Short for Primary mirror.
The primary mirror is the first mirror in a telescope to reflect the incoming light. Its focal point is called the prime focus. Its focal length length divided by the diameter is called the focal ratio. The focal ratio is mostly written as f/n, where "n" is a numerical value. (Example: f/8 means that the focal length is eighth times the diameter of the primary.)
The diameter of the primary defines:
- Limiting magnitude (light gathering capability)
- Best possible angular resolution
- Usable magnification range (indirectly)
Its focal length defines:
- Field Of View (FOV)
- Focal length of the eyepieces for the usable magnification range.
Other aspects:
- Weight & thickness
- Shape & accuracy
In order for the mirror not to sag under it's own weight (which would deform its surface, and we're talking about much less than one wavelength !) the mirror needs a minimum thickness. This thickness depends of course of it's diameter, but also on the mirror support, called the cell. The older/simpler cell system, using 3 support points, need a thicker mirror than a cell using 9, 18 or 27 support points. A mirror is said to have "full thickness" if its sagging on a three point cell will not seriously degrade its performance. The relation between diameter and "full thickness" is given below.
The weight of the mirror depends on it's volume. As can be seen from the above relation, the volume will increase dramatically for larger mirrors. This is shown in the table below.
Above 250 mm (10") the weight will severely impact the mechanical design of a scope. Is has become common to use thinner mirrors (also called "thin mirror") with a more elaborate cell. Scopes with a mirror of 500 mm (20") with a thickness of 50 mm (2") have successfully been build. Unfortunately it is rather complicated to calculate the exact position of the support points so that approximations supported by experience are used.
The Newtonian primary is often paraboloidal, for higher f/# it may also be spherical. The paraboloidal is considered the best choice for visual Newtonian. See "Paraboloidal" and "Spherical" for more details.
One of the most debated questions among ATM-ers is the question about the accuracy of the mirror surface. Just how accurate is accurate enough ?. Often the Rayleigh criterion is used to 'deduce' an accuracy figure of 1/8 wavelength. This is IMHO an error. The Rayleigh criterion talks about the TOTAL wave front error. However this total error is made up of a lot of contributions, the primary surface, the secondary surface, the eyepiece errors and last but not least the atmospheric deformation of the wave front. Just allowing the primary to account for the total tolerable error is wrong. The only available conclusion is that the mirror should be as perfect as possible.
This is of course very unsatisfactory. "Fortunately" it takes a lot of observing experience to separate a good from a very good, or a very good from a perfect mirror. It is thus that for a first mirror an accuracy of 1/8 wavelength is indeed acceptable. It is unlikely that casual observers would see a difference between such a mirror and a perfect mirror.
In addition there is the factor "diminishing return". The last 10% of performance needs 90% of the effort. This point is different for everyone, but IMHO it will be shortly after an accuracy of 1/10 wavelength is achieved.