Another dumb question

[Hood]

Ok will probably be tomorrow before I look as its 11.30pm here and I want an early night
I have often posted about Debounce Interval etc but I post quite a lot so dont remember specifics.
Hood

[dresda]

The first part of the video am am moving in MDI at 40inch/min Y-. Then I am coming back at 40inch/min in jog, you can see the difference.
Ray.

http://www.youtube.com/watch?v=TiuxWUTzrmg

[Hood]

I would say it is slower than you would expect with the parallel port so its likely due to the DSPMC and I think what Rufi said explains that, however it  will never be as instant as the jog is.
Hood

[dresda]

This first unit is for my machine, then I would like to build some for my customers but the first thing they are going to say when the see that is, What the F@4C& is that doing. I have a Milltronics with a 386SX and motion board that stops in feed hold on a dime.
Hurco,Autocon, Cinci, Centurian, all PC based with motion boards, and they stop instantly in feed hold, we are only talking 40inch/min here. If I was going to loose postion, I would loose it in jog at 120inch/min right.
My experience with PC integration sucks, but I have been around CNC machine tools for over 30 years.
Hood you have experience with steppers, I don't. Correct me If I am wrong, the reason people go with servos is because you don't loose a position.
I have enjoyed this build and yaking with you guys, but this is a major setback. It is like I am  setting up 2 systems and they are not talking to each other, I hope it's just a parameter.
Ray

[Hood]

I probably have more experience with servos than steppers. I would say the main reasons I went with servos is Constant Torque, Velocity, Acceleration and Closed Loop to motors.
I am not an expert on exactly how Mach does things but as Mach is a controller working under windows it is a buffered system. Jogging does not have a planned path, Mach just sends out pulses and when the key is released Mach slows the pulses at the acceleration in motor tuning. Mach obviously keeps track of the pulses it sends out so can update the position. With feedhold however the toolpath has already been planned and buffered so when you press feedhold you are interrupting the planned toolpath so I would think Mach has to interrupt whats in the buffer, recalculate and refill the buffer.
As said I am no expert but that is just my interpretation of what is happening.
I still think the DSPMC is slowing things down but as said before Feedhold will not be as fast as stopping a jog, so if the feedhold is too slow then likely you will have to either try EMC or go with Milltronics or similar.

Hood

[dresda]

I tried the stop program function when I was running the program at 40in/min and It stopped great, but wouldn't start again, no errors.
Ray.

Quote from RAY.L

With servos, this should not be a problem, unless you're moving fast enough that when the step pulses stop the axis coasts far enough, due to inertia, to trigger a servo fault due to exceeding the maximum allowable following error.  Barring that, it should overshoot, then quickly return to the position Mach3 thinks it's in.  On my servo-driven knee mill, I can usually hit Stop without losing position, even when doing a rapid.  The same is NOT true of steppers, however.

Regards,
Ray L.