[davidmb]

Just started out making my own automatic tool changer spindle, I have based it loosly on a scaled down version of commercial taper tooling, I have elected to use an 8.5deg taper with conventional pull stud design, I have just finished the main spindle component. The spindle will run in a pair of angular contact bearings at the nose and plain radial bearing at the top end, drive will be via the side of the main housing via a toothed belt driveleaving the top of the spindle assembly free for the pneumatic cylinder to release the taper tooling.

[TOTALLYRC]

Looks like you are off to a great start.

At some point I need to do something similar for my router as well.
Since I have another machine that uses ISO-20 tool holders I will make mine new one to match.

Mike

[tpworks]

Would wheel bearings such as Honda accord front wheel bearings work as lower spindle bearings?

[jsheerin]

A wheel bearing might work if you were okay with a max speed around 1000 rpm. But even then, they would probably not be the best choice.

[davidmb]

Hi tpworks & jsheerin,

I'm using 7402's as the main bearings, they are available in various qualities and tolerances ( and prices ! ) that run upto about 16000 rpm and a dynamic load rating of 15000N ( 1500 Kgf) should be overkill and make for a long lasting and maintainable spindle. Just found that the taper is very slightly bell mouthed by about 0.0007", suspect that the grind wheel was just ever so slightly below center when finishing, hope I can correct this.

[davidmb]

After spending 2+ hours trying to set the compound slide on my lathe to a accurate angle ( +- 1' of a degree ) using a dti I gave up in disgust and retired to the house. Pondering the problem I came up with the following idea, setting the compound slide to the test bar is ralatively straight forward, if this is done, then with a mirror attached to the toolpost and a laser and a steel rule, I should be able to set the compound angles very accurately and repeatably, the picture shows what I am going to set up and try.

[Big-tex]

I wanna see where this is going, I guess I have ATC bug

[davidmb]

Finally got arount to setting up a laser pointer to see if my idea for setting accurate angles works, longest job is setting the topslide parallel to the axis of the lathe, this is accomplished using a dial gauge and is the hardest part of the setup. (NB. the topslide needs to be properly adjusted and run without bindiing)

once this is done, setup the mirror fixture and laser so that the reflected bean is roughly vertically in line with the laser itself, now move the steel rule so that the reflected beam is in the middle of the rule. ( Once this is done, the rest of the setup is really easy, honest )

Now we need a little trigonometry to complete the setting, given the angle that we need to set the topslide to

offset = tangent ( angle ) * (distance between mirror and rule)

for my setup, I have 1665mm between the mirror and the rule, this equates to 1mm on the rule is approx 2 minutes of arc, the two tapers in the thumbnails were cut by just setting the topslide with the laser, took about a minute to change between settings, this is the best matching taper I have ever cut, the taper locks together when turned by hand. Now I can get back on and re-cut the taper on my spindle, knowing I can machine my own taper tooling. :banana:

[RomanLini]

Nice idea with the laser. I've done a similar but less accurate system on my lathe using a straightedge and a ruler, but a laser is genius.

...
Now we need a little trigonometry to complete the setting, given the angle that we need to set the topslide to

offset = tangent ( angle ) * (distance between mirror and rule)
...

Isn't the mirror angle (each taper side angle) half of the reflected angle?

So if you are measuring 10' reflected angle at the ruler, that means the mirror and taper have an angle half that, ie 5'.

[davidmb]

Hi RomanLini,

Yes the machined angle is half the reflected angle, after measuring the machined taper ( so not quite a genius - but that means it's twice as accurate as I thought ). For higher accuracy the ruler could be moved further away, a front surface mirror could be used to eliminate the reflection off the front of the glass and some way of focusing the laser spot to something real small, I can do it with a hand magnifier to get a very small spot but I would need a lens with a focal length of about 3.4m ( not found one yet but Edmund Scientific said they could make one, bit expensive though ), the idea came to me from remembering how a reflecting galvanometer works ( very sensitive moving coil meter ).

[RomanLini]

Here's a thought, if you mounted the laser in a mount with a little battery and strong magnet you could clip it on your lathe bedway, and laser mark a 0 degree reference spot on the wall. Then clip it on your crossslide and see the spot on the wall to set the angle by tangent. If the wall is a reasonable distance away it would be very accurate.

Assuming your lathe doesn't move you could even write calibrated markings on the wall, it would allow quick easy setup with a magnetic laser, the same easy way you can use a magnetic base dial gauge.

[davidmb]

Hi RomanLini,

Since it worked so well I am already making a mirror mount that will be a permanent fixture to the compound slide, should be coolant proof, I can fit the laser directly to the wall just above the 1m steel rule ( bought specially ), so I can have a permanent setup, means I can go back and remake taper tooling whenever I need.

[RomanLini]

That sounds pretty good. Provided the lathe doesn't move I would expect it to be accurate and repreatable once it's calibrated.

You may get issues with large taper angles like 30' as you would be working with a reflected angle of 60', that's going to tie up a lot of wall space...

[davidmb]

Hi RomanLini,

I hear you on the angle, but I only needed to accurately setup for an included angle of 3.5 in 12 ( 16.26 deg standard self releasing taper similar to iso and cat tooling ), the pneumatic solonoids turned up today for the pneumatic tool ejector, going to use belville spring washers to provide the tool retaining force, not looking forward to strip the machine to pull the nylon tubes into the cable chains.

[RomanLini]

Maybe you could leave a pull-string inside the cable chains, so when you add cables and hoses you just attach to the string and pull them through.

[cornbinder23]

Maybe you could leave a pull-string inside the cable chains, so when you add cables and hoses you just attach to the string and pull them through.

The same way most industrial electricians do it they pull poly yarn in with the wires for any future pulls through the raceways or conduit

[davidmb]

Is'nt hindsight wonderful, I'll try and remember for next time , when I built the machine I never even considered that I might even need an ATC spindle, and yes I have managed for 3 years now changing by hand but the class of work I am now wanting to do requires quite a few tool changes and just picking the smallest tool required would add too much time to machine.

[ger21]

Make sure you put 5 or 6 strings in there. In the last 2 weeks, I've painfully fed 4 wires through my 1/2"x1" chain. I told myself to run some string, but was too lazy. Now I need to run 2 more, and it's getting pretty tight. Takes about15 minutes to feed the 5ft.

[cornbinder23]

Make sure you put 5 or 6 strings in there. In the last 2 weeks, I've painfully fed 4 wires through my 1/2"x1" chain. I told myself to run some string, but was too lazy. Now I need to run 2 more, and it's getting pretty tight. Takes about15 minutes to feed the 5ft.

What I do is just pull 1 poly line/string and if I make another pull later on down the road I pull another one with the conductors, as I am using the old one for the conductor pull, keeps the raceway fill to a minimum and also minimizes the tangling/twisting

my2c

JTCUSTOMS

[CarveOne]

Make sure you put 5 or 6 strings in there. In the last 2 weeks, I've painfully fed 4 wires through my 1/2"x1" chain. I told myself to run some string, but was too lazy. Now I need to run 2 more, and it's getting pretty tight. Takes about15 minutes to feed the 5ft.

Why not just pull in a new string along with the new wire?

CarveOne