Controller Limitations

I shouldn't put things down in such an offhand manner, because I don't have a lot of experience with things that come in boxes.

I'm currently struggling with a Teknic controller that doesn't seem to want to let me hard limit the motor current, and which (at least I very much doubt) will let me hard limit the current input into the box.

I'm used to control systems designs that start with a big blank whiteboard, a difficult target to achieve, and a set of very clever circuit designers, mechanical engineers, software engineers, and component vendors. So anything that doesn't give me total control over the schematic and code comes across as being "limited to a frustrating degree".

I wouldn't trust a PM motor to a controller that can't somehow limit the drive current. Typically, current more than seven times the long-term rated maximum current will demagnetize it nearly instantly. Performance is greatly curtailed if the overload protection is something other than a hard limit.

Jerry

Wouldn't the current limit be in the drive? Usually the controller outputs +/- 10volts. It would seem to me the current limit would be in the drive and the voltage limit in the controller.

Don't you feel like you are re-inventing the wheel over and over. It takes a lot of effort to design a motion controller which all the interfacing, sofware and diagnostics. I find it hard to believe you can't find a motion controller some where that will do what you want.

Peter Nachtwey

When one needs to be round, another 'U' shaped, yet another rectangular with a big bite out of one corner, and the fourth trapezoidal, when they all go into a big sealed box that has to sit out on a runway in the desert and operate even when the sun is shining down...

One makes custom.

There is a surprising number of applications for closed loop control that doesn't fit into the traditional 'square box' market. Mostly it's either extremely critical applications (like the one cited above), very space constrained applications, or high-volume applications (I suspect that anything that's going to ship more than 100 units a year is getting up to the point where it's more economical to roll your own; I should see if I can pin that down).

At any rate, the actual control software is usually dwarfed by the rest of the software on a board; in a custom application you can spend as much on an interface board to translate between your systems internal communications and an 'off the shelf' part as you would just implementing the whole dang thing yourself.