mistake in OpenPnP: using nozzle N1 for all calibrations

[Evgenij Rakhno]

Dear OpenPnP team

I have been using OpenPnP for about a year to perform simple tasks.
I have a machine with four nozzles, and during this time I’ve developed a clear understanding of which nozzle provides the most accurate placement results—and unfortunately, it’s not N1.

For a number of mechanical reasons, N1 is actually the worst nozzle on my machine (the best one is N2). It has the largest runouts and tilt. However, unfortunately, this nozzle plays a key role in defining the machine’s base parameters, such as the distance to reference points, the distance between the down-looking camera and the base nozzle, and the precise location of the up-looking camera.

I believe that over time, all users will figure out which nozzle performs best, and therefore selecting a “base” (best) nozzle for calibration procedures would allow users to significantly improve their machine’s accuracy without any physical modifications.

Therefore, I believe that hardcoding which nozzle must be used for all calibration operations is a mistake

[Toby Dickenson]

The default nozzle, used for machine calibration, is defined as the first nozzle in the list. I agree it would be good to have a nice way to redefine that default.

I think the best solution today is to edit your machine.xml file in a text editor, find the <nozzles> xml node, and move your N2 <nozzle> up to the top. Then redo some calibrations of course.

--
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/58ddfe4e-5c1e-4f91-9648-257da52174d3n%40googlegroups.com.

[Evgenij Rakhno]

Thank you, I’ll try today.

The biggest blocking factor is that calibrating the top and bottom cameras takes me about one and a half to two hours, and two out of three calibrations returned “too many errors.”

субота, 11 квітня 2026 р. о 17:57:28 UTC+3 to...@tarind.com пише:

[Evgenij Rakhno]

haha. done. the machine has already gone through all calibrations, but I figured out what’s wrong with this method — earlier I thought the nozzle list was formed in alphabetical order, but it turns out they are displayed in the sequence defined in the "machine" file. So if previously the nozzles were N1/N2/N3/N4, now they are N2/N1/N3/N4. And it’s not as funny as it seems at first glance, because they appear in this order literally everywhere the nozzles are referenced, and it’s very inconvenient. 

I wonder if the machine will lose its calibrations if I swap nozzles N1 and N2 back again later. I’ll let you know once I try it.

субота, 11 квітня 2026 р. о 22:21:53 UTC+3 Evgenij Rakhno пише: