Links to all my notes (includes mini-projects and source code etc)

Make do and mend position sensor

You task, Mcguyver, is to detect the elevation of a Dobsonian OTA using parts from old computer peripherals

The essential problem with a DOB OTA is that it's 'removable'. So whilst there is 'no problem' detecting the base 'pointing' direction using a permanently fitted compass, fixing anything to the OTA axis is a real pain.

About the only thing that's going to work is a 'softish' tyre friction wheel .. the weight of the OTA ensure it 'engages' each time and there are no pesky gear teeth to 'crunch up

The problem with this is, of course, that any significant loading on the wheel is going to cause slippage.

One solution is to use an optical encoder (these can be expensive, unless you dismantle an old 'mechanical ball' type mouse which contains 2 encoder sets !)

However, difficulties in using optical encoders (which requires you detect 'forward' and 'backward' movement in order to add/subtract from the running 'pulse count' = current position) makes the multi-turn potentiometer 'sensor' my preferred solution

So long as 'backlash' is minimised, a pot is 'immune' to direction change (backward/forward) errors and always returns the 'absolute' position even after a Pi re-boot (no count required).

This means the problem comes down to the drive - if the pot is fixed directly to the axis, then '10 turns' of the sensor wheel = 1 turn of the OTA.

In fact, the OTA can only go from 'just past vertical' to 'just past horizontal' i.e. no more than about 120 degrees (1/3rd a turn).

If we use a 5 turn pot. with a drive wheel of diameter d, the max. distance that can be travelled (before the end of the pot. is reached) is 5 * Pi * d

The Dob. axis 'wheel' thus has to drive the pot. less than that distance. The Dob axis of diameter D, in 1/3rd a rev, travels 1/3 * Pi * D

This gives us 1/3 Pi D &lt; 5 Pi d, or d &gt; D/15. Since D on my ancient Dob. is about 6" we get a drive wheel dia. of 0.4" (or greater) = using a 10 turn pot. (to improve accuracy) reduces the wheel dia to 0.2" (or greater).

A soft-tyre 1/2" wheel would thus 'do the job', however with a relatively small wheel to Dob. axis 'contact area' avoiding slippage will be a real pain.

If possible, the drive wheels should be assembled directly to a long potentiometer shaft.

However, if using Meccano parts, the wheel axal diameter is 8 swg (4.06mm) whilst standard potentiometer shaft diameter is 5mm (although it might be possible to 'turn it down' (hard if it's got a flat on it)).

The need to connect a 8 swg (4.06mm) wheel axis to the 5mm pot. shaft can be an advantage if some sort of 'slipping clutch' can be used (so we don't wreak things when we run into the pot. end stops) = for example, some plastic tubing :-)

An alternative, more robust method, would be to use heat-sink tubing, however there would be some danger that the Pot. shaft might be twisted off (or the drive wheels forced to slip) causing damage to the sensor assembly.

Like most people this is where I reach for my old Meccano box.

Part 23a (Pulley Wheel with boss) has 1 1/2" diameter and will take a 'tyre'. The 23ap is the plastic pulley (with brass boss). Part 23b (or 23bp, the plastic version at about 20p each) comes without a boss, however there's nothing to stop you from epoxying it to the axis (in fact, to maximise the contact area, I suggest 3 or 4 x 23p on a the same shaft)

The problem is the 'tyres'. The Meccano ones are totally unsuitable - any you find will have gone hard (if not brittle) and, with an OD of almost 1", are way too big. I needed something 'softish' that would end up with an effective OD neared 1/2".

The best solution turned out to be a thick 'O ring' - look on eBay for packs of 5 post free for less than £2 (with ID (Inside Dia) about 10-11mm, OD 14-15mm so would squash down to about 1/2" (12.5mm)) - and avoid the usual rip-off 'UK seller' who wants 99p each PLUS postage ..)

An alternative source might be a 'mixed set' of O rings that you can sometimes find in the £1 shop (or other discount store).

To use a 10 turn pot. Meccano Part 23c, a 3/8" (0.375") dia. solid rubber 'pulley wheel', might just do the job (if it can be clamped/glued to the shaft).

Fitting the sensor.

There are two possible places where a sensor wheel could 'mesh' with the OTA = on the bearing axis itself or the bearing 'inner flange'.

Using the inner flange would mean the sensor assembly could be 'hidden' on the side edge of the 'rocker box' and would require minimal holes etc. to be drilled. However space is rather tight and it would be all too easy to break the sensor assembly whilst placing the OTA into the rocker box.

There are two ways the sensor might be placed to 'run' on the axis boss :-
1) on the mount 'shoulders' - which would again minimise holes but would require some clever spring 'bracing' to maintain running contact but still allow the OTA to be lifted on and off - and there would still be some danger of knocking the sensor whilst trying to align the OTA to the rocker box.
2) in a slot cut into the rocker box axis 'hole' - this allows the OTA to rest on the sensor wheels, so the weight of the OTA can be used to maintain contact.

After a lot of thought, I opted for the 'rest the OTA on the sensor wheel' solution as this would avoid the need to 'line up' springs etc. and (with the main components mounted on the outside of the rocker box) avoid the possibility of breaking the sensor whilst mounting the OTA.

By building the 'sensor' (bearings, shaft, wheels, Pot. etc.) as a stand-alone 'unit' I was able to align the wheels etc. 'on the bench' (rather than 'on the rocker box' which would have been a lot harder)

Meccano part "(flp16) Flanged Plate 2x3 flanges on ends" can be used as the base plate. To support the axal, part "(77) Triangular Plate 2x2x2 Hole" can be used as the end plates (standard Meccano hole spacing is 1/2", so the triangular plate 'height' is 1/2 * root[3] = 0.866" above the mean hole centers).

The assembled sensor 'module' was then placed into a segment cut-out of the Dob. rocker box axis mount with a stiff spring (so it would press against the OTA axis 'bearing').

This note last modified: 29th May 2018 14:36.[top]
Next subject :- index
[top]

Notes index links

Links to all my notes (includes mini-projects and source code etc)

Make do and mend position sensor

Fitting the sensor.

Next subject :- index