Short Run Production!

[DocsMachine]

I have, for quite a few years now, been making short runs of specialty production parts- specifically for paintball guns. Since this is not a big-bucks market to start with, and I'm trying to do it out of backwater, Left Armpit, Alaska- and on a frayed and knotted shoestring budget- I've been doing those short runs of 20 to 50 parts on plain engine lathes and fully-manual Bridgeport style milling machines.

Last year, I "upgraded", for want of a better phrase, to an actual Turret Lathe- which was still an obsolete antique, but a notable step up in speed to cranking the handles on an engine lathe.

This year, I finally upgraded to a real, live, CNC turning center.



I'm still very much the green newbie when it comes to programming a CNC, but I'm getting there. So for the time being, my workflow is still a bit of a mix- some is done on the old turret, some done on the engines, some on the new CNC. Eventually I'll get to the point where I'm making the whole thing on the CNC, though maybe still with a bit of 'second op' on a manual, but for now it's still that mix.

I've also been playing around with putting together some videos, so as I'm running a very small batch (about 30 pieces) I shot a few scenes and put together a quick clip:

Part 2 will be turning on the CNC, and Part 3 is back to the manuals for knurling and finishing.

Doc.

 

[dr_clyde]

Man, doing any sort of production on a manual machine tool is borderline masochism now. I can see making a dozen or so, but runs of 20-50 parts would make me want to jump on a pile of kitchen knives.

You will like having a CNC, it makes machining much more enjoyable for any amount of quantity, or even one offs where the geometry is sufficiently complex. Any radii, threads, tapers, chamfers, or even simple shoulders are much faster and more consistent on a CNC lathe. Once you get comfortable programming, you won't want to run parts on the manual stuff if you can help it.

That lathe looks pretty small, what are the specs on it?

 

[mogandave]

3"?

 

[DocsMachine]

Part 2 is up!

Man, doing any sort of production on a manual machine tool is borderline masochism now.

-Certainly, but up until very recently, I had no real choice. Besides the fact I didn't actually know (and still kind of don't) how to run one properly, I'm also in Alaska. There are NO used CNC machines available up here. I don't mean "there's not a very good selection", I mean there aren't any. There are some in the State, yes, but they never come up for sale.

The other half of that is my particular little market is not a high volume, high profit one. Yeah, had I known then what I know now, I might well have gone into a different line of work, but I've carved out my little corner. Before now, a "run" or fifty parts was big, in many cases, basically a couple years' supply. I'm hoping these machines can change that.

Once you get comfortable programming, you won't want to run parts on the manual stuff if you can help it.

-Well, I'm also kind of a job shop, and I like making one-off parts. The CNCs are for the production stuff- let the robot do the repetitive stuff.

That lathe looks pretty small, what are the specs on it?

-Which one?

The blue Omniturn is generally a 5C collet machine, though technically it can accept up to a 4" chuck or 4" "pot chuck" (a 5C collet with a big head.) 5HP spindle, 12" x 9" travels. It's meant for small parts, which is why I got it, as the vast majority of pieces I need to make are 1" and under.

The turret is a Warner & Swasey No.2, also a collet machine with a capacity I think to 1-5/16", 5HP motor, and the entire thing weighs roughly twice what the Omni does. It's actually a lot of fun to use, though admittedly I've never done a run larger than 50 parts on it. If I had to use it for a full 8-hour shift, yeah, I'd probably get tired of it in a hurry, but for short bits, it's still a fun toy.

Doc.

 

[mogandave]

Nice videos, thanks!

I never minded running production manually as long as it was a decent setup. My first job was at a tool & die shop, and all I did was low/no skilled grunt work.

That turret is badass, my mother ran a turret during WW2.

 

[dr_clyde]

-Well, I'm also kind of a job shop, and I like making one-off parts. The CNCs are for the production stuff- let the robot do the repetitive stuff.

You say that now...

Trust me when I say it is faster and easier to run one-offs in the CNC once you get comfortable and familiar.

The manual equipment will always have a place, but CNC is the way to go for any sort of complex part. I have a very nice manual setup, and I run one-offs in my Mazak all the time. Even for simple parts.

If you enjoy it and can make money doing it manually, sure, I guess. But you can charge the same money and be done in half the time with the CNC, and the finishes will be better.

 

[DocsMachine]

Trust me when I say it is faster and easier to run one-offs in the CNC once you get comfortable and familiar.

-I have no doubt. But that assumes the 'comfortable and familiar' part. You're ... ah, looking at the first part for which I ever wrote a program. And it took me probably a bit longer than you might think.

I've been a manual machinist for some 25 years. I've been a CNC machinist for about 37 minutes. I have about half an hour of CAD experience and none at all of CAM.

Yes, in a few years, I hope to be a lot more comfortable with running this thing, but I still have a long ways to go.

No, for those of you that have done short run production (or even long run, I don't judge. )

How do you "save a job"?

That is, you have a part. You make X number of them for a client. If the client comes back in a year and orders Y more, do you tool up from scratch, or do you have a file of saved drawings, programs, CAD and tool call-outs?

Do you just save the program in whatever storage medium you're using, and if the job comes back, rely on memory and the comments in the program to set the machine back up? Do you keep physical drawings? CAD models?

The part in that video is the first one I've made with this machine. It's one of my own products, and I may need more in two or three years. I thought about this, and whipped up a way to have both a hard copy and an electronic copy.

The program- barely over a page of printed text- is saved on the machine itself, saved in a job folder on my main desktop, and printed out. There's also a dimensioned drawing, a photo of the tool setup (gang tooled lathe) and I whipped up a 'worksheet' with the relevant data. Start and end dates, program name, client (if applicable) notes on which tools and inserts, etc. etc.

The electronic folder has a full copy, and the hard copies are put into a 3-ring binder.

If I ever need to do that job again, I should be able to set it back up again in probably less than an hour.

How do you do that sort of thing?

Doc.

 

[mogandave]

I think typically the "programming" is done offline and the programs sent out the the machines to run.

When I was programming (turret-presses, plasma-cutter & press-brakes) I'd generate all the programs in the office, save them all and just take the code out on a floppy (or later thumb-drive). In some shops the machines are all networked together and the operators just pull their programs down and run them one after another.

You should be able to save all your programs. Also, when you get another machine, or upgrade your existing unit you can just get another driver for the programming software, and generate new code for the new machine using the existing program.

Online programing/programming at the machine blows.

 

[MushCreek]

If you want to make parts fast, you use an old screw machine, like a Brown and Sharpe. You make finished parts in seconds. When I had my own shop, I had a brand new gee-whizz CNC lathe, and I had a Swiss-type screw machine made in 1953. The old Swiss finished two parts in the time it took the CNC to feed out the bar stock. I spent the first ten years of my career running various screw machines. We had eight spindle machines that spit out finished parts in 1-1/4 seconds, including drilling and tapping holes. Not practical for 50 parts, though.

 

[dr_clyde]

-I have no doubt. But that assumes the 'comfortable and familiar' part. You're ... ah, looking at the first part for which I ever wrote a program. And it took me probably a bit longer than you might think.

I've been a manual machinist for some 25 years. I've been a CNC machinist for about 37 minutes. I have about half an hour of CAD experience and none at all of CAM.

Yes, in a few years, I hope to be a lot more comfortable with running this thing, but I still have a long ways to go.

No, for those of you that have done short run production (or even long run, I don't judge. )

How do you "save a job"?

That is, you have a part. You make X number of them for a client. If the client comes back in a year and orders Y more, do you tool up from scratch, or do you have a file of saved drawings, programs, CAD and tool call-outs?

Do you just save the program in whatever storage medium you're using, and if the job comes back, rely on memory and the comments in the program to set the machine back up? Do you keep physical drawings? CAD models?

The part in that video is the first one I've made with this machine. It's one of my own products, and I may need more in two or three years. I thought about this, and whipped up a way to have both a hard copy and an electronic copy.

The program- barely over a page of printed text- is saved on the machine itself, saved in a job folder on my main desktop, and printed out. There's also a dimensioned drawing, a photo of the tool setup (gang tooled lathe) and I whipped up a 'worksheet' with the relevant data. Start and end dates, program name, client (if applicable) notes on which tools and inserts, etc. etc.

The electronic folder has a full copy, and the hard copies are put into a 3-ring binder.

If I ever need to do that job again, I should be able to set it back up again in probably less than an hour.

How do you do that sort of thing?

Doc.

It depends on how you do the programming, really.

If you have the control memory for it, most CNCs can store lots of programs on the control. Just pull it up, set the offsets and go to work.

If you're programming in CAM and either offloading to a floppy/thumb drive or drip feeding you can store the files on your computer.

Most conversational controls have some sort of program storage built into the control. My Mazak lathe has hundreds of programs stored in memory, all I have to do is put in the program number, set my Z offset, check the right tools are in the turret and go to work. I can back them up offline by floppy. Gotta love early 2000's machines.

I would HIGHLY recommend you leave tools in their respective pockets on the turret. Tool 1 is always a DNMG OD tool, tool 2 is always a boring bar, tool 3 is always a groove tool, etc. That way you don't forget and accidentally try to do a boring operation with a partoff tool. My turret is a 12 pocket turret, and I have 3 ID tools and 4 OD tools always loaded. I may change out boring bars, but the main boring bar is ALWAYS in pocket 8. Helps keep my programs consistent. Consistency is the key to preventing crashes.

Lathes are much easier to keep tooled than mills, as you typically only need a few tools to do the work. Some turning tools, boring tools, grooving tools and threading tools will usually always be in the turret in some fashion. Might as well make it easy on yourself.

No matter how you do it, you should be charging a "setup fee" or whatever for a job. That covers loading tools, teaching tools, installing work holding like jaws or collets, loading up programs and setting offsets. This is different than the programming fee, which is a separate charge.

 

[dr_clyde]

If you want to make parts fast, you use an old screw machine, like a Brown and Sharpe. You make finished parts in seconds. When I had my own shop, I had a brand new gee-whizz CNC lathe, and I had a Swiss-type screw machine made in 1953. The old Swiss finished two parts in the time it took the CNC to feed out the bar stock. I spent the first ten years of my career running various screw machines. We had eight spindle machines that spit out finished parts in 1-1/4 seconds, including drilling and tapping holes. Not practical for 50 parts, though.

Screw machines are by far the best way to make thousands of parts quickly, but they are a BEAR to get setup and running. Could take DAYS, especially if you have to make all the cams.

CNC lathes make up for the slower cycle times by ease of setup and programming, especially for small runs. Operator skill is also much less intensive for a CNC.

 

[DocsMachine]

I think typically the "programming" is done offline and the programs sent out the the machines to run.

-Certainly. My question was less the storing of the actual program, and more about the job as a whole. For a plasma cutter or press brake, the "tool" is almost always the same. For a mill or lathe, you might have to set up a specific thread tap, a specific size cutter, in the case of the lathe, an internal boring bar rather than externals, etc.

Most guys I've been hearing from say they just add that to the comments of a program, but in the case of this Omniturn, you can't have a separate 'comments' block, and can only add very moderate comments in a line, after that actual G-code, And as the tool layout may be different for each part, I wanted to come up with a way to save that information along with the program.

If you want to make parts fast, you use an old screw machine, like a Brown and Sharpe.

-I'd love to have a screw machine. I'd also love to have a product that sells so fast it's worth putting on a screw machine.

But "programming" a screw machine is even more of an esoteric art form than programming CNC. These days, you basically need a CNC mill to make the cams for the screw machine.

It's also worth noting there's precisely one screw machine in the entire state that I'm aware of. A B&S 00, probably comes with a third of a ton of cams and other tooling, and the guy that has it knows how to set it up. And even he hasn't used it in about twenty years.

Not practical for 50 parts, though.

-Hell, they're not often practical for a thousand parts. These days they're mainly used for small, simple parts at quantities of 100K or more.

And in my specific case, the majority of my products will be between 50 and 200 pieces.

Doc.

 

[Stedlin]

50 years ago I was in a position some what like you are. I was designing products and making them all on manual machines including a clausing turret lathe. I started out running the turret lathe myself and kept optimizing the production rate to the point where I was running parts almost as fast as a brown and sharp screw machine I bought later.

I didn’t buy a CNC machine until about four years later when they first became available.

About 20 some years ago I bought a couple of CNC lathes like the one you have. The best thing I can say about them was that they were the cheapest CNC lathes I ever bought.

Good luck. I hope you don’t need a service call.

 

[DocsMachine]

If you have the control memory for it, most CNCs can store lots of programs on the control. Just pull it up, set the offsets and go to work.

-That's basically the plan. For this machine, a "program" might be two printed pages of Notepad text. I don't know how big the hard drive on this thing is, but when I got the machine, it had some 240 programs already on it. You can also run off a USB key- as in, leave the program on it and run it from there- so even a tiny 1Gb key could hold thousands more programs.

Again, it's less the program itself, and more the supporting information- how big is the starting blank? Where are the zero points? Which tools were used? Which inserts? Etc. etc. It would seem that keeping physical notes is common, as is an electronic file, both with notes and drawings.

I would HIGHLY recommend you leave tools in their respective pockets on the turret.

-No turret, this is a gang-tool machine. Yes, I suspect I'll get to the point where I have a couple of OD turning tools left in place, and are consistently used from job to job, but I can see drills and internal boring tools changing quite a bit between jobs.

Lathes are much easier to keep tooled than mills, as you typically only need a few tools to do the work.

-On a gang-tool, some shops set everything up in a single bar, and swap out the entire assembly when changing jobs. That's pretty expensive, though, and for just a few dozen or even just a few hundred parts, it's likely not worth it.

No matter how you do it, you should be charging a "setup fee" or whatever for a job.

-Sure, if it IS a customer job. It's worth noting, though, that the majority of parts off this machine will be my own product.

Doc.

 

[DocsMachine]

About 20 some years ago I bought a couple of CNC lathes like the one you have. The best thing I can say about them was that they were the cheapest CNC lathes I ever bought.

-This unit is certainly not the best I could have potentially gotten. But it was relatively inexpensive (one of the main points ) and they're reasonably well regarded for speed and accuracy.

Sure, I could have gotten a HAAS ST for about $50,000, or a Doosan LEO for $70,000, or a Mazak Quick-Turn for $100K, but I don't have anywhere near the work to justify those kinds of machines. I also don't have the space, the power requirements or the floor space, to say nothing of the in-pocket money or even financing.

I ain't Boeing. I'm a one-man-band operating out of a two-car garage, making batches of parts that wouldn't even justify booting up a bigger machine.

Good luck. I hope you don’t need a service call.

-I'll have a far easier time repairing this machine on my own, than I would that HAAS or Doosan or Mazak. What you think of is low featured and poor capabilities, I think of as reliable and simple to repair if necessary. (One of the other things I had to keep in mind when making a selection. )

Doc.

 

[Stedlin]

If I recall correctly the main issues we had with the Omni turns was with the air Collet closer.

 

[dutchgray]

I do like your turret lathe, if I had the space I would probably buy one even though I haven't ever done any production work, as a just in case measure.

 

[dr_clyde]

You’re asking about setup sheets it sounds like.

Take photos of the setup, write down all the needed info. Tool types. Insert part numbers. Number of parts per bar. Chuck jaws or collets. Etc.

I have some binders with all the prints/info for repeat parts in them. You really only need to know program number, tool type and corresponding pocket number, required stickout, and approximate bar stickout from the chuck. Don’t try to keep work offsets. Set them every time you run the part.

 

[DocsMachine]

If I recall correctly the main issues we had with the Omni turns was with the air Collet closer.

-Haven't heard anything about that, but the manual does state that the air pressure for it needs to be closely watched, and you shouldn't let the oiler run out.

Omni's troubleshooting manual has a full breakdown of the closer plus repair instructions- hopefully I won't have to get to it right away, but if I ever do, it appears a pretty straightforward repair.

I do like your turret lathe, if I had the space I would probably buy one[...]

These W&S's are incredibly well-made, and personally, I find it a lot of fun to use. Even if/when the CNC stuff obsoletes it the rest of the way, for a job shop that does one-off parts on a semi-regular basis, it's very useful for drilling. Instead of cranking the tailstock handwheel back and forth (and on most home-shop size lathes, only having a 2-1/2" to 3" throw on the TS ram) it's as easy as a drill-press handle. And on this machine, something close to 8" of maximum stroke.

Doc.

 

[DocsMachine]

You’re asking about setup sheets it sounds like.

-I believe so. That's kind of why I asked- I assumed there had to be some sort of regular procedure that any production shop, however large, would need to go through. I wound up kind of "reinventing the wheel" and came up with something I think will work for me, but I wondered if there was more of an 'industry standard' sort of thing.

Doc.

 

[Ign]

Not entirely related but the worst of both worlds is a "hybrid" where you can load programs or allegedly turn the handles yourself. You've got NO feel and that's a HUGE deal - imagine driving a car where you couldn't feel acceleration, braking or cornering.

Anyway....not helpful.....but let manual be manual and CNC be CNC

 

[DocsMachine]

Not entirely related but the worst of both worlds is a "hybrid" where you can load programs or allegedly turn the handles yourself.

-Those have their place, generally in a small job-shop that occasionally needs to do a small handful of simple parts. It's kind of a stepping stone between a normal engine lathe and a full CNC- you can use the handles to "teach" it a cut, say, facing a part, and then just tell it to repeat for the next dozen or two dozen.

Faster and easier than an engine lathe, not as much trouble to set up a full CNC for it. (Especially if said full machines are already tasked making parts.)

But yeah, turning with the handwheels is kind a weird. Trying to bore soft jaws and the like with an MPG wheel or a joystick- especially a kind of poorly-located joystick like on this Omni- is not terribly fun.

Doc.

 

[jabberwoki]

i freakin love that turret lathe.
Handy *******.

 

[bugnut]

Doc as you are dipping your toes in the cnc pool the answer is documentation, like Dr suggested. Take the time to sit down and review the top ten or 20 most used/sold parts. This then should become a reflection of what tools need to be on the machine at all times. Then start out with simple but close measurements, that dictate what locations the tools are in, what inserts with radii specs are used and the detailed minutate of what goes where. Establishing a reference system that details tool and holder location based on something that will never change on the machine, say the top left corner of the crossslide. Then set up the tooling from there. A cheap digital camera taking pictures of scales/rulers to important features such as end of boring bars, length of drills and PROGRAMS form the baseline documentation to get you setup and moving next time you need to machine the same part.
If you're making 1000s of different parts, group them into families based on features ie internal threads, od threads etc.

I have done similar on many occasions for getting others started on this path.

The most important and hardest part of this is to stop making parts and to sit down and think what needs to be done first!

Feel free to pm me if I can assist you getting set up.

 

[DocsMachine]

Last of the three "makin' stuff" videos!

Now, keeping in mind I didn't make these videos to show "the best way" to make parts, or even "how I usually make parts" or anything like that. Mainly, it's just "hey, I have a couple cool machines, if you're interested, here they are in action".

Yes, I could have- and should have- done the entire part in two operations on just the CNC lathe. The "real" way to do it would have been with a bar-feed for the first step, drilling the back, knurling and then parting to length. Second operation would be with soft jaws to hold the knurled body, and turning the threaded end as shown.

But, this is my first real CNC and I'm still very much in the early stages of getting up to speed with the programming and operation. This little batch was literally simply because I didn't want to "waste" the time it took to program that first test part, so I ran a few using material I had on hand.

I didn't bother trying to set up the Omni to finish them, as I didn't have some of the tooling I'd need (knurler, a chuck I can use soft jaws on, etc.) and I wanted to move on to the first "real" part- which should hopefully be in the ballpark of 200+ pieces.

Doc.

 

[mogandave]

Nice video, thanks. It is amazing how quickly parts can be made manually with a decent set-up and quick-change tooling.

 

[DocsMachine]

Nice video, thanks. It is amazing how quickly parts can be made manually with a decent set-up and quick-change tooling.

-That turret lathe is an absolute antique, made in 1939, and has been entirely obsolete for fifty years.

But even still, I've made a ton of parts on it. It was a surprisingly effective stepping stone between my manual machines and the CNC. (I'd expected it'd be a few more years before I could get the CNC- a happy chance sped that timetable up for me a bunch.)

And even with the CNC, it'll still be handy- I tend to think better in manual than G-code, and if I just need ten or twenty parts, often it'll be faster for me to tool up the turret than write and prove a program for the CNC.

Besides, I'm stuck with it. There's not a single person in the entire state that'd give me so much as $500 for it.

Doc.

 

[DocsMachine]

Just machinin'!

The first 'real' production run on this machine, once I have the rest of the material in, should be somewhere around 250 parts.

Doc.