Why use glass in a reflecting telescope?

[DrZygote214]

I was just reading about some very large observatories being planned or under construction, and how these huge mirrors require up to 6 months to cool, or the glass will crack.

So I have to ask the question, why use glass at all? My understanding of mirrors (for reflecting telescopes) is that they put the reflective coating on top of the glass, and it's that reflective coating that does all the reflection.

This is different than a normal household mirror, which puts the reflective coating underneath the glass for protection, so then of course you need something transparent like glass. But I'm asking about telescopes, which put the coating on top.

So why not just use steel, or some other stalwart material, and put the reflective coating on that?

 

[origin]

Good question. The reason is the Coefficient of Theremal Expansion (CTE). Glass does not expand or contract much with temperature in genral and the glass that is used in telescopes has a CTE very close to zero in the operational temperature range. You can imagine that a small change in shape of the mirror would be a big problem for the image quality.

There are other materials that can be used but glass is still the best at this point.

 

[DrZygote214]

I see. I was afraid it might have something to do with how the glass is grinded and made into an accurate shape, like maybe that's more difficult with other materials? But from what I've read about glass grinding, it seems not relatively easy anyway.

This does raise other questions, however. What about aerogel? Or carbon-carbon? Too expensive?

Also, we have adaptive optics nowadays, that vary the position and angle of the individual hexagon mirrors to adjust for atmospheric seeing. Is it possible to program in adjustments for thermal expansion/contraction?

 

[Billy T]

why not just use steel, or some other stalwart material, and put the reflective coating on that?

There have been a few attempts to make parabolic reflecting surface by spinning liquid mercury. When grinding glass, the surface shape easily produced is spherical, not parabolic as a spinning liquid's is. Unfortunately, it is very hard to spin a liquid without exiting tiny surface oscillations. What is needed is "highly viscous" mercury. Also that surface is horizontal so unless willing to only look straight up, you need a flat mirror* to make light fall vertically on to the Hg mirror. Looking straight up still lets you scan much of the heavens. For example + or - ~23.5 degrees in six months ( Tropic of Caner to Tropic of Capricorn is straight up, but moving so only very short exposures; however nearly linear "star-streak" lines could be reflected onto the entrance slit of a spectrograph.)

Near zero thermal expansion glass is used because the glass must be very strong / thick if large to not distort under its own weight.** That means it is difficult to expose it to the night air's thermal changes without shape distortion - interior parts at different temperature than the surface makes for distorting forces. Thermal expansion its self is not a problem, if the temperature could be uniform through out the mass, but differential temperatures can not be avoided.

Now days, there is little need of large single piece mirrors. Automation can keep surface of many separate ones on the same parabolic surface.

* It is very cheap to make strong flat mirror - just slowly cool moltent glass. "Plate glass" can have a very flat surface if this is done.

** the back side of the glass is many hexagon holes to reduce weight, help keep interior at same temperature, and yet have very stiff / rigid mirror.

 

[origin]

Near zero thermal expansion glass is used because the glass must be very strong / thick if large to not distort under its own weight.

Most mirrors are of the 'ultra-light weight" versions so they do not have to be so thick, such as the hubble mirror. I have some knowledge about this because my wife worked on that fabrication of the Hubble mirror.

Here is a picture of a ultra-light weight mirror.

edit to add: Glass mirrors are typically ground to a paraboloid shape and not spherical.

 

[Billy T]

Most mirrors are of the 'ultra-light weight" versions so they do not have to be so thick, such as the hubble mirror. I have some knowledge about this because my wife worked on that fabrication of the Hubble mirror....

Thanks -You posted about when I was adding foot note to say the same.

PS was that job at Perk& Elmer? If so, did she known they faked the tests that would have saved US many millions. - No need for astronauts to later add the "corrector plate."

 

[DrZygote214]

Thermal expansion its self is not a problem, if the temperature could be uniform through out the mass, but differential temperatures can not be avoided.

In that case, can't some sort of heating system be applied? It wouldn't have to keep it very warm, just uniformly warm. Are observatories really so budget-constrained to be afraid of the electric bill involved with heating the primary mirror? Or maybe such heat would cause more atmospheric seeing near the mirror?

My next idea will be to use steel, but surrounded by a thin layer of glass to protect it from fast thermal variation.

Most mirrors are of the 'ultra-light weight" versions so they do not have to be so thick, such as the hubble mirror. I have some knowledge about this because my wife worked on that fabrication of the Hubble mirror.

Cool, was it the real mirror, or the backup? Also, does she know anything about how or why the main mirror got incorrectly shaped?

 

[Billy T]

In that case, can't some sort of heating system be applied? ...

Conceptually yes, but how can you keep the mirror surface both exposed to star light and uniformly warm.

 

[origin]

In that case, can't some sort of heating system be applied? It wouldn't have to keep it very warm, just uniformly warm.

That would be incredibly difficult.

My next idea will be to use steel, but surrounded by a thin layer of glass to protect it from fast thermal variation.

You would have to match the CTE of the glass to the steel or the glass would shatter when it went into tension. So now you have a glass and steel mirror that will be changing shape all the time. Remember these telescopes have to be outside. At least the end of the telescope needs to be outside (even worse for delta T).

Cool, was it the real mirror, or the backup?

I do not know the history of that particular mirror blank but I assume it was a reject.

Also, does she know anything about how or why the main mirror got incorrectly shaped?

The grinder/polisher did not take into account the deflection of the blank from gravity, so when it was launched into orbit and no longer affected by gravity the shape changed slightly and the focus point was off by a small amount and could not focus properly. It was fixed on a subsequent repair flight.

 

[DrZygote214]

Conceptually yes, but how can you keep the mirror surface both exposed to star light and uniformly warm.

Wait, are you saying that starlight, the points of light, cause a significant thermal variation on that point of the mirror? That would be hard to believe, seeing as how glass has one of the smallest CTE's known. Even for steel, that would be hard to believe. Maybe you're saying the dome interior is exposed to the atmosphere, and therefore there are variations of temperature inside?

That would be incredibly difficult.

I don't see why. Just run a bunch of wires with resistors around the underside of the mirror, and maybe the rim too. Temperature sensors would be needed also, but this does not sound hard to me either. If computer control can do something as complex as adaptive optics, I don't see why it can't do something like adaptive thermal control (adaptive thermics?).

If what I'm saying truly is insurmountable, then maybe we can imagine a completely enclosed dome with a flat windshield perpendicular to the telescope's direction. The purpose of being completely enclosed is to have a uniform temperature controlled environment so that we can use steel or something similar for the mirror. It would not be hard to rotate the enclosed dome along the horizon. It might be hard to rotate the telescope and windshield vertically while still preserving the sealed environment, but I think similar sealed rotations have been done in other devices (can't remember tho).

EDIT: forgot to mention that the dome with telescope does not necessarily need a person inside. CCDs and computers can record all images and feed them to a nearby building.

 

[Russ_Watters]

In that case, can't some sort of heating system be applied? It wouldn't have to keep it very warm, just uniformly warm.

No: the mirror must be near exactly the same temperature as the surrounding air to prevent distortion of the image due to convection (making the stars twinkle more). Many telescopes are indeed equipped with cooling systems to help accomplish that.

I retrofitted my telescope with a cooling system for that reason.

 

[origin]

No: the mirror must be near exactly the same temperature as the surrounding air to prevent distortion of the image due to convection (making the stars twinkle more). Many telescopes are indeed equipped with cooling systems to help accomplish that.

I retrofitted my telescope with a cooling system for that reason.

What size and type telescope have you got?

 

[Billy T]

Wait, are you saying that starlight, the points of light, cause a significant thermal variation on that point of the mirror? ...

No. As stated before it is exposure to the nite-time air, normally falling in temperature. A large transparent glass cover could keep that air out, but not the cooling of the inner surface contacting air in side the telescope enclosure. Like wise wind pressure on the cover would not be steady but flexing it as it too thermally warps.

 

[DrZygote214]

I'm still confused. The mirror or mirror segments are still mounted onto some metal truss frame, such that the telescope can be rotated and pointed, right? So that metal frame undergoes thermal expansions/contractions that then mess up the position or shape of the mirror...?

 

[Billy T]

I'm still confused. The mirror or mirror segments are still mounted onto some metal truss frame, such that the telescope can be rotated and pointed, right? So that metal frame undergoes thermal expansions/contractions that then mess up the position or shape of the mirror...?

Not if it is a one-piece very rigid and strong near zero thermal expansion coefficient glass. If the mirror is many much cheaper to make pieces of glass each has its own frame with automatic control. - If these multi-mirror telescopes are ever placed on the air-less moon, the individual mirror could be a kilometer apart (don't need to look thur same air refraction) - then some great stellar interferometry would be possible.

 

[Russ_Watters]

What size and type telescope have you got?

Looks like my optical tube has been discontinued, but it is an older version of this 11" SCT (Re-branded for Orion):
http://www.telescope.com/Telescopes...finementValueIds=4527&refinementValueIds=4533

But on this mount:
http://www.telescope.com/Mounts-Tri...o-Telescope-Mount/pc/-1/c/2/sc/36/p/24338.uts

In action:
http://www.russsscope.net/

Thermal management causes a host of problems that need to be addressed, even on the small scale like what I do: Air currents make stars twinkle, temperature changes overnight cause focus to drift, cameras need to be both cold and stable.

 

[Russ_Watters]

I'm still confused. The mirror or mirror segments are still mounted onto some metal truss frame, such that the telescope can be rotated and pointed, right? So that metal frame undergoes thermal expansions/contractions that then mess up the position or shape of the mirror...?

It can if not done correctly, yes. I don't know for certain, but I suspect the mounts are adjustable so the light from the different segments can be collimated.

....Oh, found this:
http://www.gmto.org/Resources/GMT-ID-01469-Chapter_8_Control.pdf

Pointing of the telescope relies on a library of standard algorithms that take as input the main axis encoder readings. Open-loop control of flexure and focus is based on observatory-generated look-up tables with the telescope elevation angle and secondary truss temperatures as inputs. The tables will be initialized during the commissioning phase of the telescope and periodically updated as required. Measurement of the pointing coefficients is done after the open-loop flexure terms are applied. The open-loop control reduces the magnitude of the corrections being driven by the closed loops, and shortens the set-up time required for alignment when slewing the telescope to a new object. There are three distinct frequency regimes for the closed-loop control of the GMT:

• The active optics regime compensates for the residual gravitational and thermal deflections of the telescope structure and optics. Active optics control is typically restricted to frequencies below ~0.01 Hz.

• The fast-tracking control regime compensates for the mechanical eigenmodes of the telescope structure, including pointing. These eigenmodes are primarily excited by wind turbulence. Vibration control frequencies are typically in the range up to 10 Hz.

• The adaptive optics regime compensates for vibration and atmospheric disturbances of the wavefront, at frequencies up to 100 Hz.

Yeah, they've thought of everything.

 

[Bells]

Looks like my optical tube has been discontinued, but it is an older version of this 11" SCT (Re-branded for Orion):
http://www.telescope.com/Telescopes...finementValueIds=4527&refinementValueIds=4533

But on this mount:
http://www.telescope.com/Mounts-Tri...o-Telescope-Mount/pc/-1/c/2/sc/36/p/24338.uts

In action:
http://www.russsscope.net/

Thermal management causes a host of problems that need to be addressed, even on the small scale like what I do: Air currents make stars twinkle, temperature changes overnight cause focus to drift, cameras need to be both cold and stable.

Holy crap Russ!

Amazing photos. You are living my dream!

 

[danshawen]

I still have the remnants of an 18" Dobsonian (reflector) in my basement which belonged to the late husband of my wife who was a project manager on the Hubble Space Telescope until his death in 2003. He was also an active amateur astronomer as well as a professional one for most of his young career. Our extended family inherited his shelves of books on lens grinding, optical tests for various kinds of lens and mirror aberrations, etc.

This inspired me to finish construction of an inspection microscope (he didn't have one of those). Although I have constructed many refractor telescopes, I never attempted building anything like a reflector telescope on the scale that David did. His sons tell me his finished work on the Dobsonian had only a tiny optical imperfection he was unable to smooth out.

I marvel at the many large ground based optical telescopes that are able to compensate for atmospheric effects that would otherwise yield only fuzzy pictures of the most distant objects observable from locations far from the light pollution nearby cities or the moon.

But as is pointed out in the introductory post to this thread, one wonders if there might not be a lighter, more adaptive means for accomplishing the job of several tons of glass for a mirror assembly and its rigid support. A silvered mylar surface and servos to shape it mounted on a rigid rotating frame in the same configuration as a (cross-section of a) mirror is one idea. When I researched it, I could find no information about whether anyone had ever tried such an arrangement. Superconducting bearings also exist, and would be one possible means for acoustically isolating the rotating mirror assembly from ground vibration (by applying an error signal from nearby seismometers). It would not be easy, but this design would make possible relatively large mirrors without all of the weight of current ones. Combined with the ability to mitigate atmospheric effects, this design would be difficult to beat.

Best of all, if the final form of the mirror had a minor optical imperfection, it could be easily repaired, unlike the very flawed design work Perkin Elmer did to the original design of the Hubble Space Telescope which required extreme measures to correct.