Renishaw touch probe calibration and accuracy

[gpcoe]

I've been working on a VF-3 (2006 model I believe) with a TR10 trunion 5 axis mounted to the table, for about 6 months now. I've learned a lot about the probing system as far as calibration and programming for part measurement and am now starting to use it for setup as well. I use a calibrated ring gauge to set probe calibration and when it sets properly it will find position fairly accurately. When I indicate the ring gauge, I will use MDI to go back to the ring and then I'll re-indicate to test the repositioning. If it is off, I'll enter in the changes on the offset page 1 axis at a time until I indicate zero within .0001-.0003 TIR keeping on the low side. This is done with an interapid tenths indicator. Once I'm satisfied with the positioning I will run a full circular calibration for ball offset and vector measure (I use them fairly often). I set the probe length with a parlec height setter and trim the ball in as true as possible at the same time before calibration. After the cal cycle I run a check on the part (boss) and again use MDI and indicator to check that it has found the center correctly.

I've had several instances now where I take the above steps and find the part off by more than .001 TIR from where the probe set it and when going back to the ring gauge (MDI) finding that the ring gauge is no longer in the same position indicated in not 10 min earlier (.0005 TIR or more). When installing the probe, I always have the spindle orient and place the shank with the circle side of the drive lug slots, towards me to ensure it's always in the spindle the same way. The ring is clamped very securely and level on a riser block mounted for secondary ops. It may take 2-3 tries to get the probe to find the gauge position accurately, but when it's finally dialed in, it works very well.

My concern is particularly the repeatablity for positioning. I can clearly see on #558 & #589 (XY cal offset) that they say .0006 and .0003 or something to that effect and it's reflected in that manor when measuring the part. When calibrated correctly I get .0003 and .00005 and measure .0002-.0003 TIR on the part at the worst. Does anyone else see this "floating" positioning when calibrating? Am I expecting to much out of this machine (I expect that I am)? Is this exceeding the machines accuracy? A lot of the parts I run on this have true position call-outs of .002 or less without MMC on features that are on the opposite end from or even cut separately from their datums, so I need this thing to go as close as possible fairly accurately. When cutting from stock (no probing needed), parts will cut wonderfully and are positionaly accurate.

Sorry for the novel. Any thoughts on this would be greatly appretiated.

Thanks
Greg

[tobyaxis]

Greg,

I just started using the Renishaw Probe on a TM3. Personally I would nver put 100% faith in a probe for keeping accuracy for more than .005. For some reason there is a hump that I can not get over about probes or CMM's for that matter.

For accurate work involving tenths I would Probe Pocket on an ID then tweak it in with the old .0001 resolution B&S. LOL, I have trust issues but when it involves a $5 to 30K part you can never be too careful.

I see the advantage of a Probe & Tool Setter for fast setups but for accurate work I use as reference only then Tweak the G50's.

I really don't like the quick rapid into the work then a slow touch off. I think they would be more accurate if the touch off were slower.

I would stick to what your doing now.

[dougtyler]

Its best if you have a calibration program that finds the center of the ring by itself. This is accomplished by have the probe measure the four quadrants (x+ x- y+ y-) without any correction values added in. Then the spindle is turned 180 degrees and the four points are measured again. Now you take the average of the two sets of points, and this is the true center on the gauge ring.

After the center is found. you can run the reagular calibration routine.

This process eliminates the human error involved with indicating.

As far as the accuracy of the machine, unless it is equiped with linear scales and the temperature is maintained within a few degrees, .0003 or .0004 is about all you can hope for.

[gpcoe]

Thanks dougtyler!! That's about as perfect a solution that I could hope for. My biggest worry has been an operator's ability to calibrate on off-shifts if necessary. I don't trust them worth a damn to indicate anything in accurate as it needs to be (if they could at all). I'll have to look into this and give it a try.


"I just started using the Renishaw Probe on a TM3. Personally I would nver put 100% faith in a probe for keeping accuracy for more than .005. For some reason there is a hump that I can not get over about probes or CMM's for that matter."

If I could only trust a probe to .005 it would be a waste of time and money. I can trust most of my parts coming from turning or other ops to be within .005 of where I need to do my work. The probing I do is fairly extensive and would be a nightmare to individually indicate in to get the accuracy I need. Don't forget this is a 5 axis machine. One part is probed at vertical for XYZ and at horizontal for XY at 4 90deg faces. Each face has features that need to be within .002 true positioned from the front face. With 300 parts to run I'd be there for a long time indicating each wpc in to where I need it for each part. Especially when they run for almost an hour a piece plus loading time.

I use both Manual and CNC CMM machines and have no other way to verify many dimensions on these parts. I can vouch for their accuracy....when they are properly programmed and calibrated, and for that matter, not abused. As for the milling probe it is very accurate...again if properly calibrated. I've gone for weeks without recalibrating and get great parts. The problem comes when it's time to calibrate and I spend an hour plus to get it to work correctly.

As for ramming it into the workpiece before measure, I believe that can be changed in one of the macro subs to move slower, though I've never had to change it. I feed 100ipm for positioning and 50ipm for measurement and it seems to work all right. I do have to say that I know I'm pushing the limits of these machines, I'm just curious what other people are doing with them.

Thanks Guys

Greg

[maweber49]

Hey gpcoe. I'm a little confused when you say that you're positioning at 100 ipm and measuring at 50 ipm. This is a Haas right? With Renishaw macros? Are you using a double touch measure (touches surface at fast feed, backs off, and then slow feed to measure the point) or a single touch (moves at a constant feed until it touches the part)?

If it's double touch the actual measure feed rate is controlled by the code in 9726 macro, F[30*#129] which works out to be 1.2 ipm. The fast feed touch feed rate is calculated by F[.6*#119]. So if #119 is set to 200 in the setup macro 9724, then the fast feed touch will be 120 ipm.

If you're using the single touch 9726 then you can control the feed rate by changing a value at the top of the macro. The default is 20 ipm with a .020" measure zone. I've found 20 ipm to be too fast and not that accurate. 5 ipm works ok for most machines.

Keep in mind that the controlling factor for accuracy is the machine, not the probe. Accuracy and repeatability are based on the machines ability to register the skip signal from the probe and then decel and stop the same way each time so that valid axis positions are written to the variables. If you're indeed cruising at 50 ipm measure speed, you're going to have some problems with accuracy and repeatability.

About the not trusting closer than .005" comment, if you keep it clean and calibrated you can hold .0002" - .0003" all day long. Even closer than that if you have a strain gauge probe like a Renishaw MP700 or OMP400

Mark W

[tobyaxis]

Originally Posted by maweber49 About the not trusting closer than .005" comment, if you keep it clean and calibrated you can hold .0002" - .0003" all day long. Even closer than that if you have a strain gauge probe like a Renishaw MP700 or OMP400 Mark W

I just don't trust probes for accurate work. You might feel the same if you have spent time on a Jig-Bore holding .00008. It is hard to change old habits but we are working on it.

[maweber49]

Absolutely. There is no substitute for that kind of precision. Ran a Swiss jig bore and a Moore jig grinder back in the 70's at Caterpillar Tractor in San Leandro, CA. I think we called that finesse.

[tobyaxis]

Originally Posted by maweber49 Absolutely. There is no substitute for that kind of precision. Ran a Swiss jig bore and a Moore jig grinder back in the 70's at Caterpillar Tractor in San Leandro, CA. I think we called that finesse.

I called it EARLY GRAY HAIR!! The only positive thing is that it is still all there, LOL.

I have to say the Probes are nice. Beats tramming a Bridgeport Head.

[Geof]

Originally Posted by maweber49 ......About the not trusting closer than .005" comment, if you keep it clean and calibrated you can hold .0002" - .0003" all day long. Even closer than that if you have a strain gauge probe like a Renishaw MP700 or OMP400 Mark W

I think too much faith is put in probes. If you look at the specs for Haas machines you find they are +/-0.0002" for positioning (accuracy) and +/-0.0001" for repeatability. These figues apply to the probes so your variation is 0.0004" and no matter how fancy your probe is you cannot get better than that because that is the limit of the machine.

[tobyaxis]

Originally Posted by Geof I think too much faith is put in probes. If you look at the specs for Haas machines you find they are +/-0.0002" for positioning (accuracy) and +/-0.0001" for repeatability. These figues apply to the probes so your variation is 0.0004" and no matter how fancy your probe is you cannot get better than that because that is the limit of the machine.

Your still awake!!! Wow your a trooper Geof. I'm working on a project and defragmenting a harddrive.

This could be the reason why my faith in probes isn't too high, LOL. I think this information is in the Renishaw Manual as well.

[dougtyler]

I have finally proven the worth of the probes, to the coworkers. Only after about 5 years of them seeing me never take my indicator out of my box, did they come around. As far as the accuracy of the probe goes, There is not an indicator that exists that can match the probes accuracy, especialy if it is using strain gauge technology. Our MP 400 probes are guaranteed to repeat down to 1/10 of a MICRON <---- thats .000004 of an inch. This far exceeds the accuracy of the machine. Our SIP jiibore machines can position within 2 MICRONS.

Put them together, And I can hold a .0002 true position all day long.

Indicating Sucks!

[Mr.Xtreme]

Guy's

After reading this about accuracy, if you tram something in and run machine for 10 mins hard, the ball screws length is growing, and you will be off location because of this.
I 've seen this especially on my SL-10 lathe ctr line for "x axis".
If you don't have scales, then accuracy will suffer thermal growth.
That said, even the best probe made is only as accurate as the machine axis it is used on.

Paul