Correct, people have used pi cams, Logitech cams etc. The current consensus
seems to be a $20 usb microscope from Amazon because it has a ring light,
which solve problems people had with uneven lighting.
On Thu, May 6, 2021, 4:05 PM Jake Palmer <82.jake@xxxxxxxxx> wrote:
According to the TAMV GitHub page, it works with an inexpensive Pi Cam
which can be had for under $10.
https://github.com/DanalEstes/TAMV
On Thu, May 6, 2021 at 12:45 PM Jeffrey B. Rogers <sgtnoodle@xxxxxxxxx>
wrote:
I'm just brainstorming since the idea of using a camera to calibrate
nozzle offsets seems pretty cool. I haven't looked beyond that YouTube
video you linked, and so maybe these aren't fresh ideas.
I assume XY nozzle offsets are very stable for any given dual extruder
carriage, even across nozzle changes. Therefore, it's a procedure that
probably only needs to be performed once or twice for the life of any given
printer.
It would therefore be relatively expensive to build dedicated optics into
a printer for this purpose, and there's only limited benefit to having a
fully automated procedure. Smartphones are ubiquitous these days, though,
and there are plenty of apps to turn the phone camera into a wifi webcam.
It seems like it would be really slick to implement a partially automated
procedure in klipper to determine XY nozzle offset via a smartphone laid
down on the print bed. It seems like the only potential hardware need might
be a macro lens, for phones that don't have great near focus.
You potentially wouldn't even need to do any feature detection, and let
the human do it in a sufficiently streamlined way. The biggest hurdle
perhaps might be the UI, since klipper itself is headless. The
implementation could perhaps itself provide an MJPEG stream and overlay
reference markers and instructions on top of the camera feed, allowing folk
to use whatever media player they want, until someone gets around to
implementing an octoprint plugin.
It seems like such a procedure could be done right now in a manual way.
Put smartphone on bed, load video up via VLC or even a video conference
app, jog nozzle one into view and stick a post-it note or something on your
screen to mark the nozzle center, then jog nozzle 2 to the same center, and
note the XY offset. The benefit of adding klipper support for it, with
feature detection, would therefore be convenience and higher accuracy.
On Thu, May 6, 2021, 12:20 PM Steve Wallace <implicitconversion@xxxxxxxxx>
wrote:
I'm working with several people to implement Danal Estes TAMV optical
tool alignment on klipper. Currently it works with Reprap firmware. It is
super accurate and quick. Check out his video demo
https://youtu.be/e_d_XHwGfRM
This currently gives use x&y alignment. He also did work on finding z
using a physical tap. He had explored using the nozzle to touch off but had
issues with plastic buildup on the nozzle causing issues.
I am planning to get back to work on it this week after I move my shop.
On Wed, May 5, 2021, 12:21 PM Dark Empathy <darkempathy@xxxxxxxxx>
wrote:
Hi All,
I have a dual extruder core XY printer, using TMC2130's with sensorless
homing in XY.
I would like to try and implement automatic Nozzle X/Y offset
calibration.
One possible technique might be:
1) Take a metal tube approx 5mm diameter, and cut 2mm off the end, so
we have a 5mm x 2mm "ring".
2) Epoxy this ring to an "unused" corner of the build plate (or any
other method - such as taking a longer length of 5mm tubing, and installing
it, vertically just off the build plate - but in an area the print head can
reach)
3) Move E0 to just above this cylinder, then lower nozzle tip so its
just inside the cylinder.
4) Then move the print head slowly, can calibrate the exact nozzle tip
position via back EMF / missed steps, in the same way the sensorless homing
works.
5) The repeat for E1, and then we will have the position offset between
E0 & E1.
It remains to be seen if this will be accurate enough for Nozzle
Offsets...
Questions:
- Has anyone done this already, or some other solution?
- If not, (I'm far from as expert as many people on this list), can
anyone advise where & how I might implement the above in Klipper?
Any pointers appreciated,
Thanks,
Julz.
