Raspberry Pi Forums

Hi all !
Mi problem is currently to make an accurate measure of the angular orientation of a 150 mm axe.
I've already made some test using a GY-80 IMU : it works fine but I finally need more precision than I thought for my functionnal use case.
I can't deeply modify the existing mechanism. So I wonder if I could use something like a flexible rack (something like this : http://www.motionco.co.uk/racks-flexibl ... 32_58.html) mounted on the axe with a little rotary switch mounted on the fix part of the mechanism.
Have someone some recommandation about rack/gears/switch to use ?
As said the axe diameter is 150 mm and I think the gear diameter should be 20 mm.

Thanks a lot !

How accurate do you need?
Have you thought of painting (sticking a printed lable) onto the shaft (axle, not axe btw), then using opto sensors?
Is there something already attached to the axle that you can measure.

We really need much more info than you have given, a picture would be great.

I have been playing with a DC motor and a 10 turn precision potentiometer (20 kohm Vishay) on the motor axis. The potentiometer linearity is 0.25 % and I seem to be able to set the position with a resolution of 0.1 - 0.2 %. The motor is not too expensive but noisy MFA Como Drills 918D1024112 (with 1:1024 gear box) and I am using a DIY H-bridge

https://github.com/oh7bf/RasPiHBridge/wiki

Limit switches are a good idea to protect the potentiometer from mechanical damage.

aTao wrote:How accurate do you need?
Have you thought of painting (sticking a printed lable) onto the shaft (axle, not axe btw), then using opto sensors?
Is there something already attached to the axle that you can measure.

We really need much more info than you have given, a picture would be great.

Hi guys and thanks for your answer.

My goal is to use any sensor to determine the relative positions of the two axis of an astro telescope equatorial mount (this type : http://www.pulsar-optical.co.uk/magnify/155.1.html).

Here is a picture I took : http://www.webastro.net/upload/images/2 ... 192865.jpg

I think I need to have an accuracy of 1 degree only (when you point a presumed stellar object position with 1 degree accuracy, you should have it in the 25x finder).
As you can see in the picure, there are some graduations on the axe (there is another one in the other axe). In concrete terms I need to get this values (to calculate some others). Naturally there will be a calibration phase to get the "zero" value.

I thought about some kind of opto sensors but I did not found any to do it. Do you think about a specific sensor ?

I thought also to something like a potentiometer line (as mentioned in the second answer) but I wonder if it should be enough strong : the telescope on the mount can be turned on the two axe with fine motion commands, but by a free hand way too on the movments should a little fast and repetitive, no ? (and the monnt is used outdoor with potentially some humidity issues)

I hope this explanation should help you to have a better idea of what I try to do ...
Thanks again.

Regards

OK, now we are in business.

First (and very important) the sensor will not need to cope with continuous rotation. Sure the mount can spin round and round, but its not used that way... So, your idea of gluing a flexible rack is pretty good.

Second. Because you need to calibrate at the start of every session (line up on a known object and set the co-ordinates) then a built in zero point is not needed (that would be difficult when gearing up to a multi turn sensor).

So, heres what you (could) do.
Go with the rack and small gear, possibly gear this up too* then use a mouse wheel sensor. Mouse wheels (and mechanical mouse movement sensors) work by shining a light through a slotted disc onto 2 slightly offset photo diodes/transistors. By monitoring when each opto is lit you can tell movement (not absolute position, but thats covered by your calibration earlier).
1 degree accuracy should be very easy to hit.

* one thing to be very aware of when gearing is backlash, namely when you reverse direction it can take a bit of movement before the gears engage. 2 ways round this are 1 only trust the calibration when moving in a particular direction or spring load the small gear.

aTao wrote: So, heres what you (could) do.
Go with the rack and small gear, possibly gear this up too* then use a mouse wheel sensor. Mouse wheels (and mechanical mouse movement sensors) work by shining a light through a slotted disc onto 2 slightly offset photo diodes/transistors. By monitoring when each opto is lit you can tell movement (not absolute position, but thats covered by your calibration earlier).
1 degree accuracy should be very easy to hit.

Whaou !!! I did not thought about such a solution : very nice !
I think I have a couple of old mouse to recycle ;o)
Tanks a lot for the idea : I think I'll try it next week, so I'll tell you if I succeeded.
Regards.