Home Made Remote Controlled Filter Wheel

[JamesF]

I wonder if it might be better to have the motor drive the edge of the wheel for more accurate control. I can't remember enough schoolboy physics, but once the wheel is rotating won't it be easier for it to continue to turn the motor once the power has been stopped if it's driving the centre rather than the edge?

It did also strike me that if you're using optical/IR sensors for feedback on the wheel position it might be useful to have the "holes" in pairs, so you slow the motor on the first and stop it on the second. I'm not totally convinced they're the right way to go though as they may be error-prone and in the long term you need the wheel to stop absolutely predictably every time. Some system that positively locates the wheel and cuts motor power at the same time would be the ideal I'd have thought.

James

[Bizibilder]

You could use a geneva wheel to turn the filters: Geneva drive - Wikipedia, the free encyclopedia it is the mechanism used to move "movie" film through a camera one frame at a time. Also used in clocks (my other hobby!) to move datework one "day" at a time. It works for both fast and slow mechanisms and is a precise method of moving / stopping your action. Simple(ish) to make as well.

[Gina]

I wonder if it might be better to have the motor drive the edge of the wheel for more accurate control. I can't remember enough schoolboy physics, but once the wheel is rotating won't it be easier for it to continue to turn the motor once the power has been stopped if it's driving the centre rather than the edge?

Yes, I've come to that conclusion having been thinking about it for a while.

It did also strike me that if you're using optical/IR sensors for feedback on the wheel position it might be useful to have the "holes" in pairs, so you slow the motor on the first and stop it on the second.

That's one idea, I agree. I'm thinking though that the wheel could turn relatively slowly - after all if I'm taking 30s or greater exposures, to take 2 or 3 secs to rotate the wheel is no problem.

I'm not totally convinced they're the right way to go though as they may be error-prone and in the long term you need the wheel to stop absolutely predictably every time. Some system that positively locates the wheel and cuts motor power at the same time would be the ideal I'd have thought.

Yes, I've come back to that idea too. Such as roller micro switch dropping into a notch on the edge of the wheel - a notch corresponding to each filter position.

[Gina]

You could use a geneva wheel to turn the filters: Geneva drive - Wikipedia, the free encyclopedia it is the mechanism used to move "movie" film through a camera one frame at a time. Also used in clocks (my other hobby!) to move datework one "day" at a time. It works for both fast and slow mechanisms and is a precise method of moving / stopping your action. Simple(ish) to make as well.

Yes, I have thought of that but thought it would need some precision engineering tools to make it. When well made, it certainly locks the wheel in each position very precisely. That would be a very elegant solution. I wonder if it would be practical - I'll give it some thought.

[Gina]

I've been looking at numerous motors and drives and what voltage the motors run at. With a 12v supply (13.8 actually) a 3 to 5v motor is not so good as a 12-24v one. This has brought me to the type of motor/gearbox I'm using for the remote focussing unit. The gear ratios are very easily changed and the unit is pretty rugged from my experience of using these to rotate CCTV cameras plus reasonably immune to dust and damp. Not that I expect to have much of either in my scope room, but it does mean it doesn't need boxing in for it's own protection. Furthermore, I have one to spare.

I have also been looking at the control electronics and sensing. The idea I proposed of having a separate encoding disc I can now see isn't possible. Whether inside or outside the casing, it would interfere with the focuser or camera parts unless it were very small (2" or less, certainly not CD/DVD size).

So I'm back to using the filter wheel itself as is done in commercial units. Some use little magnets between the filters and hall effect devices to sense. I've looked for hall effect devices which seem a bit thin on the ground except from Hong Kong/China or the USA. So I may try reed switches as in one of my original ideas. With the exact positioning being covered by a micro switch roller dropping into a slot, the position sensing is not a position critical measurement as long as it identifies the correct filter. With 8 hole units, they use three sensors to encode the 8 positions into 3 bits of data. The data bits being determined either by the presence/absence of magnet or it's polarity. Presence/absence could be used with reed switches. 9 holes would need 4 data bits. I already have a couple of dozen 6mm diameter by 3mm thick high-power magnets, so that's no problem.

[DIYASTRO]

Gina,

Some nice and interesting suggestions for your filter wheel construction coming through, as i have said before within another thread that we have exchanged ideas on. I still am of the opinion and agree with chris (Yes Yes), that a blank position has its uses.

In terms of drive. I also looked at the Geneva drive ( since it was my first project at engineering collage after my initial two years) yes its a brilliant drive system but takes up a lot of space within the 'wheel' itself. The internal geneva is simply wonderfull when you get it acurate enough. The problem I found was that in order to get a fixed possition for a filter the drive pin needs to be at some located position within the slot so the problem of accurate indexing is still with you since the motor drive still needs to be stopped perfectly. So really no gain. As you know I have gone for the stepper / gearbox and toothed timming belt option. ( As you have seen in my pancake / filter focus thread) I would have prefered the filter wheel to be a huge spur gear with direct drive from the stepper / gearbox but too expensive.

You expresed a liking for a center drive shaft. Good option if you can keep the drive gearing to the outside of the 'filter sandwich' so to keep the depth / thickness down to managable size.

If you have a look on the Yes Yes webcam mod thread ( perhaps you have seen it) I managed to get the motor and worm drive within the webcam itself complete with the stepper motor drive circuit to revolve the filter wheel and it is very accurate. (See page 10 and top of page 11 on the Yes Yes wb cam mod thread for pics)

What i am trying to say is that you have to do an awfull lot of work to get the same result as a proggrammable stepper motor driver and well engineered gear drive.

The proposed sandwich filter wheel construction that you intend is also one that I have tried beware! not all filters from different manufac. are the same thickness best to try and keep to one make.

The servo that i have tried that rotates through 360 deg was from hobby tronics I think at around the prices that you have from everyone else but very powerfull in torque.

regards

Boyd

[Gina]

Thank you for your input Boyd I'll go through your post and answer in detail tomorrow. I'd just like to say... I thought the idea of the Geneva motion was that the Maltese cross was locked over three quarters of the indexing wheel - that is the "fixed" position. The motion can be stopped with the peg anywhere outside the slot. It's moving the cross wheel onto the next position when it's in the slot. Of course, with more than 4 positions, it isn't a Maltese cross but more like a star with more points. Probably why it's now more often called "Geneva Motion".

With 4 positions it locks very solidly but with say 8, it would be less solid as there would be less of the surface in contact and instead of 45 degrees of arc it would only be 22.5.

[Gina]

Gina,

Some nice and interesting suggestions for your filter wheel construction coming through, as i have said before within another thread that we have exchanged ideas on. I still am of the opinion and agree with chris (Yes Yes), that a blank position has its uses.

There will be several blank holes while I'm building up my filter sets. I shall start with the Baader parfocal LRGBC filter set and add Baader NB filters later. However, it would be possible to have an empty hole if I go for 9 holes. I've done the geometry and 9 holes would be OK with the size of disc I have on order but 9 goes from 3 data bits to 4 and increases the electronics complexity. Microswitch and 3 filter ID bits makes 4 which fits nicely with 4bit data chips. However, I'm still thinking about how I'll implement the control electronics and PC interface.

In terms of drive. I also looked at the Geneva drive ( since it was my first project at engineering collage after my initial two years) yes its a brilliant drive system but takes up a lot of space within the 'wheel' itself. The internal geneva is simply wonderfull when you get it acurate enough. The problem I found was that in order to get a fixed possition for a filter the drive pin needs to be at some located position within the slot so the problem of accurate indexing is still with you since the motor drive still needs to be stopped perfectly. So really no gain. As you know I have gone for the stepper / gearbox and toothed timming belt option. ( As you have seen in my pancake / filter focus thread) I would have prefered the filter wheel to be a huge spur gear with direct drive from the stepper / gearbox but too expensive.

See my post above. A huge spur gear would be great but not practical for a cheap build.

You expresed a liking for a center drive shaft. Good option if you can keep the drive gearing to the outside of the 'filter sandwich' so to keep the depth / thickness down to managable size.

I've now gone off that idea, preferring edge drive for more control.

If you have a look on the Yes Yes webcam mod thread ( perhaps you have seen it) I managed to get the motor and worm drive within the webcam itself complete with the stepper motor drive circuit to revolve the filter wheel and it is very accurate. (See page 10 and top of page 11 on the Yes Yes wb cam mod thread for pics)

What i am trying to say is that you have to do an awfull lot of work to get the same result as a proggrammable stepper motor driver and well engineered gear drive.

Yes, I remember that but I want to keep the weight down - with the frac I'm already well out on a limb with the camera etc. on an 80mm extension tube. The thickness of the filter wheel unit will mean the draw tube will be further in but I'm still concerned with the weight overhung on the scope. I will need to add extra weight on the objective end of the scope for balance as it's already as far as it will go in the rings and on the mount.

The proposed sandwich filter wheel construction that you intend is also one that I have tried beware! not all filters from different manufac. are the same thickness best to try and keep to one make.

Yes, I intend to stick to Baader filters.

The servo that i have tried that rotates through 360 deg was from hobby tronics I think at around the prices that you have from everyone else but very powerfull in torque.

Ah yes

[yesyes]

I think I would go for a stepper motor rather than a servo or DC motor.

With a stepper you would only need one index position on your wheel (kind of like a home position). From there you can move the stepper by number_of_steps_per_revolution / number_of_filters to advance to the next filter. Also, the holding torque of the stepper would hold the filter in position. No need for any brake mechanism or similar.

The control electronic would be a bit more complex but I'm confident you'd manage that ..

[Gina]

I think I would go for a stepper motor rather than a servo or DC motor.

With a stepper you would only need one index position on your wheel (kind of like a home position). From there you can move the stepper by number_of_steps_per_revolution / number_of_filters to advance to the next filter. Also, the holding torque of the stepper would hold the filter in position. No need for any brake mechanism or similar.

The control electronic would be a bit more complex but I'm confident you'd manage that ..

That would be fine if I had a way to positively connect the stepper motor to the wheel such as teeth on the wheel or a toothed belt drive. The large pulley would have to be either quite small to clear the filters or large enough that the filters could go inside. The small size pulley would not give accurate enough positioning but I don't know if a 6" pulley would be available - I'll have to check.

I haven't found details of any commercial filter wheels that use this method. Most seem to use friction drive to the edge of the carousel and either hall effect or optical position sensing. For example, here is a photo of the nine hole Atik EFW-2 Filter Wheel carousel. This appears to have a friction ring around the edge and holes in the wheel for optical sensing. This carousel has one index hole for each filter and one other, presumably to sense the home position (filter no. 1). This gives me an idea for simpler ID sensing.

[yesyes]

I was thinking of a stepper like this one:

5V Micro Mini Reduction Stepper Motor for PIC 51 AVR | eBay

This is geared and has 4096 steps / rev (after the gearing). You could mount the wheel directly onto the stepper shaft. The shaft is off-centre, so the stepper would be out of the way but still near the centre of the filter wheel.

[Gina]

I was thinking of a stepper like this one:

5V Micro Mini Reduction Stepper Motor for PIC 51 AVR | eBay

This is geared and has 4096 steps / rev (after the gearing). You could mount the wheel directly onto the stepper shaft. The shaft is off-centre, so the stepper would be out of the way but still near the centre of the filter wheel.

Looks a possibility, there wouldn't be any clearance problem, but would it be strong enough? There's bound to be some backlash in the gearing so it would still need a more accurate positioning mechanism (such as the micro switch).

[Gina]

This is a rough diagram showing a 9 hole wheel with micro switch and 1 of 9 notches in the wheel. The reference for filter 1 is a magnet and reed switch.

[steelfixer]

Gina

What about using a toothed belt but turn it inside out so the teeth are on the outside. Fix that to the outer rim of the wheel and drive it with a stepper motor mated up to a corosponding gear wheel.

[Gina]

Gina

What about using a toothed belt but turn it inside out so the teeth are on the outside. Fix that to the outer rim of the wheel and drive it with a stepper motor mated up to a corosponding gear wheel.

That's a nice bit of lateral thinking - thank you

[Gina]

Just one little problem... finding a belt exactly the right size.

[steelfixer]

Gina - what is the diameter of your wheel?

Scrub that question I have just found it 110mm.

[steelfixer]

Gina check out this link.

It is for a flexible gear strip.

motionco_Plastic Gears_Flexible Plastic Gear Racks_Buy Online

I konw its 10mm wide at its widest but the teeth are only 4mm and you can soon trim off the excess.

[Gina]

110 was for the 6 hole wheel, now I'm considering more holes and the next size up is 145mm and will take 8 or 9 filters. I'll probably design for 9 holes.

I've been looking at the design of a Geneva drive. To get a decent locking action the slot and notch would need to be quite deep, taking up a lot of "real estate" at the edge of the wheel. It seems to need about a quarter of the radius. A possibility is to use two rotations of the smaller wheel to give one position change. That would make the smaller wheel half the size and therefore slot and notch half as deep. So for 9 holes it would need 18 slots and 18 notches. And one eighth of the radius taken up with the drive. I'll see how this works out.

[Gina]

Gina check out this link.

It is for a flexible gear strip.

motionco_Plastic Gears_Flexible Plastic Gear Racks_Buy Online

I konw its 10mm wide at its widest but the teeth are only 4mm and you can soon trim off the excess.

That's very interesting Not cheap though - I would need 500mm (actually 455.53mm) which costs about £25. It's not designed for continuous rotation and the join would need thinking about to allow full rotation.

[steelfixer]

Gina you would not need to have a joiner in it as you would glue it all the way around the disc.

[steelfixer]

Gina

I have just found this. Toothed belt 455 mm long - £2.00.

5mm HTD® Timing Belts [4575] : Beltingonline.com, Supplier of power transmission and conveying products

Enjoy

[steelfixer]

Gina me again.

Here is a link to just about anything you could want.

Component, Materials and Tools for Kart, Buggy, Robot and Model Engineering Builders

[Gina]

Yes, that's a possibility. I wonder what the chances of getting the teeth to match up would be. Oh I can would out the mathematical probability if I knew the "stretchability" of the belt.

OTOH if it's a question of gluing the belt on I could use one of these and cut it to fit. These belts have a finer tooth pitch as well as being much cheaper :- Synchroflex Timing Belt T2.5x6mmx480mm

[steelfixer]

Gina the link for the toothed belt is the correct size for your disc 455mm so you will not need to cut it, and it is only £2.00