[SCzEngrgGroup]

OK, kiddies, here's a machining test:

You want to make six pieces of the part pictured below, but with the following restrictions:

1) The finished parts must look exactly like the model - NO differences.
2) You must make all six pieces, from a single piece of stock, in a single, un-interrupted run of a single g-code program. Once you press CycleStart, the program will must run, non-stop, with NO operator intervention, until all six pieces are completed.
3) There must be NO second ops - the parts must be ready to go straight from the machine into service - no sanding off tabs, removing glue, cleaning up clamp marks, etc.

How ya gonna do that?

Regards,
Ray L.

Attachment 291580

[RCaffin]

Hi Ray

That part needs machining from all 6 sides - as far as I can see. You are not going to make 6 of those in one program from one bit of stock on one machine in one setup.

You may not like my answers:
* I would redesign the part so it can be made in-house on a production basis. (Been there, done that, many times.)
* Alternately, I would have it done externally by injection molding. There are costs ...

Cheers
Roger

[SCzEngrgGroup]

Hi Ray

That part needs machining from all 6 sides - as far as I can see. You are not going to make 6 of those in one program from one bit of stock on one machine in one setup.

You may not like my answers:
* I would redesign the part so it can be made in-house on a production basis. (Been there, done that, many times.)
* Alternately, I would have it done externally by injection molding. There are costs ...

Cheers
Roger

Roger,

Actually, I already HAVE made them, exactly as described. I just want to see if anyone can figure out how to do it. For production, they will be injection molded.

Regards,
Ray L.

[RCaffin]

Hi Ray

Ah - extra info: they have been made. OK!

The two problems I can see are the undercut which is not shown adequately, and the clamping. More views please?

Do we have to assume straight bar stock with zero prep, or can we have some prep beforehand?

Cheers
Roger

[AUSTINMACHINING]

OK, kiddies, here's a machining test:

You want to make six pieces of the part pictured below, but with the following restrictions:

1) The finished parts must look exactly like the model - NO differences.
2) You must make all six pieces, from a single piece of stock, in a single, un-interrupted run of a single g-code program. Once you press CycleStart, the program will must run, non-stop, with NO operator intervention, until all six pieces are completed.
3) There must be NO second ops - the parts must be ready to go straight from the machine into service - no sanding off tabs, removing glue, cleaning up clamp marks, etc.

How ya gonna do that?

Regards,
Ray L.

Attachment 291580

1.Install atc


Can't see the backside, but make holes, secure with fasteners, machine all features, perimeter,undercut with your saw or keyway cutter. Done.

[RCaffin]

OK. A bit wasteful on metal, but maybe the metal is cheaper than the machining.
I will add that I do something similar but 10 at a time, with a number of jigs.

I will describe for one unit: repeat N times along X axis. I normally machine all 10 units with each tool before going on to the next tool.
Order of machining operations not super critical for the most part.

Start with large rectangular bar mounted upright (ie not flat) along X axis. Bar to be significantly higher than the part.
Drill big hole to slightly over needed depth.
Drll small hole ditto
Machine out cavity
Machine external contour to required depth or part ht
Bevel or chamfer top edge
(Note: bottom edge is NOT chamfered)
Using T-slot cutter, undercut protruding lug, including the rounded corners
(At this stage, part is still attached to bottom part of bar)
Using large slitting saw, slice off the part from the lower part of the bar

I have reservations about whether a little tumbling or deburring might be needed for the bottom edge.

Cheers
Roger

[Hirudin]

I think one of your previous posts telegraphed the answer...

Clamp the stock down, do some 2.5D milling, perform undercut, use slitting saw to cut parts off of stock. I'd also do some back-chamfering on the edge of the undercut for aesthetics/deburring.

Without dimensions it's tough to see how you might fixture the stock in the first place, but my guess would be a somewhat thick fixture with Mitee-Bite clamps, something tall enough to allow the slitting saw arbor to get to all the parts without hitting the table/vise and Mitee-Bite clamps to allow the stock to be held only at the bottom so there isn't too much waste. If the part is 12mm tall you could probably make them out of 3/4" stock taking into account the 1/4" or so of stock lost to clamping and the kerf of the saw.

[tmarks11]

...You must make all six pieces, from a single piece of stock, in a single, un-interrupted run of a single g-code program....with NO operator intervention

Are you using a vertically mounted rotary table? Because it looks like both the top and bottom required machining,

EDIT: We are missing a critical piece of information: size. RCaffin suggested using a t-slot cutter to do the undercut with-out flipping, which would work up to a certain size part. I was thinking this was a larger piece.

"Lift Sensor Bracket". So I am guessing this is indeed small. Did RCaffin win the prize?

[SCzEngrgGroup]

The part is small - 0.85" x 1.2" x 0.5". The overhang on the bottom is about 0.380", the holes are 0.375" and 0.156".

I'll explain what I did in a while, in case others respond. There was NO prep of the material - I cut a single piece of stock to rough length with my table saw, mounted it on the machine, pressed CycleStart, and walked away. When the machine stopped, I had six fully finished parts waiting for me. I will say that what I did was far from the fastest way of getting these parts machined, but the fact that I didn't have to be there to baby-sit the machine, mount new blanks, turn them over half-way through, etc. I'll make that trade-off any day, as I can be off doing other useful work while the machine is working away.

Regards,
Ray L.

[SCzEngrgGroup]

Roger came closest to what I do, though I think my way is much more material-efficient. I use 1.5" round stock, mounted vertically in a 3-jaw chuck. The part is machined, with the under-cut and parting performed using a slitting saw (a dovetail cutter would be better for the under-cut, but I don't have one). Then, I use G52 to shift the Z origin down, and repeat 5 times. The slow part is cutting the under-cut, because the saw I have is pretty thin (0.075"), so I have to cut one slot, then step down and repeat about 6 times. I do the slitting at 0.07" DOC, 0.050WOC, 500 RPM, 100 IPM. But, despite being slow, I get a pile of very nice parts, with no human intervention. I find this is often the most efficient, and easiest, way to make small parts that may otherwise be difficult to hold.

Regards,
Ray L.

[Hirudin]

Interesting. I've thought about cutting parts like that before, but never actually tried it.

[RCaffin]

Hi Ray

So what's my prize?:cheers:

Yes, I did think about using bar stock on HMC with a 5-axis head on the side ... :banana: :bat:
But the method I described is something I could actually do.

The slitting saw for the undercut must be awful slow. Have you thought about using a T-slot cutter instead? After all, you have an ATC there ...

Cheers
Roger

[SCzEngrgGroup]

Hi Ray

So what's my prize?:cheers:

Yes, I did think about using bar stock on HMC with a 5-axis head on the side ... :banana: :bat:
But the method I described is something I could actually do.

The slitting saw for the undercut must be awful slow. Have you thought about using a T-slot cutter instead? After all, you have an ATC there ...

Cheers
Roger

Roger,

The slitting saw is not fast, but it's not all that slow either, since I'm cutting at 100 IPM. And that's taking it easy. I'm sure I could get a lot more aggressive. Since the under-cut is about 0.4" deep, that would be a pretty big T-slot cutter. But it would be much faster than the saw.

Regards,
Ray L.

[RCaffin]

Hi Ray

OK, 0.4" is a bit of an undercut, but it is feasible. Multiple passes I think.

MSCDirect: $150 to $600. Cheap, aren't they? Actually, I think the MSC prices are utterly ridiculous. I don't care what their quality is like: their markup is just out of sight.

Amazon: a search on T slot milling cutters brought up a suprising amount of rubbish - enough to send you away from them. Cheaper though: the Magentoo brand (=Chinese I am sure) was under $20 for something usable in HSS.

eBay: lots of stuff ex Hong Kong for <$20.

But then, if you can walk away and do something else...

Cheers
Roger

[engnerdan]

Roger came closest to what I do, though I think my way is much more material-efficient. I use 1.5" round stock, mounted vertically in a 3-jaw chuck. The part is machined, with the under-cut and parting performed using a slitting saw (a dovetail cutter would be better for the under-cut, but I don't have one). Then, I use G52 to shift the Z origin down, and repeat 5 times. The slow part is cutting the under-cut, because the saw I have is pretty thin (0.075"), so I have to cut one slot, then step down and repeat about 6 times. I do the slitting at 0.07" DOC, 0.050WOC, 500 RPM, 100 IPM. But, despite being slow, I get a pile of very nice parts, with no human intervention. I find this is often the most efficient, and easiest, way to make small parts that may otherwise be difficult to hold.

Regards,
Ray L.

Not a bad way to do it with the volumes you run.

A friend of mine, who runs large production jobs, does a very similar process on parts like this. He will "5 side them" (cut the top and 4 sides) out of a bar of material, maybe 5 parts per bar. Instead of using the slitting saw he would make a set of soft jaws to hold all 5 parts at once and remove the remaining bar stock and free all 5 parts during the second op. But for him he would be making 500 of a part and the faster cycle time of not using the slitting saw pays off.

[SCzEngrgGroup]

Man! This is starting to feel like real work! I've finally got the production PCB design completed, and will go out for fab on Monday:

Attachment 294152

The PCB is 13.5" in diameter, and mounts inside the upper shroud for the carousel. This location minimizes cable lengths (most will be on the order of 18" long), and negates the expense and hassle of a separate enclosure, or very long cables and conduit needed to extend it into the machines E-box.

I now have all the forms for the vacuum-formed parts completed, and will be making the first actual production pieces next week. I am absolutely thrilled with how the plastics are coming out. Now that I have the hang of the vacuum forming, the parts are very quick and easy to make, look great, and should be nearly indestructible.

To save a little time, the first production units will use cast urethane instead of injection molding for the small plastic parts, so I hope to get those done in the next week as well, using one-piece silicone rubber molds made from aluminum models. I will likely use this same method to make some custom molded connectors. This will also enable me to eventually make all the sensors on the PDB user-replaceable modules as well, rather than being epoxied into place as they are now. Not that there have been any failures to date, other than through abuse...

I've also finally ungraded my MiG welder from gasless to gas, which will give me better looking welds (for the mounting brackets).

Things are finally starting to fall into place, so there are no significant obstacles to ATC production now - a lot of work still to do, but most of it is now pretty routine stuff.

Regards,
Ray L.

[mklotz70]

That's awesome things are going well!! I'm sure I'm not the only one that's happy to hear this!

[SCzEngrgGroup]

Made my first silicone mold today, for casting the small plastic parts for the ATC! Came out pretty much perfect on the first try. The detail is amazing - even the finest machining marks are perfectly transferred to the mold! The whole process is very simple, fairly quick, and relatively inexpensive.

Here is it, along with the machined aluminum parts used as patterns:



Tomorrow I'll make the first cast urethane parts, and, hopefully, complete assembly of the first set of "production" shrouds.

Regards,
Ray L.

[SCzEngrgGroup]

Made the first cast urethane parts today! While hey didn't come out quite perfect, they are perfectly usable, and better than I expected on the first try. The detail is amazing - all the machining parts that transferred from the aluminum patterns to the silicone rubber molds are just as sharp and clear in the molded parts. The process is really very easy, and quick. I will definitely be using this process for other things, where appropriate.



Update - Now that the urethane has fully cured, I can't believe how hard, and tough, these little parts are! This stuff is impressive!

Regards,
Ray L.

[SCzEngrgGroup]

One full set of sensors done! This process works really well - MUCH faster and easier than machining the little housings, and they look great. The real production ones will look even better, in black, rather than the natural "tan" urethane. In the first photo you can see the lock sensor on the left, and the carousel tool and home sensor on the right (which mounts out-of-sight on the underside of the carousel top plate. In the second photo is the arm home sensor (near the bottom right) and the two lift limit sensors on the left.

Attachment 294504

Attachment 294506

All the plastics issues are now well under control, with at least one full set of "production quality" prototypes on-hand. With luck, in about 10 days I'll have the first new PCBs, which is the last major missing piece. Once those are checked out, it should be clear sailing to the first true production ATCs, with luck well before end of the year!

Regards,
Ray L.