hexagon logo
Register Sign in
  • Replies 9 replies
  • Subscribers 18 subscribers
  • Views 133 views
  • Users 0 members are here
Options
Related

Falsifying PDF Reports

acgarcia
I just found an instance to where someone copied a passing report, pasted it and changed the file name to a failing part number. Does anyone have anything program wise to prevent this?

Details:

Operator scans a barcode the inputs a report comment with the serial number. Then I use assignments to take that input and use it to print a pdf with the serial number as the filename. If a part pass the inspection, it goes to a passing folder and it goes to a failing folder if there is 1 or more out of tolerance dimensions. So I have some redundancy with the serial number being present in the report comment and as the filename. Our production software checks to see if the serial number is in the passing folder in order to move it on to the next step. So if it fails, it cannot move forward and it has to be scrapped or reworked. Pretty simple.

Machinist had 2 parts pass the CMM and the 3rd failed last Friday. I have a CNC programmer troubleshooting the CNC machine to make adjustments and he wanted to re inspect the bad part. Problem was we couldn't find it in the scrap bin and didn't know where it was. So he wanted to see the good reports and as we looked at the filenames (serial numbers) we saw the failing serial number in the passing folder. We opened the report and the report comment had a different serial number in the field. It appears the machinist opened the passing folder, copied a passing report, pasted it, and changed the file name to match the failing part. This would allow him to move the part through. So I have 2 identical reports with the exact same time and deviations but with the same report comment serial number and different filename. Luckily we caught this before it left the building.

Now I have IT working on removing filename edit access so now we can't edit filenames but you can still copy a failing report from the failing folder and paste in in the passing folder. They are still working on that part.

Anyway, does anyone have any advice or preventative measures so this doesn't happen again utilizing pcdmis?
Parents
  • The guy that did it was let go yesterday. In the past, I thought about not printing bad reports but then the operators wont be able to see whats bad on the report. I've asked the operators if they would like the program to stop at the moment an out of tolerance condition has been found and go straight to a print command but that would leave out the rest of the part and they wouldn't know if another feature is out until the next part. The problem with access is more of a windows setting and pcdmis. I was just wondering what kind of controls for this type of scenario is being used out there.

    How do you all handle re inspecting the same part? Say I have 1 dimension that fails by 0.001" out of tolerance. The operators here would re clean the part and most likely re-run the probe calibration program cleaning the probes with a rag and cleaning the tp20 magnet with the putty before starting. Usually that brings that does enough magic to bring that feature just back in to tolerance. Sometimes not. Is that acceptable?

    I know there are a lot of variables to consider. Probing methods, number of points, and probing hardware. My machines are shop floor models so they aren't supper accurate. The shop is very clean but there are particles in the air and shop dust settles on the CMMs. CMMs are located near CNC machines so some vibration is a factor too. There is also temperature fluctuates in the shop. But when you run 20 good parts in a row and the 21st is bad, its like stop the presses. They want me to "redo" the math or fix the CMM. I'll go out there and clean the scales, re-home the CMM, re-run the calibration program and then the part. Sometimes it helps, sometimes not. ¯\_(ツ)_/¯

Reply
  • The guy that did it was let go yesterday. In the past, I thought about not printing bad reports but then the operators wont be able to see whats bad on the report. I've asked the operators if they would like the program to stop at the moment an out of tolerance condition has been found and go straight to a print command but that would leave out the rest of the part and they wouldn't know if another feature is out until the next part. The problem with access is more of a windows setting and pcdmis. I was just wondering what kind of controls for this type of scenario is being used out there.

    How do you all handle re inspecting the same part? Say I have 1 dimension that fails by 0.001" out of tolerance. The operators here would re clean the part and most likely re-run the probe calibration program cleaning the probes with a rag and cleaning the tp20 magnet with the putty before starting. Usually that brings that does enough magic to bring that feature just back in to tolerance. Sometimes not. Is that acceptable?

    I know there are a lot of variables to consider. Probing methods, number of points, and probing hardware. My machines are shop floor models so they aren't supper accurate. The shop is very clean but there are particles in the air and shop dust settles on the CMMs. CMMs are located near CNC machines so some vibration is a factor too. There is also temperature fluctuates in the shop. But when you run 20 good parts in a row and the 21st is bad, its like stop the presses. They want me to "redo" the math or fix the CMM. I'll go out there and clean the scales, re-home the CMM, re-run the calibration program and then the part. Sometimes it helps, sometimes not. ¯\_(ツ)_/¯

Children
No Data
© 2024 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement