Can't figure out camera to needle calibration

mucek
Posts: 40
Joined: Mon Oct 26, 2015 7:06 pm

Re: Can't figure out camera to needle calibration

Post by mucek »

Today me any my colegue tried to make Liteplacer work properly, but found error in needle calibration. It is done on proper height (needle down) and focus is great. When using "by the book" calibraton routine, needle always had few milimeters offset (so much, that pickup of 0805 was like 50% :( ). With manual fixing of calibration valves, 1206 components could be picked up quite properly, but again not the 0805 and 0603 (where every second falled off, as it was only partially held).

I've read about this - will be checked tomorrow.
You can check it by measuring needle (uplooking camera setup, needle, click measure), move the needle a known amount (shift+F key is 1mm), measure again; the difference should be 1mm.
Any other ideas?

Regards,
Gregor

p.s. Replaced also creepy up-camera ... Had plenty of crashes with the one, that was delievered with LP. First top cam was replaced with C270 and problems on top cam suddenly disappeared! now also bottom camera is replaced (Adonis) and program seems to run with much less crashes then before ;) (though it still crashes sometimes for no reason ... I am using latest version)

p.p.s. Another thing to add in manuals: when mounting up-looking camera - how shall it be turned? Moving PCB (over it) left, shall move image on the screen left or right?! ;)
mrandt
Posts: 407
Joined: Mon Apr 27, 2015 10:56 am
Location: Stuttgart, Germany

Re: Can't figure out camera to needle calibration

Post by mrandt »

mucek wrote:Today me any my colegue tried to make Liteplacer work properly, but found error in needle calibration. It is done on proper height (needle down) and focus is great. When using "by the book" calibraton routine, needle always had few milimeters offset (so much, that pickup of 0805 was like 50% :( ). With manual fixing of calibration valves, 1206 components could be picked up quite properly, but again not the 0805 and 0603 (where every second falled off, as it was only partially held).
Is pixel-to-mm ratio / box size set correctly? Try to measure it by moving needle as described here or try RMOD / Ver2 with auto calibration and copy & paste determined values into your LitePlacer Ver 1 / Main version.

http://liteplacer.com/phpBB/viewtopic.p ... t=10#p1443

Big advantage is that this way the ratio is also determined for correct Z-level. Much harder to do with a ruler and if you measure at slightly different Z-level with HD camera you get quite a relevant difference due to zoom error.
mucek wrote:I've read about this - will be checked tomorrow.
You can check it by measuring needle (uplooking camera setup, needle, click measure), move the needle a known amount (shift+F key is 1mm), measure again; the difference should be 1mm.
Any other ideas?
If the measurement is off, it is a good pointer that box size / pixel-to-mm ratio is not set correctly. See above.

If that is not the case, I would guess either mechanical issue or camera-needle-offset - but suggest you try one thing at a time.
mucek wrote:p.p.s. Another thing to add in manuals: when mounting up-looking camera - how shall it be turned? Moving PCB (over it) left, shall move image on the screen left or right?! ;)
Find picture of default orientation here:
http://liteplacer.com/phpBB/viewtopic.p ... tion#p1276
JuKu
Site Admin
Posts: 1114
Joined: Thu Feb 14, 2013 3:06 pm
Location: Tampere, Finland
Contact:

Re: Can't figure out camera to needle calibration

Post by JuKu »

I'm just putting a note here taht I'm following this, although I don't have much to add to what mrandt already has said.

The needle calibration is done 1mm up from PCB level, on purpose. My thinking is that needle is not going down all the way to PCB level, the component is in between. Average small component height is maybe 1mm. But as noted, 1mm difference in Z should make only very slight difference, if any. If Z level makes significant difference, something is badly misaligned. It could be Z axis itself or more likely, the uplooking camera mounting or the chip inside the camera. If the camera is at fault, I'll replace it, of course.
android
Posts: 25
Joined: Mon Oct 12, 2015 12:30 pm

Re: Can't figure out camera to needle calibration

Post by android »

@mrandt
android wrote:
But needle is exactly centre at the height it used for wobble calibration(26.56mm) my pcb level is 27.56.


But you mentioned before that it moves out of center if you move it to 27.56 - or did I misunderstand that?
What I meant is that after wobble calibration>needle to cam>needle down(to PCB level), needle tip is not at the centre of up cam.Like Juha said wobble calibration is not done at the PCB level but 1mm up from PCB level(In my case PCB level is 26.57mm , so wobble calibration happens at 25.57mm). So if I raise my needle 1mm from PCB level(25.67mm) tip is at the centre of upcam.
There might be some alignment issues with either of my up cam or Z axis or both, because slight movements in Z axis makes noticeable shift of needle tip from up camera centre.I will check on that.
mucek
Posts: 40
Joined: Mon Oct 26, 2015 7:06 pm

Re: Can't figure out camera to needle calibration

Post by mucek »

Well, probably the most significant factor was wrongly turned up-camera ...
When I did mrandt's test (see his previous post) I noticed, that camera was flipped 180 degrees.

Well, I must now proudly say, that first boards are already placed and baked! (still hot ;) )
There's still plenty of things to improve, most important make trays for components: sticking them is really annoying and very quick it happens that some components jump out :(

Regards and thanks for all your help so far,
Gregor
mrandt
Posts: 407
Joined: Mon Apr 27, 2015 10:56 am
Location: Stuttgart, Germany

Re: Can't figure out camera to needle calibration

Post by mrandt »

mucek wrote:Well, I must now proudly say, that first boards are already placed and baked! (still hot ;) )
Congrats! :-)
Post Reply