Intermittent probe problem, please help

[raj]

For the last three or four days I have been having trouble completing any scans on parts with the probe. After a few hours at some random point the probing process seems to come to an end. The message I get is that there has been a clearance problem. However ther seems nothing apparant why this would have happened. The probe is a few months old and I have had good success with it for about 25 scans, some of which were many hours long. Does anyone know what might be the cause ? is there a way to try to get the probing to continue from where there was a clearance problem so that all the rest of the scan is not lost.
Should the probe be oiled or anything like that ?
Any insights would be appreciated. Thanks

[bill.young]

Hi Raj,

It's hard to diagnose these sort of things after the fact, but here are a few things to look at if it happens again...


* Is the input switch indicator on or off when the routine stops? Is the probe tip in contact with the surface?


* Has the z-axis moved to the maximum clearance value for the file? If not, does the z-axis readout say it has?


* If the input switch is ON and the tip is above the pattern surface when it stops, does the input switch go OFF like it should when you touch the tip?

Bill

[raj]

Hi Bill,
Each time the problem happened, the probe had risen to what appeared to be the safe Z height, well clear of the surface that was being probed. The input indicator was off. Touching the probe did not cause any change to the indicator. I had to get out of the "paused" routine when the indicator light worked as expected, going on and off when it was touched. I have been having a tough time getting any scan to go to completion last few days. Maybe making a change to the code so that when there is a clearance problem the program waits till the user is back and then does another try at continuing where it had paused would be one way to solve/diagnose the problem.
After four or five attempts I may actually get a completed scan, that is what happened on the previous ornament I was scanning, but I am in a bit of a bind with this. Any ideas what I might try to get to the bottom of it.

[billp]

Raj,
How large of a "safe Z height"do you use in your files? I normally use .5" so it does not go too far from the uppermost point on my model. The only time I ever had symptoms as you are describing is when I had either failed to "Z zero" off the top of my material first, OR I had screwed up the safe Z height somehow...
Do you allow your probe to actually touch your table as you are scanning?
I have found that by accurately measuring the height of my model, and then editing the probe file to only allow my tip to go within .05" of the table surface I get less chance of drag,distortion, and it also speeds up the overall probe time tremendously since the probe is not stopping to touch the table every .02". If I need to scan right to the edge of my model I elevate it slightly on a piece of clay or some other similar material so I can get the full model, and still not touch the table..

[bill.young]

Hi Raj,

Does the input switch come on as soon as you exit out of the stalled probe routine, or do you have to move the stylus a little to get it to come on?

Bill

[raj]

Bill,
The scan that I am currently trying to get done is a crown molding with a leaf design. Optimistically, at first I went for broke to scan six feet by 5 inches or so. I used 1.2", 1.5" even 2" as Z clearance. All those would easily clear the molding. The molding varies in height by about .75 inches. After the probing failed a few times at different places I thought I would do a much smaller scan of about 7" by 5" and do the full length by joining pieces using software.
Each time I did a Z3 to zero all x,y and Z before starting the scan. The shop bot software routine usually worked very reliably and likely has been tested out quite well. In this particular scan there are no "flat parts", if you can imagine a leafed cornice, so the probe never goes down as far as the table surface. I used overkill values for Z clearance (2") just to try to get the scan to complete, but the problem probably lies elsewhere. I avoided using any altered probing routines which although faster haven't had the testing and mileage put on and I didn't want to complicate the problem. Even if it took four or five hours extra, I just wanted to get to the end of the scan in a reliable manner ! On my last scan, the probe was paused above a smoothe part of the cornice. The probing had worked for all the leafy ornamental part of the cornice ! I was reluctant to add changes to the probe program because it would introduce another set of issues into the problem. When you are a small business like mine, beauty, elegance and speed in software ends up secondary to getting the job done. While I keep working on "would be nice" issues, even if a probing takes 20 hours, if it works reliably is a real plus and something I can live with until I find a better more elegant alternative solution. Not working reliably is a killer and one sure way to sleepless nights ! That's where I am slowly creeping towards ! A slow probing process can be a practical, "path of least resistance way" if it works reliably. I wonder if any one else ever experienced anything like it and what was the cause ?!

[raj]

Bill,
That is a very good question and I should have paid closer attention to the light when I got out of the routine, but I didn't note the status immediately after I got out of the routine. I'll run the routine again and get back to you on your question. The failures have been fairly frequent. At least the last five or six attempts never completed on this particular ornamental cornice.

[raj]

Bill,
O.K, maybe the biggest clue : the routine failed after about an hour this time. The light was off. After getting out of the routine the light was still off. When I touched the probe I was able to make the light come on and off. Does this mean there is a "sticky probe" problem ? Probe not making contact when lifted (occasionally). What would be the solution ?!

[bill.young]

Hi Raj,

I wonder if some dust (plaster?) or other crud has gotten inside the probe which is occasionally causing a bad contact. You might try blowing it out with some of that canned air and see if that helps.

Bill

[raj]

Hi Bill,
Thanks very much for all your help. I think I do have a contact problem. Donna at Shop Bot walked me through how to open up the probe and check it out. Hard to know exactly where the contact problem is, but to me, one of the contacts looks suspicious. I am sending the probe for her to take a look. Thanks again.