Page 1 of 1
Downward Camea fixed to Z-axis?
Posted: Tue Feb 28, 2017 2:07 pm
by Mart
The downward camea, can this one be fixed to the moving z-axis. That it goes up and down along with the nozzle.
Has this too much influence on the operation of position determination?
Re: Downward Camea fixed to Z-axis?
Posted: Tue Feb 28, 2017 6:09 pm
by mrandt
Hey Mart,
with normal lenses the size of the projected image and thus the px to mm ratio is dependent on viewing distance.
So moving the camera up and down will make optical measurements impossible with the current software.
Also, you may run into problems with focus; as the camera is only sharp for a given distance.
TLDR: Moving camera up and down -> bad idea
Regards
Malte
Re: Downward Camea fixed to Z-axis?
Posted: Wed Mar 01, 2017 9:20 am
by mawa
Hy Mart,
one of the most important calibrations is the pixel / mm ratio of both cameras.
Due the the lens characteristics every mm distance can bring a change.
Just try for yourself:
Take a ruler, place it right on the table surface and let the video draw a grid. Measure the distance between the ruler ticks.
Now put a PCB unter the ruler. That brings the ruler approx. 1.6mm closer to the camera. Measure again and you will see there's a significant difference!
The better you calibrate this value the better your optical homing, fiducial and other calibrations will work.
That is also the reason why everything should be on the same z-level. Best choice IMHO is PCB Thickness above table surface. If you use Juhas "glue the tape to the table" method then you are almost right on typical paper tape and plastic tape with a pocket depth of approx. 1.5mm.
Re: Downward Camea fixed to Z-axis?
Posted: Wed Mar 01, 2017 11:13 am
by Mart
Thanks for the comments.
Software performs only measurements if the nozzle is raised? Or do I see this wrong?
Would like the nozzle and camera fixed together, so that there distance is a constant. Even though the z-axis slightly skewed.
Re: Downward Camea fixed to Z-axis?
Posted: Wed Mar 01, 2017 12:29 pm
by JuKu
> Software performs only measurements if the nozzle is raised?
Correct.
> Would like the nozzle and camera fixed together, so that there distance is a constant.
This makes good sense. However, you are really asking for the nozzle tip Z axis and camera optical axis to be aligned with each other, and we don't achieve that by moving the camera to the nozzle. Instead, we need a camera mount that has two axis tilt trims. With a clever design and a 3D printer we might come up with something usable. Calibrating that might be rather difficult, though.