Which Brands to Consider?

    Hello,


    Just joined the forum :smiling_face:


    I have been tasked with buying a robot to replace an existing mechanical system for feeding a stamping press. We are a university doing research (mostly in warm/hot sheet forming applications), so continuous production of the same part(s) is not part of our daily life.


    This will be the first robot installed in our labs, so we are essentially starting with a clean-slate in terms of brand preferences and compatibility. We will be connecting this to an Allen Bradley PLC via Ethernet/IP (I assume all major robot manufacturers will have this comm available).


    What I am looking for help with is which brand would provide the best features for low volume prototyping/research? We will be continuously changing parts, geometries, paths, etc, so I don't want to end up with a cumbersome interface which makes these sorts of changes more difficult. Without having any prior robot experience, it's hard for me to predict what features for a particular brand are going to be beneficial vs ones that are detrimental.


    I have been looking at Fanuc (mostly due to user base), ABB, Yaskawa, Kuka, etc. I have been playing with ABB's RobotStudio, which seems like a powerful piece of software. I also like the RAPID style of programming (it's my understanding that not all brands have this high-level programming language?).


    Our robot capacity that I am aiming for would be 6 axis, 40-60 kg rated payload, and a reach of 2.0-2.5 m.


    Thank you for the advice.


    Cheers

    Hi


    My personal opinion and based on what you wrote go for ABB or KUKA.
    I enjoy all of them, most of my experience is actually with Motoman and Fanuc, but, again based on what you wrote go ABB or KUKA. You will not be disappointed

    Retired but still helping

    Are you buying new or 2nd-hand? That has an effect as well -- due to Fanuc's corporate policies, for instance, buying 2nd-hand has definite hidden costs.


    My personal favorite robot for most "hacker-iffic" work is KUKA, but I wouldn't mind a newer ABB (the older S4-generation series had, in my limited experience, lots of hardware issues). They both have quite good motion planners, in my experience -- your path suffers very little change between runs at low-speed and high-speed. Fanucs, IIRC, need an add-on option to match this performance.


    OTOH, for a "simple" material-handling, machine-tending application, an even cheaper robot might do, but it's hard to say if/when you might run into their limitations. ABB, KUKA, or even Fanuc might be over-kill for what you're trying to do, but over is always better than under.


    Are you planning on lots of hands-on, teach-pendant programming? Or more programming done offline, then downloaded to the robot? Lots of simulation? Any "exotic" motions, like really time- or dimension-critical paths (things like spray paint, or very high-quality MIG welding, etc)?


    If you're planning lots of critical decision-making code in the robot (as opposed to making the robot a dumb slave of the PLC), then if you look at Fanuc, you need to be aware that Fanuc's higher-end programming language (Karel) is a paid option, not a "natural" part of the robot (as opposed to most (all?) of the competition). I'm not trying to steer you away from Fanuc, but the way Fanuc tends to make so many things add-on options is something any first-time Fanuc buyer needs to be aware of going in.


    (OTOH, to be fair, KUKAs require you to buy E/IP, ProfiNet, etc, as options, and you need/want to buy them as part of the initial robot order, since getting them after-market is always more expensive).


    You also need to pay attention to local/regional support. A less-perfect robot with better, faster local support is probably better than a "perfect" robot whose support has a 3-day turnaround time. You don't always have to get support from the mfgr, either -- most robot brands have "system partners" and "certified integrators" located in lots of places that can provide support and training. In fact, finding a good local integrator (many of them will provide first-visit advice to schools for cost, or even free sometimes) can save you from a LOT of pain and aggravation. They can help you choose a good robot brand for your application, and help you avoid pitfalls in interfacing, option package selection, mounting issues, etc.

    Thanks for the feedback.


    To clarify, we will be buying new. It also seems like once we buy a piece of equipment we will have a bigger/faster requirement a year down the road, so this is what we tend toward buying more than we actually need.


    I anticipate most programming will be done offline, then just tweaked into position at the robot. We like simulations here :icon_mrgreen: and it will certainly be useful for EOAT designs and verification. There will not be any "exotic" motion required. Essentially pick and place.


    It sounds like Fanuc robots are similar to Fanuc CNC controls in that everything is an option and everything is expensive :uglyhammer2:


    I'm not sure yet if I will handle some of the I/O and decision making at the robot or if I should just do it all in the PLC. In the end it will probably be a bit of both. I assume things will generally be quicker if I let the robot handle the code?


    Support is good, but not usually critical. Waiting a few days/weeks generally isn't a problem. We are located near Toronto, Canada.


    Cheers

    Take a look at Kawasaki, also. You may be surprised. Effective Pendant Programming ("BLOCK programming") for "simple tasks" AND a robust programming language ("AS") for "more sophisticated tasks." Programs can be a mixture of BLOCK statements and AS statements, very fluid and useful for applications that are easy to turn over to the techs so that they don't call you in the middle of the night.

    I have to agree -- Kawasakis are suprisingly capable for "cheap" robots, especially for relatively simple tasks like material handling. OTOH, I can honestly advise you to avoid Nachi like the plague from personal experience, and every colleague I have who works with Comau robots has nothing good to say about them.


    Keep in mind that, if you want to use simulation to generate your robot programs, the simulation software tends to be very $$$$ -- I've heard very good things about Visual Components, which is a lower-cost option compared to Process Simulate, without giving up too much capability, but even that is something like $20k/year. And trust me, you don't want to know what Process Simulate costs (don't even ask about Catia/Delmia!). On the really low end, there's RoboDK, but I don't know how capable it is in outputting OLPs.


    Also, your Simulation package will also have to have a brand-specific postprocessor option for your robot. Also, on the higher-end packages, there's usually an additional "Realism" option if you want highly accurate time results from your simulation ("RRS" in Process Simulate), and these are also brand-specific and additional $$.


    So, just keep in mind, there are some non-obvious costs involved that can sneak up on you.

Advertising from our partners