Messages

2521

General Mach Discussion / Re: Motors stalling /growling

« on: May 06, 2019, 03:38:37 AM »

Hi,

I can get the motors to run smooth when I jog, but when running the demo in mach 3 they sound terrible.

First lets dispel a myth, Mach3 Demo and Licensed Mach3 run identically. The only limitation that need concern you is that Demo
is limited to 500 lines of Gcode at a time. The kernel frequency of Mach3 Demo is limited to 25Hz, but this need not concern you,
unless, as previously explained, you choose to set a too higher microstepping regime; if you keep to 8 microsteps per full step
or below the limited kernel speed will have no effect.

The fact that the motors run smoothly when jogging tells us that the steppers and drivers are good, it also tells us that the BoB
and Machs parallel port can produce the required high quality pulse stream. The fact that the same steppers/drivers/BoB run poorly
when processing code suggests to me that Mach's parallel port is running poorly when the Gcode interpreter is commanding
motion.

My guess is that when the Gcode interpreter is running it introduces a heavy demand on the PC's interrupt system which in
turn causes excessive jitter within Mach's parallel port driver, itself an interrupt driven timer.

Proving that this is so is not straight forward.

Would you please post a screen shot of DriverTest.exe. Its not conclusive but is still a fair indication as to the health or
otherwise of the parallel port driver. Note that some PC's seem to do a good job with a parallel port and another machine,
possibly with much higher specification, does poorly. There is no way to tell short of installing Mach and trying it.

Have you considered an external motion controller such as an Ethernet SmoothStepper or a UC300? One of the
advantages that an external motion controller offers (vs the parallel port) is very high quality, high speed pulse
streams WITHOUT being dependent on PC hardware. Should you consider this option you should budget somewhere
in the $150-$300 range, if you buy cheap Chinese rubbish you'll have even less luck with it than the parallel port!!!

Craig

2522

General Mach Discussion / Re: Motors stalling /growling

« on: May 05, 2019, 03:02:09 PM »

Hi,
it is possible to set your microstepping regime higher than 8 microsteps per full step (1600 steps per rev)
but you get little advantage.

It is seductive to believe that you get increased resolution as the number of microsteps go up, unfortunately
that does not happen in practice. In fact no genuine increase in resolution is achieved beyond half stepping (400
steps per rev). If no increase in resolution then why bother? The answer is smoothness of motion....the higher the
microstepping the smoother the motion.

If however you increase the microstepping too high then the required pulse rate to spin your motors will exceed what your
controller is capable of.

For example if you set your microstepping to 64 microsteps per fullstep (12800 steps per rev) and you want a maximum
speed of 600rpm from your stepper the required pulse rate is:

600 / 60 =10 (revs/sec)
Pulse rate =10 x 12800
              =128kHz

Machs parallel port in default is only 25kHz and there are only a few external controllers that can generate pulse stream at that
speed. If however you choose 8 microsteps per fullstep the pulse rate is:
600 / 60 =10 (revs/sec)
Pulse rate = 10 X 1600
               =16kHz

ie well within Machs parallel port even at default speed.

As it turns out the smoothness of an 8 microstep regime is very nearly the same as a 16 or 32 microstep regime
ie there is little to be gained beyond 8 microsteps (1600 steps/rev). As a consequence 4, 5, 8 and 10 microsteps
per fullstep are the norm among Mach CNCers.

Craig

2523

General Mach Discussion / Re: Mill or Turn?

« on: May 05, 2019, 07:21:12 AM »

Hi Dave,

i can fully understand why but Mach4 is still way behind

Have you tried it recently? By my estimate its Mach3 which is behind.

I suspect you would be very disappointed with CSMIO Mach4 plugin though.....its pretty damn buggy and CSLabs have
done nothing about it for months and months.

Craig

2524

General Mach Discussion / Re: Can't Install Mach3 on my PC

« on: May 05, 2019, 01:15:40 AM »

Hi,
Mach3 and Mach4 are available as a free download from the NFS website, you don't need a disc.

https://www.machsupport.com/

What won't be included is a license file for Mach3. It seems unlikely that the disc would include a license file unless it is a
pirate copy, very common from Chinese suppliers.

Mach4 has a different licensing arrangement with each individual PC licensed and that is done online on the NFS website
and could in no case be supplied on a disc.

Craig

2525

Mach4 General Discussion / Re: Step pulse duty cycle

« on: May 04, 2019, 04:15:19 PM »

Hi,
I am of the opinion that closed loop stepper drivers offer too few advantages over open loop drives to justify the extra expense.

It is my opinion that the manufacturers of these things play on the insecurity of customers with phrases like 'never lose steps'.
If an open loop stepper loses steps because of overload (too high accel or too high velocity or too high inertia) then a closed loop
stepper will lose steps as well, and for exactly the same reason. A closed loop stepper is still a stepper, it has the same torque/speed
characteristics as an open loop stepper. Its true that the closed loop stepper will try to recover....but if its overloaded it will
fail. If it gets too far behind its commanded position it will fault 'following error', ie warning you that its lost steps. That is an
advantage over an open loop stepper.

The better path is to select a stepper, it matters not whether its open loop or closed loop, that can do the job required
WITHOUT losing steps, ie operate WITHIN its torque/speed area. This requires a certain minimum holding torque but more
importantly places an upper limit on the inductance of the stepper.

In experienced stepper buyers look at the holding torque only reasoning that the higher the torque the better the motor.
What they don't realize is that a high inductance motor will lose all its torque at speed whereas a low inductance stepper
retains more of its holding torque.

If a manufacturer or supplier cannot provide you with an inductance specification or torque/speed diagram.....walk away,
they are trying to hoodwink gullible buyers.

The second issue is the voltage of the supply/driver driving the stepper, the higher the voltage the faster the stepper will go
without undue loss of torque. Gecko drives are typically capable of 72V-80V and Leadshine AM882' are 80V capable.
Leadshine AM882's are pretty good value. Pair that with a low inductance (2mh or less) stepper and 80V power supply
and that stepper will sing!

The other alternative is genuine servos. Over recent times the price of genuine, good quality AC servos have come down in price.
Have a look at DMM:

http://www.dmm-tech.com/

They are a Canadian company but manufacture in China. Good quality at a very good price. There are cheaper 'no brand'
Chinese servos...don't go there!

Servos still command a premium over steppers (open or closed loop) but they are EXCEPTIONALLY CAPABLE.
Once you have used them you'll never go back to steppers.

Craig

2526

Mach4 General Discussion / Re: Mach4 Timer Script

« on: May 04, 2019, 03:35:21 PM »

Hi Tony,
I think that list of poppabears is out of date.

May I suggest opening the register diagnostic and pinning it to the top.
Observe the range including the timer variables. They are populated and change as time passes. This is the latest build of
Mach4 but these variables have been enacted for several years at least.

Craig

2527

Mach4 General Discussion / Re: Dust hood

« on: May 04, 2019, 05:28:15 AM »

Hi,

If MACH4 wants to be industrial, it cannot just remain at hobbyist quality.

There's no satisfying you....in one post you are complaining that Mach4 is so configurable that is to too complex for
a hobby product...and then in the next you are saying that is survival requires better than hobbyist quality.
I wish you'd take up knitting.

Just I case you didn't notice I just said that you can modify m0, m1 etc to do as you wish and emulate Fanuc or Hass if
you want.

Craig

2528

Mach4 General Discussion / Re: Dust hood

« on: May 04, 2019, 03:40:52 AM »

Hi.

M6, M7, M8 and M9 are are already KNOWN to the "MACH4 GUI" modules. User defined Macros are NOT.

This is a red herring. m6, m7, m8 and m9 are predefined.Please not the DELIBERATE use of lowercase, that's how Machs
Gcode interpreter expects all m codes to be expressed.....use lowercase or you can expect some curly errors.

When the Gcode interpreter encounters a m code call it in the first instance searches for the m code macro in the macros
folder of the current profile. If it does not find a macro that matches the call is searches up the directories until it finds
the required macro.

Lets say you program calls an m8. If you do not have an m8() macro in your macros folder Mach will use the default
NFS supplied m8() macro. If you decide for whatever reason that you want to include some custom behavior in the m8()
macro then you would write a m8() which by necessity duplicates the expected default behavior plus whatever extra functionality
you require.

My understanding is that it was not possible to write an m0(), m1(), and m2() for instance. That is no longer the case,
although why you would want to modify these m codes is beyond me.

The important point is that NO macros, be they built in or user written are co-routines.

Have you investigated the noise issue? 24V does not preclude noise, careful attention to circuit impedance is also required.
Post a circuit of the output of the BoB, including what you know about of the BoB output the output driver and the input
circuit of your dust hood mechanism.

Craig

2529

Mach4 General Discussion / Re: Dust hood

« on: May 03, 2019, 04:56:31 PM »

Hi,
the fact that the line of wxMessageBox code does not occur suggests that it is not a software error.

If the dust hood is triggering it is doing so without being instructed to it suggests EMI noise.
Even 24V signaling is subject to noise if the circuit impedance is high.

If for instance the on state current is 100 uA at 24 V for a power of 2.4mW and impedance of 240kOhm. Therefore a noise impulse
averaging 2.4mW could falsely trigger the circuit.

If the on state circuit current is 10mA at 24V for a power of 240mW and impedance of 2.4kOhm then a very much
larger noise impulse, 240mW, is required to trigger the circuit.

A higher signaling voltage does not in itself ensure noise free communication.

Craig

2530

Mach4 General Discussion / Re: No GCode Movement after pressing stop

« on: May 03, 2019, 04:49:48 AM »

Hi,
not sure why it is that m6 appears to throw a spanner in the works. I'm guessing it that is because he
CycleStop() function does not execute gracefully when its in a macro, in this case m6.

As I said at the beginning of this thread <Stop> is poor operating practice, the correct procedure is <Feedhold> followed
by <Stop>.

Try <Feedhold> then <Stop> while an m6 is executing and see if the behavior differs.

Craig