Rfc - rc servos + mylar as telescope primary

After a trip to mitzpe ramon and seeing saturn’s rings, moons of jupiter and andromeda in a 30cm telescope I got onto a telescope binge.
This article is interesting as it gets past conventional mirror/lens tech. But back in the world of imaging optics I was pipe-dreaming about feasibility of a large mylar mirror with multiple servos stretching it to some approximation of a parabola, e.g. with a servo every 5 or 10 cm behind a sheet of lets say 2m diameter.
Realistic at all?
Good telescope surface quality is apparently 1/20 wavelength which for 0.6micron is .03micron. 1/4 wavelength or raleigh limit is also some sort of standard.
Anyway to get 0.03micron movements out of a 200steps/rev stepper with a shaft diameter of 3mm (or 0.05mm/step at shaft) I’d need to gear down by a factor of >1000 which already is sounding rather unlikely . Maybe , if harmonic gears can be 3d printed and not bought.

Alternatively a servo , e.g. with feedback keyed on a calibration step? I did see an estimate of 0.05 degree for both servo and stepper.

Anyway the density of servos/steppers is another limiting factor - assuming the sheet is flat between servo points, i’ll lose the parabola shape maximally halfway bet. servos or every 5cm for 10cm spaced servos .
Assuming f/10 which is apparently good for small field-of-view stuff like planet viewing ,
focal length / mirror diam = 10 so focal length would be 20m . Lets put aside the practical matter of dealing with a 20m long focus for the time being ;
The parabola formula for that focal length would be y=1/(4f) x^2 = 1/80 x^2 in meters
The worst error is (i think) near the bottom of the parabola , if the first servo is at origin and next is 10cm away it should fix a height of
y(0.1m) =1/80 * 0.1^2 m = 0.12mm and halfway there should be at
y(0.05m) =1/80 * 0.05^2 m = 0.031mm
but will actually be at 0.06mm (half the 0.12mm height fixed at 10cm), which is already an error 1000 times the 0.03 micron desired.
So it seems servo density is a major limiter here, unless the mylar stretches naturally more towards a squared curve than straight lines between points. I suppose this depends on tension of the mylar membrane, with less tension tending towards more curvy.

Hive mind, what say you ??


This is very interesting , with upcoming neurolinks makes you wonder when they’ll skip rendering for eyes altogether , go straight for feeding the perception.

The pinching method seems too awkward, too many highly sensitive degrees of freedom and moving parts for everything to go wrong (or rather, for at least 1 something to go wrong, and while trying to find what it is break 10 more things)
Perhaps vacuum formed is more realistic in terms of (fixable) aberrations

(vanguard from Oregon doesn’t approve)
Also might be worth looking into this guy (haven’t yet myself)

What if instead of Servos you would use vaccuum? wouldn’t you get more parabola-like curve?

Googles for this idea and this is one quick link:

a video of me from 30 years in the future has surfaced

apparently i changed my name and also figured out how to send videos to the past

piezo actuators seem to be the right kind of thing for this as well as adaptive optics - they’re also pretty cheap. a big array of piezo actuators in a rough parabola would be good start , coated with some kind of stuff that could be figured down to the right shape like maybe pmma or other plastic, and then silvered , would give you a slightly deformable parabolic surface of more or less arbitrary size

Did you see this professor’s keyboard? amazing it still works… :sweat_smile: