Hi,

Mach4 threading support is complete as described in the manual, BUT it is dependent
on the motion controller to implement threading features.

As of 22 January 2018, and as best I know, threading is fully implemented by
Hycon, PMDX SmartBOBs, and PoKeys.

That information is very valueable! Thank you!

When you read propaganda from hardware builder, than it looks like everyone has best product on market. So its very hard for those that don't know that hardware, find out differences. And if you ask at forum, you have to separate answers from fanboys and wise guys.
Would be great, if artsoft would publish a table, that compares hardware support for special mach4 themes like threading. Support for closed loop and pid parameters could be another question. Or hardware support for feed override during motion ...

I guess, such comparision could speed up competition of hardware builders

Unless the controller can report back to Mach and Mach can accommodate moves that don't go where they are supposed to then Mach has no chance of controlling the machine. That requires an out of band axis which can be altered WITHOUT affecting a coordinated move of the remaining axes.

Sure, out of band axis is one requirement. But if mach4 should recalculate speed of turning axis, it need to process motion feedback of an axis (during motion). At least, if speed depends on position.
So the position has to be sampled (if using an encoder) and transmitted quite fast.
If you look at the 40Hz of an ESS: an axis moved at 12m/min moves 5mm between to communication triggers, so no chance to calculate any finegrained speed changes.

Reinhard

Hi Reinhard,

But if mach4 should recalculate speed of turning axis, it need to process motion feedback of an axis (during motion). At least, if speed depends on position.

I think the situation is even more complex than that. If you wish to alter one axis, in this case we are talking about the rotational
speed of the A axis, but any axis that is one of the coordinated axes, it would be necessary for Mach to abort the remainder of the
move, hoping against hope that the sudden deceleration doesn't screw up your homing, read the actual location of the controlled
point and then recalculate the move for all axes including the variation of the one axis that you wished to alter.

I suppose its not impossible to imagine a control system that would do so but that is far away from where Mach4 is. Once the
position/velocity/time data is dispatched to the controller it can't be changed by Mach. The controller can but as you know the
circumstances where the controller effectively 'overwrites' the PVT table are to support specific functions, eg THC, homing etc.
No controller that I am aware of can alter a coordinated axis velocity during a coordinated move. Maybe PoKeys or Galil with
their internal programming could be coaxed into doing so but not otherwise.

Craig

Sorry guy's,

that i have to disturb.

you are talking here about 200-300-400$ Software, not about "real" (sorry to use this word),
CNC controllers (10000-20000$ aeria).

Tom

PS that is all "nice to have" option

Tom,
Not disturbing at all and only wish more  of the 25000+ members would chime in......
The real ones are expensive, but, frankly no one from a hobby point of view is going to shell out that kind of money. SO  will just echo the following:

 Mach 4 hobby works for every command that is in the current manuals and part two is the motion controller
can provide for the instructions given it by Mach4 for every command. One without the other is useless.
Maybe a better spec would be specifically what  a motion controller does not currently do relative to Mach 4 software.

Steve
Thanks for the info about which external motion controllers provide  for threading.
If a external controller  does not work "yet" then it is of no use use. Promises don't count any more!

RICH

If a external controller  does not work "yet" then it is of no use use. Promises don't count any more!

RICH

Hi Rich,
Well said. By this time if the controller can't do what Mach4 can do then Houston, we have a problem!

Hi,
all of the readily available controllers do all the basic stuff.

There are a number of functions which require specific support/behaviour from the controller to be enacted.
The simplest is homing, but literally if a controller can't support homing I can't see it being on the market either.

The features which determine whether a controller is mature or still under development are probing, THC, backlash
compensation and lathe threading.

All Mach4 ready boards that I'm aware of support probing, so much so that it should be considered entry level.

The rest are all over the place. Vital Sytems Hicon board has always enjoyed a reputation for leading the development
curve. Given the cost of a Hicon board perhaps it should come as no surprise, Vital Systems probably enjoy a development
budget greater than all the others.

Amongst the 'value' manufacturers PMDX and PoKeys seem to be offering a greater range of features than the ESS say,
and the CNCDrive products (UC100, UC300, UC400) are only so recently released with a Mach4 plugin it comes as no
surprise that the features offered lag behind so of the more experienced Mach4 manufacturers.

When external motion controllers were first contemplated and then manufactured to suit Mach3 there was a wide spread
in what was offed as extra features and quite some years passed before all manufacturers implemented all of them.

That situation is repeating itself with Mach4. One comment in this thread was that NFS should maintain a list of compatible
boards and features offered. It would be nice but NFS is absolutely not the right organisation to do it...they're too close
to the manufactures concerned and any suggestion of bias or tardiness of behalf of NFS to update it list could result in a nasty
stouch behind the scenes.

A suitably respected and/or qualified individual could maintain such a list but it would rely on the support of the manufacturers
to provide the information. I could well imagine that some or even all manufacturers might be very reluctant to pass information
to another person whom would have a significant role in wether the market took up their products or not.

In some respects I think the manufacturers should be responsible. For instance Steve has chimed in that his PMDX smartBoBs
support threading and encoder spindle control. Why the hell should I be unsure, I mean I've been to the PMDX website on
many occasions yet had not gotten that message. SHOUT OUT THE GOOD STUFF Steve!

A comment earlier in this thread suggested that the third party hardware supplier model has made uptake of Mach 3 or 4
somewhat more difficult than it might have been. That overlooks the advantages to customers re competition for those
same products. Notwithstanding the confusion and varied approach to extras that comes with multiple manufacturers I
feel overall that I, as a customer, am well supported.

Craig

Hi,

No controller that I am aware of can alter a coordinated axis velocity during a coordinated move.

Really?
For me its such a basic requirement, that I can't think how to live without it.

I worked on industrieal cncs and there feed-override is vital. Every cnc-job will be tested by using feed-override.
Well, I build a hobby cnc, but I'm used to professional work and I'd like to do so with my hobby cnc too :O

you are talking here about 200-300-400$ Software, not about "real" (sorry to use this word),
CNC controllers (10000-20000$ aeria).

That's right, but I think, mach4 is not the problem

The features which determine whether a controller is mature or still under development are probing, THC, backlash
compensation and lathe threading.

Well, probing is so essential, that I think, a controller that can't do probing or which causes problems at probing should be retired from market and blamed any time you write a new message.
In the feature-list I miss the feed override by hardware during movement!
I heard about hobby controllers, that support it, but it was not a controller for mach4

THC and lathe threading might be the same software problem - but as already said, I'm convinced, that its not a problem of mach4 but of the controllers.

backlash is separate goal.
From my point of view, there's a path from entry-level-cncs (driven by simple steppers) until professional cncs driven by what ever motors controlled by closed loop with absolut positioning system (which for me is an industrial issue).
For me, the path looks like

• simple steppers
• steppers with closed loop drivers
• encoders moved from motor to ballscrew - closed loop by application (mach4)
• absolute encoders and closed loop by application

I start with level 2 (closed loop stepper drivers) on my cnc and I'm interested in level 3 - which (for me) is the answer to backlash and the like.

But to enable these functions, you need a different interface between mach4 and motion controller. For me, the refresh rate between mach4 and motion controller should be at least double the supported steprate. As most hobby-level stepper drivers support a max. steprate of 200kHz - so 400kHz sounds like a reasonable limit for hobbylevel. Industrial solutions might require refreshrates of about 1MHz or more, but that could be a feature of industrial mach4.

I don't know the amount of data, that has to be exchanged between mach4 and a motion controller, when mach4 controlls the closed loop, but I think, 100MBit Ethernet might not be fast enuf. So may be, those controllers need to be attached to workstations internal bus (like mesa 6i24 anything I/O)...

Reinhard
[/list]

I'm interested in level 3 - which (for me) is the answer to backlash and the like.

I can't imagine any controller at any level being the answer to backlash. Backlash is simply a mechanical problem. No controller can truly fix anything it doesn't have control over (which is the definition of backlash).

Just my 2 cents.

No controller can truly fix anything it doesn't have control over (which is the definition of backlash).

Agree. So looks like I didn't understand the meaning of backlash

Reinhard

Hi Reinhard,
if you attempt to change the pulse rate of one axis which is part of a coordinated move the eventual endpoint will no longer be the point which the trajectory planner
calculated.

All the other axes will reach their desired end points but the axis you altered mid flight will not. How does your software now plan the next move?
Feedrate on the other hand can be changed because ALL axes involved in the move vary and by the same proportion, so the move may be slower but at any
given point along the trajectory all axes are coordinated.

Craig