Controller Step input questions - any Einsteins out there?

Hi,

I am newbie and am looking to build Controllers for a DC Servo on a Bridgeport. I am thinking of using Mach4 and a CNC4PC breakout board.

Can someone point me in the direction of a GOOD reference to understand the various protocols / standards involved?

As far as I can see the Controller inputs, for each axis, are Step and Direction; with a Controller Error line output.

The Step is a set of 5V pulses - what exactly does this represent. Can't quite get my head around it yet.

Presumably Mach4 sends out the pulses to match no. of encoder steps to reach a new position from the current position? But what influences the rate the pulses are sent; is it as fast as possible (and the controller has full velocity control) or does Mach4 control the velocity by the step rate? How does Mach4 decide when to send the next position change; does it depend on the Controller being faster and keeping up using its own velocity profiles to judge when to send the next position? Does Mach4 have some kind of delay to ensure the machine has completed its move or is it the relationship between the Mach4 (slower?) and Controller (faster?) velocity profiles that ensures steps are completed in a timely manner. The Controller profiles are presumably tuned for best machine response. My 'confusion' arises because the encoder feedback goes to the Controller (great for PID control) but not Mach4 which provides the step/position control. Is there a 'danger' of a stop/start path movement if as Mach4 does not know the actual position it must play 'safe'?

Or am I a total idiot and misunderstood!

Thanks in advance for your intelligent replies.

Dick

The axis motor velocity is controlled by the step pulse rate created by Mach4 which is determined by the feed rate and the vector of the movement.
The actual positioning of the motor shaft is done by the controller and the step and direction software is blind to the actual position.
The PID algorithm is in the Controller, it cannot take a break.

Hope that helps

Super Tech & Associates Main Web Home Page

Hi Dennis,

Thanks.

That makes sense. Mach4 must control the feed rate (dependent on tool and material) and the Controller must have the encoder data for the PID.

It just seems a little wrong that the closed loop is only to the Controller and Mach4 doesn't know the actual position and the Controller doesn't know the final step position. No doubt it all works; but it seems a shame that each part of the system may not be optimal. I suppose it makes the system parts more 'independent' so each can 'focus' on its job. Set up is important though to ensure the two parts do not get out of sync - e.g.. Mach4 starts changing to the next position before the prior position has finished or Mach4 has to stop at each position for the machine to catch up! I would have thought an additional encoder feed to Mach4 would actually make the user side simpler as Mach4 could make decisions based on actual position.

Dick

Servo motors are always playing catchup or wait it's ahead.
The P is for proportionally how far off it is.
The I is a sum of the P's over Time.
The D is a reverse of P over time hence it dampens quick moves.

I do not know actually how Mach4 works. I made my own Host App and USB servo controller, ergo SuperCamXp.

I found it is critical to have Error by being off the intended position. 256 is a natural number so I went with that.

I also found it critical to have a Over Current Error, motors nudged up to an obstruction can over amp if it can't get to position.

With stepper motors you can fly blind. Servo motors have a blind race car driver focused on the speedometer.

Super Tech & Associates Main Web Home Page

Hi Dennis,

.. and thanks again.

Thanks for the overcurrent tip. Your website looks interesting - you have put a lot of work in!

Does anyone out there know of 'hobby' software, like Mach4, that uses encoder feedback to fly less blind?

Dick