This site will not recommend a specific unbranded part, and the reason is not snobbery. Every length here was read from a manufacturer's documentation, and a clone's length is a claim nobody has checked. If the stack we give you is built from a tube that is 16.2 mm rather than 16.5, the number at the top of the page is wrong for you — and you would have no way of knowing which of us was at fault. What this page can do instead is show you how to check one yourself.
What actually differs
- Thread pitch. M42×0.75 and M42×1 share a diameter and will not mate. Branded astronomy parts are almost always ×0.75; general photographic adapters often are not, and forcing them damages both.
- Machined length against stated length. This is the one that matters, and the one nobody publishes a tolerance for.
- Concentricity and squareness. A tube whose faces are not parallel introduces tilt, which looks exactly like a spacing error in the corners — and stacks worse the more pieces you add.
- Whether the threads bottom out properly. A ring that does not seat fully adds an unknown amount of length, and it will not be the same amount twice.
The measurement that counts is not the obvious one
Put a threaded spacer in calipers end to end and you will read too much. The male thread screws INTO the next part until the two shoulders meet, so the length it contributes to your train is the distance between its two mating faces when seated — not its overall length including the protruding thread.
On a typical adapter that difference is four or five millimetres, which is roughly ten times the tolerance you are working to. Measure shoulder to shoulder: the flat face that meets the part in front, to the flat face that meets the part behind.
The reliable check is to assemble the stack and measure the whole thing in one go, rather than measuring pieces and adding them up. Errors that hide individually show up once.
When a clone is fine, and when it is not
- Fine: a single pass-through spacer, on a slow scope with a generous tolerance, where one tenth of a millimetre either way disappears into the margin.
- Fine: anything you have measured and found true. A checked cheap part is better than an unchecked expensive one.
- Risky: fast optics. Tolerance tightens as the focal ratio drops, and at f/4 you have very little room.
- Risky: one of four pieces. Four parts each 0.15 mm out in the same direction is 0.6 mm, which is outside most correctors' tolerance on its own.
- Risky: anything carrying a thread transition. An adapter that is also the joint between two thread sizes is doing two jobs, and squareness matters more there than anywhere else in the train.
The honest summary
Branded parts are not magic — you are paying for a length that has been held to a tolerance, and for faces that are square. On a forgiving setup that is worth little. On a fast astrograph with four pieces in the train it is worth quite a lot. Either way, the thing that actually removes the doubt costs about the same as one branded spacer, and you only buy it once.