The term “adaptive optics” sounds like something that should be really complicated and really expensive. And in general, the ability to control the properties of optical elements is sufficiently difficult enough that it’s reserved for big-science stuff like billion-dollar space telescopes.
But that doesn’t mean there aren’t quick and dirty adaptive optics that are suitable for the budget-minded experimenter, like this thermally deformable mirror. As [Zachary Tong] explains, this project, which started quite some time ago, is dead simple — a 4 by 4 array of through-hole resistors stand on end, and these are attached to a glass coverslip that has been aluminized on one side. An Arduino and a couple of shift registers make it possible to individually address each of the 16 resistors in the array. Passing a current through a resistor heats it up a bit, leading to thermal expansion and a slight deflection of the mirror sitting on top of the array. Controlling which resistors heat up and by how much should lead to deformation of the mirror surface in a predictable way.
The video below shows some of [Zach]’s experiments with the setup. Unfortunately, he wasn’t able to fully demonstrate its potential — the low-quality mirror didn’t cooperate with his homebrew interferometer. He was, however, able to use a dial indicator to show deflection of the mirror in the 2- to 3-micron range by heating the array. That alone is pretty cool, especially given the dirt cheap nature of the build.
As for practical uses, don’t get too excited. As [Zach] points out, thermal systems like this will probably never be as fast as MEMS or piezoelectric actuators, and many use cases for adaptive optics really don’t react well to added heat. But changing the shape of a mirror with air pressure is another thing.
Thanks for the tip, [smellsofbikes].
I’m not sure you want to heat the mirror of an catadioptric system. Generally, you can to cool it down as much as you can to limit thermal noise and atmospheric ray bending.
No, you don’t want to cool down your terrestrial telescope’s optics as much as possible because that would then certainly be below the dew point of the ambient air and you would get dew or frost on your optical surface in question. And you would create additional turbulence in the air layer between your cooled optical surface and the ambient air above it, from convection.
As optical surfaces that are pointed towards the cold sky will radiatively cool (so their temperature can and will drop below the ambient air temperature), it is not uncommon to have anti-dew heaters on telescopes, and/or fans that very gently (without causing turbulent flow) circulate ambient air over those surfaces to prevent the temperature falling too much below the ambient air temperature and to prevent said turbulent convection flows. So heating indeed has a place in telescopes, for professionals and amateurs. It is the imaging sensors that you generally want to cool down as much as possible to reduce noise.
Depends on whether your mirror is exposed to ambient air. Flood the tube with e.g. dry nitrogen (a small SCUBA tank filled 100% N2 – ask the dive store real nice for ‘NITROX 0’ with a nudge and a wink, and be very persuasive it will always be clearly labelled as not breathing gas – will last for years) and condensation is not a problem. Still have the problem of air turbulence, but active sub-ambient cooling can at least get you down to temperature a lot faster than just flowing night air through. If you are less sparing with the N2, you can use the expansion to chill down the mirror during purge.
Gonna go out on a limb here and guess that these issues aren’t going to be the bottleneck in this case
It’s my understanding that the mirror isn’t going to do much radiating; it isn’t a significant source of system thermal noise any more than the physical temperature of an antenna or lead-in wire in a radio receiver. Astronomers cool their sensors to lower system noise.
If you need dry nitrogen just go to any HVAC supply house or welding supply store. It is used to purge the copper lines for brazing and/or used for pressure testing prior to filling with refrigerant. The welding supply will charge a tank rental fee but most HVAC supply will give you the bottle and charge only for the gas. If the HVAC supply ask if you work the trades just say yes.
Interesting. I worked at Lennox building, filling then brazing AC units and we didn’t use nitrogen.
In EU nitrogen leak testing is mandatory for F-gas containing HVAC systems, but this is not the case in every country.
If you were filling the unit then brazing you would have serious problems. The point of purging the tubing with the dry nitrogen is that when brazing the copper oxidizes and forms particles that will then flow with the refrigerant through the system. Dry nitrogen has no oxygen so no oxidation occurs. It is also used when tig welding stainless etc. You purge the non welded side of the material with dry nitrogen and no oxides form. https://www.youtube.com/watch?v=ZjpnP9uhxyk This link shows exactly what I’m talking about.
Typically the shielding gas for TIG welding is a blend of Argon and CO2. Nitrogen will promote the formation of nitrides which are just as undesirable as oxides in the weld zone.
CO2 is not used for TIG welding as it damages the tungsten electrode, 2% N2 balance Ar is sometimes used for TIG welding, though pure Ar and Ar/He blends are more common, at arc welding temperatures CO2 and N2 are both gases that exhibit reactivity with some metals, this can be either an advantage or disadvantage depending on the materials being welded.
Most welding places here want you to buy the tank, then they will swap it with a full one whenever you need a refill. They handle the inspections.
How do the HVAC places get their tanks back? I’d think most HVAC guys would have a few empties laying around.
Very nice try. I would propose to use the piezoelectric effect in some smd ceramic capacitors. It would require testing their expansion vs voltage for different types and values, but the nice thing would be that no or much less heat would be involved and virtually no power consumption is needed to keep the mirror in a specific shape.Also with smd the number of ‘pixels’ could be dramatically increased or the tilt angle of the mirror segment between 2 neighbouring pixels could be greatly enhanced, even if the expansion is less than with the thermal expansion. Downside could be that much higher voltage could be needed but with a low side switch and pwm driving that should not be a big problem.
I though about piezo actuators but the price is about 50-100€ each and indeed needs higher voltages so I did not share initially. I did not thing we could use simple capacitors but why not.
TDK FAQ about capacitor singing states about 10-25nm displacement @1V amplitude. It’s far from the 2-3µm provided by the resistor expansion. It could do a low but very high precision deformation.
I wouldn’t call this adaptive optics. The game in adaptive optics for an earth based telescope is to compensate for atmospheric effect, i.e. what astronomers call “seeing” and that requires a way to deform the mirror at rates a few orders faster than could be achieved by thermal effects.
But using something like this to adjust the mirror figure in a more or less static sense seems reasonable, I just wouldn’t call it adaptive optics.
For homebrew I could see this being an easy way to accurately compensate for a number of defects that change more slowly.