Touch probe questions

Hey All, I'm having a bit of a head scratcher concerning some odd things I'm seeing with my Tormach digital probe, the one they call the Digitizing probe in PathPilot.

Had some problems getting good results on a project a couple of weeks ago, and found the the stylus had actually started to unscrew from the probe body. I set it aside to finish the project and just got back to it yesterday.. After re-tightening the stylus, I ran the probe calibration routines in PathPilot and got it setup to within a tenth on the machine.

However in double checking with a 1 inch gauge block I saw a consistent error of minus about .0018". This stayed constant within a tenth or two, no matter whether probing in X or Y and with many variations of probe orientation in the spindle.

I was puzzled trying to reconcile the calibration results with the block measurements. I tried other sizes of gauge blocks, up to 3" and the error stayed constant, didn't increase with length of the block.

I was kind of stumped trying to understand what was going on, then today I probed a precision ring with an accurate bore diameter of 1.1107" and still got the same error. I know that stylus tip ball diameters have an "effective diameter" than is different from their physical diameter, to account for things like stylus flex and others.

So I started changing the tip diameter in the tool table, trying to get a measurement that matched the know diameter of the ring bore, and eventually I got a setting that gave the correct reading within a tenth on the DRO. I then measured a few different gauge blocks and got good results with them also.

So the point of this long-winded post is that to get the proper readings on the test pieces, I had to enter a diameter that was less that the physical diameter of the nominal 5 mm ball, 0.1947" vs 0.1965", as measured with a good micrometer.

So, bottom line, I'm asking if I'm fooling myself seeing good measurements now, is it just luck, or is the diameter adjustment causing accurate measurements? The reason I'm concerned about it is that I have never heard of the "effective diameter" being less than the measured diameter.

Ant thoughts on the subject would be much appreciated.
Terrt

Quote:

---

Originally Posted by MFchief

..the "effective diameter" being less than the measured diameter.

---

Hi Terry
How fast are you probing? Trigger delays will (if I've got this the right way around) allow the tip to move further before PP detects the trigger position. This will give the impression that the tip is smaller than measured. I'm still don't have my "real" probe operational so I'm not sure how much difference this might make. Try reducing the feed rate, perhaps by half, and see if this difference reduces.
Step

Hi Step,

There is some difference, I had been probing at 20 ipm and the 1" gauge block reads 0.9999 in the DRO. This is a Shars block set, but it measure 1.00" with my Mitutoyo mic that is supposed to be good to 0.00005".

A quick check with the same block at different speeds gives:

5 ipm 1.0003
10 ipm 1.0002
20 ipm 0.9999
30 ipm 0.9996
40 ipm 0.9995

So there is a definite change, that seems linear, if not constant, Keep in mind that when I adjusted the tip diameter in the tool table I was using 20 ipm, so it makes sense that that is the closest.

Terry

P.S.
Sorry the list looks goofy, but apparently the software here doesn't like extra spaces in a line of text.

OK, I'm off too test mine!

Along similar lines though, I found that the stated height of the ETS was inaccurate. In order to get it to agree with my height gauge (a mitutoyo one), I had to adjust the height setting of the ETS in it's setup.

Hi Terry
That's just one tenth per 5ipm - measurable but a long way from .0018".
On the other hand I've been too hung up with probing routines recently and forgot the mechanics. Keen did some measurements a while back and found a significant flex of the probe tips before they triggered. I dug out the thread with his results:
https://www.cnczone.com/forums/torma...ml#post1763824
He's using the standard passive probe as a reference and seems to be measuring a significantly larger deflection at the trigger point than you are. Presumably because your probe is more responsive.
Step

Yes you are discovering 'pretravel' (the flex before trip) ....and there is also pre travel variation.

Yes you need to set a 'effective diameter' yes smaller than the actual tip diameter - that is close to tip diameter - 2x average pre travel.

The amounts will be less than on the Tormach passive probe, but the same problem occurs with all tri swing arm electro mechanical probes without full compensation software.


Cheers Cliff



https://www.youtube.com/watch?v=Qtg6K27h_eI&t=4s

Thanks for the replies,

Step I had previously seen Cliff's video and was thinking about it when I was checking my probe also. That's the main reason I rotated the spindle to check other orientations for the "swing arm", and why I probed in Y as well as X. That really was the head scratching part, none of those changes affected the readings more than a tenth or two.

Cliff, I suspect that the pre-travel effect you described in your video is what's going on, but I didn't expect it to be that large.

Dannirr, I would be very interested in you results, thanks.
Terry

Terry - well, mine is consistently out 8 thou! in X and Y. I had not noticed this before, because all I ever do with it is set my work offset - and I use middle of the piece about 99% of the time - where it would make no difference as it would self cancel! Now I'll experiment with tip diameter as you describe. I was probing a gauge block, which was certified but also checked on my height gauge.

I think I'm, going to test it against the Haimer next - I know that's properly calibrated.

Thanks for bringing this to our attention.

Thanks for checking yours, I'm sorry you found it to be out so much. Did you try the calibration routines, if so, did it improve?

Good :Luck

Terry

An update. I found my probe to be out by about 8 thou, but not consistently in the same direction. Furthermore, it disagreed with my Haimer when measuring things. I decided to recalibrate the probe - which I had done on purchase - but found it was not running true (concentric). Using a dial gauge, I recalibrated it so it was exactly the same in all rotation. I then tried the actual routine in PP and it was identical too.

But even so, it was not the same as the Haimer. I measured the probe tip diameter - it was as advertised. I then experimented with altering the probe tip diameter in offsets, and found that if I changed it from 0.1181 to 0.1008, it was exactly the same as the Haimer in finding zero, and in measuring things.

So now I am more confident in it.

I use the Tormach passive probe on a non Tormach, but I do exactly what others are doing. I went to a larger beefier styli to reduce flex. I then set the tip diameter in my control using the measured diameter. When I calibrate I adjust the tip diameter via tool wear in my tool library until I get DRO measurements that represents the known measurement. It's extremely consistent from there on out. I do a double probe, one at 15ipm and a second touch at 5ipm (not sure of PP does this?). I find that you have to keep the probe contacts well coated in dielectric grease. When the get dry not only do I encounter issues with coolant, but I find that the probe becomes inconsistent.

Seems everyone is eventually going to the same place, the need to get an effective tip diameter that works. I asked some unrelated questions at Tormach the other day, but while I had them on the line I talked about getting the correct tip diameter, and they told me they were working on a procedure to release, which is pretty much what we are all doing, measuring a known accurate diameter and adjusting the setting to get the correct diameter.

The biggest remaining issue for me is that when you adjust the probe by turning the set screws, the amount of screw movement becomes very small as you approach the correct settings. After that, I am never comfortable that all the screw are tighten enough not to back out because of vibration. I think this is inherent in the design for the mounting for the probe, and don't know how to address it short of "fiddling" with the screws forever.

Terry

You could hit the top of the screws and side of the housing with a drop of CA/superglue once you've set them.
Or you could try some loc-tite green, which is supposed to penetrate already-made thread engagements and seat them more firmly.

It would be nice if that lower plate was made out of like .5" material. If that was the case you could torque the through bolts against the set screws so that the system was tight. It would also reduce the amount of flex in the system making the entire thing a lot easier to true up, and that would be quite a large improvement IMO. I had plans on doing this, but I have like 500 other projects in line before that, so it may never happen. That said, once I get it good and trued up I've never had issues with it changing. Even after decimating a probe the other day it was still dead on.

I don't see any reason why it would loosen. Its not subject to rotation, or much vibration in probing routines.

I'm not even certain that the screws loosen, it's just that the needed movement is so small, that after getting to the point where the adjustment is good, I'm hesitant to go around and tighten the other screws.

Although you're right about the lack of rotation and vibration during probing, it's possible I might be experiencing vibration because of the way I keep my probe stored. As you can see from the photo, I keep it in the enclosure, already plugged in and ready to use. I might have to re-think that. LoL

Terry

Mine lives in there too. I was able to tighten all three firmly without a change in calibration.

Perhaps I just haven't been patient enough with the adjustments, I'll try tighting all the screws better after I get a good setting.

Terry

I have both the Torch Digital Touch Probe and ETS tool setter. Both of these have given me similar issues as everyone else is seeing. I changed my Digital probe tip diameter to account for the trigger delay. The ETS has another issue also related to trigger delay. The routine for touching off the work offset or spindle nose uses different probe speeds than touching off tools. The second touch-off at the slower speed will differ between tool touch off and spindle nose touch off. I was seeing ~ 0.0004 variances. I have to set the probe speed at 40ipm (or is it 50ipm) for the nose touch off and 20ipm for the tool touch off. This will result in both having the slower touch off speed matching exactly and eliminating the variance.

Quote:

---

Originally Posted by TurboStep

Hi Terry
How fast are you probing? Trigger delays will (if I've got this the right way around) allow the tip to move further before PP detects the trigger position. This will give the impression that the tip is smaller than measured. I'm still don't have my "real" probe operational so I'm not sure how much difference this might make. Try reducing the feed rate, perhaps by half, and see if this difference reduces.
Step

---

Quote:

---

Originally Posted by draper-ballou

I have both the Torch Digital Touch Probe and ETS tool setter. Both of these have given me similar issues as everyone else is seeing. I changed my Digital probe tip diameter to account for the trigger delay. The ETS has another issue also related to trigger delay. The routine for touching off the work offset or spindle nose uses different probe speeds than touching off tools. The second touch-off at the slower speed will differ between tool touch off and spindle nose touch off. I was seeing ~ 0.0004 variances. I have to set the probe speed at 40ipm (or is it 50ipm) for the nose touch off and 20ipm for the tool touch off. This will result in both having the slower touch off speed matching exactly and eliminating the variance.

---

Sorry for such a late response but I guess I'm still the first :)
All probing routines I've checked in PathPilot use a fine feed rate equal to 1/20th of the initial feed rate EXCEPT the Move & Set Tool Length routine on the Offsets page. This is the only one I've found which uses a fixed fine feed rate of 2.5 in/min. This correlates with your results for 50 in/min initial feed.
Was this intentional or an oversight from Tormach? I can only guess.
Step