Another tab for WebCam

Yes, that must be it.

I had the same git pull on all 3 machines (quite an old one, but very stable).

About a month ago I did a new git pull on that machine to update something else, didn't realise the tabbing code had been incorporated.

Will test later, but I am sure you have it.

regards

Hi

Removing the tab embedding code from .axisrc worked with 2.6~pre. run on Ubuntu 8.04

Next problem was that emc halted with errors related to being unable to find emc.halio.
I had already removed the emc.halio from the chain in camviewcfg.

I edited emc-camview to prevent it from including emc.halio in the default internal chain, but it was not finding
camviewcfg for some reason, despite it being specified and being in the working directory.

Eventually found that the .ini entry required a fully qualified path to the config file, as in
EMBED_TAB_COMMAND = camview-emc -C /home/bob/emc2/configs/Easimill/camviewcfg -w {XID}
(even ~/emc2/configs/Easimill/camviewcfg did not work)

Now fully up and running on 8.04. If I need the DRO values visible whilst using it, I can create pyvcp displays
to show them.

Thanks again for your work

Hi

Had the camera out today, this is what the camera I recommended looks like mounted in a holder.
Holder is 36mm dia with 28mm bore for camera and a 20mm stub for mounting in a collet or collet chuck.


regards

Hi

Just an update to testing, if anyone is thinking of using this camera.

The focus/zoom ring is not exactly concentric to the body, this means that altering the focus to sharpen the image will move the cross hair centre.

I have roughly focussed to the distance I want to work at and then super-glued the focus ring. I can make fine adjustments to focus by moving the Z axis up and down.
This solves one problem but not all.

You will never be able to remount the camera holder exactly as previously in terms of rotation within the spindle collet. Because the focus ring is not concentric, this means there will never be repeatability between mountings.

The overall effect is that the camera is not suited in this mounting to absolute positional measurement, but is perfectly good for incremental / relative measurement or positioning.

By which I mean, you cannot mount the camera go to a reference mark on the mill bed and be at the same co-ordinates every time.
There will be small inconsistancies no matter how careful you are.
You would think that you could go to the co-ordinates for the mark and tweak the camera rotation until it lines up perfectly, but it never seems to quite work like that.

But having established a reference point and zeroed co-ordinates to it, all subsequent measurements or moves are very accurate.

I will assess how much use I actually make of the camera and if it becomes integral to my working methods, will have to consider mounting within a rectangular block fixed to the side of the head, with parallax calculations made to arrive at a true centre point.

For now my use has been in making accurate measurements of existing components to duplicate them and it works well in this role measuring from an established datum point.

regards

Just put mark on camera and the body of the spindle. And when need absolute measurment's just be sure to align marks before it.
But to be shure check if the camera is paralel to the Z axis.
Put some mark on the table and move camera as close as possible to the mark. (have mark in the centre of crosshair). Now go as far as possible (Z axis) and check if the mark is still in crosshair centre.
I have camera permanently in body of spindle and use constant offset.

Hi,
It's a R8 spindle so I have a locating groove plus my spindle has a locking latch for tool removal.
So if I bought another R8 20mm collet and left it permanently mounted in it, that might improve matters.

The inconsistancies are not huge, but enough to prevent any confidence in accuracy.

regards

Hello ArcEye,

It's a R8 spindle so I have a locating groove plus my spindle has a locking latch for tool removal.
So if I bought another R8 20mm collet and left it permanently mounted in it, that might improve matters.

That is pretty much what I did with a touch probe. Works well, but I had to make sure the shaft was a good fit for the collet (not slightly too small) to be accurate.
I also mounted the probe in the holder with 4 set screws 90 deg. apart so I could adjust the probe to center and lock in place.

Nice work.

Rick G

look at www.miketreth.mistral.co.uk/centrecam.htm

he has a way of centering the crosshair such that axis inaccuracies are corrected

I think software compensation is the way to go with cheap cameras and optics

-m

Hi

Yes it's interesting, just a pity it is windoze only.

Looking at psha's crosshairs.c in the camview-emc sources, building an alternative plug-in with adjustment capability should not be beyond the wit of man.
Might take a bit of experimentation to work out what size offset units are required, I suspect they will be very small.

First I am going to try to iron out as much as possible mechanically.
I have ordered another 20mm R8 collet and will go for a permanently mounted holder in a collet.
Between the R8 collet locating groove and the spindle lock, that should give as repeatable mounting as I am going to achieve.

I may also go for Ricks 90 degrees apart set screws, camera to holder, to give a little adjustment plus extra solidity and see where I get too.

regards