What is the most advanced/impressive application you have ever implemented?

Well, the large-scale fuselage assembly with robots mounted on OmniMoves and tank-drive platforms, with accuracies down in the 0.030" range, was pretty impressive... http://www.boeing.com/boeing/Feature…e_07_14_14.page

Or there was the GM press-linking system, with the robots "following" each other using dynamic speed control, and self-optimizing on the run to keep firing the presses earlier and earlier until they were barely missing the robot on the exit move....

I've done many jobs tending CNC machines, presses, done palletizers I call them "the precise job". Most of them as challenging as they are, most of the motions are horizontal and/or vertical. I like these job because you use many resources from the robot. From your own routines doing some math, to the massive handshaking with the PLC and the machine in question. I think this job probe how good (accuracy, etc, etc) more than teaching skill.

I think the most challenging job a waterjet/laser cutting. First of all, they have very fast motions, a lot intricate motions, decelerations, accelerations that really show how good you are teaching path. I usually work with 4 robot sharing the same area. They are very close to each other and continuously running on different directions with sudden moves.

In the last 5 years I've dozens of jobs with vision and I really don't enjoy them. I hate to depend on camera values to program my path, but I guess that the future

I hope to one day do projects like these. I know I have a long way to go but I am determined to get there. I have learned a lot working with Kawasaki robots and next month get to start learning Fanuc. I have come a long way with not having any formal schooling. My hat is off to you guys and these awesome projects. I love robotics!!

Unfortunately the most advanced/impressive ones are the ones people aren't allowed to talk about
There's many cool projects being worked on for the military and aerospace industry, hence the nda agreements.

The highest level of accuracy job I worked on was TIG welding the engine shroud for space rockets.
This involved some 5000+ (yes five thousand) 50mm / 2" long welds, all with laser search/tracking,
automatic exchange of torch neck / tungsten electrode, tcp-calibration etc.

Another fun job was welding valve seats for marine engines, so we're talking BIG valves.
The program was created so it was fully parametric and in the end the operator had to enter a
"recipe" with ~8 values and the robot would take care of path generation, multilayer, searching, etc.

Biggest one is probably a project with three inverted robots mounted to its own Gantry with a 4m hoist/elevator,
3m slide and a ~30m track. Two 10T single axis positioner with a height/length adjustable tailstock and a 5T 2ax positioner.
All in all, 33 axis ( not on the same controller though ).

Dear Friends,
I had done a very interesting application of printing the graphics on the fuel tank of motor cycle using normal industrial printhead.
Challenges were like:
Accurate path teaching.
Maintaining the gap of 3 to 5 mm between fuel tank surface and printhead.
Fuel tank is having the curved surface.
Integration with Printhead.
Development of Inks.

Biggest Challenge: Fuel tank is made through seam welding and the surface gets deviated by 10 to 12 mm and still i have to maintain the gap of 3 to 5mm between the surface and printhead.
I have done this with a simple method where LVDT is used to identify the deviation and same is given to robot in forms of XYZwpr as offset.

Outcome of the project is we have reduced the costing of graphics by 97% as only ink is used for graphics and eliminating costly stickers.

Hope you liked the project.

Regards
Sachin

going from offline to finished welds without a dry run was one of my most satisfying, and technically satisfying jobs ... although at the time I hated it.

My top 2, in no particular order:

(1) Deployed a high-precision cutting operation at a customer site. Curvy path, path points required to be perpendicular to surface at all points and maintain <target offset dim> +/- 0.05 mm path tolerance (as measured by destructive testing of the workpiece). As the robot moved the cutter tool around one portion of the path, the acute angle between Link2 & Link3 decreased in negative motion, stopped at an inflection point, and then reversed to positive motion. This inflection caused an additional 0.15 mm to path tolerance over about 30mm of path length. We made a transcontinental conference call with the robot company's Applications Engineering Department and subsequently modified the Joint2 & Joint3 servo parameters and kinematic model to reduce that additional aberration and maintain the +/- 0.05 mm path tolerance. About 2500 path points.

(2) Implemented a high-precision dispensing application for a heater grid on a workpiece with complex curvature. EOAT was an ink-jet type of dispensing apparatus requiring workcell & program integration. Paths were all created in offline programming with all point vectors normal to surface. The workpiece had no useful datums to establish a frame for path localization. Designed, fabricated, and implemented frame locator hardware on the workpiece to establish a frame on this part. Workpiece had significant geometry challenges affecting the multiple paths. Each path required customization of speeds, accels/decels, and coordination with the dispensing frequency of the jet head. This was necessary to produce heating grid lines with predictable specific resistance (ohms/mm) values critical to the heating grid performance. About 5500 path points.

Quote from mark

I hope to one day do projects like these. I know I have a long way to go but I am determined to get there. I have learned a lot working with Kawasaki robots and next month get to start learning Fanuc. I have come a long way with not having any formal schooling. My hat is off to you guys and these awesome projects. I love robotics!!

I'm just like you !!!!

Quote from SkyeFire

Well, the large-scale fuselage assembly with robots mounted on OmniMoves and tank-drive platforms, with accuracies down in the 0.030" range, was pretty impressive... http://www.boeing.com/boeing/Feature…e_07_14_14.page

Or there was the GM press-linking system, with the robots "following" each other using dynamic speed control, and self-optimizing on the run to keep firing the presses earlier and earlier until they were barely missing the robot on the exit move....

SkyeFire that was programmed using CENIT software! Very cool project indeed!

Van Every