Remastering a Motoman.

    Heya Everyone!

    Hope everyone reading this is happy and healthy!

    I have a question although I suspect the answer is a definite "No".

    We have implemented an Offline programming software called RobotMaster to work with our Motoman XRC Robots without having to shut down the robotic cells to program new jobs. We've come across unique issues with our robots in their setup which have largely gone unnoticed due to online programming being unaware of the issues.

    We've noticed across all our robots that a combination of the TCP and Axes values are out, sometimes rather dramatically. I've been trying to convince management to get Yaskawa technicians in to Remaster and align the axes professionally which will only benefit our offline programming capabilities (as far as I am current aware) but they are concerned that as soon as we do that all of the currently programmed jobs on that robot will be professionally out.

    Every robot we have sent off to have the cell designed within RobotMaster we've been told its has mastering issues. We are refining the TCP values offline but that can only get us so far without having to face the elephant in the room: The Robot mastering.

    Out of curiosity, Is there a program or something we can run that will prevent us from having to reprogram all the current jobs on our robots once its remastered? We currently use a program called DynaCal I believe to assist in the transition of programs jumping between robot cells.

    I suspect that if we update the settings in DynaCal we can simply just "convert" the programs to the new values to minimise the amount of work that will be required if we do remaster the Robots professionally but every time I question this way of going about it I'm simply told no without any real explanation from our lead programmer.

    any and all help would be appreciated.

    Robot Type: Motoman XRC UP50

    Controller Type: JZNC-XPPO2B

    Controller Cable Type: RS232C

    I don't know if the above information is helpful but I figured I'd throw it in to assist in anyway shape or form.

    Thanks Everyone!

  • Go to Best Answer

    Just curious what your industry is? In my experience, every time I've received programs which were written off-line, usually designed with motosim, they always required at least one touch up pass ie: they aren't perfect.

    Also after zeroing out any axis, I usually end up touching up every critical position. I'm not sure if there's a way to avoid this after you re-master a robot. I think there is some kind of sensor based system that records before and after TCP positions and offsets things accordingly, but how well these work in reality I don't know. Is it a big deal just to touch the programs up? Not sure about your facility/industry and how things operate there. I'm sure someone here will have some better answers for you.

    Quote from Robottech

    Just curious what your industry is? In my experience, every time I've received programs which were written off-line, usually designed with motosim, they always required at least one touch up pass ie: they aren't perfect.

    Also after zeroing out any axis, I usually end up touching up every critical position. I'm not sure if there's a way to avoid this after you re-master a robot. I think there is some kind of sensor based system that records before and after TCP positions and offsets things accordingly, but how well these work in reality I don't know. Is it a big deal just to touch the programs up? Not sure about your facility/industry and how things operate there. I'm sure someone here will have some better answers for you.

    I work in automotive and our robots are primarily milling robots working with plastics.

    I mean, I expect some form of touch ups (nothing is perfect, I understand that) but due to the nature of the issues it's a lot more than what I'd think is reasonably acceptable. Especially when I've seen the finished results of programs designed off RobotMaster (our installer uses it in his own private business) and he doesn't have half the issues we face. Management see the downtime of touching up every program due to remastering (even on the robot cells that don't get touched for weeks at a time) as a waste of time and money which i can understand but biting that bullet would make things a lot easier for offline in the future.

    I can get the programs done Offline but depending on what robot cell I'm using to get the job done, it'd honestly be quicker to online program most jobs.

    Also probably doesn't help that the Robots we are using are 20ish years old.

    Quote from Dreamawake

    Every robot we have sent off to have the cell designed within RobotMaster we've been told its has mastering issues. We are refining the TCP values offline but that can only get us so far without having to face the elephant in the room: The Robot mastering.

    Out of curiosity, Is there a program or something we can run that will prevent us from having to reprogram all the current jobs on our robots once its remastered? We currently use a program called DynaCal I believe to assist in the transition of programs jumping between robot cells.

    I suspect that if we update the settings in DynaCal we can simply just "convert" the programs to the new values to minimise the amount of work that will be required if we do remaster the Robots professionally but every time I question this way of going about it I'm simply told no without any real explanation from our lead programmer.

    I think you are going to have to do touchups. Any software used will be garbage in, garbage out. Two issues going on the tcp calculations are based off the mechanical home position calibration data. If the calibration data is messed up trying to teach a tcp will be messed up.

    If you have Yaskawa come in and "master" your robots, try to get them to use MotoCal not MotoCalV. MotoCal is the DynaCal software (least was 15 years ago) but with the ability to calculate link lengths. DynaCal will only get new abso data. MotoCal will get new abso data and link lengths. I don't know if Yaskawa does MotoCal any more since MotoCalV was released. Neither one will compensate for a 20 year old drive with wear and tear.

    I know a thing or two, because I’ve seen a thing or two. Don't even ask about a third thing. I won't know it.

    Quote from 95devils

    I think you are going to have to do touchups. Any software used will be garbage in, garbage out. Two issues going on the tcp calculations are based off the mechanical home position calibration data. If the calibration data is messed up trying to teach a tcp will be messed up.

    If you have Yaskawa come in and "master" your robots, try to get them to use MotoCal not MotoCalV. MotoCal is the DynaCal software (least was 15 years ago) but with the ability to calculate link lengths. DynaCal will only get new abso data. MotoCal will get new abso data and link lengths. I don't know if Yaskawa does MotoCal any more since MotoCalV was released. Neither one will compensate for a 20 year old drive with wear and tear.

    I always knew I'd have to do touch ups, but I just wish I could minimise the amount. At the current moment on one of the robots, if I do a curve movement, it pulls the cut bit out as it rotates. I've had to split tool paths off line and sometimes even move tool paths 15-20mm to get them within tolerance. Our RobotMaster technician has tried his best to even out the errors to make it as accurate as possible but it just feels due to these issues we can utilise the OLP to it's full potential which is frustrating to say the least.

    I'll definitely keep MotoCal in mind during my next meeting when I push for updates to the robots. I had a suspicion that using modern software with 20+ year old drivers could be a big issue but our management team have the mindset of "We'll deal with that issue when it arises" and even then they refuse to update to newer or even second-hand newer Robots which i can only imagine would benefit the business in the long run.


    Quote from Lemster68

    I think that is a very good and important point there. Like, even if it was mastered well, what you might be seeing as your issue is wear. Or, at least, part of it.

    I had a suspicion that would be the case. Every time I bring up wear, our maintenance manager shuts that idea down immediately stating their 6 monthly maintenance prevents that from occurring but the evidence speaks to the contrary.


    Thank you everyone for all of your help and support, It is always greatly appreciated! :smiling_face:

    Hi, I've been doing some research between a few OLP softwares and also came across RobotMaster.

    I was wondering if Yaskawa is supposed to come and calibrate your cell before RobotMaster does his thing?

    I've been told no from RobotMaster but i'm a bit skeptical about that point.

    Quote from Loki

    Hi, I've been doing some research between a few OLP softwares and also came across RobotMaster.

    I was wondering if Yaskawa is supposed to come and calibrate your cell before RobotMaster does his thing?

    I've been told no from RobotMaster but i'm a bit skeptical about that point.

    Hey mate,

    I think that depends primarily on the condition of your robot. When we supplied the backups of our Motoman robots to Robotmaster to calibrate and generate the robot/robot cell they could see our mastering issues within the back up itself.

    I'd suggest staying in close contact with whoever you are dealing with during the creation of your cell design within RobotMaster and if they suggest getting someone out to calibrate the robot to minimise the inaccuracies I highly suggest going through with their suggestion just from my personal experience as I am currently living the nightmare of having to OLP with inaccurate Robots and management refuse to try and fix the issue.

    All-in-all I think RobotMaster is a fantastic program is utilised correctly.

    Hope that helps, mate.

    Quote from Loki

    Hey, thanks for your answer.

    Funny thing is, I think i'll go toward Motosim after a demo they showed us. So calibration will be part of the process anyways!

    Heya, No worries! Glad to be of any and all help if possible :smiling_face:

    Ooh, Colour me interested, I'll definitely have to have a look into Motosim!

    Late answer to this post.

    Unfortunately are robots quite poor (not only Motoman) considering about absolute accuracy. The robot suppliers always state the repeatability of the robot and that is often very good, typically 20-60 micrometers. When you use tools such as RobotMaster this is of no help. You need the robot to reach a specific coordinate since the target has never been taught on the specific robot individual you apply the program on.

    The first step is a good mastering. Then you typically can reach accuracy +/- 2 mm (sometimes worse than this). Manufacturing and assembly tolerances of the robot prevent better accuracy.

    I am working in a company addressing this problem. By using external optical measurement system, it is possible to create a model of the robot describing the differences against the "perfect" robot (both distances and angles between joints a.k.a kinematics but also elastic characteristics to have better accuracy for different payloads). We can then reach typical accuracy of +/-0.2 mm for a mid sized robot. We can then perform a post-processing step, using the program generated from for instance RobotMaster and calculate "fake" targets in a new program that will make the robot move as intended.

Advertising from our partners