Go Down

Topic: Tilt only sensor for telescope. (Read 2503 times) previous topic - next topic

Bergie

My first post, so please don't be to harsh! :)

I want to build a tilt sensor for a telescope. I have Maksutov Cassegrain and it suffers from "mirror flop." I.e. when you look at something low in the sky, the focus changes when you swing it to something higher in the sky. Thus a constant need to refocus arise when you change the angle of the scope.

I have fa-dangled a home made focuser from and old servo motor, and some flatbed scanner gears. To focus, I have a simple toggle switch to go in whatever direction I choose to focus, it focus sharp, but is slooooow. So I want to save time while the telescope is swinging around by having the focuser change focus a bit. No need to focus sharp, but at least give the focus a nudge in the right direction for a couple of seconds. (Yes, its that slow.  ;) ) Now the plan is to add a tilt sensor and roughly calibrate it to the tilt.

Summing up: If the scope goes up by "x degrees", the focuser run clockwise for "x time."

I know that x - y tilt sensors is available, (not in South Africa though, and very expensive.) But maybe there is something simpler and cheaper and at least a bit more accurate than a "ball bearing in a can." I thought a slotted wheel with a weight will do, but it messes up the scope since its too big and awkward, and not much more accurate than two canned ball bearings. I want something small and simple... Or am I asking too much? Any ideas? It only has to go through 90deg.

robtillaart

You might use a "stretch" sensor connected between moving and fixed parts.

Create a lookup table how much stretch is generated @ which angle. You can do with 5 or 10 values and interpolate them by means of - http://arduino.cc/playground/Main/MultiMap - if it is not linear.

Another trick could be some hall sensors with different resistors in parallel and a magnet triggering one sensor at the time.


+5V ---< R0 >-----< HALL >-----+  0 degrees
                                |
+5V ---< R1 >-----< HALL >-----+  10 degrees
                                |
+5V ---< R2 >-----< HALL >-----+  20 degrees
                                |
                                .
                                |
+5V ---< R9 >-----< HALL >-----+  90 degrees
                                |
                                |
                                +----- arduino AnalogRead()

Rob Tillaart

Nederlandse sectie - http://arduino.cc/forum/index.php/board,77.0.html -
(Please do not PM for private consultancy)

Grumpy_Mike

Why not have a rotary shaft encoder measure the movement on an axis? Or a series of tilt switches (I think in SA you can still use the much better mercury type) arranged at different angles.
However it strikes me that actually fixing the scope might be a much better option.  :)

Bergie

Ah, fresh ideas I never thought of myself.

There is unfortunately no "fix" for mirror flop. The mirror is mounted on a base plate that is in turn geared to the focusing knob, so the whole mirror moves up and down. I think a faster focusing-mechanism should help!  :)  It takes about 30 second for one revolution...

I am also toying with adding a wheel to a pot that can run on the shaft, but a mechanical connection like this may make it too much of a hassle, especially when the dew builds up...

Speaking of which. I have build a PWM dew controller, which I plan to hook up so the Arduino reads temp sensors and control the amount of power automatically. Then do some focusing control etc. and when I grew up, I will try and build a wireless (infrared) controller to focus my scope with. When I glue the Arduino to my scope, I don't want it just to blink and LED!  :) It should work for a living. There is a GPS module in the scope mount... wonder if I can access that...

Keep the ideas coming!

RuggedCircuits

Why not use an accelerometer? When stationary the 3 accelerometer axes reported (X/Y/Z) feel only the force of gravity thus the relative magnitudes give you a 3-dimensional vector pointing straight down (after some math!) -- hence you have tilt.

--
The Gadget Shield: accelerometer, RGB LED, IR transmit/receive, speaker, microphone, light sensor, potentiometer, pushbuttons

mowcius

An accelerometer will not be accurate enough to measure the tiny accelerations - do you mean a gyroscope? :)

Senso

A gyroscope to read tilt?
A gyro gives you the rate of change but not the actual value, so when not moving it just gives you 0, so its of no use, a gyroscope with a resolution of 3.9mG/LSB like the ADXL345 is not enough?

mowcius

Oh I always get confused with them...

RuggedCircuits

A low-cost accelerometer will be noisy and not too precise but the OP said "the plan is to add a tilt sensor and roughly calibrate it to the tilt" so I figured a rough measurement from an accelerometer would be good enough.

--
The Rugged Motor Driver: two H-bridges, more power than an L298, fully protected

retrolefty

#9
May 24, 2011, 03:38 am Last Edit: May 24, 2011, 03:40 am by retrolefty Reason: 1
Quote
I have Maksutov Cassegrain and it suffers from "mirror flop." I.e. when you look at something low in the sky, the focus changes when you swing it to something higher in the sky. Thus a constant need to refocus arise when you change the angle of the scope.


Is this a commercial telescope or a DIY scope? You really should try and solve the root problem and fix the mirror flop. I've build a couple of reflector scopes in the distance past and a decent mirror mount design should not allow for primary mirror movement enough to cause focus length to change. Most mirror mounts are required to be 'loose' enough to not allow any stress points on the mirror as temperature changes, however not so loose as to flop around.

Hot glue is a one method that will allow a stress free attachment of a mirror to it's holder.

Lefty


willnue

If you have an old mouse laying around you could harvest the ball wheel encoders or scroll wheel encoder from it and use it to track the tilt.

Convert a PS/2 Mouse to a Dual Wheel Encoder (Good pictures)
http://www.linuxpcrobot.org/?q=node/3

Understanding a Mouse Scroll Wheel (Arduino Example with video & code)
http://www.stevekamerman.com/2010/12/understanding-a-mouse-scroll-wheel/

Or use build one yourself with and IR sensor:
Homemade wheel encoder
http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1257038219

Hope it helps,
willnue

Bergie

Been offline a bit.

Thanks for info to try and fix the scope, Lefty. As I said before, no cure in hot glue. The mirror on Maks is adjusted up and down to focus. The play comes from that adjusting bit. It seems even high end Maks suffer from it. A common cure is to fix the mirror in place and use a Crayford external focuser.

Accellerometers is too expensive for me (even the cheap ones I can find locally) and overkill.

The "bearing in a can" tilt sensor is is not quite what I want, although a couple of them could do the trick. At a guess I would need to adjust the focus for every 15 deg or so of tilt. I have no hope to get it perfectly accurate in any case. What I thought of is the kind of tilt sensor thingy you get in digital level meters. I used them over the years to check downtilt on GSM antennas. But I discovered it uses the expensive X-Y accellerometers...

Willnue's idea is great, and I did not explore that one, simply because I can't find any ball mouse anywhere, so I chucked out the idea of a mouse. I simply never considered the wheel have the same principle.

The output shaft of the telescope mount have notches I can use with a switch or sensor, but I have no idea how such a sensor will be able to tell which direction the shaft is turning.

willnue

If you don't like the "bearing in a can" version, check out the RPI-1031 4-direction Tilt Sensor from ROHM, which is more like a bearing in a box. You still might need a few of them to get the result you are looking for, but they are pretty inexpensive. I haven't used them, but I have used the "bearing in a can" and it doesn't take much to beat that.

Quote
The RPI-1031 employs the principle of ROHMs advanced photo interrupter technology through the integration of one IR LED and two IR photo transistors. It allows identifying the gravity direction in equipment turning up, down, left and right.


Info at ROHM with nice diagram of how it works:
http://www.rohm.com/eu/highlights/product-news/tilt-sensor-rpi-1031.html

FYI - They are verrrry small SMD parts, and you can find them online for <$5 USD with or without breakout boards.

That being said I still think the mouse wheel encoder would proabably be your best bet. Start looking around for old broken PCs...

willnue


willnue

#13
May 26, 2011, 05:52 am Last Edit: May 26, 2011, 05:59 am by willnue Reason: 1
Just thought of something else.

If it only has to go through 90deg, what about a using a potentiometer with a gear on it that you add into the gearing you already built? A potentiometer should be easy to find just about anywhere and is easy to read with an Arduino.

Reading a Potentiometer (analog input)
http://arduino.cc/en/Tutorial/Potentiometer

willnue

Bergie

Thanks for the ROHM link willnue. Its a very interesting device.
As for the pot, it was my more or less my original plan. I'm trolling for fresh ideas. Folks can come up with incredibly good ideas. Sometimes a nudge in the right direction is all that is needed.

I remember many moons ago I fixed an old electric organ. The volume control was as expected on the foot pedal. What I did not expect was that there was a lightbulb on one side with a triangular slit on the foot pedal, and on the other side there was a photo resistor. As you push the pedal deeper, the wider section of the slit moves past the bulb and more light falls on the photoresistor, thus increasing the volume. I would have used a pot if I designed it, but this solution was quite novel. (Of course in the good old days a major problem was scratchy sound when you turn a volume control. The cause was dust inside the pot. This volume control clearly would never scratch due to dust.) Of course I don't want light when stargazing, but maybe an infrared led, receiver and a slit in a piece of plastic... Perhaps explore along that line? Don't know what I do want to do. I've got time to explore other ideas.

If I knew how to get the output on an optical mouse to be interpreted by the Arduino, I may stick the sensor on the shaft, and use that output to drive my tilt sensor. I think I love the new ideas more than the actual solution...  ;)

Go Up