[storm2313]

This defines the power of Pumotix motion control software

[ZASto]

Just curious: What are you going to machine at 585 in/min (14859 mm/min)?

One thing is spinning motors on desk, another pair of sleeves is having motors movin iron

Don't get me wrong, I am running Mach3, and PUMOTIX (PLCM-E3 + PLCM-B1) as well as GRBL on Arduino Nano.

[dazp1976]

Anything is faster on a decent pc especially with SSD.
Mach3 is generally pport or usb used so no surprises there.

If talking machine/motor speeds then, ALL programs can be sped up easily (desn't mean you'll actually cut at 'silly' speeds).

[shineworld]

Multi-axis speed depends by many factors.
Remove hypotetical any difference in how different CNC creates motor commands. All works same for following notes.

The things change a lot if CNC support RTCP or not RTCP.

Without RTCP
There are CNC which manage rotational F as you declared in G-Code, without any elaboration.
You put G01 B30 F100 and indipendently where the TCP (tool center point) is the B axis rotate at 100mm/min.
There are CNC which manage smartly F declared in G-Code depending by distance of B axis rotational center.
In this case is a smart simulation of what speed is necessary to tool TCP when the tool si far from rotational point.
More is far and more F slow down to compensate increasing radius distance.
This solution is better than first, the CNC run slowly than first but in some mode the working tool point uses the declared F to remove chips.
Unfortunately this solution can define a single F for a single LINE but interpolation works between previous line to current line with an infinite (almost) serie of positions points, so you can a more better than first case result but not satisfying.
This way give very poor info to Look-Ahead to keep an harmonious TCP speed. Is just a workaround used by more smart CNC.

With RTCP
The more valid way is to have a true RTCP multi-axis CNC.
In this case every interpolation sub-point, usually with timing of 1ms or inferior time, which exists between two or more G-Code lines enters in interpolation feed planning, look-ahead management, and you HAVE ALWAYS a true
TCP programmed Feed in tool center point, indipendently how far you are on rotational axes fulcrums.

This is the true kindom where to compare CNC, because they can differ on smooth, look-ahead, and smart algoritms to keep F fluent, continous and gain a well finished product.

I've tested many CNC boards, from cheap to very expensive, multi-axis behavior, with RTCP or without RTCP.
Without RTCP is always how to count rings after a rock fall in the lake..
Finishing without RTCP is terribly slow and with roughly result, but overall a very stress for tools and mechanincs.
With RTCP you get a fluent work, without stress, overall if CNC have a good Jerk Control engine.

[storm2313]

I believe in displaying real world results as an engineer, and the videos display this live below.

Theories are only relevant on paper, and not valid with robotics until proven in real time.

Here's two more videos reflecting facts Pumotix is faster.

Here's a 6040 CNC that went from Mach3 to Pumotix, and the end user experienced roughly a 75% increase in speed covered in this video.
https://youtu.be/Ee01X_kTzDY

Here's a side by side comparison of Mach4 getting beaten by Pumotix in the Roadrunner G-code from Artsoft.
https://youtu.be/vezyDfKlB04

[shineworld]

Good test !

Using another CNC board and software with same axis parameters I've got 4'18"

XYZ Axis parameters:
Max Velocity: 147.638 in/min
Acceleration: 25 in/s2
Roadrunner G-Code from https://raw.githubusercontent.com/sy...adrunner.gcode

Much depends by CNC internal algorithms