CNC Lathe with automated parts loader recommendation

Hi,
We're looking to purchase a CNC lathe with an automated part loader/unloader to perform a simple secondary facing operation on a part that's approximately .375" OD x 1" long. Target quantities are 2 million parts/year therefore part-to-part cycle time is important (sub-five second target). One machine that we're considering is an Omni-Turn GT-75. I like the simple, quick loading approach but am not overly thrilled with the 300 in/min rapids. What other options are available and what would you recommend? Thanks!

If it's just a simple facing operation 300 in/min rapids would well down my list of concerns.

Limy Sami said:

If it's just a simple facing operation 300 in/min rapids would well down my list of concerns.

Click to expand...

Limy, with a sub-5 second takt time, 300 IPM rapid is a serious concern. You need every 0.1 second you can get out of the machining cycle.

OP, you need to also consider spindle accel/decel times and power. While the part may only require a few hp from the spindle, the ramp needed for these times dictates a much more powerful spindle. And the control also needs to be considered. For example, the cycle time on a QTN200-II Mazak (circa 2004) with the 35 hp spindle vs a QTN200 circa 2000 with a 20 hp spindle (Both with 640 Fusion controls) for the exact same program, part and parameters is 1.5 seconds faster and it's all in the spindle accel/decel. And there is a further 1 sec drop to a new QTN200 (circa 2014) Matrix-II controlled machine with a 35hp spindle.

When you get into this realm of cycle times, everything matters. (Lots of experience speaking here...)

I'd be curious to see pictures of what this looks like. Would you use something like a vibratory tumbler (minus the media of course) to move the parts into a ramp/slot to align them to pick up? Sorry, probably explained that like crap haha. I once helped make a machine that used the bowl part of a vibratory tumbler-like mechanism that vibrated rivets up along a spiral ramp on the outer edge of the bowl then into a slide that fed into the riveting machine. Crystal clear now right? In all seriousness I am curious how this would work for such small parts. I have seen a robot loader moving bigger (3-4" dia) parts that were stacked on end in a tray. Obviously this is a little easier to program a robot to pick them up. I know on the finished part Royal makes those units to catch the parts...

Maybe - forget one machine ?
Its going to be very expensive to do this consistently, at those speeds.
If cost matters, something like the relatibvely cheap omniturn, and just add as many machines as you need.

I would expect that something that can do 5 secs takt, 24x7x365, will end up costing 500-800k.
Cost/part == 0.25$ + maintenance and electricity.

Loader will be a big part, imo. Robot ?
Typical robot loaders are slow.
Very slow, for those rates.

So, you might need something like a 2-spindle lathe, with the robot continuously shuffling parts.
Even then, I dont think 2 spindles + robot will do that 5 secs cycle.
The load/unload is just too slow.

Two robots feeding 4 omniturns, maybe.
20 secs/spindle might be realistic.

Also, the spindle acceleration takes time, as was said, and uses lots of power.
100% spindle load till its upto speed, and again to slow down.
So, using 40kW @0.15$ = 6 $/hr.
x4 = 24$/hr = 576 $ /day.
== 21 large for power.
This may matter, or not.

Also, maintenance.
Figure 500$/spindle (robots are one) == 3 grand/month.
Total costs maybe 21 + 36 == 57 large per year.

There is no eternium, and everything wears out and has issues.
A spread-system of 4 x 2 is a lot easier to maintain, and wont fail 100%, and some parts can easily be taken out for service, while production keeps running.

Parts are a lot cheaper/easier to come buy for e.g. omniturns vs some super special exotic japanese or german machine.

MMcCawley said:

Hi,
We're looking to purchase a CNC lathe with an automated part loader/unloader to perform a simple secondary facing operation on a part that's approximately .375" OD x 1" long. Target quantities are 2 million parts/year therefore part-to-part cycle time is important (sub-five second target). One machine that we're considering is an Omni-Turn GT-75. I like the simple, quick loading approach but am not overly thrilled with the 300 in/min rapids. What other options are available and what would you recommend? Thanks!

Click to expand...

2 machines ????


Peter from Holland

Further to my last.
4 spindles @40 kW = 160 kW.
To get 30 kW at spindle the machine might need 40 kW.

Do you have enough power ?
Also, 160 kW 24 x 7 is a hell of a lot of heat (= 30 saunas).

Something to take into account.

These details can easily be the critical path that determines viability and or profit.

Tonytn36 said:

Limy, with a sub-5 second takt time, 300 IPM rapid is a serious concern. You need every 0.1 second you can get out of the machining cycle.

OP, you need to also consider spindle accel/decel times and power. While the part may only require a few hp from the spindle, the ramp needed for these times dictates a much more powerful spindle. And the control also needs to be considered. For example, the cycle time on a QTN200-II Mazak (circa 2004) with the 35 hp spindle vs a QTN200 circa 2000 with a 20 hp spindle (Both with 640 Fusion controls) for the exact same program, part and parameters is 1.5 seconds faster and it's all in the spindle accel/decel. And there is a further 1 sec drop to a new QTN200 (circa 2014) Matrix-II controlled machine with a 35hp spindle.

When you get into this realm of cycle times, everything matters. (Lots of experience speaking here...)

Click to expand...

Tony, while bowing to your knowledge and experience, my thinking on rapids was along the lines of ;- it depends how far the tool has to move to clear the load / unload op, i.e. keep it all as tight as possible and it cuts down the need for high rapids.

I agree totally with your ramp up and down times, ..........might save a bit of time if you can swop parts on the fly - slowing the spindle rather than stopping.

Without seeing the part, might you be better off milling? Five seconds per part seems hard, but twenty seconds per square inch in a mill sounds glacial. A 10x20 fixture could possibly hold 800 of these. A bowl feeder to a SCARA would easily populate an 800 up fixture in 4000 seconds. That's about an hour. Put that fixture in any mill, face it off, then if you need to pack them nicely a second robot. If not, engineer a flip and dump into your fixture and repeat.

Maybe this is feasible, maybe it isn't. Depends on the part geometry. But spinning a mill up for five minutes once an hour seems a lot easier than spinning a lathe up and down 800 times an hour.

While 300 IPM rapids is painfully slow for today's modern machine tools, one has to consider how far the tool must move to clear the part. We have a gang tool machine with the same blazing 300 IPM rapids and with tools set up properly, 1 linear inch of rapid travel tool to tool or less is easily achievable. In the right environment, it will blow the pants off our newest Mazak.

I have no experience in production of these quantities but I suspect one operator manning several Omni turns maybe worthy of consideration. The cost of ramping up and down a 35HP spindle 24/7/365 for a 3/8" part may buy you a couple of Omni turns alone.

Why not a machine that does the 1st and 2nd op complete? Rotary transfer? Hydromat?

Comatose said:

Without seeing the part, might you be better off milling? Five seconds per part seems hard, but twenty seconds per square inch in a mill sounds glacial. A 10x20 fixture could possibly hold 800 of these. A bowl feeder to a SCARA would easily populate an 800 up fixture in 4000 seconds. That's about an hour. Put that fixture in any mill, face it off, then if you need to pack them nicely a second robot. If not, engineer a flip and dump into your fixture and repeat.

Maybe this is feasible, maybe it isn't. Depends on the part geometry. But spinning a mill up for five minutes once an hour seems a lot easier than spinning a lathe up and down 800 times an hour.

Click to expand...

That is what I was trying to describe earlier!

I vote for mill !

Did not occur to me, but I think a largish mill is the answer, with a custom fixture/air chuck.
Something like a haas vf6 or so, cheap, reliable as hell in alu.
Check the chip evacuations.
Fill it twice (.. n times) a day, even manually if needed.
Maybe hold thousands .. you do the math.

You avoid all the complications and all the spin up/down power wasting, loads and waste go down 95%.
Finish will be great.
Likely custom diamond tooling like in auto work.
A 3000$ tool lasts a year, according to the guys in charge making auto parts, 24x7x365. One tool. One.

A good fixture will be plain on the surface, so the chips just slide off.
You can put in a washdown system with a 500$ pump, if needed, and just blast the chips off.

Comatose said:

Without seeing the part, might you be better off milling? Five seconds per part seems hard, but twenty seconds per square inch in a mill sounds glacial. A 10x20 fixture could possibly hold 800 of these. A bowl feeder to a SCARA would easily populate an 800 up fixture in 4000 seconds. That's about an hour. Put that fixture in any mill, face it off, then if you need to pack them nicely a second robot. If not, engineer a flip and dump into your fixture and repeat.

Maybe this is feasible, maybe it isn't. Depends on the part geometry. But spinning a mill up for five minutes once an hour seems a lot easier than spinning a lathe up and down 800 times an hour.

Click to expand...

I vote for 2 machines myself. 10 seconds / machine is easily doable (well, if you know what you are doing -- 8.5-8.8 seconds is a bit more difficult but still very doable depending on what machine you are using) with a 6-axis, 10kg robot. (I have some running 8.5 sec.)
If you do it correctly, part exchange should be around 2.5-3.2 seconds including orienting the spindle, but this depends on several factors, clamping time is one of the factors - can't do this with a standard 3-jaw with 3 mm of radial clearance to the jaws - clamping time will kill you. A situation like this calls for a collet if at all possible.

OmniTurn


If the link works, there are numerous videos showing Omni-Turns with auto-loaders. Some pretty nice applications for high volume jobs. (if link doesn't work, just google Omni-Turn and look for the listing that says 'videos' in the title.) Good ideas here even if only viewing for future reference.

I think Omni-Turn is much more coordinated with turnkey load/unload systems these days. We initially had some problems with parts getting scuffed up when ejected from chuck at higher RPMs. It was eventually solved. Ours had bowl-feeders for going to the loader. It was years ago. Much better engineering from what I can tell from the few videos I recently viewed.

Regarding volumes, these are very economical. As others have already mentioned, I would just add machines as needed to meet production requirements.

Vertical lathe? Infeed conveyor, spindle picks part, machines, drops on outfeed conveyor. No robots few moving parts.

My apologies everyone for the delayed response. I had to take a last minute business trip and am just now getting caught up.

To answer some of your questions:
Mike1974- You are essentially correct, the parts would be “dumped” into a vibratory bowl that would position them in such a manner that an automation sequence would load/unload them from the lathe.

Hanermo- Great breakdown on the estimated consumption costs. Yes, the overall cost is important so these are certainly factors to consider. By your analysis they certainly aren’t minimal.

Comatose/Hanermo- Milling (and grinding) are approaches that we have considered (and not completely ruled out). We’ve made fixtures very similar to what you’ve suggested that hold ~750 parts. While this operation works, it still takes ~5 seconds per part to machine which does not include loading and unloading (both time consuming). It is important that we have a completely burr free edge which we have not been able to achieve without deburring with a wire wheel after machining which of course adds to the time.

DaveN- The first operation is not a machining operation, it is a forming operation so it’s done on a separate machine.

Flubber- Thanks for the link. I agree that the OmniTurn is a good candidate, I just wasn't sure if there is a different machine or operation that may be better suited for this job.

GaryHLucas- Do you have a video or link explaining the vertical lathe with infeed conveyor?

Thanks everyone for your input!

We have added bowl feeders to Brother machines on several occasions. One of these applications is a facing operation. About as fast as you are going to get to load, unload and machine is 7-8 seconds. That is with a machine with 2000 ipm rapids and 2g acceleration. Of course that is with segregating the machined parts and unloading them somewhere cleaner than the chip bin.

I would look at Hardinge GT-27.HardingeUs.com > ProductCatalog > Product Much better machine faster and long term more reliable. Also if you could axially load the parts through the spindle you might be able to cut down on your load time.

I secound the loading through the spindle, picking them up and putting them in is just nuts, vibratory hopper them into some kinda shoot setup then use a typical bar pusher to advance + eject the last one from the chuck. Nothings stopping you haveing 30 or so in a spindle liner just being pushed forwards one at a time, collet relived so it only grips the one being worked, reloading the spindle liner from the hopper whilst the next parts facing. Think a simple 1" stroke ram as your bar feeder. With a collet and suitable parts catcher, you would not even need to stop the spindle. Part done in just a couple of secounds should be easy, forget tonys 8-10 secounds + the complexity’s of robot tending them.

As to spindle speed, go to a form tool, not a single point tool and you only need to hit a couple of hundred RPM + in-feed but a few thou and the parts done, single pointing is for dinosaurs. A gang lathe with nothing but a stop, parts catcher and skive like facing tool setup would only need to move about 2" total, rapids even at these cycle counts won't matter, hell if anything, you need to extend the moves to get - avoid ball screw - way fretting wear caused by the lube being pushed out due to but partial wiggle of the way trucks balls.

Gang lathes are bloody fast because they don't have to back off to index turrets, they can damn near stay right in the action zone. don't know what the turret index time is on tonys QTN200, but 300 ipm gives you 5" of movement on a 300IPM gang lathe in a second, on a 3/8" part, and say you gang a stop, cutter and shoot at 1" spacing or probably less, you only need 3" of movement total, + a bit of z movement, so say 5" total which is but 1 second, solve the spindle ramp time by keeping it spinning and your productivity goes through the roof, and your 35hp spindle can drop down to a old 1/2 HP washing machine motor with still plenty of spare duty cycle :-)