nozzle tip calibration Misdetects

[Developer Algo]

Quick question: What is the recommended number of allowed misdetects during the  Nozzle tip calibration ?

i have a strange issue. When i do Nozzle tip calibration through I & S it fails with error "........... too many misdetects ........." 

But when i do it through the Nozzle Tip Calibration  tab and only change the allowed misdetects from 0 to 6 and leave everything as is, it calibrates as in the the screenshot 

[Developer Algo]

here are screenshots 

[Toby Dickenson]

Your screenshot can't be far off the native camera resolution, and your nozzle tip is only 6 pixels diameter!!!!!

What is the pixel size of your bottom camera? My apologies if this has already been discussed in a previous message.

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/openpnp/972ce1e7-3214-4da7-982b-17a82c42a86bn%40googlegroups.com.

[Developer Algo]

Do yo mean Units Per Pixel ? The camera is set at  1280x1024 @ around 245 fps @ exposure time 0.004 ms ~ 31 microseconds, saturation @ 50 . Everything else is set to default.

[Developer Algo]

From the Mind-Vision  MV-SUA133GC Camera Datasheet the camera pixel size: is   4µmx4µm

[tonyl...@gmail.com]

Please post your latest machine.xml file and your log file.

[Developer Algo]

here is my latest machine.xml

[Developer Algo]

Here is the logs 

(1) Using I & S  (2) Using the Nozzle Tip Calibration Tab , in this case both fail. I have had some success using  Nozzle Tip Calibration tab but it is not repeatable most of  the time

[jbussmann]

Reading through the logs, I find "measured distance 0.6173551696084875 exceeds offsetThresholdLength 0.5" as the result from almost every try. Since each the x component is the dominant contributor by far, this appears to be something tat doesn't rotate with the nozzle or the tip. So I would suggest you recalibrate your bottom camera position. 

[tonyl...@gmail.com]

Or just increase the Offset Threshold

[Developer Algo]

Hey Jbussman or Tony ,

Which is the recommended option (1) recalibrate the bottom Camera position ( i have not performed the advanced Camera calibration yet )  (2) increase the Offset Threshold e.g 0.65

or in this case it does not matter so much which option i use ? Just want to make sure i follow the recommended way of configuring OpenPnP items 

[jbussmann]

You can use either method, however I wouldn't recommend simply increasing the offset threshold as I would consider this bad practice.

- It only treats the symptoms, not the cause.

- The most effective threshold is as tight as possible and only as loose as necessary.

- I like to have this threshold at 0.1 mm (easily reachable with the available camera and nozzle offset calibrations) and it being triggered means something has shifted and I want to know this.

You can of course increase the threshold and check if the calibration runs as expected. However after that I would suggest you go back to the bottom camera calibration (no advanced). Most likely you will have to ckeck "include solved" at the top (because you already solved this solution somewhen in the past) and then reopen and rerun that solution. Unfortunately I can't tell you how it is called precisely but it should say something like "Determine the up-looking camera BottomBackCam position and initial calibration."