Backlash compensation

[sean89]

I have noticed a issue when backlash is used 0.01mm I get stagged holes

I machined a 14mm Dia hole in 1mm step downs to 10mm deep and by the end it s about 1.5mm wider then it should be and it only staggers in the x- direction. And then the set datam point has moved about 1.5mm in the x axis

if I turn backlash off it s fine just goes straight down and can repeat over and over again and perfect then turn backlash back on and problem is back

I originally thought there was a shaft coupler slipping But there all perfect tight

any ideas

 

[breezy]

Sean,

Could you supply a copy of the program and machine configuration settings, so that MASSO Support can test this.

Regards,

Arie.

 

[sean89]

Hi Arie,

I have attached the files i used.

 

[masso-support]

Hi Sean,

I have tried loading your settings file for testing but I am having trouble with it. Either it is corrupt or I am not loading it into the right software version

Can you please confirm the following.

Are running the Masso G2 or G3
Are you running 3, 4 or 5 Axis Mill software
What version software do you have loaded on your Masso.

Cheers Peter

 

[sean89]

It s G3 5 axis version

 

[masso-support]

Thanks Sean,

You must be running old software as the settings file for a G3 is no longer uses that format.
You should be running version 3.47 that was emailed out on the 25th of June.
What version are you running?

Cheers Peter

 

[sean89]

Hi Peter,

Sorry i attached the wrong file

find the new one attached

 

[masso-support]

Hi Sean

Thanks for the new file. I loaded it and had a look.
I was not able to use your settings as the high pulse per revolution is beyond my machines capability. I imagine you must be using servos.

I ran your file twice. The first time with 0mm backlash and the second time with the 0.01mm backlash you are using on the x axis.
I am not getting any staggered moves in the X axis as you describe.
I measured each hole in both the X & Y axis directions and they all measure the same dimension.

Backlash is also non cumulative so should not cause such a large X axis shift. It sounds like you might be missing steps.
You could try reducing the acceleration on the X axis and see if that makes any difference.

Cheers
Peter

 

[sean89]

Hi Peter,

thanks for taking a look.

I cant imagine im loosing steps because yes i am running Servo motors and there Delta Brand
I run the machine with it off and there is no problem

i dont think turning the acceleration down should make a difference as Y axis is not effected
i did however slow the feed rate down on a test to about half and still had the same result

It really does seem like its accumulating

So i did a little experiment, at the masso controller i switched the X and Y outputs plugs so now Y axis is driven by the X axis output of the masso
and the problem followed there so now the stepping is in the Y axis on a completely different motor on another ballscrew so I can confirm
there nothing mechanically wrong and the motor is fine. There is a problem in the masso

 

[masso-support]

Hi Sean,

Thanks for the additional testing.
I will do some more testing and see if I can simulate it. The key might be the high pulse per revolution

Cheers
Peter

 

[masso-support]

Hi Sean,

Just to let you know I have reproduced the problem you are having with backlash.
It is tied up with the pulse rate and lowering it to something like 30000 to 35000 stops the problem using your settings so if you are looking for a temporary way to use backlash please reduce the pulses per revolution.
Let me know if you need assistance with doing this.
In the meantime I have passed the problem to have the software look into.

Can you let me know your Masso serial number and I will add your serial number to the ticket I created so we have it on hand.

Cheers
Peter

 

[froes-billie]

Yeah,

Great info, same issue for me, just finished converting an old lathe. Really big backlash is both axis. More than 1mm, this gives positioning problems. After jogging around Z+ and Z- a few times and moving to the set zero point (in the same direction from which the zero point was set, very important) I find that my marking on the motorpully is shifted a couple of degrees.

Might be because of the high backlash value? I donno, maybe masso rounds of the 2 biggest digits or something?

Will look for an update, for now, I dont use backlash, just have to program my backlash in the CNC program (which are very simple programs in my case).

To bad, but without backlash the motorpulley always goes back to same spot at a set zeropoint. With backlash compensation it will change after a few axis movements.

Greetings Jordy

 

[nootje]

I have also issues With the backlash compensation. Right now it makes the machine useless for high tolerance workpieces.



My backlash settings are 0.05mm on X 0.01 on Y and 0.02 on Z-axis. As soon as I try to mill with backlash settings, the axis aren't moving smooth anymore.

At the point where one of the axis changes of direction, it's like the machine stops for a microsecond.

After a couple of 100 command lines I get a shift on the X-axis of almost 1mm.

I turned off the backlash compensation, restarted the program and no more shifts and back to smooth paths.



I'm running a DIY CNC setup dual Y axis, with the B-axis as slave on the Y axis. Lichuan 3phase closed loop steppers. 5000 steps/ rev and 10mm pitch ball screws.

Files are created with fusion360 and I never had this before I moved on to Masso.

To me it looks like the controller is having a hard time with the addaptive toolpath in combination with the backlash compensation(picture atached, +/-100K lines for the complete program)

A solution would be welcome and great

 

[cncnutz]

Hi Nootje

When you turn on backlash compensation you are asking the axis to stop and reverse direction to take up the slack in your drive train everything you reverse the motor direction before moving on with the next cut.
This is noticed on curves when moving from one quadrant to another. When it reaches a new quadrant one motor has to stop and reverse before it can continue on. This causes the small stop you notice.
It is even more noticeable on adaptive tool paths because they are predominantly curves.

The drawing below shows where on a circle an axis has to stop and reverse before it can continue.

With regards the loss of position can you please post a copy of your printable settings file please.

Hope this helps
Cheers Peter

 

[nootje]

Quote from CNCnutz on March 28, 2021, 11:24 am

When you turn on backlash compensation you are asking the axis to stop and reverse direction to take up the slack in your drive train everything you reverse the motor direction before moving on with the next cut.
This is noticed on curves when moving from one quadrant to another. When it reaches a new quadrant one motor has to stop and reverse before it can continue on. This causes the small stop you notice.
It is even more noticeable on adaptive tool paths because they are predominantly curves.

thx for the reply Peter.



I did understood the way backlash is compensated, but I didn't realize the effect on the smoothness would be that great. I can live with the unsmooth effect, I just turn backlash off at the roughing part and back on for the finishing steps when I'm at time stress. I've been looking into the fusion setting, but no way (at least for me) to implement the backlash over there.

The shifting with anti backlash is a thing I can't work with. I did a repeating test and the shifting appeared to be the same direction and same distance. Without backlash compensation no shift, so I ran the jobs from today and yesterday with the X and Y backlash compensation off.

enclosed the printable settings.



thx again

Cheers,

Chris aka Nootje

 

[cncnutz]

Hi Chris,

I'm not seeing the printable settings file.
You may need to zip it as the forum only accepts certain file types.

Cheers Peter

 

[nootje]

Quote from CNCnutz on March 29, 2021, 5:44 am

Hi Chris,

I'm not seeing the printable settings file.
You may need to zip it as the forum only accepts certain file types.

Cheers Peter

my bad... here's another try.



Chris

 

[pinoy-cnc]

Hi Sean,

Just to let you know I have reproduced the problem you are having with backlash.
It is tied up with the pulse rate and lowering it to something like 30000 to 35000 stops the problem using your settings so if you are looking for a temporary way to use backlash please reduce the pulses per revolution.
Let me know if you need assistance with doing this.
In the meantime I have passed the problem to have the software look into.

Can you let me know your Masso serial number and I will add your serial number to the ticket I created so we have it on hand.

Cheers
Peter

Hi Peter,

Is this also applicable on my issue? I have this issue only on my x axis.
In between 0 to 1200. I have off of 0.5mm to maximum 1mm.

I am using 4000 pulse/Rev.
My motor is closed loop stepper motor

Thanks

 

[zombieengineer]

@pinoy-cnc

Backlash can be caused by a many different causes (free-play, bending, "sticktion" come to mind).

I believe the pulse rate issue listed above is due to the stepper driver missing steps (electrical problem rather than skipping a step due to load).

To eliminate this as a potential cause I would recommend placing a mark on the stepper motor / pulley to represent "top dead centre". Move axis to "TDC" record the position, move away and then back (at varying speeds) to the position representing "TDC". If the axis stepper motor returns exactly to the same angular position then the error must be between the stepper motor pulley and the rest of the motion control system (gear boxes, gear rack, etc).

Without a detailed description / photograph of the axis it is hard to give suggestions on what to check next.

4000 pulses/rev is starting to get into the "excessive" steps/rev for an open loop stepper motor however it is likely to be OK for a closed loop stepper motor. For an open loop stepper motor at about 10 micro-steps / full step is about the limit of interpolation before the micro-step is smaller that the interpolation error. If the closed loop feedback encoder has sufficient resolution (fairly easy to create an encoder with sufficient resolution) then the interpolation error is considerably smaller.

 

[pinoy-cnc]

@pinoy-cnc

Backlash can be caused by a many different causes (free-play, bending, "sticktion" come to mind).

I believe the pulse rate issue listed above is due to the stepper driver missing steps (electrical problem rather than skipping a step due to load).

To eliminate this as a potential cause I would recommend placing a mark on the stepper motor / pulley to represent "top dead centre". Move axis to "TDC" record the position, move away and then back (at varying speeds) to the position representing "TDC". If the axis stepper motor returns exactly to the same angular position then the error must be between the stepper motor pulley and the rest of the motion control system (gear boxes, gear rack, etc).

Without a detailed description / photograph of the axis it is hard to give suggestions on what to check next.

4000 pulses/rev is starting to get into the "excessive" steps/rev for an open loop stepper motor however it is likely to be OK for a closed loop stepper motor. For an open loop stepper motor at about 10 micro-steps / full step is about the limit of interpolation before the micro-step is smaller that the interpolation error. If the closed loop feedback encoder has sufficient resolution (fairly easy to create an encoder with sufficient resolution) then the interpolation error is considerably smaller.

Thank you..

But the strange thing is I have 2 machine that have X axis problems one with ballscrew with panasonic motor installed and this rack and pinion system with closed loop motor installed.

Y axis has no problem on this 2 machine. I don't encounter this problem with the old version of masso firmware. I done more than 15 cnc machine I don't encounter this problem.