Help: Preparations for unsupervised CNC lathe operations

[MadMax]

I'm contemplating setting up my Daewoo CNC lathe for overnight unsupervised operation. Besides getting a bar loader and running a chip conveyor, how does one set up for unsupervised operation?

I can see using a tool presetter to do a series of touch operations to detect insert wear or damage, but how does one detect a more catastrophic insert failure during a feed and stop the spindle? Is it possible to use spindle load as an error detection criterion?

I'm not as concerned about making an evenings worth of bad parts so much as having a huge wad of aluminum chips welding into a ball and causing a bad crash due to a failed insert or a ball of unbroken chips. How does industry detect these conditons?

[Superman]

Originally Posted by MadMax I can see using a tool presetter to do a series of touch operations to detect insert wear or damage, but how does one detect a more catastrophic insert failure during a feed and stop the spindle? Is it possible to use spindle load as an error detection criterion?

Things to look for or have on your machine may be :

Load monitoring
If available would give an alarm if normal expected torque climbs over a preset limit on that tool for a preset amount of time and you would also set the machine on how to handle that alarm ( ie. immediate all stop, finish the pass and stop, or put the alarm light ON an keep going, and so on.)

Tool Life Management
Automatic changing of worn tooling ( deemed unusable by the autogauge cycle or has completed a set number of parts or exceeded a set cutting time) with a backup tool in the turret

Autogauging
Automatic offset adjustment of wear offset to a set amount before being deemed unusable
Breakage detection, if the difference is over a set limit
( auto toolgauging may be hard if you have one of those manual fold-down tool measures, you may have to use an auto-probe to actually measure the part to adjust the tools , $$$ big dollars there )

[beege]

I know someone who's done this, and one thing to consider, too, is coolant level replacement and over-fill detection. He only had coolant replacement in place, and flooded the shop one night. I think that most contingencies also somehow were connected to a system that called his cell if something went wrong.

[MadMax]

Thanks for the replies!

I've taken a look at my load monitor. I had hoped to use it for tool breakage detection, but I haven't figured it out yet. I'll slog thru the Fanuc manual tomorrow.

I guess I need to get an automatic presetter to do a periodic tool assessment.

As to coolant replacement, I was thinking of hooking up a toilet tank float/valve in my coolant tank to autofill. I was also thinking that I could work out a leakage detection circuit that would shut off a solenoid valve and kill the coolant refill in case I had a bad valve that resulted in an overflow. I'm not sure if I'm overthinking this though. I don't mind making sure that the coolant level is high enough before closing for the night and my machine seems to do fine for a couple days of running without refilling.

I'm a bit nervous about the whole venture. I'm thinking of stuffing a webcam inside my lathe so I can log in and take a peek inside my machine whenever I like. I'll probably need some sort of lens cover with a formula 1 type wiper

[Delw]

Done it for many years on the supermax cnc lathes.
Brass I have no problems with. alum is iffy 12l14 is no problem.

heres some other things to look at. if you get any rats nest in your parts during the day DONT do it until you get that part fixed even the slightest rats nest.
if you can run 2-3 days with out any problems and 0 attendance you will be fine to run at night, if you cant your going to have a problem.

web cam is a good idea used it in the past however its no good if you dont have a way to shut the machine down remotely or you live across the street.
were the parts going to the bottom of the lathe or a part conveyor, webcam on the finished parts works the best.

[kcoste]

The link below is for a product announcement on cnczone.
http://www.cnczone.com/forums/showthread.php?t=91131
It is a sensor which automatically measures the length and type of tool immediately before every cutting operation. If tool is broken, it will measure short and the sensor will command an alarm to shut down the machine. It can also be configured to call, page or email you when the alarm occurs.
Please let me know if you have any questions.
Website for the sensor is:
http://amasst.com/products/Products_TVSA.htm

Keith

[kcoste]

One more comment, the sensor operates optically in less than ~1 second, so there are no time delays with typical touch sensors.

[nicko17]

You going to need something to catch the parts too if you dont have anything.

Coolant shouldn't be much of an issue. Just fill it up before you leave. It may not be worth the trouble of setting up something for that.

Depending on your tolerances for the parts , your going to need to set up some auto offset/wear. I'll be interested to see how this turns out.

[roundman]

This is a good subject. My boss said it would be no problem just load up bars of material in the barfeeder prove the program and go no problem. Asking how he would capture all nights worth of parts and other resonable issues which have been discussed here are just stupid questions to him. Of course he has never done this himself and even though we do not have a order large enough to attempt it I would find it a challange to at least think it through. I find some jobs run fine yet at times chips will build on a tool or something and I usually just blow it off with the air nozzle no prob but if I am not there it just keeps running and building up. I guess running it while watching it to see if it would be feasible would be a first step basically proving it for a set time. I guess we could move in that direction incrementally. Now we do not have any large repeating orders to mention and not enough work like we are used to having. I am keeping an open mind to this.

[nicko17]

Originally Posted by roundman This is a good subject. My boss said it would be no problem just load up bars of material in the barfeeder prove the program and go no problem. Asking how he would capture all nights worth of parts and other resonable issues which have been discussed here are just stupid questions to him. Of course he has never done this himself and even though we do not have a order large enough to attempt it I would find it a challange to at least think it through. I find some jobs run fine yet at times chips will build on a tool or something and I usually just blow it off with the air nozzle no prob but if I am not there it just keeps running and building up. I guess running it while watching it to see if it would be feasible would be a first step basically proving it for a set time. I guess we could move in that direction incrementally. Now we do not have any large repeating orders to mention and not enough work like we are used to having. I am keeping an open mind to this.

I dont understand how they can be stupid questions. You going to need alot more than a bar feeder to run it automatically. It would be interesting to know what he would have in mind to catch the parts. Unless you want to pick them all out of the chips bin in the morning.

[roundman]

He is extremely close minded. He claims to be a master machinist. We do not believe him.

[Superman]

It happens all the time,
If you are a top-notch gun on the machine,
.........there is no way they want you to climb the ladder

( to wrap it up, to be a boss.... you are not a gun )

But there are places that actually put the gun in a supervisory role, so that his skills can be passed on to those that really need to learn them