Using a potentiometer to measure arcseconds.

[skatefast08]

How should I pick the correct type of potentiometer that will measure an angle to the nearest arcsecond (.000278 degrees)? What parameters and their values for a potentiometer would be important to consider for this kind of application?

I believe if I used a random potentiometer, it may not have the same resistance value at the same position (.000278 degrees difference) from the position it was before. Maybe I need to consider noise? Or some kind of tolerance. Maybe even a larger potentiometer that works up to 100 kOhm would have more resistance values and therefore reduce the angle error. What would do?

 

[barry]

Forget it. There’s absolutely no way you’re going to find potentiometer, even a multi turn one, with that kind of resolution and repeatability. You need some magnetic or other type of sensor.

 

[andre_luis]

Not clear, do you want to achieve a resolution of 0.000278° on a full circular scale of 360° ? That would result on a required resolution even smaller than 1ppm, indeed impractical. Even the best-performing rotary encoders hardly reach anything less than ~1/10,000 of a complete circular turn (e.g ~0.043° for an 8,192 encoder)

 

[skatefast08]

No, I mean on a 90 degree turn. Is there any accelerometers that I could use for this application, or inclinometer that would measure to the arcsecond (I want to be cheap with this)? I want to measure an arcsecond to measure the distance to stars from a telescope using the parallax theorem. What would you recommend?

 

[betwixt]

For such tiny measurements, use a graticule in the eyepiece or a large optical encoder. Even if you could get such precision from a 'cheap' potentiometer, the temperature and environmental stability would make the reading useless. I would guess the best you will get with 90 degrees sweep on a potentiometer is about 1 degree, you need 500 times more accuracy.

Depending on the size of your telescope, a large optical encoder might be an option. You would move the telescope to center the star and note the rotation of the optical disk. With a quadrature encoder system you should be able to achieve something close to the accuracy you need without too much cost.

Brian.

 

[Nik_2213]

Hi !
IMHO, measuring stellar parallax is HARD, which is why it had to wait until the mid-19th Century for large optical instruments and, after the first successful determinations, photography to allow multiple comparisons of positions six months apart.

https://en.wikipedia.org/wiki/Stellar_parallax

The work is beset by complication and pitfalls.
Murphy's Law rules.
The analogous hunt for exoplanets, bravely begun by Peter van de Kamp, came to grief when his data was found to be fatally compromised by telescope maintenance. Think how Space Hubble's mirror was off by a whisker due calibration error...

Similarly, now Gaia data is pouring in, some of the Hipparchos data has been confirmed as wildly wrong. Some diagnosed due to a star flare offsetting apparent centre, some due un-resolved binaries / hot jupiters, some due a background star in just the wrong place. Yup, that's Murphy's Law !! Some, IIRC, 'still don't know'...

If you're planning to study short-ish period binary position angles, rather than full-on stellar parallax, that's probably practicable...

 

[barry]

There’s no optical encoder that will come close to the OP’s requirement. Probably the best you will find is 65536 counts/ rev. And even that is going to be expensive.

 

[betwixt]

I've got a 16" (800mm) telescope with opto-electronic guider. I'm not sure what positional resolution it has but as a guess it is around 0.1 degree so far short of one arcsecond. It's fine for finding objects but not for scientific measurement.

Brian.

 

[andre_luis]

I do not have any experience with telescopy, but if I were to implement the angular control, this would not be properly by measuring the angle of inclination with rotational encoders from the center of curvature, but rather with a flexible incremental linear encoder tape tied along the entire 90-degree internal perimeter of the aperture of the dome, where the end of the telescope would be anchored on a circular rail. These obviously are not off-the-shelf items, and would have to be manufactured on demand upon design specifications, but it is what I can imagine as being possible to be done with the current resolution technology achieved at least with most encoders available on the market. You may be realizing from now that your stringent requirements can not be met with low cost solutions. Anyway, another possibility to try, could be using a rotary encoder rolling along the curvature, but not sure it it works, something like this:



This way you would somehow multiply the actual resolution of the encoder for a factor as wide as the diameter of the wheel on which the encoder axis is attached, compared to the total course along the dome.

 

[summitville]

Renishaw has +/- 0.5 arc-second absolute encoders, not cheap
Renishaw


Put a 10,000 Pulse / Rev Optical Encoder on the Drive Motor shaft.
Add 130:1 reduction gears between the motor shaft and the telescope shaft
After your remove the backlash, you can measure arc-seconds between two stars
$110.00 / each ?
ENCODER

 

[andre_luis]

Just remembering that you would actually need 2 of these sets of motor+encoder to provide the desired position, one for the vertical incline, and another for the cardinal direction on the horizontal plane.

 

[kripacharya]

..... want to measure an arcsecond to measure the distance to stars from a telescope using the parallax theorem. What would you recommend?

Doing this with potentiometers etc will be virtually impossible.

Here is an alternate method which doesn't require micro-angle measurements, assuming you have access to a fairly large telescope with good magnification --

Use astrography, and a little bit of a priori knowledge.

Take pictures of your target star along with surrounding stars (around 5 arcseconds worth) throughout the year.
Also simultaneously take pictures of Alpha Centauri (or some other choice of close / bright star) throughout the year as well.

Locate those 2 pics where Alpha Centauri has moved the greatest. Measure that moved distance (relative to background stars). We know from other sources that this star is ~4.24 light years away.
Locate 2 pics of YOUR star where it has moved the maximum. Compare this distance with Alpha C - a ratio is good enough - and do the calculations to determine the distance !

Would this work for you ?

 

[kripacharya]

I've got a 16" (800mm) telescope with opto-electronic guider. ...

Seriously ?? I would KILL for an 800mm... making do with a 115mm Newtonian as of now. Celestron. Added a home-brew laser pointer and HATE the manual focus. Hoping to someday make it remote control smooth.
Which one do you have ?

Am posting here because your PM is disabled

 

[betwixt]

Sorry, slip of the brain cells, 16" is 400mm not 800mm but still quite a beast. Takes two people to lift it and a ladder to reach the eyepiece. The only thing it lacks is a gadget to make the clouds go away. As I'm in the rainiest part of Europe you can appreciate the problem!
I'll enable PM for you.

Brian.