Automatically measuring spindle probe runout

Matt@RFR · Jul 15, 2021

I'm exploring an idea to automatically measure the runout of a spindle probe every cycle. It would then measure locations and update work offsets.

I have failed to find an indicator with output that would work at all. There are some that would do what I need, but they are either not sealed well enough to survive that environment, or they have too much deflection pressure; Ie the indicator would deflect the probe, not the other way around.

Machine is Brother R650 and probe is Blum TC-52. Budget is $3,000 or so to outfit two machines and preferably have a spare tool.

Requirements for this tool is:

- Coolant proof to survive inside a CNC mill full time
- 1nm measuring force max (probe deflection force is 1.5nm)
- Be able to measure TIR and report if out of a given tolerance: IF [TIR GT .0003] SEND SIGNAL ... type of thing
- Preferably hard wired for power, but batteries would be ok if the device can be made to alarm the machine if the batteries die

If such a tool exists, I'll need to know:

- How to trigger the start of measurement from within a program
- How long the TIR cycle will take in seconds
- How to deal with I/O to be able to use macros to decide what to do with the information gathered.

Milland · Jul 15, 2021

Just clamp a gauge block somewhere to use as a reference contact for the probe, and make any needed adjustments from that reading. If needed a small air blast nozzle can be aimed at the block to keep the reference surfaces clean of chips and coolant.

What you propose with the indicator will be problematic from chips and dried coolant making it unreliable.

Matt@RFR · Jul 15, 2021

Milland said:
Just clamp a gauge block somewhere to use as a reference contact for the probe, and make any needed adjustments from that reading. If needed a small air blast nozzle can be aimed at the block to keep the reference surfaces clean of chips and coolant.

What you propose with the indicator will be problematic from chips and dried coolant making it unreliable.

The gage block will move around as the machine thermally grows and shrinks, so I won't be able to tell the difference between an axis that's shifted and runout in the probe itself.

Chips will be dealt with with a dedicated air blast, and our coolant never dries out.

EDIT: I missed your meaning with the gage block my first pass. I could hit the gage block four times with the spindle oriented 90º each time and then compare those results to each other. Problem solved, thanks very much Milland.

CNC Hacker · Jul 16, 2021

Yeah this is generally handled with a datum artifact that can be checked in cycle, like a bored hole or tooling ball. You probe it for size and if it measures off more than your tolerance you know the calibration is out, which would usually mean the stylus is running out. Renishaw compensates for some level of stylus runout with their software calibration, does the Blum probe not do that?

BROTHERFRANK · Jul 17, 2021

Matt@RFR said:
I could hit the gage block four times with the spindle oriented 90º each time and then compare those results to each other. Problem solved,

I was going to suggest the tooling ball as a standard as CNC Hacker brought up. If you go your route use the M variables with your cycles. You can choose to leave the probe on for multiple cycles and then choose to turn it off on the last one. Saves lots of time.

PROBE · Jul 17, 2021

Matt@RFR said:
I'm exploring an idea to automatically measure the runout of a spindle probe every cycle. It would then measure locations and update work offsets.

I have failed to find an indicator with output that would work at all. There are some that would do what I need, but they are either not sealed well enough to survive that environment, or they have too much deflection pressure; Ie the indicator would deflect the probe, not the other way around.

Machine is Brother R650 and probe is Blum TC-52. Budget is $3,000 or so to outfit two machines and preferably have a spare tool.

Requirements for this tool is:

- Coolant proof to survive inside a CNC mill full time
- 1nm measuring force max (probe deflection force is 1.5nm)
- Be able to measure TIR and report if out of a given tolerance: IF [TIR GT .0003] SEND SIGNAL ... type of thing
- Preferably hard wired for power, but batteries would be ok if the device can be made to alarm the machine if the batteries die

If such a tool exists, I'll need to know:

- How to trigger the start of measurement from within a program
- How long the TIR cycle will take in seconds
- How to deal with I/O to be able to use macros to decide what to do with the information gathered.

I am trying to understand the problem you have. The significant change in runout of the probe at each cycle can be caused only by not repetitive clamping of the tool at each tool change. If this is happening only when you change the probe, change the probe shank. If this is happening with any tool - why to update the probes offsets if each tool change causes change in its geometry. Check the machine tool clamping mechanism.

I was coping with the problem of updating the probe offset data in machines which had no tool orientation option.
In this cases the calibration ring was constantly present on machine table, placed as close as possible to machine reference point in order to minimise the thermal expansion influence. Each probe tool change started with calibration routine updating the XY offsets of the center of the probes sphere. The coolant rather then compressed air was used to clean the ring before use. Compressed air is the most aggressive machine tool enemy and should never be used for "cleaning" purposes.

Stefan

Matt@RFR · Jul 18, 2021

CNC Hacker said:
Renishaw compensates for some level of stylus runout with their software calibration, does the Blum probe not do that?

No, so process accuracy is controlled by runout on the Blum probing routines. I changed how we handle our Haas probing to match the Blum setup, since it's really quite easy to remember "NO RUNOUT", vs mounting a gage ring and running calibration cycles periodically.

BROTHERFRANK said:
I was going to suggest the tooling ball as a standard as CNC Hacker brought up. If you go your route use the M variables with your cycles. You can choose to leave the probe on for multiple cycles and then choose to turn it off on the last one. Saves lots of time.

Yep. Took me awhile to remember the M codes are not in order! M1 to turn on, M3 for no switching, then M2 for turn off. I'm working with ken now on how to turn the probe on and off while it's still in the carousel to save the time waiting on it to turn on/off. On the job we're running now, the on/off cycles are fully half the probe total time.

PROBE said:
I am trying to understand the problem you have

There's no current problem, only foreseeable problems that I want to eliminate. The random chip on a tool taper, or the probe adjustment screws coming loose. I only probe for locations, and those locations are not measured on every part, so one bad tool change can scrap a part and we probably wouldn't catch it.

I've had those adjustment screws come loose one time in many years of probing lots of stuff, but it caused enough havoc and scrap parts and free rework that it's worth it to take the few seconds as a sanity check.

Vancbiker · Jul 18, 2021

Probing a ring gage and comparing current values to new values would tell you if a chip had gotten into the taper or adjustments had moved.

cameraman · Jul 18, 2021

PROBE said:
I am trying to understand the problem you have. The significant change in runout of the probe at each cycle can be caused only by not repetitive clamping of the tool at each tool change. If this is happening only when you change the probe, change the probe shank. If this is happening with any tool - why to update the probes offsets if each tool change causes change in its geometry. Check the machine tool clamping mechanism.

I was coping with the problem of updating the probe offset data in machines which had no tool orientation option.
In this cases the calibration ring was constantly present on machine table, placed as close as possible to machine reference point in order to minimise the thermal expansion influence. Each probe tool change started with calibration routine updating the XY offsets of the center of the probes sphere. The coolant rather then compressed air was used to clean the ring before use. Compressed air is the most aggressive machine tool enemy and should never be used for "cleaning" purposes.

Stefan

PROBE said:
Compressed air is the most aggressive machine tool enemy and should never be used for "cleaning" purposes

In the Makino manuals "they" stress to never use a compressed air / air gun to clean a table off/ parts etc. in the machine.

Sort of tangential to core discussion,

I know old school Bridgeport style machining to go - (pssst, psshh pst psssssst pft ) with the shop air - that's pretty common; but wondered what the whole set of reasons or reasoning might be to not do that on a VMC especially a mold machine that might be cutting dry anyway + other ?

I figured Vancbiker-man would also know.

It's one of those things where they say "Don't' or "Never" but never really explain precisely why ?

CarbideBob · Jul 18, 2021

Matt@RFR said:
EDIT: I missed your meaning with the gage block my first pass. I could hit the gage block four times with the spindle oriented 90º each time and then compare those results to each other. Problem solved, thanks very much Milland.

Gage block not needed, any reasonable finish works.
This four hits may not give you the highest possible runout but for sure gives a very good clue.
Ideally every side touch gets the four and average but that is crazy expensive in time and very rarely done.
CMMs are different and this off center is no big deal at all even if off .050 or more. One can still do microns.
People mix probing on a cnc tool and CMMs. :nono:

Probes on a machine tool can be the best things since sliced bread. Yet sometimes they fail and the why becomes a source of arguments.

Acknowledging that some runout or error is always going to be in play is a good step IMO. Few would think or play with this problem. Is Kudos the right word?
Bob

5 axis Fidia guy · Jul 22, 2021

I'm not sure I really understand the problem. Does your probe cycle orientate the spindle every hit? I always work from a hole or block centerline, the probe could run out a few thou easy and it would still be accurate because it always rotates the same direction for each hit. Every few weeks I check it with a ring gauge and it corrects the runout.

Matt@RFR · Jul 22, 2021

CarbideBob said:
Gage block not needed, any reasonable finish works.
This four hits may not give you the highest possible runout but for sure gives a very good clue.

Thanks Bob, both excellent points.

5 axis Fidia guy said:
I'm not sure I really understand the problem. Does your probe cycle orientate the spindle every hit? I always work from a hole or block centerline, the probe could run out a few thou easy and it would still be accurate because it always rotates the same direction for each hit. Every few weeks I check it with a ring gauge and it corrects the runout.

We probe centerlines just like you, but neither my Renishaw nor my Blum probe routines orient the probe like you're talking about for their standard stuff. Other things like 3-point-bore and calibration routines, yes, but we don't use those in cycle. I could implement that functionality, but then I'm writing my own probe routines and I'm not ready to go that far yet. At least, I don't think I am.

Vancbiker said:
Probing a ring gage and comparing current values to new values would tell you if a chip had gotten into the taper or adjustments had moved

I could come in and hit the same spot on the ring gage at different spindle orientations to determine runout, and then hit the gage with a standard bore routine to determine thermal movement in the machine. That would work and the probing a ring gage and using G52 to shift everything has been on my mind for a long time.

Help me think through this: The ring gage would have its own work offset, call it G59. When the job gets setup originally, I have to probe all the other work offsets because jaws and fixtures don't go in exactly the same every time, and probe G59 at the same time, and set G52 to 0/0/0. Then, each cycle, I store G59 XYZ in variables, hit G59 again and compare the two. Then I use that result to move my G52 and shift the whole setup each cycle.

Search

Largest Manufacturing Technology Community on the Web

Search

Automatically measuring spindle probe runout

Matt@RFR

Titanium

Milland

Diamond

Matt@RFR

Titanium

CNC Hacker

Cast Iron

BROTHERFRANK

Stainless

PROBE

Hot Rolled

Matt@RFR

Titanium

Vancbiker

Diamond

cameraman

Diamond

CarbideBob

Diamond

5 axis Fidia guy

Stainless

Matt@RFR

Titanium

Similar threads