Calibration Help

[flytyree]

I continue to have calibration issues. I envy individuals that have 0.001 accuracy.

Attached are screen shots of my homing, x-axis, y-axis, z-axis, spindle, and main screen.

These are the steps I take.

1. Home machine
   
2. Jog the machine to a location that allows me to interface with my dial indicator. (1-inch travel with the indicator)
   
3. F1 and then to the axis I am calibrating.
   
4. Select Wizard
   
5. Select Zero
   
6. On main scree insure that the axis I am working on is also zero
   
7. Using MDI select the travel distance, e.g., 0.9 inches.
   
8. In the F1 screen enter travel distance noted on the dial indicator
   
9. Select complete.
   
10. I repeat the above until I have a minimum of 3 consecutive readings on the dial indicator.
    

After these steps I am lucky to have a RANDOM 0.020 repeatability.

I am using Gecko stepper driver.

Clearly I am doing something wrong.

 

[breezy]

@flytyree

To get higher accuracy you need to move a greater distance than 1 inch. Ideally use nearly the full length of the axis that you are calibrating.

Your X axis is set to 2.37 inches / revolution therefore 0.9 inches will only rotate the motor ~140 degrees or 0.4 of a revolution which will require MASSO to issue ~800 pulses.

Now if you moved 36 inches that would be about 15 revolutions and 30,000 pulses, which will allow MASSO to better calculate the number of pulses required to travel a given distance and hence percentage of error is reduced. If you were 0.001 out in 0.9 that is 0.0011% error, where as 0.05 in 36 is 0.0013% error, yet the measurement error was 50 times larger than in the short distance measurement.

The other factor you need to take into account after/during calibration is backlash. When calibrating only move in one direction during the calibration measurement move, if you overshoot, backup past the measurement point and re approach from the same direction again.

Peter produced a video on his CNCNUTZ channel on how to calculate backlash using a small gcode file to move the axis back and forth to produce a backlash measurement.

Regards,

Arie.

 

[masso-support]

as @breezy explained, there also those factors and please also keep in mind that the wizard can take you very close to finding the right value to help with calculations but, in the end, you should manually change those numbers a bit to do final calibration of the axis.

 

[flytyree]

I have attempted to have repeatable excursions of 1 inch with the dial indicator. Then send the axis 20+ inches via MDI. Then call this new location "zero". Then in wizard establish a new zero. Go back to MDI and send the machine 20.5 inches. This will engage the dial indicator 0.5 inches. If it is short (or long) I enter this new value into the wizard. Is this an acceptable way to increase the precision of the PPR?

Thank you.

 

[breezy]

@flytyree

I have attempted to have repeatable excursions of 1 inch with the dial indicator. Is this an acceptable way to increase the precision of the PPR?

Sorry, I think that will introduce errors into your measurement, each time you move the dial indicator the chance of error is introduced.

The way I did it on the Bicton Men's Shed router was to tape a tape measure to the bed, lined up a pointer in the spindle with one side of the 100mm mark, set zero in the wizard, jogged using the rapid & step keys on the keyboard to the same side of the 1100mm mark and entered 1000 into the wizard and calculated the result. I then returned to 100mm mark zeroed the DRO and issued a 1000mm move command in MDI and checked accuracy. Then I tweaked the last digit in the Travel / Rev figure up or down depending on which way the error was. I was able to get the precision to the thickness of the major mark on the tape measure in 1000mm, I reckoned that was close enough for a router that was going to used on wood.

If you want to use a dial indicator you need to use some fixed length device to compare against, something like a 123 block or several of them stacked in a line.

Regards,

Arie.

 

[flytyree]

Good morning from New Mexico. Thank you Arie.

I am trying to discern the significance of using the keyboard to move the pointer and THEN using MDI. I am not sure I understand the significance. But, I will try that method.

As a point of clarification, in my procedure I did not physically move my dial indicator only the pointer.

 

[breezy]

@flytyree

I am trying to discern the significance of using the keyboard to move the pointer and THEN using MDI.

We don't have a MPG so using a wireless keyboard allows me to creep up on the measurement mark, because using the mouse on the screen is awkward.

After the wizard has calculated the travel/rev figure, it is a case of moving back and forth to fine tune that figure which can be done using MDI.

Regards,

Arie.

From sunny Western Australia.

 

[flytyree]

Thank you from the sunny Southwest.

I continue to work my issue and I greatly appreciate your input.

Clifford

 

[keymont]

I think the thing you're missing is that you aren't using a large distance to calibrate initially. Using the keyboard or MDI makes no difference, really, but it is that you want to zero, and then move your axis to as far away as you can on your machine, measure that distance as accurately as possible and enter it into the wizard. Then, after you've got it "rouged in", you can use your dial indicator to fine tune the number of steps per inch/mm or whatever.

Re-read Breezy's post above. He explained it well. It's all about using a large distance to give the wizard more data to do its calculations. More steps to measure = more data.

Good luck!

- Mike

 

[flytyree]

Thank you for your input Mike.

Clifford

 

[evermech]

@flytyree

You could also consider doing the motor calibrations mathematically. If you know the driver settings and the mechanical details of your axis you should be able to figure out exactly how far one pulse from Masso will move the axis. Maybe someone smarter than myself could design and post a work sheet, that when specific details are filled in, could help users to figure this out. When you know this information then you will know the accuracy of your machine which can be valuable information.

Guy