Messages

3161

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 27, 2019, 01:17:53 AM »

Hi,
where about are you setting your Homing parameters? From your description:

also i have changed my homing to 40% but there is next to no movement still, any ideas?

You are using Configure/Control (Mach)/Homing-SoftLimits tab. That is not correct, at least for the approach and backoff
speeds. They are set in the ESS plugin. It is a bit confusing but some of the Mach settings are ignored in preference to
the ESS plugin. Speeds and index homing certainly. Softlimits, Home in Place and Home Order are all still set in the Mach
plugin, not the ESS plugin. You will have to experiment some to determine which settings are applied.

Note that the ESS has two speeds, the approach speed, which exactly as described is the speed at which it approaches the
home switch. It can afford to be reasonably fast, provided it doesn't hammer the switch with overrun. The back off speed
can be much slower and therefore more accurate. When the axis backs off the switch, that is the moment that the machine
coordinates are reset. Accordingly the accuracy is determined principally by the back off speed.

Craig

3162

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 27, 2019, 01:02:55 AM »

Hi,
acceleration of 3500 is pretty impressive. When machining parts the accuracy of the corners is very much determined
by the acceleration the machine can attain. For this reason I recommend the max acceleration your motors and machine can handle.

I would push it to 5000, then 7500....until you find the limit. It may well be that your machine starts careening around the workshop
and that will constitute the effective max acceleration.

If it transpires, and from your description it sounds probable that the mechanics of the machine will limit the acceleration
but that will allow you to bump up the max velocity.

The very real fact is that steppers lose torque at high speeds, closed loop steppers are no exception. It is normal for instance that
the low speed torque, or holding torque, is reduced to half at 500 rpm and reduced to  a quarter at 1000 rpm. There again
being closed loop does not change those characteristics. My suggestion was to use 500 rpm as a starting point for your
max velocity. Now that you have really upped the acceleration and are probably approaching the effective max acceleration
now you can start turning up the speed.

I would try 10000 mm/min, that is equivalent to the steppers doing 2000 rpm. Particularly if you have to limit the acceleration
to prevent the machine from rocking around you may find 10000mm/min doable. That would be a very impressive result  for
steppers.

Craig

3163

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 26, 2019, 10:16:22 PM »

Hi,

also just a quick question when i press "ref all" all 3 axis barley move to the home switches any idea what i might have done

That has to do with homing also call referencing and it is adjustable. Don't bother with it until you have the motors
dialed in. If you cannot move a precise distance then you are going nowhere.

Try this MDI:

Code: [Select]

g0 x100
x0
x100
x0
x100
x0
x100
x0
x100
x0

Does it move backwards and forwards, 100 mm per move? Does it move flat stick?
If so now try the Y axis:

Code: [Select]

g0 y100
y0
y100
y0
y100
y0
y100
y0
y100
y0
y100
y0

If the Y axis is moving backwards and forwards OK then try some linear interpolated moves:

Code: [Select]

g0 x100y100
x0y0
x100y100
x0y0
x100y100
x0y0
x100y100
x0y0

If all these work OK then you can start thinking about tuning your motors. By tuning I do not mean fiddling with the
Steps per Unit. Those settings are now made and unless you change motor settings or ballscrews you shouldn't ever have
to touch them again.

I would concentrate on acceleration first. Acceleration (linear) is equivalent to torque (rotation) so to explore the
torque capacity of your motors at various speeds is where you want to go. You might achieve a high speed say, but have
to dial back on the acceleration in order to avoid losing steps. I think that is a mistake, increase acceleration as much
as you can and THEN increase max velocity until the onset of losing steps.

Try quadrupling your acceleration to 1000 mm/sec². Run the same test as above. How did it go?
If it went well double the acceleration again, now up to 2000mm/sec². It is now very likely your machine
is flopping around like a fish. Industrial production machines have accelerations of 1g-5g that is 10,000mm/sec²
up to 50,000mm/sec². Hobby machines cant usually get within a bulls roar of that and they try to throw themselves
around the room if you try.

What you are trying to achieve is the highest acceleration your motors can deliver consistent with the machine not flexing
alarmingly AND your heart rate  stay below 200BPM! You might chose to back off from max by 10-20% and allow some headroom.

Now you want to start increasing the maximum velocity until it starts losing steps. Given that your steppers are closed loop
the drive will try (try being the operative word) to catch up on lost steps. Usually what happens if it starts losing steps the drive
faults 'following error'. You know then that you hit the limit. Back off 25% and run your tests again. What you are trying to
achieve is the maximum RELIABLE speed with the steppers having still plenty of torque to accommodate both  good acceleration
and cutting forces.

A thorough investigation of your motors and machine capabilities will probably take a couple of days. Try to be methodical about it.
Often, once you tune your machine, those settings will stay the same over the life of the machine. It really does pay to explore
the limits. So many machines around the globe are actually way more capable than the owner realizes and he will merrily be
using but a fraction of the machines capability because of his inexperience/impatience/disorganization at the time he set those
limits.

Craig

3164

Mach4 General Discussion / Re: Finial switching over to Mach4 amd Ihave a few question.

« on: January 26, 2019, 09:11:14 PM »

Hi,

https://warp9td.com/index.php/kunena/7-general-discussion/7748-gantry-squaring-slave-sensor-not-registering#19732

Craig

3165

Mach4 General Discussion / Re: Finial switching over to Mach4 amd Ihave a few question.

« on: January 26, 2019, 08:59:58 PM »

Hi,
I don't have slaved axes so I can't answer definitively.

Slave motors are assigned to each axis on Configure/Control/Axis Mapping.

Each assigned slave motor will require its own Step and Direction pins and if you want to square the gantry each slave
will require its own home switch to complement the master home switch.

The gantry squaring routine is carried out by the ESS not Mach. Further details on how it works are probably best found
on the Warp9 site.

Craig

3166

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 26, 2019, 08:45:28 PM »

Hi Dale,

ran g0 x100 again and got 93mm am heading in the right direction

Not quite sure that is the case. In the previous instances it was a setting that caused a predictable variation between
actual and what you wanted. 93mm vs 100mm we actually expected is much harder to understand.

You can correct it by adjusting the step per unit until it comes right. It is not correct however. The calculation is so
straight forward having to deviate from it means that a fault elsewhere is being overlooked.

One possibility is that the stepper is losing steps at speed. If you try MDI g1 x100 f250 what happens?. The axis will
move but more slowly. If the discrepancy is caused by trying to go too fast then this test should highlight it.

Another possibility is that you have a loose coupler, its surprising how many get caught with such a simple fault.

Craig

3167

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 26, 2019, 07:33:49 PM »

Hi,
I still think you have the switches wrong....its easily done.

There is one other remote possibility and that is the difference between PPR and counts per rev.

Encoders have a certain number of lines which is the same as PPR. In the case of your steppers they have 1000
line encoders. When the two encoder channels are combined, they are in quadrature to each other, you end up getting
four times the number of distinct counts per revolution, ie 4000 counts.

I would expect, and am led to believe, the dip switch settings are counts per rev, not lines per rev. I am suspicious that
the actual travel is exactly 1/4 of what was expected.

Please go and double and triple check the dip switches. If they are indeed set to 2000 pulse per rev as you believe
then set your 'Steps per Unit' to 1600 and try the same test. The ESS can handle the extra pulse speeds no trouble.

Craig

3168

Mach4 General Discussion / Re: mcCntlGcodeExecute best practice

« on: January 26, 2019, 07:03:05 PM »

Hi,
re-reading your post a little more closely and I'm thinking my previous rely is 'out to lunch'.

If I understand what you are asking is whether you can test a signal or other input WHILE a move is in progress?

If that is the case the GcodeExexcuteWait() will not work because the Gcode interpreter is hogging the CPU and there is
no way for Machs Core to do any other work.

If however you use GcodeExecute() then the function will return once it has issued instructions to the motion controller.
Thus Machs Core could presumably do something in the mean time. I'm not quite sure how it would work.
Maybe:
rc = mc.mcSignalWait(
      number mInst,
      number sigId,
      number waitMode,
      number timeoutSecs);

would work.

Craig

3169

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 26, 2019, 06:46:42 PM »

Hi,
sorry, in that last pic I highlighted the settings to make 5000 pulse/rev, please disregard my highlighting.

Craig

3170

Mach4 General Discussion / Re: Motor Tuning in Mach 4, Very Stuck

« on: January 26, 2019, 06:44:07 PM »

Hi Dale,
then something is wrong. It could be that the ESS is issuing 2000 pulses per rev for 20 revs to get your 100mm
travel but the stepper is missing steps. Its possible but unlikely, I would expect there to be unusual noises.

I think it more likely that the settings in the stepper driver are wrong.

I'm guessing that you have the dip switches back to front and you have accidentally set the steps per rev to 3200
not 2000 as you intended.

Now you issue an g1 x100 instruction so Mach/ESS dutifully sends 100 (mm) x 400 (pulse/mm) =40,000 pulses.

If I am correct your stepper driver took those 40,000 pulses and applied them to your motor but at 3200 per rev.
So your motor would turn 40,000 /3200 = 12.5 revolutions and at 5mm per rev =12.5 5 =62.5mm.
Does that sound correct to you?

Have a close look at the bottom of page 14. Compare with the settings required to make 2000 pulse/rev on page 15.

Craig