jmkasunich: I remember that when we were testing the tp, we saw that the Xacc signal was very "noisy". I have been trying to convince myself whether a 'dddt' block, which produces both a first and second derivative, and holds the temporary first derivative value as a double, would be any better.
I am leaning towards doubting it, because the trouble is that the 'input - last' difference may not have too many significant bits when calculating the first derivative
I just had an interesting idea
'loadrt ddt names=dx,dy,dz'
it would create instances named ddt.dx, ddt.dy, etc
seems like its something you'd either want to do across the board, or not at all
my "someday" goals for HAL are "on request" instantiation of blocks, rather than using "count" or something like it to make N of them up front, and then being stuck with that quantity
if/when that happens, it seems only natural that the "newblk" command would allow the user to name the new instance
if you are working interactively that would be ideal
perhaps the "names" thing should be deferred until then?
if you are assigning names, you might want the user assigned part to be a prefix, not a suffix, for example
so halcmd show X would show you X.ddt.blah and friends
[02:38:46] <jepler> http://linuxcnc.pastebin.ca/166621
ah, so you completely drop the ddt and use the user specified names only
that could work
then *ugh* you could implement 'newblock' by doing a partial 'save', 'unloadrt', 'loadrt names=....,newname', and then 'load'
ugh is right
newblk (and newthread, and a few other things) needs deeper changes
having undoubtedly broken the tree, it's my bedtime...
what do you mean by "provide fperiod"
bedtime for me
do you maybe know how to install a certain version of a deb?
apt-get install ppp version ?
thanks chris ;)
* cradek gently reminds alex_joni about the apt-get manpage
I DID look there
but too briefly ;)
[13:43:18] <alex_joni> http://dsplabs.utt.ro/~juve/blog/index.cgi/projects/01157974398
cradek: how does that look?
very interesting! welded and powder coated?
have you measured how upright it is yet?
any plans for a spindle?
welded and powder coated
didn't measure yet, not sure I have with what
I think I'll put a probe on it ;)
or a very small dremel
wanted to ask you one thing though
on the PMinMO design there's a variable resistor to set current through the stepper
how do you set that up?
I can think of a few ways - easiest would be to put a scope across the current sense resistor
or just set the reference voltage according to the datasheet (I think there is even a test point on the board for that)
does it use a 1-ohm power resistor? If so, set the voltage out of the variable resistor to 1V per amp
(I don't remember how I set mine)
there are 2 .5 Ohm resistors
those you mean?
ok, so 1V / amp ?
no, not with .5 ohm resistors
the .5 ohm resistors are used to sense the current
this is compared to Vref (the var resistor)
so it would be .5V/amp
cradek: found a small problem with the install script
just wanted to install it on a friends PC (dapper), but he had no new line at the end of /etc/apt/sources.list
which cause the deb http://www.linuxcnc.org/
dapper emc2 line to get added to the end of the previous line
text files that don't end in a newline are broken
instead of "fixing" the script, this user should be chastized so harshly for having such a sources.list that he switches to Windows XP
I halfway agree with jepler
I don't think that's a problem with the install script
>> is a well-established way to append a line to a file
do you know what put the last line in his file? was it a ubuntu program, or the user?
but maybe we can echo an empty line before the deb.. ?
it really can't hurt
or maybe a comment : #emc2 repository
I guess I don't mind if you want to add an empty line
but you should also tell your user to not do that...
I would like to set up something to automatically put the latest PDF and HTML documentation on the website .. built nightly or so
I can run the build on one of my systems, but I'd need ssh/rsync access to the web server
is this doable? Who would I ask?
I think I have access to do that
you'd have to ask swp
I wonder if it would be better to have jmk do that from the compile farm
I think he's already set up to put stuff on the webserver (the farm output)
*** Fatal Error --- but easy to fix ***
Cannot have '.' in file-name prefix, else dvips fails on images
Change the name from ../docs/src/code/Code_Notes.tex and try again.
huh I haven't seen this error before
(trying to build docs on a fedora machine)
[15:56:43] <alex_joni> http://www.linuxcnc.org/docs/html/
is that an autocompile?
no, it was done manually
but I hope it will be automatic in the near future
yeah, sounds nice
huh some of the images are just not right at all: http://www.linuxcnc.org/docs/html/hal/rtcomps/index.html
jepler: did you try optional stop again?
alex_joni: no, not yet
jepler: ok, no problem
* alex_joni will be away for the next 2 days
travelling again? you seem to do that a lot
yeah, short trip to germany
I'm back on thursday though, so it's ok
have a goodt rip
jepler: only tomorrow morning
tonight I'm still around :)
appears that the EMC version of classicladder-0.7.3 was run through a
formatting program of some sort, because when I do a diff on the two
directories, I get large sections of basically every .c and .h file
cradek: that is true
it was part of the old build system
how many times do I have to say it!?
what? that formatting is bad?
I feel bad for him
I think it's nice of him to try to help on that part :)
it's a tough thing to do though
yes, and made 100 times harder by some person running indent
maybe he can run indent on both versions now
to get something consistent
then apply the changes on the untouched latest version
yes that might be the least painful thing to try...
anyways.. I'm off to sleep.. early start tomorrow
not sure when I'll be back online
have a good trip