[ninefinger]

Richard,

If your playing with a rotation of the carousel then don't look at the circumference, look at the angle. IE your going for 10 tool holders spaced equally, then that is simply 1/10 of a rotation per tool. Using the 10x micro stepping then that is 200 steps per tool position. Keep the number of tools an integer that is divisible into the number of steps and everything is cool.

Now I'm not sure what the positional accuracy of a stepper is but you may need to put reduction in to account for that -again keep it an even multiplier. The other reason for a reduction might be 1) to reduce the velocity of the revolving tool holder 2) increased torque so a smaller stepper can be used - inertia of 10 tools at a 10" radius may not be that trivial...

Great work - I am following along with keen interest.

Mike

[rwskinner]

This is the stepper I have: 495 in-oz, 1.8 step angle, 200 steps per turn.
http://www.kelinginc.net/KL23H2100-30-4BM.pdf

The belt pulley combo I'm looking at is a 0.080 MXL pulley set up with 1/4" belt.
Motor = 20 tooth
ATC Spindle 120 Tooth (Not in stock)
6 to 1 ratio
139 tooth belt for a 2.416 center.

2000 Steps per Rev * 6 = 12000 steps per rev.
33.13 / 12,000 = 0.0027" per step
OR
360 Degrees / 12,000 steps = 0.03 degrees

Looking at it the way you mention, 360 degrees / 10 tool holders is a tool every 36 degrees. 36 / 0.03 = 1200 steps between tools.

Richard

[Bubba]

Richard,
Referencing ninefingers post above, I seem to remember specs on several motors indicate a +/- 5% accuracy and further Mariss once said that you cannot rely on micro stepping for absolute accuracy and should consider your design to no more than half steps. So while you will use your 1200 steps as noted above, your guarantee will only be to a half step position +/- 5% on the stepper motor itself.
Hope I haven't confused the heck out of this.

[ninefinger]

Art,

Yes, the positional accuracy of the steppers is what I was vaguely referencing but I didn't remember the actual number. But if its +/- 5% of a full step that is not too bad at all 0.05 x 0.027" (full step) using Richards calcs works out to +/- 0.0013" that would seem to be close enough I would hope and that is worst case. Using a half step then your positional accuracy would be half that so 0.0007" rounding up - I wouldn't be worried about it at that tolerance. Of course this is with the reduction gearing. At a direct drive it does start to become significant (multiply by 6 = +/- 0.0078")

Looking good Richard - keep it up.

Mike

[rwskinner]

Almost all modeled up. I still have the column brackets, air cylinder, and covers. Alibre is weird and it seems to loose motion on certain parts when you start adding sub assemblies.

I'm building a spindle cartridge that houses two 32x15mm AC bearings. I protrudes down 2" for the TTS shank clearance. The slide is a 18" slide with 12" of actual stroke.

My nice little control box with PDB button, solenoid, and ESD will have to be removed. for the clearance or I could mount the ATC on the right side.

Richard

[Starleper1]

Looks like a pretty solid design. I'm curious to see how well the nylon grippers hold. I was throwing around some designs for my 30 taper ATC, with both spring loaded arms, which equals a lot of extra hardware, and the plastic grippers like you have there.

I guess you have to find that happy medium where they hold tightly but are not too difficult to get around the tool holder. At any rate, nice work!

[rowbare]

Wouldn't you want a bit of compliance in your grippers/turntable say 50 thou or so? If everything is totally rigid then you will spend more time tweaking and tuning than changing tools. I imagine that is the reason behind the chamfer on the new TTS adapters. If things are off a touch, the chamfer will guide the holder into the collet. If the turntable is too rigid things would jam up.

bob

[Starleper1]

I haven't looked, does the TTS ATC use plastic grippers?

Also if you look at a commercial tool changer from say, Haas. They use spring loaded arms that would be considered a rigid setup. So it does work, it's just getting everything setup properly like you said. But once you set it, there should be no need to touch it again.

Also with the TTS tooling there is no drive keys in the tool holders, so it doesn't matter how the tool is oriented when its picked/placed. With tools that have these keys you need to make sure the tool doesn't rotate at all in the tool changer when it's spinning.

Who knows maybe the plastic arms are enough to keep everything where it should be. A little R&D is necessary to figure out the best solution for the application.

[rwskinner]

It's going to take some testing for sure. Hoss, Tormach, and Z Bot uses the one peice plastic grippers.

I started to go with a gear setup so there would be a little backlash but ended up with belts. I may have to make some design changes here and there.

I di not like how far out the entire slide sits. I may increase the length of the carriage bracket so the slide sits back a little further.

Ahh, the splindle and motor mounts will have 6061 side plates, as a stiffner and to serve as a cover.

Richard

[small_toys]

I found these while researching grippers and plan to use this design when I make my ATC for BT30 taper.

[rwskinner]

Thanks, that is actually a pretty nice design. The holes weaken it just enough to flex in the proper places.

Richard

[bogiestl]

Uh... I'm working on something sorta similar (but not a tool changer...).

Delrin for the "bearing surfaces," and duh... steel for the springy... Make the grabbies expendable, and mill a bunch at one time. We have multiple component parts available. Let's use them.

[rwskinner]

I have all the materials for the ATC and cut a few small parts but I'm installed Central AC and Heat in my shop so it's side tracked the ATC for a little bit. Holidays are slowing a few things down as well. I'll post something when I get started again.

Everyone have a Merry Christmas and a Safe and Happy New Year. If you don't do the Christmas deal then enjoy your day off.

Richard

[BobWarfield]

Nice article Richard!

I like that gripper you have pictured.

Give some thought to where your design can tolerate inaccuracy and where it has to be pretty right and what your strategy will be to deal with that.

For example, let's consider 2 potential accuracy areas, since indexing the tool was already discussed.

First, what about the Z-height of the tool when the spindle takes it up?

If you have a drawbar, that accuracy need not be great as the drawbar will get the tool pulled up into the taper. If you are going to use an R8 collet-style setup, that accuracy is potentially a big deal as it impacts the Z repeatability of your machine on a toolchange.

If in the latter situation, I would consider some kind of spring load of the carousel with a microswitch. Lower the head until the microswitch closes and you know the tool is pushed up into the collet with at least the force of the springload.

Second, what about Z height and the "fingers" of the plastic tool retainer? The reason I like the one pictured is I could imagine the thin tab being machined slightly smaller than the actual groove in the toolholder to provide some tolerance.

BTW, like the linear rods. I was thinking myself of rail and bearings with V-grooves, similar to what the router guys use. Clamp the V-groove bearings above and below the rail, either 4 or 6 of them. I've seen commerical VMC's use that approach and it seems cheap, cheerful, and strong. The carousel can be deployed either with an air cylinder, stepper, or scew motor with microswitch stops.

Another possibility is the robodrill-style carousel that pivots down.

Cheers,

BW

[rwskinner]

Hello Bob.
Good points.

I currently use the TTS holders and collets and use a power drawbar. The slot in the TTS holders is in the same place for all the holders. The Steel Fork supports the weight and keeps it aligned and the nylon gripper keeps it pulled back into the ATC so it doesn't move around or come flying out.

The TTS holders have like a 30 degree taper if I remember correctly that also helps guide the tool into the collet. This is where it's helpful to have a little slack in the drive mechanism so it can self center. That shouldn't be a problem for me, everything I build seems to have a little slack in it

Merry Christmas,
Richard

[vidio1]

Attached is a pdf of a tool holder.......

[Starleper1]

I think to some degree this is one of those projects that can look perfect on paper, but is going to take some real trial and error to figure out where it's weaknesses are. When you take something like this on, there are so many different issues that could arise it's tough to work out the kinks in a CAD system. Hopefully everything works out for you first try but I've learned not to get my hopes up.

[rwskinner]

Yep, but generally that is the way most of my projects work out. Build a prototype and modify what needs modification.

Richard

I think to some degree this is one of those projects that can look perfect on paper, but is going to take some real trial and error to figure out where it's weaknesses are. When you take something like this on, there are so many different issues that could arise it's tough to work out the kinks in a CAD system. Hopefully everything works out for you first try but I've learned not to get my hopes up.

[rwskinner]

Yesterday I was able to find time to machine the spindle shaft, spindle cartridge, and spacer hub. This afternoon I machined the center disc and assembled.

I'm going to try and get at least a few metal forks made this evening. I made one out of aluminum and it worked wonderful. I still have to make one of the nylon grippers to see how well it works before I get too carried away.

Let me get a little work done and I'll try to post some pics this evening.

Richard

[rwskinner]

All the fingers including the Nylon fingers machined today and bolted up. They grip pretty good. Maybe a little too good. Once I get everything together where I can test I can always slot or relieve the nylon some. They may be fine the way they are.

Pictures....