Re: [ATM] undercorrecting?

["Arjan te Marvelde" <arjan.te.marvelde@hetnet.nl>]

Hello Mike'

> Sorry, I did not realize that the article
>    http://home.hetnet.nl/~artm/atm/articles/cooling.html
> was meant to analyze the effect of radiative cooling (I can't find
> where it says this) - I thought it was from convective cooling.

Well, actually it is both; I should make a more precise version.

> Below are a couple of messages from you.  The first one seems to
> indicate that the effects of radiative cooling are negligible.  I also
> heard someone else quote e = 0.01 rather than 0.04, so maybe that
> value should be verified.
>
> In the second message you seem more convinced that there is an effect
> from the radiative cooling.  Is that correct?  If so, what made you
> change your mind?

Yeah, I know. At first I tried to feed my gut feeling (i.e. no radiation
effect) with some quick math, but it turned out that the problem deserved a
bit closer examination to be sure about the orders of magnitude. Also there
were some errors in the first model, but I'm quite confident the current
version is OK.

> I am curious what other factors could be in play with regard to the
> mirror mentioned on the Dutch mailinglist.

Dunno, could be anything, really.  Maybe it's stress in the glass? I just
heard from Jan van Gastel that this specific mirror has been (re-)figured a
couple of times.

> What seems to be overlooked here is how the front and back of the
> mirror are prepared.  The front is polished and coated with aluminum
> and silicon dioxide (what's the effective emissivity of the
> combination?), while the back is probably not.  What difference does
> that make for the emissivity of glass?  Is it larger than the effect
> of radiative cooling on the mirror?

Ah, that's easy. I found a number of web pages listing emissivity of various
surfaces. Just Google 'emissivity' or go directly to
http://www.electro-optical.com/bb_rad/emissivity/matlemisivty.htm
You will see that polished Al is 0.04-0.06, and CVD Al with SiO overcoat is
set at 0.04. For the various kinds of glass you can take 0.95. Polishing
does not change that much.

I did not take radiation from the back into account, since the ground
temperature will be approximatly equal to the mirror temp. If this is not
the case, the high emissivity of bare glass will quickly make this a factor
(either heating or cooling).

> Along that line of thinking, how about the insulating effect of the
> mirror cell's support points?

Good point, if you have high-school physics/math level I think you should be
able to work with the FlexPDE model. You just install the program, obtain a
student license, and load/run the model from my website. Changes can be
easily made in the model in text format, e.g. by providing a profile for the
heat transfer parameter k, instead of the constant it currently is. For the
nifty things there's a help file...

> Seems many other factors could account for very small temperature
> gradients in the glass.  0.015K is not much.

This is true. But if you only consider mirror temperature wrt. ambient
temperature, you will always end up with the same equilibrium. Radiation is
a factor that can disturb this equilibrium, and sustain a heatflow. But as
my paper shows it is cancelled almost completely when you apply enough
forced air cooling (or warming...). I'd still like to know how you can do
without fans by choosing the right design of your scope.

> This has been a great topic by itself, and it also started a great
> thread about how eyepieces affect the star test.  Lots of good stuff
> being investigated!

Yes!

Cheers,
   Arjan

_______________________________________________
ATM mailing list http://www.atmlist.net/