Position signals

[Seadog]

I was under impression that the turret rotates 360 degrees and back, but just to be clear, are you saying that it can do many complete revolutions?

That's the part that I don't think is clear. I would assume it will be and can't see why it shouldn't be. Every pot with a 270° movement that I've ever come across has stops.

 

[Hood]

To be honest I am tring to get something simple like switches but compact. I am not really wanting to have to have loads of circuitry/arduinos etc as that is getting beyond my ken and adds complexity.

 

[ajlelectronics]

That's the part that I don't think is clear. I would assume it will be and can't see why it shouldn't be. Every pot with a 270° movement that I've ever come across has stops.

Not when you gear it down 2:1. The only time it falls down is if the rotation is ever greater than 540 degrees. However a further reduction in the gearing could accommodate that of course. Resolution would reduce, but we don't need that level of resolution anyway.

Actually there is a fly in that of course, the software would need to accept the possibility of two values for position, so that it wouldn't get all confused when it goes "over centre".

 

[ajlelectronics]

If the turret can potentially go many revolutions eventually, seeking the shortest route between tools, then my idea doesn't work as we need the absolute position regardless of how it got there. However, the servo can only drive the turret a finite distance so I think it would work. Even driven from the servo belt maybe.

 

[Hood]

A sensor that has a stop in its rotation is not any good as the turret rotates fully and could keep rotating for as many times as is required.
In reality it will never move more than 180 degrees in one tool change as it is programmed to take the shortest route but the shortest route may always be in the one direction so for example.
Tool 1 0Deg
Tool3 135 deg
Tool 6 270 deg
Tool 2 45 degrees
Tool 5 225 deg
tool 8 etc etc etc so in theory depending on tool calls it could be rotating many times past the zero deg and keep doing so even though it will only ever move a max of 4 positions in one tool call.

 

[ajlelectronics]

tool 8 etc etc etc so in theory depending on tool calls it could be rotating many times past the zero deg and keep doing so even though it will only ever move a max of 4 positions in one tool call.

Yes, that is what I have realised, so we need to sense absolute position because the "zero" of the turret may creep a long way over a few days of use. I will have a coffee and see what I can think of! I wasn't thinking about the fact that the servo has no stops either, hence my confusion.

 

[ajlelectronics]

I'm thinking rotary encoder currently.

 

[Seadog]

I'm thinking rotary encoder currently.

At last

 

[Hood]

Yes rotary encoder is basically what I am wanting but home made to suit the space and don't really want loads of additional circuitry to be required.

I have used OPB917 slotted optos in the past and could use them again but they take up more room than the switches would as the disc diameter to accomodate 4 OPB917's would be quite large. That is why I was wondering if there was anything similar but with 4 channels in one compact housing.

I am not good with electronics so simple is best for me

 

[ajlelectronics]

I have an idea forming. Just need to find the device to show you...

 

[Hood]

I have an idea forming. Just need to find the device to show you...

Hopefully a simple idea for a simple mind

EDIT

Just to clarify the simple mind is mine, just in case it was taken the wrong way.

 

[pressbrake1]

My suggestion is used on rotary transfer machining centres so should be good enough
I take it there is a locking mechanism so what positional accuracy does the lock need

 

[ajlelectronics]

I can't see one illustrated on a quick search. It is a triangular plastic package containing an IR LED and an IR sensor. This will switch when a reflective surface is brought close enough to the business end. You would want four of them in a stack and white marks on the turret (on a dark background) to produce the switching. 4 devices will give you the 8 positions and the output is one that you are familiar with. I would still drive an Arduino with it though.

 

[Seadog]

You could use something like this (ignore every other position) - https://docs-emea.rs-online.com/webdocs/0ff6/0900766b80ff6d94.pdf and mount it on the motor mounting plate in your drawing. Drive would be by another toothed wheel which is driven by the belt. Extremely compact since it occupies dead space between the turret and drive motor and uses the same mounting plate.

 

[Hood]

My suggestion is used on rotary transfer machining centres so should be good enough
I take it there is a locking mechanism so what positional accuracy does the lock need

The problem with using 2 prox's instead of 4 is the space required is not much less. It will be less on the width but prox switches tend to be quite long or at least the cheap ones are. I would then have to work out how to do counting in the CNC control, if it was possible.

The positioning needs is exact as it uses a Hirth coupling when clamped so the turret will always be exactly the same position on tool 2 as it is on tool 2 the next time it clamps there etc etc.

The position sensor doesn't actually need to be deadly accurate as it is not actually positioning things, it is simply a feedback to confirm it is in the position it thinks it is, think a snapped coupling or belt and the servo drive would think it has moved the turret to thr correct position but the turret plate has not actually rotated and just reclamped in the same position.

 

[Hood]

You could use something like this (ignore every other position) - https://docs-emea.rs-online.com/webdocs/0ff6/0900766b80ff6d94.pdf and mount it on the motor mounting plate in your drawing. Drive would be by another toothed wheel which is driven by the belt. Extremely compact since it occupies dead space between the turret and drive motor and uses the same mounting plate.

I am really wanting something to be directly coupled to the shaft that is on the turret plate so that if a belt snaps or coupling breaks the actual position of the plate is known, or rather the position is known not to be what it is meant to be

 

[ajlelectronics]

Still struggling to find the actual product "Ir proximity sensor" . They work something like these though...

https://i.ebayimg.com/images/g/mGMAAOSwn3VavKvz/s-l1600.jpg

 

[ajlelectronics]

It needed the word "reflective" in the search term! 5mm range.

 

[Hood]

You could use something like this (ignore every other position) - https://docs-emea.rs-online.com/webdocs/0ff6/0900766b80ff6d94.pdf and mount it on the motor mounting plate in your drawing. Drive would be by another toothed wheel which is driven by the belt. Extremely compact since it occupies dead space between the turret and drive motor and uses the same mounting plate.

Looking at that link howeever it may be possible to couple that directly to the gearbox shaft. It still leaves the oldham coupling in the train and so potentially that could break but the likelyhood of that is minimal.

 

[Seadog]

I am really wanting something to be directly coupled to the shaft that is on the turret plate so that if a belt snaps or coupling breaks the actual position of the plate is known

If the belt snaps the switch does not move. It has positive detents so will still signal the position it's at.

Or drive the switch directly from the turret pulley by meshing the two.