#linuxcnc-devel Logs

Nov 20 2018

03:09 AM rmu: self-modifying bitstreams!
03:41 AM rmu: I have a strange problem with the recently activated automatic toolchanger, perhaps somebody can point me where to look. I'm using gmoccapy as gui.
03:42 AM rmu: The toolchanging process uses all sorts of inductive sensors for feedback and checks if pocket to put tool is empty, pocket is empty after pickup etc... , because of the design of this thing, it is easy to crash something if not carefully monitoring and waiting for these feedbacks
03:42 AM rmu: the strange thing is that programs that involve a toolchange seem to look at these feedback pins even while generating a preview
03:44 AM rmu: so given tool 0 is in spindle and program wants to change to tool 1 and then back to tool 0 it complains that pocket for tool 0 is empty
03:50 AM rmu: toolchanger is implemented with remapped M6
03:50 AM rmu: stdglue python prolog/epilog and rest is ngc files
03:51 AM rmu: second strange thing is that M73 in this ngc file doesn't work
03:51 AM rmu: while preview generation
03:52 AM rmu: i would have asked on #linuxcnc, but that channel seems to mostly contain noise
03:56 AM rmu: communication with toolchanger hardware happens via motion digital-io pins and M6x codes
03:57 AM rmu: programs execute fine per se, but i can't get a preview
04:03 AM rmu: 2.8 master, pretty recent
06:52 AM rene_dev_: rmu the macro needs a check for #<_task>
06:52 AM rene_dev_: otherwise it also does things in the preview
06:55 AM rene_dev_: o100 if [#<_task> EQ 0]
06:55 AM rene_dev_: (debug, Task is null)
06:55 AM rene_dev_: o100 return [999]
06:55 AM rene_dev_: o100 endif
07:28 AM rmu: rene_dev_: thanks. exactly what i was looking for. not sure why i didn't find it, should get more coffee perhaps
11:03 AM jepler: quick reboot on the forum coming up
04:34 PM rmu: i got tinkerboard "running" with hm2_spi and 7i90. needs some changes to hm2_spi or probably a driver directly for rockchip spi hardware like rpi, currently watchdog bites about once per minute
04:46 PM pcw_mesa: what changes would be needed for hm2_spi? its pretty generic
04:48 PM seb_kuzminsky: iirc most of the spi latency is in the kernel driver
04:49 PM rmu: pcw_mesa: 32bit transfers are not supported, so some byte-reordering and is needed
04:50 PM pcw_mesa: might be better to bit bang it if the SPI interface is so broken
04:50 PM pcw_mesa: or byte/nibble bang
04:51 PM rmu: like epp interface?
04:52 PM pcw_mesa: on some SOCs, GPIO can be toggled to 20 MHz or more
04:54 PM pcw_mesa: since reads are synchronous anyway, you dont gain much with autonomous SPI hardware (you gain a bit on writes)
04:55 PM rmu: datasheet says hardware supports 16bit transfers, and has 32 "location" fifo
05:05 PM rmu: the raspberry pi GPIOs could be toggled with 100s of MHz iirc, somebody built a software-defined fm transmitter that consists of a rpi and a piece of wire as antenna
05:55 PM Tom_itx is now known as Tom_L
06:29 PM skunkworks: We just found out that you cannot copy users and group files from a 16.04 to 18.04... (migrating users and groups and samba...)
06:30 PM skunkworks: but - if you copy the files from 16.04 to 16.04 it works -> dist upgrade to 18.04
06:30 PM skunkworks: it works
06:34 PM skunkworks: once that worked - I stopped looking for reasons why it didn't work from 16.04 to 18.04. The format of one or all the files must have changed..)