Owing to difficult sky conditions, only one session has been possible with the 3 nm filters installed. Below are some notes that I made primarilly for myself, but I am happy to share these with you. There are two other issues that I would wish to comment on, and your response would be appreciated:
1. The physical outer diameter of the filter housing seems to be slightly larger than the Baader ones that are being replaced, this has meant that it has not been possible to fully screw them into the filter carousel - especially the Halpha one.
2. I carry out fine focusing using a wide-band IR/UV filter, and assume that the characteristics of the narrowband filters are identical. On examining my raw images collected, I feel as though the focus is just a tiny bit off. Could it be that there is some variation of substrate between manufacturers? Carrying out fine focusing through a narrowband filter will take a good while, which I would wish to avoid.
Initial Comparison of 3nm and 7nm Optical Filters

These notes refer to imaging of the object NGC7023 using the Hɑ 3 nm optical filter in a 1.25 inch format. (This filter and the two others were purchased from FLO, order PO261345). The telescope is a 1200mm F4.7 Newtonian having a Baader Mark-III MPCC Coma Corrector (Also purchased from FLO). The exposure duration was 300 seconds per raw image, and guidance was controlled by the PhD Mk2 software, and over a period of 3 hours - the RMS error for the duration on both axes was 0.8 arc-seconds.

Two sessions using the same target were executed during which 10 raw images were collected for the original 7 nm filter, and then 10 images using the new 3 nm filter Antlia 3nm H-alpha Pro Astronomy Filter. The optical train was unchanged, save for the filter substitution. The camera was an ASI 1600MM PRO with the sensor cooled to -20° C.

To establish the gain from using the 3 nm filter, specific measurements were made on one raw image selected from the middle of each set. The following parameters were obtained using Pixinsight 1.8.9-2:


(all values are x10E-2) 7 nm filter 3nm filter

(a) Mean of entire raw image 2.93 1.30
(b) Mean of a small non-star region (background) 3.06 1.25

(c) Mean of a small region of nebulosity 5.42 3.40


There is literature that sets out the typical spectral distribution of the current generation of LED street illumination, and it shows that the spectral density is fairly uniform across the visible band. I strongly believe it is the scattering of this source that constitutes the predominant background interference at the location of my observatory. This locality is classed as Bortle 6, and is the main reason for investing in the 3nm filter set – the 7 nm set which I have used for many years has not delivered satisfactory images at this location. Consequent;y it seems not unreasonable to expect that the “noise” component of my images to be reduced in the ratio of the respective bandwidths i.e. 7:3 or 2.33 times. Looking at (b) above we see that the measured ratio of the background noise is 2.45. This seems pretty close to expectation.

We can make a further deduction from the data above, on the basis that the background (b) could well be added to those regions where a wanted signal exists, for example, the nebulosity. If the background (c) is subtracted from the nebulosity signal (c), we obtain an estimate of the uncontaminated wanted
7 nm filter 3nm filter
(d) Estimate of uncontaminated nebulosity (c - b) 2.36 2.15
(e) Ratio of wanted to background (d/b) 0.77 1.72

We see some improvement in the signal to noise ratio, which was the objective. When suitable sky conditions return further analysis will be carried out'

IW 15/11/2023.