Repeatability Issue

[DLAM]

I am recently having an issue with repeatability on my ACORN Equipped lathe. This is a EMCOTURN 220P so it is a retrofit CNC lathe that has good and solid mounts and ways not a manual machine conversion. The machine has been running great and with no issues with repeatability over the last couple years after conversion to ACORN. I just had to replace the ball screw and nut on the X Axis. After the repair I went through the calibration of the X Axis and everything worked out fine (so I think). Running the first run of parts since the repair and I am having issue with repeatability. It will run the program fine with parts in tolerance and no issue. When I go to run the program again the x axis reference has shifted up (+) about .010” - .015”. You can see the deflection in the drill (.070”). I have found that if I reference the axis’s back to home position between each cycle start then there is no problem and parts are in tolerance. The home position is set by proximity sensor and ZRI in the drive motor.
I thought there was a mechanical issue with the axis causing the x axis to loose position. I went through the X axis and could not find any interference, binding or anything that could cause the axis to loose position. The axis is driven by DYN4 Servo’s in step and direction. The Z axis is having no issues.
Anyone having any ideas? Any and all help is appreciated. I am attaching a current report from this morning.

[cncsnw]

After it has run a cycle or two, send X back to home using the same M92 and M105 moves that you normally use for homing, but do not run the M26 to re-zero X.

Then check the Machine Coordinates position. If it is not within a couple counts of zero, then you have lost steps.

The issue is probably with the command signals going from the Acorn to the drive. Either it is dropping steps (so the drive does not perform steps that the Acorn commanded), or it is picking up noise (so the drive is performing steps that the Acorn did not command).

[DLAM]

M92 and M105 are giving me an unknown parameter error. I will have to dig into the routine that does the reset home command to to see what routine I need. I see when I reset home the x Axis is off about .005" before it resets the value. Now I guess the Easter egg hut starts. trying to figure out where the noise is coming from causing it to loose steps.

[DLAM]

The X axis is consistently being .0058" off by running the program 1 time. I am not seeing any other errors.

[cncsnw]

If you run the program multiple times, without rehoming, does it lose an additional 0.0058" each time (i.e. does the error accumulate)?

From your homing macro, "cnct.hom":

Code: Select all

M92 /X L1 ;Move to X+ home swith, backoff until it clears.
M105 /X P4 F.0001 ;Move X minus at .0001 ipm until input 4 closes

[DLAM]

Yes the error is accumulative. I ran the program 5 times and the error was about .030".

[cnckeith]

if accumulative random value = loose pulleys, lost step signal and or motor stalling

.0001 ipm? whoa that is really slow even for a zri.

and while not related to your issue. you are using a very old version of cnc12 (v4.50)
you do have some pc resend errors, i would run the Centroid PC Tuner on the PC.
make screen shots of the Wizard menus.
and install the latest version of CNC12 (here. https://www.centroidcnc.com/centroid_di ... loads.html)
configure with the new Wizard.
run the stress test.
so we are all using the same version of CNC12 and can help you better.

i see in the error log lots of x overall turns ratio changes being made.
what is the ballscrew pitch? and how is the motor connected to the screw?

[DLAM]

Keith,
I thought of mechanical issue as well and went through things and have not been able to identify any mechanical issue. The pulleys are keyed on to the ball screw with a nut against a shoulder. The pulley on the motor shaft has 2 set screws and keyed. the only thing I could think of is the belt slipping teeth. it is a 3/4" wide timing belt, I tightened it as much as I could and there is still room in the adjustment slots to take up more if belt stretches
The .0001" ipm for the ZRI is a little slow luckily it doesn't have to move very far at all. When I first setting up the machine after conversion to Centroid I had a problem missing the ZRI pulse with higher rates so I kept it at the .0001" ipm.
The new ball screw is 4 mm pitch and the old ball screw was 8 mm pitch. I had a problem finding a direct replacement for the ball screw I didn't have a lot of room for the nut so I was able to keep the same diameter just had to change the pitch. All the changes in the turns per inch settings was trying to get the axis dialed in with the standard that I have.
The machine was working great and is backlogged with work is why I have not upgraded CNC12. know that it is sort of down I need to get that done so I can hopefully get the machine back up shortly. It sounds like there may be some new features and utilities that will help the process.

[cncsnw]

Somehow I doubt it is really moving at 0.0001 in/min. At that rate, a full revolution of the X servo motor would take over 13 hours. A half turn would take over six hours.

I am not sure what the maximum speed (or minimum time per encoder count) is for catching the index pulse via a ZRI signal to a PLC input. It probably depends on a lot of things, including the encoder resolution, index pulse gating, servo drive response, and PLC input response time. But I expect even a feedrate of 0.18 in/min (which would be about 25 encoder counts per second, or two PLC scans per encoder count) would be more than slow enough.

[cnckeith]

Keith,
I thought of mechanical issue as well and went through things and have not been able to identify any mechanical issue. The pulleys are keyed on to the ball screw with a nut against a shoulder. The pulley on the motor shaft has 2 set screws and keyed. the only thing I could think of is the belt slipping teeth. it is a 3/4" wide timing belt, I tightened it as much as I could and there is still room in the adjustment slots to take up more if belt stretches
The .0001" ipm for the ZRI is a little slow luckily it doesn't have to move very far at all. When I first setting up the machine after conversion to Centroid I had a problem missing the ZRI pulse with higher rates so I kept it at the .0001" ipm.
The new ball screw is 4 mm pitch and the old ball screw was 8 mm pitch. I had a problem finding a direct replacement for the ball screw I didn't have a lot of room for the nut so I was able to keep the same diameter just had to change the pitch. All the changes in the turns per inch settings was trying to get the axis dialed in with the standard that I have.
The machine was working great and is backlogged with work is why I have not upgraded CNC12. know that it is sort of down I need to get that done so I can hopefully get the machine back up shortly. It sounds like there may be some new features and utilities that will help the process.

what is the pulley ratio, what kind of timing belt? photos please.
i am suspect of your overall turns ratio number.

don't over tighten timing belts bad things will happen a good timing belt doesn't need a lot of tension to work properly.