that would put more energy into the spikes. But at the cost of lower throughput. Does the extra energy in the spikes more than compensate for the lower throughput? That's why I tinkered with the FFT PSF. The normalized PSF certainly confirms Dennis' argument, but when you allow for the energy loss, the spikes are slightly fainter, roughly by a factor of root 2.

Dennis has pointed out Pavel found that the lower throughput did not result in fainter spikes. However I'm not convinced by his argument. That's why I've nominated our local astro. soc. imager, who has run off a mask pattern using the svg file generator, to make his own Bahtinov mask, and then my 'Y' mask, using the same bar width. Maybe the Bahtinov mask will yield brighter diffraction spikes, maybe not. Certainly the 'Y' mask will yield finer spikes, which ought to look more intense as a consequence, plus you have about 90%+ throughput as opposed to 50%-.

What some of you maybe overlooking though is you don't have to make the bars in the 'Y' mask the same width as the equivalent Bahtinov mask, you could make them slightly wider, which makes the spikes shorter but brighter, but still plenty long enough for you to see the same effect as the full Bahtinov mask.

Maybe in average seeing (ą2" - ą4") the bright knots that make up the diffraction spikes are more noticeable. But I would have thought that finer spikes would still appear finer in average seeing, and enable a better judgement as to focus. Bear in mind you have a depth of focus dependent on your 'scope f/ratio and the wavefront error of the objective in any case. It will be interesting to see how the two masks compare in excellent, good and indifferent seeing, on the same telescope.

Anyway, I hope my little Fraunhofer diffraction primer has cleared up a few misconceptions about diffraction and how patterns are produced, and how the Bahtinov mask functions. Whichever focusing mask you prefer, just bear in mind it is the bar/slit width to aperture ratio that determines the brightness and length of the spikes. If you make your Bahtinov mask with wide bars/slits you will get shorter, brighter spikes (with fewer knots) than if you make it with narrow bars/slits. If the slit pitch is twice the slit width, as it ought to be for a coarse diffraction grating, the more slits you have, the greater the number of orders produced, and the greater the extent to which they overlap.