TP20 Calibration Problems

[drice]

We have 2 machines that use the TP20. An Excel 9.15.9 & a Global Status. They are in separate locations. The problem we are having, is that when I calibrate a probe configuration on one module, then calibrate another configuration with a different module, I am either hitting the sphere or missing it. I have looked at each setup & they are the proper lengths. We are using MR 3.7-2. I guess, what I am wondering, is this an indicator, that the modules have run its course or is the probe file been corrupt & if so, how? This has happened to me this week on the 2 machines. We also have a Global Image, with a TP200, running MR 3.7-2, but I have not had any problems in regards to this. If anyone can shed some light, it would be greatly appreciated.

[phil george]

Try deleting that probe file and recreate it. They do get corrupted.

[drice]

That is what we ended up doing. I'm just trying to find out why.

Thank you for your quick response & help,

Dave

[Wes Cisco]

That is what we ended up doing. I'm just trying to find out why.

Thank you for your quick response & help,

Dave

I suspect it is called an "enhancement". LOL

I have found that many times after I calibrate the first module, when I bring up the second, if I only calibrate A0B0 by itself first, then go back and calibrate "mark used", then the other angles will calibrate. HTH

[#2#]

I have a feeling Wes is on the right track. It sounds to me like the two probes are not relating properly to each other.

Some things to think about...

When it asks if the calibration sphere has moved...

Yes - you take a manual hit and it uses this manual hit to guess where the sphere is for the rest of the cal.

No - no manual hits necessary, it thinks the sphere is exactly where it was for the last calibration (for that probe or any other probe). Only answer no if you are absolutely sure the sphere has not moved AND you have not re-homed the machine as home floats around ever so slightly.

In order for it to relate one probe to another and/or one angle to another it needs the following two things.

1)A0B0 must have been calibrated on both probes.
2)The sphere can not have moved between the calibration of each probe. I think this is technically not true but it makes life so much easier if you follow that rule.

I always calibrate A0B0 on a probe even if my program isn't going to use that angle. A0B0 is used to relate one probe to another!

Your best bet is to calibrate the first probe with all angles used including A0B0 (even if it isn't used). When you do this answer yes to the sphere moving, take your manual hit and then let it go. Then without moving the sphere, calibrate the other tip with all angles used including A0B0. Answer no to the sphere having moved.

Now, if the second tip misses the sphere, as it sounds like yours is doing, your probes are not relating properly. In order to bring them back do the following.

Calibrate probe 1 with all angles used and A0B0 answering yes to the sphere moving.

Calibrate probe 2 with all angles used and A0B0 answering yes to the shpere moving.

Calibrate probe 1 with all angles used and A0B0 answering yes to the shpere moving. Then calibrate probe 2 with all angles used and A0B0 answering no to the sphere moving. It shouldn't miss the sphere now.

One more thing to consider. In the calibrate screen there are 4 little options for how to calibrate

1) Manual - all hits for the cal have to be taken manually.

2) DCC - the first hit must be taken manually if you answer yes to the sphere moving.

3) Manual+DCC - I'm not sure how this differs from DCC. Maybe you always have to take the manual hit even if you answer no to the sphere moving.

4) DCC+DCC - This is the fun one and the one I always use. For each tip angle it will take 3 'sample' hits to get an idea of where the sphere is. It will then take the number of hits specified in the calibration window. These 3 hits are kind of an iterative thing so that if your manual hit is way off center, they will figure that out and your prehit/retract will measure the same for all your calibration hits - it is kind of like doing a double calibration. This one will still ask for the manual hit if you answer yes. If you answer no, it will take 4 'sample' hits and proceed automatically. This one is especially usefull if your probes and/or angles aren't relating properly.

[#2#]

OK, one more thing. Take your time on tip qualification. Yes, it is time consuming but if it isn't done right, your results can not be correct - especially if you are comparing two features measured with different tips.

[Troy Tatters]

I have been doing this too with success. I'm using 3.5 MR2 and before I began doing this regularly The probe file would flip tip vectors around causing the probe to try to go through the artifact to get to it's prehit distance.

I've edited bad tips vectors instead of replacing probe files in the past, especially large ones. Now I back up my probe files seperately so I can always recall a known good one.

[Troy Tatters]

Always qualify A0B0 on new probe files and/or after moving your artifact. I've rarely had problems since going this route.

[kpayne]

I have a feeling Wes is on the right track. It sounds to me like the two probes are not relating properly to each other.

Some things to think about...

When it asks if the calibration sphere has moved...

Yes - you take a manual hit and it uses this manual hit to guess where the sphere is for the rest of the cal.

No - no manual hits necessary, it thinks the sphere is exactly where it was for the last calibration (for that probe or any other probe). Only answer no if you are absolutely sure the sphere has not moved AND you have not re-homed the machine as home floats around ever so slightly.

In order for it to relate one probe to another and/or one angle to another it needs the following two things.

1)A0B0 must have been calibrated on both probes.
2)The sphere can not have moved between the calibration of each probe. I think this is technically not true but it makes life so much easier if you follow that rule.

I always calibrate A0B0 on a probe even if my program isn't going to use that angle. A0B0 is used to relate one probe to another!

Your best bet is to calibrate the first probe with all angles used including A0B0 (even if it isn't used). When you do this answer yes to the sphere moving, take your manual hit and then let it go. Then without moving the sphere, calibrate the other tip with all angles used including A0B0. Answer no to the sphere having moved.

Now, if the second tip misses the sphere, as it sounds like yours is doing, your probes are not relating properly. In order to bring them back do the following.

Calibrate probe 1 with all angles used and A0B0 answering yes to the sphere moving.

Calibrate probe 2 with all angles used and A0B0 answering yes to the shpere moving.

Calibrate probe 1 with all angles used and A0B0 answering yes to the shpere moving. Then calibrate probe 2 with all angles used and A0B0 answering no to the sphere moving. It shouldn't miss the sphere now.

One more thing to consider. In the calibrate screen there are 4 little options for how to calibrate

1) Manual - all hits for the cal have to be taken manually.

2) DCC - the first hit must be taken manually if you answer yes to the sphere moving.

3) Manual+DCC - I'm not sure how this differs from DCC. Maybe you always have to take the manual hit even if you answer no to the sphere moving.

4) DCC+DCC - This is the fun one and the one I always use. For each tip angle it will take 3 'sample' hits to get an idea of where the sphere is. It will then take the number of hits specified in the calibration window. These 3 hits are kind of an iterative thing so that if your manual hit is way off center, they will figure that out and your prehit/retract will measure the same for all your calibration hits - it is kind of like doing a double calibration. This one will still ask for the manual hit if you answer yes. If you answer no, it will take 4 'sample' hits and proceed automatically. This one is especially usefull if your probes and/or angles aren't relating properly.

Very well explained Goodluck... this is basiclly the method I use.

Maybe you should post this under "PCDMIS Sample Code"


Kevin

[#2#]

Very well explained Goodluck... this is basiclly the method I use.

Maybe you should post this under "PCDMIS Sample Code"


Kevin

Thanks! Glad I wasn't wasting my breath (finger movements?).

[drice]

Thank you to all who have replied.

Dave