[nmorong]

Craig- A while ago I looked into and am not really interested in the open loop controllers. I want to end up with the most robust and customizable setup, not the simplest and cheapest. A major goal is to add a 4th and 5th axis, which is much easier if I can work with whatever I can find. I also need servo control for my spindle, replacing the (10kW?) spindle drive with something that closes the loop at the drive would probably make the Kogna look cheap in comparison.

I kinda want to give the Dynomotion stuff a try. In my experience most stuff aimed at the hobby market is not as good as the professional stuff, while most professional stuff purposely limits you to sell you options/upgrades. The success that some people have had with Kflop retrofits on actual industrial machines, combined with seeing how much work Tom is putting into support makes me think it can get me the results I want, and upgrades are limited by my ability to figure it out rather than figuring out how to pay for it.

Tom: I actually already have a Kflop board, but I think I’m leaning towards buying the Kogna for this project. I’ll find some other use for the Kflop. I need to find the time to plan the whole system and make a purchase list.

-Nate

[joeavaerage]

Hi,
a machine with closed loop servos but with an open loop motion control is every bit as good, and in some case better than a closed loop motion control.

The bottom line with closed loop feedback is bandwidth. The higher the bandwidth the better the mo0tion of any servo and its response to a command.
Modern AC servos, even the el-cheapo Chinese ones have (position) closed loop bandwidths of 200Hz to 500Hz, and the top tier Yaskawa's and siemens 1kHz and better.
No closed loop motion control is going to match that.

As an example I have an Allen Bradley AC servo from circa 2005. Included in the drive is a feed forward term to model the magnetic saturation of the servo. This means that less feedback need be used
to linearize the servo resulting in greater bandwidth. This is possible because Allen Bradley have such detailed and intimate knowledge of their own servo that they characterize the magnetic saturation.
No closed loop motion controller can match tha....not because they could not, but no user has the mathematical details of their servos magnetic properties. Without the intimate data you cannot make
a feed forward correction. Thus the closed loop bandwidth of this servo will always be higher with this servo when the loop is closed by its matching servo drive than with any motion controller that closes the loop..

Second issue that you might like to ponder is that all the best CNC machine manufacturers are going to distributed motion control using a bus like Ethercat, Profibus, CANOpen and others. Such trajectory planners,
the central PC are not feedback types, the trajectory is planned, communicated to the servo drives and EACH servo drive is responsible for its own portion of the overall motion control effort and IS closed loop.
It seems good enough for DMG Mori....then its probably good enough for me.

You said at the beginning of this thread you wanted to understand the broad principles of the choices that you might make, then is a good example. Fully closed loop controllers were de-rigor back in the 80's and 90's,
but they are steadily losing market share to open loop controllers (with closed loop servos) and distributed motion control (with closed loop servos). I would guess the market share has shrunk from 90% to something like 10%.
Consider this statement 'fully closed loop motion controllers are old school'. Do you agree or disagree, and what would motivate someone to say such a thing?

Craig