Advanced Search

I have just purchased a amazon bob. It looks similar to the one you mentioned. As far as I understand most of them have the same pins on the db25 port.

For just a straight cut-off operation I think this is doable in hal.
(the unroll is done in hal, then the cut-off is just at the beginning of the Gcode)

If you want diagonal or other fancier cuts that gets trickier. I also expect that
the unroll operation would be less accurate than normal joint motions and
possibly unidirectional. It also needs to rehome when done.

 

It has been done (using interpolated position) but depends on very good
speed control of the spindle vs load, since you are essentially "flying blind"
for 1 whole turn.

With the software encoder you would connect the external index pin
to both 'A' and 'index' of the encoder component and set the encoder
component to counter mode.

A small improvement to deal with just one sensor would be to have a 1/2 disk
so you can get 2 counts per turn.

tommylight wrote:

Keep us updated, if possible, and take care.
 

+1

An index pulse alone is not enough you also need a pulse train signal like an encoder channel to track the spindle velocity:
linuxcnc.org/docs/html/examples/spindle.html

Can roller with 0-10v inverter be V axis for example and unrolling handled in gcode?
Any better appproach?

Moved the post with the second batch of images in here.

Yes, you don't need the hal encoder comp at all and unless you have a base thread
(not practical with an Ethernet connected device), the hal component will be quite slow.

Oh dear..My poor old brain.
So, the MPG is connected to the correct pins, but the input designation I should be using is perhaps hm2_7i92.0.7i73.0.2.enc0.count to get access to the high speed values?

So take the encoder count stuff out of the hal thread and instead couple
net mpg-out ilowpass.0.in <= hm2_7i92.0.7i73.0.2.enc0.count