Motoman K6SB Retrofit

Hi Guys,

What is life without multiple unfinished projects and here is another one, retrofitting my 1991 Yaskawa Motoman K6SB.



This robot was used on the Holden (GM) production line in the Elizabeth South Australia plant welding Commodore exhaust systems according to the guy I brought it off who stripped it out of the plant when it closed.

This will be a slow build as I have big plans for the old girl!

First I will replace all the old propietary AC servo motors with Leadshine closed loop stepper motors.




I have brought six Leadshine EtherCAT closed loop stepper drives to power the motors,



I will build a new control box to house the drives and I will need to buy a power supply that can power them all. I will also use this Brunner EtherCAT I/O module that has two encoder inputs that I hope to connect to two THCAD's. One will used as a welding seam sensor and the other a touch sensor.

I brought these motors because the input drive for the harmonic drives fits these on these motor shafts. I will need to make a new motor mounting plate to mount the motors to the frame.




Does anyone have any advice on what I should do with the old grease in the harmonic drives? I am thinking I tear the frame down and strip all of the old grease away on all the joints and install new grease,

thefabricator03 wrote: Does anyone have any advice on what I should do with the old grease in the harmonic drives? I am thinking I tear the frame down and strip all of the old grease away on all the joints and install new grease,

That, do that !

Nice.
I'm a bit jealous as I have a Mitsubishi arm that I had hoped to refit for running linuxcnc but the motors are built into the casing and the bigger ones have hollowed shafts to feed the cabling through. All custom designs using a proprietary interface for the encoders. I'm still debating on what to do with it.
Where did you get that Brunner EtherCAT I/O?

I brought the I/O unit from Brunner, www.brunner-innovation.swiss/product/ethercat-io/

I had to contact them and they gave me pricing, I think I paid $440 Euro with shipping from Switzerland to Australia.

Robots are funny. Everything each manufacture does is HIGHLY proprietary. I have a few newer units and if they brake down I cannot get new parts and second hand parts are very pricey.

Have you already used EtherCAT in LinuxCNC or why did you decide to go that route?
How do you home the 6 joints? (I presume those drives don't have absolute encoders)
I found that one problem in linuxCNC is the missing capability to check for joint limits when using cartesian coordinates and non-trivial kinematics. Are you using simulation before executing the paths to check for limit violations and/or singularities?

Aciera wrote: Have you already used EtherCAT in LinuxCNC or why did you decide to go that route?
How do you home the 6 joints? (I presume those drives don't have absolute encoders)
I found that one problem in linuxCNC is the missing capability to check for joint limits when using cartesian coordinates and non-trivial kinematics. Are you using simulation before executing the paths to check for limit violations and/or singularities?

I still have to work those things out. The motors have incremental encoders and brakes. I was thinking I save the last known position of the encoder in LinuxCNC and use that to home the axises. Dont know if that will work well or not but I will see. When I repaint the unit I will remake the home position stickers and use them to set the home position.

I plan to work with some developers and get any features I need added to create a LinuxCNC robot configuration. I hope it to be what PlasmaC is to plasma cutters but for six axis robots. I want the config to have singularity detection built in much like my Motoman controllers and also dynamics control. It will be a lot of work but I am up for it. Hope to get it working in the next couple of years depending how I go with other projects I am working on.

I have also brought a copy of RoboDK to plan out my work cells in 3D. I will use that when I get to the stage of creating programs but like PlasmaC, I want it to be able to work on its own and not be tied to any one CAM program. This will be a pretty big project but one I hope will pay dividends in the future.

Also with EtherCAT, I went that way because it is the fastest modern fieldbus. I can use it to link external axises that are spaced around a work cell with normal Ethernet cables. I have not used it so it will be a learning curve.

Cool.
For a while I was working on converting that Mitsubishi arm I mentioned and got as far as working out the kinematics and how to switch between joint and cartesian modes as well as movements in tool coordinates but all just in a LinuxCNC simulation. I'm now stalled because I can't find drives that work with the motors and the encoders. STMBL drives might work but they are being redesigned.
So I play around with my other Mitsubishi robot. The vendor lock-in has dampened my enthusiasm somewhat though.
Anyway, I'm interested to see how you are going to make out with the encoders on your drives as it seems that in these robot joints the encoders sometimes have some insanely high resolution to achieve the desired accuracy.

I am very interested in the accuracy of the Leadshine 1000 line encoders as well.

But more interested in seeing how accurate is good enough. For a welding application you dont need precision of 0.01 of a mm.

thefabricator03 wrote: For a welding application you dont need precision of 0.01 of a mm.

In some cases even 0.5 will do just fine with a 1mm CO2 wire. Tested and confirmed with a Kuka KR150 welding 2mm wall thickness 50x30mm oval tubing.