Strange Happenings With 30mm/min Feedrate

hpdrifter · Post by **hpdrifter** » Tue 01 May , 2007 18:01 pm

Hi All

Has anyone observed this before??

I have been machining some metal components at a feedrate of 25mm/min successfully and decided to change this to 30mm/min to speed up the process.
I noticed something wrong and observed what appeared to be loss of datum in the x plane worsening as I the process progressed. I suspected stalling of the stepper motor although the forces were small.
I then tried the operation as an empty cut (ie no forces) and the same thing happens.
I then machined an item in plastic at 100mm/min no problem.
I suspect that some sort of resonance at 30mm/min is making the motor jump??
Any coments would be welcome.

Martin · Post by **Martin** » Tue 01 May , 2007 23:29 pm

You have not said what machine you are using.
& what version of software you are using.

Post by **Denford Admin** » Wed 02 May , 2007 9:02 am

I suspect that some sort of resonance at 30mm/min is making the motor jump

I would agree - I seen very strange things on stepper motors in the past, and its just down to the physics of the motors.
I've seen a motor change direction randomly when we drove it with a larger current than it was rated for.

Maybe a microstepping drive would solve your problem - upgrading could be very easy - depending on what machine it is ?

hpdrifter · Post by **hpdrifter** » Wed 02 May , 2007 10:39 am

I forgot to say in the excitement the machine is a TRIAC and the software is Denford Triac V2.4 dnc.

Do you have any details on the upgrade??

Thanks

Post by **Denford Admin** » Wed 02 May , 2007 10:53 am

Some info on the upgrades is here:

viewtopic.php?t=116

Is it PNC controller ? If it is - we can't offer an upgrade

I don't think you would ever get a PNC to output enough steps for microstepping, even if you improved the drives.

Its still possible that new stepper motors, and possibly new drives would help with the step "jumping".
Are there any current settings on the drives at the moment ? it may be worth playing with those to see if the problem gets better / worse

hpdrifter · Post by **hpdrifter** » Wed 02 May , 2007 12:33 pm

Yes, it is a PNC controller.

This machine has always done everything I want with no problems.

I think the obvious thing to do is avoid feeds between 27 and 33 mm/min.

hpdrifter · Post by **hpdrifter** » Thu 03 May , 2007 18:41 pm

Am I the only one that has had this problem then??

Post by **Denford Admin** » Fri 04 May , 2007 8:39 am

Am I the only one that has had this problem then??

I've not heard of it being a problem recently - I'm guessing not many people use that feedrate often.

I'd guess that different (newer) stepper motors driven at different currents would either be better or resonate at different feedrates

Steve · Post by **Steve** » Wed 23 May , 2007 12:13 pm

Stepper motors do have resonance bands at some frequencys that result in variations in the torque available. This would cause a problem cutting an arc as the each axis would have to pass through that frequency at some point.

The Resonance bands are minimised by the type of motor connection (parallel or Series) and by selecting full or half step.

These you have no control over.

To resolve resonance affecting the torque curve of a motor the only other thing that can be done is to damp the motor. (some motors are fitted with counterweights.)

In your case I suspect the problem will be caused by the slideway being too slack and as a result the motor is running without a load to damp the resonance.

Can you try adjusting (tighten) the Gibb strip to increase the damping of the axis?

comet · Post by **comet** » Wed 23 May , 2007 23:15 pm

Hi,
its strange how there is never a resonance problem with printers or
plotters ???
comet

Steve · Post by **Steve** » Thu 24 May , 2007 7:56 am

Printers and plotters tend to run at fixed speeds per axis so the frequency issues can be avoided. also due to the low load the stepper motors have a lot lower current and size than those required to move a machine.

davidimurray · Post by **davidimurray** » Thu 24 May , 2007 9:10 am

Have a look at this -

https://www.engr.colostate.edu/~dga/vide ... ration.wmv

Cheers

Dave

Steve · Post by **Steve** » Thu 24 May , 2007 9:21 am

Thats a good demo!

With a machine tool we have to run the steppers throughout the speed range at all frequencies.

The loading on the motor damps out the resonance to a point where no steps are lost (hopefully).

If stalling occurs it can be because there is too much load or too little.

comet · Post by **comet** » Sun 27 May , 2007 19:51 pm

yes, I think this Video gives you all the reason you need to use gecko g340's and servo motors,I wouldnt entertain using steppers on a machine build/conversion now-there nothing but trouble. yes servos can be more expensive,but there are numerous ways to reduce this (attach cheap HEDS encoders to any dc motors) and then you have a proper CNC machine!

Steve · Post by **Steve** » Tue 29 May , 2007 9:23 am

While steppers can have problems with resonance and if they stall you loose position its not all bad!

If the stepper is designed correctly and running within its step frequency the machines are extreamly acurate and reliable.

A stepper motor will be in position to within 0.8% of a step all the time. With 400 half steps/ Rev this is generally more acurate than a servo which is constantly jumping about trying to hold position.

comet · Post by **comet** » Tue 29 May , 2007 16:02 pm

Steve,
my encoders have 2000 counts(with quadrilature) I have never noticed them jumping about,even using the 5xmultipler (400 counts) dithering is not noticeable.
I am sure steppers are fine for training machines or positioning non loaded objects.
The differance being,I suppose that I earn my living to an extent on my machine.
I now have up to 4000 mm per minute feeds,as aposed to a safe 500mmpm.
Now my machine stops dead if it goes out of position more than .2 mm. The PID motor control will keep on loading up the amps untill its at the correct position,if not it faults.
No more do I waste a vast amount of time tuning it to be reliable,only for it to go out with as much as a change in temperature.
I can also trust my machine,so that I am not constantly listening for the change in tone when they misstep and scrap the job and usually the tool.
I have converted three machines from stepper to servo,and every time it transforms it out of recognition,to as I said a proper CNC.
One other point I have found Mach3 to be the best software availabe for controlling,plus the back up as with Gecko is second to non.
So well worth considering!

Mr Magoo · Post by **Mr Magoo** » Sat 02 Jun , 2007 9:35 am

Mmmm - this opens up a whole new can of worms...

A traditional closed loop servo system relies on servo error (or servo lag) in order to generate the command voltage to drive the servos. The more advanced servo systems can apply a 'feed forward' offset to the drive to try and bring this errro back to zero.

Generally your CNC system will allow you to monitor this servo lag value and so allowing you to adjust the accaleration, and the position and velocity loop gains to optimise this. A poorly setup system risks overshoot at corners (eg. gouging on internal rectangular pockets) and excess servo lag (eg. mis-shaped circles and arcs)

A system setup to have a permissable servo error of 0.2mm could mean overshoots and circles having upto this amount of error before being detected if not setup correctly.

Regarding servo drives that take stepper signals as input and close the position loop within the drive, the CNC Control doesn't have any feedback as to where the axes really are in these systems and so position accuracy is only as good (tight) as the position loop is in the drive.
For example: a servo lag of 0.1mm at 1000mm/min could give an error in a complete circles diameter of upto 0.2mm with this system

Steppers have none of these problems.

I'm not saying one system is better than another, just pointing out the importance of choosing and setting up the correct servo system if you 're thinking of switcing from steppers.

comet · Post by **comet** » Sun 24 Jun , 2007 22:34 pm

"Permisable error" is misleading,the servo drives will push to try and get to a zero error. if not then you have a problem.
as far as "overrun is concerned I have always found that the problem is more likley to be the axis slowing down or braking under friction and having the ramp up/down set to long.
In real life a simple servo system is far superior in every respect to steppers,apart from cost,the difference being the price of aquiring and fitting the encoders.There are plenty of cheap brushed dc motors and drives about.
I cant think of any proper CNC machines that use steppers.they are just not a viable option in industry.My guess that there new found popularity is
down to the availability of cheap chinese motors and drives along with
machines to retro fit them to.Plus of course the excelent Mach3 controller.
In short if your thinking of converting to servo's you will not regret it.

Denford Software & Machines

Denford Software & Machines

Strange Happenings With 30mm/min Feedrate

Strange Happenings With 30mm/min Feedrate