Thread: Robot & HMCs... Tony?

I see a lot of robots tending lathes and tapping centers, but HMCs... not so many.

How does the investment of a robotic cell compare to the same machine with a pallet pool? I'd imagine that the additional equipment (automated rotation of the 2nd pallet station, pallet rinsing, raw material conveyors, finished workpiece conveyors, etc) adds up quick.

You can't additionally cost the raw material presentation, finished goods removal, fixture rinsing, etc in this comparison as those are required for *any* robotic machine tending cell. The only real additional cost is in the pallet rotation and the issue of getting your power to the pallet for clamping mechanisms because of the rotation issue. This can be done using several methods, from a simple index cylinder and stop pin to a full blown servo system. Personally, I'd go with servo so that it would be a simple program change to go from a 2 sided tombstone to.. say a 8 sided tombstone...... or you might have a variety of different pallets in the pool with unique loading requirements.

We use to have Mazak H-400's with robotic loading. These were gantry-style robots, but robots nonetheless. The very first system used an air motor and stop pin to index the pallet (circa 1991). That proved somewhat problematic in reality in that the motor would sometimes overshoot/undershoot and the pin would miss, but for the most part, it ran well. The next version we used a servo motor on the pallet rotation and that worked wonderfully for many years. We had 6 of these running 24/7 for many years.

When you step up to a 6-axis robot, you must size the robot to the expected maximum part weight. Since HMC's often see a lot of larger parts, you may need a large capacity robot and associated feed/removal hardware. However, other than that, it can be a much easier job than working inside of a VMC / Lathe because of clearance issues. Some of these integrations with regular VMC's/Lathes can be a real PITA when it comes to working clearance inside the machine.

Thanks Tony.

We'll be adding at least two HMCs to our current lineup in the next 2 years and will likely go the robot route. Even with optimized toolpaths, toolchanges, etc, our cycle times never dip below 30 minutes on a sparsely loaded 500mm 2-sided tombstone (2 parts) or 45 minutes on a 500mm 4-sided tombstone (4 parts). We could probably get a single robot to tend 4+ machines. Have you ever setup anything like this?

I was thinking about attaching the 6-axis to a gantry, keeping the floor space clear. Machines would be positioned face to face to minimize gantry length.

Radar, how many different part numbers are you talking about? Those longer cycle times are ideal for a regular ol' off the shelf FMS. How long does it take for a guy to unload/load one of the 4 sided pallets that runs 4 parts? Could you build fixtures that hold more parts for even longer cycle times? Is your demand high enough that you want to run 24/7, with 16 hours of that lights out?

The gantry idea is a good one and doable as most robots are designed to be floor, wall or ceiling mounted. The issue is that the gantry needs some serious engineering and build. With 6 axis of freedom, there can be some *substantial* moment loads to deal with in *any* direction depending on robot size and payload. I prefer floor mounting when possible because of this reason. Most all robots are available with a 7th linear axis to be able to move between machines.

As for doing this, we've done many cells where the robot tends more than one machine. Even tried some gantry systems, but it proved exceedingly difficult to get it stiff enough for the application in the limited space we had.

Originally Posted by Joe788

Radar, how many different part numbers are you talking about? Those longer cycle times are ideal for a regular ol' off the shelf FMS. How long does it take for a guy to unload/load one of the 4 sided pallets that runs 4 parts? Could you build fixtures that hold more parts for even longer cycle times? Is your demand high enough that you want to run 24/7, with 16 hours of that lights out?

About 15 part numbers. All of them start out as raw stock and are completed in two operations hitting six sides. Many of them have the same width and height dimension, so they can gripped with the same end effector.

A 5-8 carrier pallet pool on each machine would in fact be sufficient to keep the machines running 24/7 with a single weekday shift, so that is our default option at this point. We probably wouldn't install a full FMS since we'd have an assortment of 400 and 500mm machines, and possibly a 630 thrown in the mix.

Max part weight is 150 lbs, so we would need sizable robot, but the expense would be justified if we can eliminate the man-handling required to muscle these parts onto the tombstone... not an easy task even with crane assistance. It would also be nice if the robot can completely automate part rinsing, part drying, and oiling, as that always seems to require a lot more labor than expected.

Originally Posted by Tonytn36

The issue is that the gantry needs some serious engineering and build. With 6 axis of freedom, there can be some *substantial* moment loads to deal with in *any* direction depending on robot size and payload.

Hmm... maybe not then since we would probably use a heavier robot.

150 lb raw stock weight, or 150 lb finished part weight? You'll have to size the robot the next one up because of the end effector weight. If the raw stock is 150 lbs, an 80 kg can handle that, but when you add in a gripper big enough to handle the parts and associated mounting brackets and arms, you would probably be pushing it, if not over. I'd opt for a 100 or a 130 kg in this situation, depending on reach needs.

Schunk makes some long stroke grippers with the moveable elements on linear rails for superb rigidity.

One thing to watch out for and you may need to account for this in your tombstone fixtures. Once you get out of the medium-heavy payload robots (in Kawasaki robot terms.....the RS series) into the heavy payload (i.e. Kawasaki Z series) the accuracy of the robot is less. For instance, the repeatability of the JT6 flange on an RS is +/-0.07 mm, while the same for the Z series is +/-0.3 mm. This is true for any robot manufacturer that I'm aware of. This can be accounted for in the fixturing though.

150lb would be the raw stock weight.

Is there typically any "float" in the end effector? 0.3mm wouldn't be a problem with hydraulic edge clamps, but how would one reliability position a workpiece flat against the tombstone with no gap?

Originally Posted by Radar987

150lb would be the raw stock weight.

Is there typically any "float" in the end effector? 0.3mm wouldn't be a problem with hydraulic edge clamps, but how would one reliability position a workpiece flat against the tombstone with no gap?

a) You design a loading helper to hold the stock after the robot lets go, but before it clamps.
b) You design such a device into the end effector
c) You use a compliance device between the end effector and the robot JT6

a or b would probably be the easiest if you have locating pins or another device to align the part in plane. You just need a spring loaded pin or pneumatic cylinder the robot can push against the stock with after/as it opens the gripper(s). As we are primarily concerned with cycle time in our applications, we typically use method a. This way the robot can be moving out of the working area while the clamping is being performed. Since cycle time doesn't seem to be your primary concern, b will also work. With your loading, option c probably isn't an option because I don't know of a compensator with that load capacity. ATI's largest unit only has an 88 lb tension capacity, well under even your part weight.

Hey Radar, I just got a spam this morning that made me think of this topic. It was from usedmachinetool.com but it looks like they don't have this particular machine listed yet. It's a 2010 Mori NH4000 with auto pallet indexing, hydraulics through the pallet, auto doors, and a robot interface for $270K. The machine looks brand new. They claim the auto stuff and the robot interface was $50K. In the pictures, the machine looks like it's still in service, and they definitely never had a robot hooked up. With the shocking price that Mori charges for stuff like spindle probe, tool probe, Consep conveyor, etc - I wouldnt' be surprised if somebody paid $400K for this machine.

Here's a copy/paste of the spam:

__________________________________________________ ____

**** This is the HIGH PERFORMANCE model with all the features -- not the new "stripped down" machine ****

2010 MORI SEIKI NH4000DCG Horizontal Machining Center
UNDER 900 Cutting Hours

PRODUCTION READY

FULL 4th AXIS
TOMBSTONES WITH HYDRAULIC FIXTURES

THRU PALLET HYDRAULIC COUPLING with
AUTO DOOR AT SETUP, AUTO PALLET INDEXING STATION ($50k NEW)

ROBOT INTERFACE

ETHERNET INTERFACE

SPINDLE AND TOOL PROBES

AND MORE..................

CHECK OUT THE PICTURES - THIS IS LIKE NEW -

Pallet Size:15.75" x 15.75", drilled & tapped
Number of Pallets:2
Pallet Change Time:6 Seconds
Pallet Index:Full 4th Rotary
Pallet Weight Capacity:880 Lbs.
X-Axis Table Travel:22.04"
Y-Axis Spindle Head Travel:22.04"
Z-Axis Column Travel:24.80"
Spindle Nose to Pallet Center:3.9" - 28.7"
Rapid Traverse Rate:1969 IPM
Spindle Taper:CT 40
Spindle Speeds: (Two Gear Ranges)14,000 RPM
Spindle Drive Motor:25/15 HP AC
Automatic Tool Changer:60 Positions
Automatic Tool Changer Typeual Arm, Side Mount
Tool Change Time:0.9 Seconds
Max Tool Length:15.7"
Max Tool Diameter:5.5"
Max Tool Weight:17.6 lbs
Machine Weight: (Approximate)21,120 Lbs.

Equipped With:

*Mori MSX-701 IV (Fanuc 31i) CNC Control equipped with: MAPPS IV, 14" Color LCD, Graphics, Custom Macro, USB port, PCMCIA port
* Coolant Thru Spindle interface
* Full Fourth Axis Rotary Table
* 60 Tool Magazine
* Overhead Coolant Shower
* Rigid Tap
* Shower Coolant
* Helical Interpolation
* 48 Work Offsets
* EtherNet Interface
* Robot Interface
* Tool Life Management
* CON-SEP Chip Conveyor System - Dragger/scraper, drum filtration, lift-up conveyor - rear discharge
* In-machine measuring system (table) touch setter,
* Spindle Probe Automatic Measuring System,
* Through Pallet Hydraulic Coupling Interface.
* Automatic Pallet Indexing Setup Station
* Auto Door at Setup Station

***$269,900,000*** FIRM

Originally Posted by Tonytn36

a) You design a loading helper to hold the stock after the robot lets go, but before it clamps.
b) You design such a device into the end effector
c) You use a compliance device between the end effector and the robot JT6

You've given me much to ponder about. Thanks!

Originally Posted by Joe788

Hey Radar, I just got a spam this morning that made me think of this topic. It was from usedmachinetool.com but it looks like they don't have this particular machine listed yet. It's a 2010 Mori NH4000 with auto pallet indexing, hydraulics through the pallet, auto doors, and a robot interface for $270K. The machine looks brand new. They claim the auto stuff and the robot interface was $50K. In the pictures, the machine looks like it's still in service, and they definitely never had a robot hooked up. With the shocking price that Mori charges for stuff like spindle probe, tool probe, Consep conveyor, etc - I wouldnt' be surprised if somebody paid $400K for this machine.

Interesting... I wasn't even aware that they sold the machine with a built in robot interface with indexing pallet station. Much more appealing than a retrofit.

Also interesting how they spec'd that machine... all of that additional equipment but only a 60 tool ATC. Would've expected at least 120.

Thanks for that!

Originally Posted by Radar987

Also interesting how they spec'd that machine... all of that additional equipment but only a 60 tool ATC. Would've expected at least 120.

Thanks for that!

Ever seen what Mori wants for that 120 ATC?

Originally Posted by Joe788

Ever seen what Mori wants for that 120 ATC?

Good point.

And come to think of it, our tool list per machine would decrease drastically if we were to split up the part numbers running on 2 machines among 4+.

Most MTB's offer a standard robot interface to CNC machines now. Whether that is what you actually need or not is another question. Some are very sparse on the signals, some are rich. Most of them think very highly of this interface ($$). Since we have 9 very experienced EE's on staff, we typically just create the interface ourselves. Most companies don't have this option though.