What's new
What's new

Going Lights Out: from a management standpoint

thunderskunk

Cast Iron
Joined
Nov 13, 2018
Location
Middle-of-nowhere
Evenin' folks,

I was hoping to pick at the pool of tough-won experience and well-practiced shop management. What does it take to get to lights-out machining and inspection from a management standpoint?

I have a slight idea of what it takes for hardware: dual station vises, tool probing, tool life management, possibly part probing, a DCC CMM with lots of table space, etc. Input on these things would be welcome as well, but my focus is on the "other stuff" such as scheduling, employee culture, resource planning, maintenance planning, finance, etc.

In my mind, it's the way to go. We're a small shop with only 6 machinists, and they are indeed machinists. I'd almost call them all toolmakers, but one is a bit crustier than the rest, so we reserve that title for him. While there are plenty of other issues to settle before we get to that point, I think it's a great goal to strive for: Why do we need to have ERP working smoother? Well, it would help get things up and going for lights-out machining. Why do we need to invest in more workholding hardware? Why do we need to upgrade office software? Etc. Etc.
 
I worked in a shop with a couple of Okuma Multuses running full time. two employees per shift were assigned to run other stuff and just keep an eye on the Multuses. I know that's not "lights out". But when the weekend came, the machines kept running until the bars ran out, or a part probed out-of-spec, or a tool wouldn't change, or the job was completed. One thing that comes to mind is a proper coolant management system. This one would refill a tank that had gotten low (and did frequently). One time the tank overflowed, and when Monday came, there was a huge mess on the floor. A sensor was installed that would dial out if tripped. Just a little experience and stuff to consider
 
Three things I would add;

Training, I can’t emphasize this enough. Shop I worked at essentially told the guys on the floor we needed to do lights out, make it happen. We were a tool room, not high volume, but most of our work in final stages was hard milled. This took a lot of time. A detail I ran on the Hermle took 3 days. Most of the guys, the really smart ones did their own research and reading, mostly outside work. We progressed. Most of what we were trying was over my head. This went on for months. Finally they brought a guy in from or associated with Mazak to do macro training. I am sure 4 days of classes for nearly 30 guys wasn’t cheap. But our progress was exponential. We started running long jobs over the weekend and short jobs during the week. Friday you could setup a job to run all weekend, Monday morning come in and have a few hours of run time left to check on what the weekend shift did, see what the emergencies were for the week, and start lining up your work.

Another consideration, tooling cost. In the previous scenario management wanted “lights out” (even though we ran 4 shifts including the weekend) however they did not consider the added costs of that spindle time. In a weekend machining a detail, you could burn through a lot of tooling hard milling a 10” DC53 due detail. As in thousands of dollars. Same amount as normal, but you could do it more often.

Culture, to me, is huge in lights out. The above shop had lots of problems with culture. I’m not bragging, just using my self as a bench mark. I liked to get my machines up and running, then line up the next few jobs and tooling. Then I liked to start setting or running something else. Many guys would setup their machine and then sit. Even on a lights out job. Some people are like that.

So what I poorly described, is managements got to support the guys in the shop, machining and production times will INCREASE in the run up to “lights out”, consumable costs will increase, if your training is lacking, and even when it isn’t, scrap prices can be crippling, but the reward should be happier less stressed guys in the shop who have more time to focus and organize their work, and take time for the little things.

OH! And don’t forget, tiny mistakes are exponential in lights out!
 
For the kind of work I do I can do it comfortably. I have touch setters on all of the machines I run lights out and run a tool break check on each tool. If you have enough tool storage in the machine you can program with redundant tooling and set your tool life conservatively. I could have my machines send a message to my phone if it alarmed but I keep my machines off line for security reasons.


I only run proven and stable programs. I've got part that is pretty tight tolerances stainless part and takes about 8 hours per part. By programming conservatively and swapping cutters(.04 and .05 10x-12x dia)via tool life counter I was able to gang 4 parts on a tombstone and get a true 32 hours of unattended run time. The nice thing about it was I didn't even need to warm the machine up on Monday morning because it was just wrapping up The job came back this past week in a quantity that will require 8 days where it would have tied up a machine for close to a month the old way of doing it.

I don't work in any notoriously flammable materials so I don't have to worry about that. My worst scenario is I come in to a broken tool.
 
Only things that I would say that are "management" related - and some, if not all have been listed already.

A) Scheduling long jobs for night and weekends.

B) Good tooling. Redundant tooling where applicable. Piles of inserts won't matter 'till morning. But expect to have 3x on hand as prior.

C) Programming - if this is considered a "management" application - program with more concervitive S/F/DOC's.

D) For milling - a tool probe would seem to be the first step.
Barfed lathes generally are saved by servo barfeeder that knows that the bar didn't feed to the expected length, but you could run junk all night.
Hard to probe tools in cycle on a lathe.

E) Start (maybe even stick with) with jobs with low material cost to machine time cost ratio!

F) Expect junk, and be pleasantly surprised if you have all good parts in the morning.\

G) ???


------------------

Think Snow Eh!
Ox
 
OP, are you the said manager? If so, good on you to ask, if not, go talk to them before even 'wasting'* your time trying to do this!! :scratchchin:

I don't have anything to add really, Ox made some very good points already.

However, make sure the manager is (if not you)

1) completely on board with this
1A) understands there will be a big learning curve
1B) understands tooling costs will likely sky rocket in the beginning
1C) has realistic expectations

2) is willing to help/empower you and the shop to 'get er done'

3) see #1 above :D


*I don't think it is a waste of time by any means, but I do know firsthand sometime management think it takes a week or so and all good, lights out, profits triple, blah blah. I actually had this conversation with a former manager AFTER they bought a used hori against my recommendations. He thought we were just going to program it, load up a bunch of tombstones and go home and be done with it*. (< I'll get back to this) After running it for a bit we realized we had some real problems - coolant lines getting plugged, poor chip evacuation, pallet problems, and so on.

*So upon reporting these issues, and brainstorming on how to fix them, we finally convinced him that, yes we could work towards lights out, but NO it would not be easy and/or cheap. And that was that, didn't even get the machine really running anything worthwhile before they sold it off...
 
I do almost the exact same as Yoke. Redundant tooling, broken tool detection, and stable programs. The only thing to add is that I do automatic stock probing before it starts.

When things are cooking, I leave one of the machines running at night probably 3 or 4 times a week. Doesn't work as well for programs that only take a couple hours, but that's still two or three (almost) free hours every night.
 
I don't think I saw this mentioned yet...

The purpose of running lights out isn't to make the absolute maximum # of parts in the
minimum amount of time..

The purpose is to make the maximum amount of good parts in the time that nobody is
there tending the machine..

I guess the key word(s) is "Stable Process".. And that doesn't necessarily mean
longevity. If you have a guy standing at the lathe, and you are getting a little
stupid and you know you are going to blow a corner on the 4th or 5th part every time.
You have the guy rotate that rougher every 3rd part. As long as the time savings
justifies the cost of 1/4th of an insert. Trying to save a nickel by pushing that
corner through the 4th part can sink you.. it may or may not fail. A re-run, risk
of damaging the tool holder etc... 3 parts is stable, 4 is not..
 
I don't think I saw this mentioned yet...

The purpose of running lights out isn't to make the absolute maximum # of parts in the
minimum amount of time..

The purpose is to make the maximum amount of good parts in the time that nobody is
there tending the machine..

I guess the key word(s) is "Stable Process".. And that doesn't necessarily mean
longevity. If you have a guy standing at the lathe, and you are getting a little
stupid and you know you are going to blow a corner on the 4th or 5th part every time.
You have the guy rotate that rougher every 3rd part. As long as the time savings
justifies the cost of 1/4th of an insert. Trying to save a nickel by pushing that
corner through the 4th part can sink you.. it may or may not fail. A re-run, risk
of damaging the tool holder etc... 3 parts is stable, 4 is not..

Not exactly related to lights out, but getting the boss to understand this could be problematic. We ran parts on a lathe in a very similar fashion - we knew we could get 8 parts per edge, 9 was sometimes, and 10 was really really pushing it. Makes sense to change the insert on part 9 right? Nope. Boss always wanted to push it and rely on the guy running it to have lightning reflexes when he heard it let go. :(
 
Not exactly related to lights out, but getting the boss to understand this could be problematic. We ran parts on a lathe in a very similar fashion - we knew we could get 8 parts per edge, 9 was sometimes, and 10 was really really pushing it. Makes sense to change the insert on part 9 right? Nope. Boss always wanted to push it and rely on the guy running it to have lightning reflexes when he heard it let go. :(

And then the insert cracks, and you lose the corner underneath it.. So instead of 16 good parts on that corner
of the insert, you got 8 and a half, and a the time wasted to re-run...

It amazes how often people will trip over dollars trying to pick up nickels.
 
And then the insert cracks, and you lose the corner underneath it..

And the insert base breaks, jamming into the screw, bending it and pressing it into the hole, ruining your only holder.

Yep. . . . Worked for one of those.
 
usually easier to have a person watch 2 or 3 machines during the night. obviously you will need
.
1) more tools and tool holders in each machine. can easily add $10,000 to $100,000 for each machine. and tooling costs can increase 200% of normal for reliability this can easily add $100,000 in extra costs per year
2) you will need cutting parameters reduced often to 30 to 70% of normal for higher reliability
3) you will need to monitor delay causes. that is if part normally takes 10 hours average per 100 parts and you have 2 parts taking a extra 10 hours obviously have to work on causes of delays
4) machine malfunctions happen. if nobody around it can be just a machine stopping up to machine damage costing over $100,000 in repairs. that is if machine making loud noises a operator can usually stop it in a minute. if nobody around it can be far worse
5) many will put tool changer in manual mode and cnc will run til the next tool change call. this can be 5 minutes to 5 hours. at least that limits it to the current tool running
6) part repeatability. that is part variability especially hard spots in metal can cause sudden tool failures. drill breaks so no hole than tap breaks cause no hole. if small drill stuck in hole a bigger drill that follows will deflect off it and put a out of position extra hole and or big drill break too.
.
usually cheaper to have a operator keep an eye and ear out watching many machines. that is if costs increase $100,000 per year going lights out and it cost $70,000 for a person to be there than its cheaper to have a person there. many times full automation is far more expensive than keeping a few people around. obviously intermittent catastrophic failures might take awhile to realize this cause the failures often dont happen for weeks or months at a time
 
We run lights out on our horizontals and it does take a lot of time getting this set up but once it is set you rarely have to change anything over a course of time. We have dedicated tools for our machines and we check those all the time. We have also created a tooling probe program that checks each tool if needed for breakage. If there is a broken or worn tool the machine will kick the pallet out of the machine and set up the next job. What is really rewarding is coming in to work the next day and either having a machine still running or pallets of machined parts complete.
 
It seems like less people running very trouble free proven programs to run at night is a good solution.

Then a employee is there to monitor when problems arise preventing scrap from problems which seem to occur.
 
I run Ti bone plates overnight on a Haas with a trunnion, 4 up. I use tool breakage detection after every tool, so if there's any issue it'll just stop and wait for me.
 
In order to succeed at lights out during the night, one should run lights out during the day. It's a culture that should be embedded at all levels of the company, starting with the programming.

Most lights-out problems happen with the tooling. Tool selection and programming should be performed such that a tool should almost never fail unexpectedly. It doesn't matter if the tool life is 15 minutes or 600 minutes, as long as the life is repeatable, and practical.

Redundant tooling goes far beyond just having a duplicate of the same tool. Sometimes you may want an identical, alternate tool to cut a slightly different material, e.g. 1018 steel vs 4140 hardened. Tool life will differ dramatically between those materials and your timers won't be effective if you can't track time in material separately. There are situations where we have duplicate tooling to cut different features in the same material and even the same part. A cutter doing helical toolpaths will wear the tip, whereas the same cutter doing HSM will evenly wear the flutes. Perhaps not important enough to warrant separate tools in softer material, but certainly a big deal in hardened materials.

Chip and coolant management is very important as well.

Part probing is important especially as a pre-flight before running a cycle. Even if everything is perfectly dialed it, someone will inevitably load and run the wrong program. If we run identical parts in different materials, e.g. aluminum vs steel, we'll go as far as to sawcut the different materials to different standard lengths. For example, when we make 6" vise jaws, we'll cut steel blanks to 6.125" and aluminum blanks to 6.187". Our automatic bandsaw is generally accurate to under 0.005", worst case 0.010" if the blade is going bad. That's more than enough for the probe to resolve the difference between 1/8" and 3/16" stock allowance.
 
Technology is getting to the point where "running lights out" procedures should be the norm and "lights on" style working can be reserved for special cases.

Potentially makes life so much easier when jobs are normally set so they can just be run lights out if that makes for a sensible schedule . Doing such is also an effective process improvement driver as it will focus attention on why the more special jobs cannot currently be run lights out and what changes could be made so that they can.

In many ways running lights out is the direct spiritual successor to old style manufacturing with ranks of dumb machines with unskilled operators pulling the handles supported by skilled tool-setters. Probably some inspiration to be gained there. The big mass production firms pretty much jumped straight over refining their processes as they went. But for the smaller shop the process / management changes in going CNC have been more like turbocharging (with nitrous injection and afterburners too) prototype and toolroom shop procedures.

Clive
 
Culture, to me, is huge in lights out. The above shop had lots of problems with culture. I’m not bragging, just using my self as a bench mark. I liked to get my machines up and running, then line up the next few jobs and tooling. Then I liked to start setting or running something else. Many guys would setup their machine and then sit. Even on a lights out job. Some people are like that.
What is in it for them to do what you do?
 








 
Back
Top