A Motorised Filter Wheel

Filter wheel on telescope39Kb The filter wheel on the telescope.

I have finally obtained a filter wheel for my colour filters. It was made for my by a retired machinist friend to my basic design. He did all the hard and clever work of turning my design into a working system. Thank you, Eric.

The main consideration when designing the wheel was to keep it as thin as possible because I use a Newtonian for my imaging and back-focus distance here is at a premium. The wheel is designed to hold 8 filters. At present it holds my 5 UBVRI 25mm diameter ones; a ¾" square H-alpha interference filter; an open hole; and an opaque position for dark frames. I also wanted a deluxe wheel - so I made it motorised and added a digital encoder with remote read-out so that I can be sure of which filter is in the beam.

I can't give complete, detailed plans for building one of these, but the following description and photographs should suffice. Contact me if you have any questions.

The wheel is 114mm in diameter and 6mm thick. The holes for the filters are 25·4mm in diameter with a 1mm ledge at the bottom of each hole to rest the filters on. Cut-outs between each hole allow a small clip to hold the filters in place. (The ¾" square filter is mounted in a special cell which fits in a slightly larger hole.) Two 3mm thick aluminium plates provide the mechanical support for the structure. They are separated by 4 spacers, although 2 of these near the mounting of the CCD were soon replaced by a solid piece of aluminum which goes completely along the top edge to minimise any flexure between the plates. The wheel is supported between 2 small ball bearings on a short axle.

Eric added a port through which the filters may be changed (or cleaned) without disassembling the unit - something I hadn't thought about. He also made some clever way that the T-thread mount for the CCD can be rotated so that the CCD can be oriented in any position and then locked.

The wheel is driven by a 16 RPM geared DC motor from Edmund Scientific (number A41,854). A small rubber O-ring (approximately 15mm diameter) mounted on the motor shaft is pressed against the edge of the wheel and so drives it. It takes a bit under 4 seconds to move between positions, a compromise between speed and minimising overshoot.

The one problem with using such a cheap motor is that it has no internal bearings and so can't take any side load. This is overcome by supplying a support, in the form of a brass bushing, on the other end of the motor shaft. It can be seen in several of the photographs below.

A simple control circuit consisting of a direction-select switch and pushbutton on the handset, and a micro-switch on the wheel are used to position the filters. To change filters, the pushbutton must be pressed long enough for the motor to drive the wheel out of the detent and enable the micro-switch. The button is released but the motor remains driving until the micro-switch falls into the next detent and so stops the motor.

Motor drive circuit

In the above circuit you can see how a DPDT switch reverses the polarity going to 2 relays, while their associated diodes control which relay is activated when the pushbutton is pressed. The top circuit shows how the relay contacts are wired. When not activated the motor is shorted (as shown in the diagram), thus ensuring the motor stops quickly. When one relay is switched current flows one way to the motor; and when the other relay is activated it goes the other way. Depending on the direction, one or other of the diodes lights showing that the motor is driving. I use 2k2 resistors wired to red and green coloured LEDs, but if you had one of the 7-segment displays which had 2 decimal-point LEDs then you might use these instead.

The rim of the wheel is divided into two (unequal) halves. One strip is recessed slightly and has the detents for the micro-switch, while the other half is smooth for the friction drive. A small roller which fits into the detents actuates the micro-switch. Eric took care of these fiddly details.

The encoder circuit is as simple as I could make it, and yet be reliable. It uses 4 Hall effect switches mounted on the body and 13 small magnets embedded in the wheel to produce BCD-encoded signals which drive a 7-segment display on the handset. It requires 2 ICs (apart from the Hall effect switches). While the wheel is moving, the display goes blank until the magnets on the wheel are again close enough to activate the Hall effect switches and display the position.

Motor drive circuit

The above circuit shows how the Hall effect ICs (3120U) are wired. The 1K pull-up resistors are mounted in the handset and help reduce noise over the line to the ICs. The signals need to be inverted (I used a 7404 as I had one in my junk box, but any inverter will do) before being fed into a 7448 BCD to 7-segment decoder/driver which does all the work. This device drives common-cathode displays - if you have a common-anode display then use the 7447 device instead.

The magnets were amongst the most expensive components - especially when you consider their size. They are Samarium Cobalt sintered discs, 3mm in diameter and 2mm long and cost $2 each. They are very powerful - probably more powerful than necessary - and can activate the Hall effect switches from over 5mm away. In practice they are separated by less than 2mm.

Not shown on the first 8 pictures is the arrangement for the encoder as this was added later. I'd make a lousy engineer as I wanted to see the unit assembled before I worked on the encoder arrangement. It would have been better had I done it all in one hit, but in the end it didn't matter.

The whole unit draws very little current and could probably be battery powered, but as I only use it from my observatory I power it from the same supply as my CCD. Edmund sell a 3V motor which could be used insted of the 12V one I use.

A thin sheet-metal cover goes around the outside to light-tight it. On the bottom (not shown on the first 8 photographs) is a 15-pin Cannon D-type connector which takes the power and control signals to and from the handset along a 4-metre cable.

The cost of the components for the wheel was (in Australian dollars):
machining $80
motor $35
13 magnets $29
Hall effect devices $20
4-m 12-core cable $12

The other parts I had in my junk box, including pieces of aluminium plate, bearings, ICs, LEDs, relays, etc. They would have added a little more to the cost of the unit. The biggest cost is the machining, but Eric charged me a tiny portion of the time involved in making this. About 40 hours of his time was involved, but he charged me just $80. I simply could not have done this if I had to pay commercial rates for the work. (And NO, Eric won't make one for you at that price - so don't ask.)

These 8 pictures were taken before the encoder or 15-pin connector were added. The modifications are visible in the next series of pictures.

25Kb A view of the assembled unit. Note the 1¼" nose to fit
standard focusers, and the cut-out in the bearing cover to
allow it to get as low down over the focuser as possible.
24Kb The CCD side of the wheel. T-threads allow the CCD to
be screwed onto the unit. The motor is mounted on this
side, well clear of the CCD housing.
21Kb Side view of the filter wheel showing how thin it is. Not
shown is the light-tight cover which goes around the unit.
21Kb Another side view with the CCD attached. The unit is just
14mm thick along the CCD axis.
43Kb The two plates and wheel before assembly. You can see
the brass insert which helps support motor shaft.
30Kb The wheel on one plate, showing the micro-switch
19Kb The friction roller on the motor. The shiny aluminium
plate has been partially blackened.
15Kb The wheel in position, showing the detents which
define each position on the top level of the wheel.

These next few pictures were taken after the encoder was installed.

42Kb The encoder. The 4 Hall effect switches are mounted on a
separate plate. The wheel is at position 7, so the magnets
for bit positions 1, 2 & 4 are visible.
37Kb Another view of the filter wheel mounted on the telescope.
Note the 15-pin connector at the top, and the orientation of the
27Kb The handset.

For more information on the use of filters for tri-colour imaging, look here, while here I discuss coloured glass filters in general.

home back to ATM

Page last updated 1996/11/28
Steven Lee