What's new
What's new

Automating low volume/high mix lathe work.

Gobo

Titanium
Joined
Jun 4, 2013
Location
Oregon, USA
We are a product shop that does LOTS of R and D with a huge mix of low volume production pars. Management wants automation on all new equipment we purchase. I have been instructed to begin researching a machine which will be purchased mid 2019. In particular, a CNC lathe in the 15 -20 HP spindle range, main and sub spindles (probably 8 and 6 inches) 2 to 2.5 inch spindle bore, Y axis and live tooling. No cheapo equipment wanted, something in the line of Mori or Mazak quality. No difficult material to machine, maybe a little stainless but mostly mild steels and aluminum. 20ish inch Z axis work envelope. Cost is, as always, important, but not a major driving factor. We currently have 17 CNC lathes some with barfeeders, some use barpullers. Several have part catchers. One has an integral unloading arm and parts accumulator. One machine has a gantry loader. We recently purchased our first UR10 and are awaiting delivery.
I cant help but think the low/volume - high mix situation will pose a challenge to automation. Other than the aforementioned automation we are already using, what would you recommend we consider? What words of wisdom are out there?
I do realize that automation solely for the sake of automating might not be the best approach, but this is what I have been told to do.
 
You have 17 CNC lathes already, and you are not sure how automation will work for your facility?

I was hoping to get some input into what may be some technologies or strategies other than what we are already using. In the past, automation was not a big concern for management. Recently, it has become THE driving factor.
 
automation was not a big concern for management. Recently, it has become THE driving factor.

Let me guess. The idiots in suits went to a seminar or some shit, and now they are getting their wee-wees to tingle
by throwing around buzz words.


There is absolutely nothing wrong with "Automating".. However you do need to do the cost/benefit thing..

If you are running prototypes and you can use 2.5" 12L14 in that machine 24/7, and regardless of what size the part
is, you can keep your same setup, it could work.. If you can keep shoveling the machine programs, without changing
the setup it could work..

Cost/benefit again.. What costs more. A foot of 2.5" 12L14, or the time it takes to swap the machine over to save
a few dollars on material cost???

I would guess it would come down to "simplification".. That machine runs 1" material, That machine runs 2" material..
The machine with the gantry loader runs 4" material.. The other machine with a gantry runs 6".

One of the manufacturing projects I did in college.. I'm pretty sure it was a manufacturing course, not a
management course.. They crossed over quite a bit.. But the basic idea was that by the time you SIMPLIFY
to make automation easier, or possible in some situations.. You've already dropped your costs so much that its
simply not worth it.

My project (group) was a staple remover.. By the time we simplified the design for automated assembly, it became so
fricken easy and fast that trying to automate it did not pencil out in the slightest..

Just tossing some stuff out to think about. What you are trying to do IS NOT EASY.. And I've put some
serious time into thinking about how to automate a job shop... And none of it has ever penciled out.
 
Can you give us a better understanding of what you mean by low volume/high mix? Lots sizes and such...

I’m jelly of your UR10 incoming.
O.K. Are you ready? Daily quantities. Whatever the number of assemblies we need to build in one day - we have to make all the components for that number of assemblies in one day. That can be as many as 50 to 60 different lathe parts. Quantities from 6 to 6 dozen. Plus one-off R and D stuff. As I said, we are a product shop with tons of R an D. We have dipped our toe in robotic welding and got burned. Bad taste still in mouth.
 
Direct to Original Question - most CNC lathes hold tools in turrets, which limits the online inventory of tools (typically 12 to 24), and typically limits the inventory of live tools even more.

For a BIG step up in price, along other issues (search PM for "b axis lathe issues") you could go to lathe with a chain or large wheel tool changer. (Think the B-head on an Okuma Multus, Or an Integrex, etc.) I recall machines that could hold 40 or 60 capto mount tools - more probably available now.

NOW - assuming the machines works, and keeps working, rather than having say 12 tools, with a fixed inventory of expensive live tool holders, you now, in theory, have 40 or 60 or ... tools, and they can all be live if you want. You also get B-axis.

(Clearly a machine with b-axis on a turrent doesn't have this property.)

So, if you could swallow, and digest, all of that, it would be a machine with a much larger inventory of online tools (like is often done with large milling machines, especially horizontals.) This larger inventory of tools could make setup for a rapidly swapping family of parts easier. In fact, it might save more time than always using the same bar size - since you might well be able to adjust chuck jaws faster than you can swap turret tool holders.

I will now note that I have never chosen to buy such a machine, and if you search around you can find posts by people who gave up on them.

But if you want a lathe that can make a maximized family of parts with minimal changeover - it would be something to look into.
(You would HAVE to have a tool setter either online or offline, and some level of probing and above all a really strong and mature CAM package to exploit this.)

See the Edge Precision youtube channel for an example of the class of machine I'm talking about. (It's a rather large mazak)
 
We do have 4 Mazak Integrex turn/mills, so are intimately familiar with that type of machine. Fantastic equipment, but I may have a hard fight convincing them that is the machine for this need.
 
What is the plan for the UR10? Seems like it won’t take much to make a larger-than-20 pound part on the new lathe?

Having one UR5 and 2 UR3’s, I’ve learned you need to take high level view of the process when integrating the robot. Make the raw material/finish part jigs work for the whole range of size parts you will make, consider grippers that can handle the same size range or can be swapped out quickly for the right gripper.

Man, running a new part every day and expecting to get the robot tuned in to that part as well... kinda gives me anxiety.
 
To make the robot / cell change over friendly, you have to do some development on the front end. A UR is not big enough to handle the size of parts you can be running on the new machine. I would expect you to need a 30 or 50kg robot for that machine. That also means a fixed robot enclosure with a locking door. A 30-50kg robot won't even know if you are in the way, will go right through you and keep trucking.

Robot program development is the key here. You have to develop a parametric program with position arrays that allow you to enter a few dimensions on the teach pendant and the robot adjusts to the part. Or you have it remember your taught dimensions for each part and take the hit on the first set-up. You cannot spend time teaching the robot every time you change over.
 
What is the plan for the UR10? Seems like it won’t take much to make a larger-than-20 pound part on the new lathe?

Having one UR5 and 2 UR3’s, I’ve learned you need to take high level view of the process when integrating the robot. Make the raw material/finish part jigs work for the whole range of size parts you will make, consider grippers that can handle the same size range or can be swapped out quickly for the right gripper.

Man, running a new part every day and expecting to get the robot tuned in to that part as well... kinda gives me anxiety.

The vast majority of the parts we run through our lathes will be easily handled by the UR10. We chose that one more for the extra reach, not the weight capacity. We will be using it for machine/loading.

And I get your anxiety thing.
 
To make the robot / cell change over friendly, you have to do some development on the front end. A UR is not big enough to handle the size of parts you can be running on the new machine. I would expect you to need a 30 or 50kg robot for that machine. That also means a fixed robot enclosure with a locking door. A 30-50kg robot won't even know if you are in the way, will go right through you and keep trucking.

Robot program development is the key here. You have to develop a parametric program with position arrays that allow you to enter a few dimensions on the teach pendant and the robot adjusts to the part. Or you have it remember your taught dimensions for each part and take the hit on the first set-up. You cannot spend time teaching the robot every time you change over.

Allow me to clarify- the UR10 will handle the vast majority of parts we process through our lathes. If this works out for us, we hope to use robots for loading/unloading on our mills also.
 
Robot program development is the key here. You have to develop a parametric program with position arrays that allow you to enter a few dimensions on the teach pendant and the robot adjusts to the part. Or you have it remember your taught dimensions for each part and take the hit on the first set-up. You cannot spend time teaching the robot every time you change over.
Robot dummy here. I was under the impression that once you teach a robot a part, then you would just call up that program and be ready to run the next time you run the same part, just like our mills and lathes. No? Obviously this assumes the robot has not moved, or has a permanent mount it can be relocated to if it has moved.
 
Given those part quantities, I think you'd be better suited towards automating/simplifying the setup process as opposed to load/unload cycles. Depending on your part sizes, a barfeed and parts catcher is simple enough to integrate. I believe the Haas barfeed can be directly linked to the CNCs main controller and could be macro controlled. The EDGE servo unit supposedly had a feature to do the same, but the tech said it was finicky and unsupported. Throw in some extra thought into your tooling blocks (extras and a good system for resetting them), tool presetter, fixturing (collet chucks are awesome), and setup macros and you can cut your change-over time down to minutes.

We had our OKUMA setup for similar work. Barfeed was the edge servo unit, would take 2 min to reprogram, plus whatever time to load/unload the unit. Collet chucks took about 2 min to change the pair. About half the tool slots were standardized, the rest we never had a good solution for changing quickly. Look into macros to speed those up. All of the work offsets were automated via macros. It even chose how to transfer parts via those macros. Assuming we didn't have to change a bunch of tools, we could change programs in about 5 min. Past that, the barfeed and oversized 12" parts catcher could run for hours without attention.
 
Let me guess. The idiots in suits went to a seminar or some shit, and now they are getting their wee-wees to tingle
by throwing around buzz words.

Sounds like they had lunch with a Mori salesman.

Where I used to work they started buying Mori's with Y axis, live tooling, sub spindles, etc. It was pretty cool seeing a part go in as barstock and pop out complete with all the turning and mill work. However, the setup times were astronomical. You need to have some decent volume.
 
Sounds like they had lunch with a Mori salesman.

Where I used to work they started buying Mori's with Y axis, live tooling, sub spindles, etc. It was pretty cool seeing a part go in as barstock and pop out complete with all the turning and mill work. However, the setup times were astronomical. You need to have some decent volume.

I agree. The small quantities are going to murder us. Currently, the shop is broken up into "value streams". We run similar parts on different machines, because those components go into different product lines. Duplication of effort is rampant. Families of similar parts should be run on one machine. We might make four different pistons (very similar features) on four different machines, just because they go into four different product lines. Can you feel my frustration? The answer to me is obvious, but I don't make those decisions. But I do get to lay my hands on some neat equipment, so I do my best.
 
Robot dummy here. I was under the impression that once you teach a robot a part, then you would just call up that program and be ready to run the next time you run the same part, just like our mills and lathes. No? Obviously this assumes the robot has not moved, or has a permanent mount it can be relocated to if it has moved.

Yes, that is correct, you can. However, you need to write a new program for each part. If you have a parametric program, then you don't have to write a new program for each part, you simply re-teach your array positions. This isn't every position, it is only the positions that matter, such unclamp and clamp and approach positions. There is no need to spend a bunch of time teaching positions between the pallet and the machine that are fine for every part you run.

So, you do an array position such as unloadgrip[av] in your program. In the background, av == 1 ..... or av == 3 or some such. Anything not taught for each part does not have the [av] at the end.

Now you have 1 program for all parts. You simply tell it which av number to use for the particular part in question. When you set up a new part, you don't spend time teaching points that aren't required, which reduces your setup time significantly.
 
Low volume high mix... is it going to be slug work or bar work?

If its slug I'd look at a Nakamura NTRX300. Should be able to get 120 tools and the B axis allows simple roughing turning holders so many clearance angle positions. I wouldn't mess with a B axis with a lower turret for low volume high mix. I dont think the setup time / managing the lower turret tool positions is worth the hassle for this application.

if its bar I'd want the biggest bar capacity you could get within reason so I'd choose the Mazak integrex i100 bartec machine.

The key here is reducing the human input to change over jobs. Do you have the capacity to dovetail your raw stock? This would allow you to load slugs of many different sizes / lengths with 1 fixture that wouldn't change and it would give you great gripping force with maximum clearance. Its a lot cheaper to leave a vise setup in a vertical and prep some small quantities of blanks than change chuck jaws IMHO
 
devoting a costly capital resource to one part because it's part of one product suggests an imcomplete understanding of "value stream", and i will be bold and suggest a less than ideal understanding of accounting. glad i don't hold stock there....

if you want to do fast change overs, with a constantly changing pool of parts, you MUST invest in procedures - meaning software, setup grids, a whole "bureaucracy" - in the GOOD sense - the infrastructure to make correct programs, prove those programs in simulation as much as possible, derive setup and tool clearence data offline as much as possible, preload tools, etc.
 
If a UR10 will handle the parts for load and unload, then they are almost certainly less than three inches in diameter and less than a foot long.

If they are less than three inches in diameter and less than a foot long, a standard bar feeder and parts catcher is much, much simpler "automation" than a robot.

What am I missing?
 








 
Back
Top