Max performance

Hi

Is it possible for EMC to control modern servo's with high rapid rates say 2400IPM and high feedrates, 1200IPM, I assume no chance with the ltp port but with a plug-in card/usb/cat connection?

A very high sample rate is not really required as the mechanical bandwidth of the system will likely be in to 100 Hz region.

Also the motor drives will have bandwidth limits in the couple hundred Hz to ~1 KHz region due to motor inductance (and random delays due to unsynchronized PWM switching) This in itself make sample rates over 10 KHz or so not of much use for medium to large size systems

Another way of looking at errors vs sample rate is by seeing what the maximum error you can get to with a full scale step in torque load. Say you have the 1200 IPM cutting speed and a fairly aggressive 0.1 second acceleration to full cutting speed (200 IPS/S) if we use this max acceleration for maximum load step size (this splits the load/acceleration torque budget of the motor 50-50) we can calculate how far our position can be off in single sample. With a 1 ms sample time the error distance traveled will be 1/2AT^2 or .0001" . Notice the inverse square of the sample period, so this gets smaller very fast when you get above 1 KHz.

A higher sample rate _will_ improve the loop phase margins, allowing more PID gain

The Mesa HostMot2 encoder counters with filtering on count to 4.4 MHz (5I20/5I25) or ~5.3 MHz(all other FPGA cards) Normally the encoders will be the limiting factor

To be honest I haven't cut at that speed yet but it's advertised on the brochure. I have used the 2400ipm rapids though, well a bit less I think 2126 but I know some machines are 2400, with linear motors I've heard of 3500 or so... and was just curious to be honest, I have no need to replace my newer machine controls, only the old ones.

Nevertheless I've had a look at some replacement modern AC Servo's and they are over 100K steps per revolution on the encoder, in the faster systems I believe these are coupled to 20mm pitch ballscrews.

So that's roughly 3000 RPM at rapid speeds so ~5 MHz count rate ( 5MHz = 50 RPS*100,000)
The HostMot2 counters can do that easily with the filter off (assuming these are quadrature encoders)

Well it's actually just over 130K but you would probably use it with a 2 or 4x electronic gear or more...

Staying on this topic, what about the acceleration curve? Can EMC be set to accelerate the motor like a modern control which uses an S curve or "bell" curve? Part of getting high performance is not the top speed but the acceleration, with mach one can only have a linear acceleration, which of course then limits one as well.

Thanks, well I'm not a pc programmer, more a user so my experience is related to what I've used so excuse my simplistic explanations / questions. That being said when I compare a 10m/min fanuc 5t machine to a 54m/min fanuc 31i machine the difference is massive, the 31i control has very good jerk control and velocity curves, the machine moves along without causing much stress in the frame, the 5t however clearly has very poor acceleration control and can be heard when starting and stopping, the same I've noticed on a 6t as well... When fitting these older machines with modern controls I would like to have the smooth motion of the 31i if possible, even if the speed remains at 10m/min.