damn I hate leaky capacitors...
$0.50 parts destroys a $3000 radio
leaky cap ate a hole in the PA board
the brown slime eats right through the soldermask and the copper
sometimes u can repair that with a good iron and magnifying glass
not sure if cleanup with alcohol will stop corrosion effects
so there used to be a tarball of the cvs-head version... did those go away?
make a paste of making soda and a lil water, use a acid brish or wtip and scrub awat the corrosion
maybe, but I'm not sure if the slime is caustic or acidic
rinse clena with denatured alcohol
baking soda should clean both and act like a mild abrasive
Anyone familiar with dacport IC's? Lets say this one for example. http://www.datasheetcatalog.org/datasheet/analogdevices/AD7248AQ.pdf
* jmkasunich is making steel wool
actually, making a steel "cup" that is 0.875" diameter, 1" deep, with 0.020 thick walls and bottom
dang small walls
hidden compartment or something?
yeah, something like that
I been running the index b60 all night
one piece slips over another, and looks like there is only the one
i'm running trunk and noticing that halscope is segfaulting frequently often when a button pushed or cursor moved, hard to reproduce.
sometimes there are messages before it crashes: http://pastebin.ca/1347450
when it faults, the halscope application cannot be restarted (unloadrt cant unload it) so I have to restart machine
there has been a bug lurking in halscope for a while, I think since roll mode was added. are you using roll?
(darn, I wish those assertions meant something to me)
i sometimes use roll and often start with roll but the fault can occur when not in roll mode
i updated libglib and libgtk today -- nohelp
ack, I need to go to a faster machine, brb
this evening i was testing with emc-machine in garage and running halscope with
its X display directed to my laptop in the house. i can still get some
assertion errors (they may be different): http://pastebin.ca/1347455
but it doesn't crash (as much? not sure)
so you think it might crash more with a slow X display?
if you get a sig11 even remotely repeatably, you should run it in gdb, or get a core dump
or xsyncing --
... a backtrace is worth a lot
do you know how to do that?
pretty sure it's already built with debugging
i did that once but there were no symbols in the lib,
oh, hm. the sig11 is inside some gtk library?
yes libgtk i think (not sure i'm remembering right, it was a few days ago i ran with gdb)
i installed a debug package: libgtk2.0-0.dbg but haven't understood how to use it
the stuff is /usr/share/doc/libgtk2.0-0-dbg is no help at all
yeah -- that's what i thought too
it seems the problem must be recent because i
never had a problem with halscope before running trunk (i've been using 2.2.6 i think)
the gtk libs are in such sorry shape that valgrind is useless
Do any of you use rfqwork.com?
I been quoting some jobs for my index-werke b60 but the jobs are few and far between mostly less than 200 pcs. I been looking for an alternative Request For Quote website, hoping someone in here would know of one.
I have this machine retrofitted with a CNC, if any of you guys need some production parts done. http://www.youtube.com/watch?v=AdxQLEOqed4
running halscope with gdb just now (remote X display):
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb744a6c0 (LWP 13265)]
0xb777fe66 in ?? () from /lib/tls/i686/cmov/libc.so.6
roll mode was not involved in this instance
would you put the whole backtrace on pastebin please
ok, another occurrence, no roll mode involved, with backtrace: http://pastebin.ca/1347483
yuck, that is a very screwy stack
years ago i did a lot of debugging of threaded code (solaris) and sometimes the error reported is not that close to the real fault
yes very true
but half the time you get lucky and the backtrace shows exactly the bug
I wish it was more than half...
==14771== More than 100000 total errors detected. I'm not reporting any more.
it's pretty clear to me that scope is attempting to draw unitialized/uncaptured data sometimes
what is this error report from?
trying again with gdb, same setup as 12 minutes ago, can't make it fault so far
but can stimulate similar assertion errors
got this backtrace and more closing the app : http://pastebin.ca/1347503
wondering if i am the only one seeing problems with halscope?
no, I've seen it crash a few times too.
I'm 100% confident it has something wrong
it's just very obscure
ok, i was wondering if i had messed up something on my machine
cradek: it's the flux capacitor that is the issue
it's flixing instead of fluxing
dgarr: I'm not finding it... maybe jmk or jepler will have some insight.
any way you can find to trigger it reliably would be very helpful.
(as I'm sure you already know)
its hard to reproduce now.
i know -- these are difficult problems
thanks for looking at it
welcome - I'll keep trying.
hey all. i have a crazy neighbour that yells at her tv. i have an exam tomorrow, and i really need to get to sleep, except that i can't because all i hear is her tv blaring and her screaming at it. is there any way i can interfere with her coaxial signal somehow? the wire runs down the side of the building, and i have access to the bit that services her directly.
i started cutting through it, but then thought against it. i was thinking maybe there was a way to short it out? an aluminium pin, perhaps?
no, it doesn't work that way
they're transmitting RF on the copper
plus you know, messing with everyone else on your local catv ring
i suggest not messing with it
best way to interfere with tv screaming is to call the cops
always worked for my neighbors when we were having a party in college
toastyde1th is now known as toastydeath
archivist btw not cnc but have u seen http://www.sdr-kits.net/VNWA/VNWA_Description.html
- looks quite an interesting kit
I like all the warnings :)
they are dealing with hams!
bunch of amateurs!
I would love to test against the real thing
shame about the windaz only
well again, its for hams - the art of self improvement / learning / experimentation has been totally removed from the hobby
did you go for the cheap 8410 system on fleabay I pointed to you?
archivist no i didnt, i found out that the guy locally who had one give it to the local ham radio club, so i can get access to it there
plus no one else there knows how to use it
even cheaper I suppose
I wonder if one could use http://emcar.sourceforge.net/
and parts of it with that project
ah yes i saw that before, an excellent project
i did a similar one in turbo pascal a few years back
btw archivist on your machine, do u have emc controlled spindle speed?
I need it
drive has 0-10 volt in so it wont be too hard
yeah i'm currently using a variable freq pulse train output from the PC to drive my vfd
but gonna try the pwm output instead
see if its a little smoother
i need to put a breaking resistor on the vfd too, but my last experiments resulted in a crater forming in the switcher module in the vfd
I have a different drive that I could use with a braking R
yeah as far as i understand, i take the DC bus voltage / motor current to give R
then find a high wattage R of about the right value
i did have a selection of those big green ceramic R's about 6" long and 1" dia, but buggered if I can find them
gah TV dropper resistors
heh not seen those for a few years
* archivist was a TV engineer early 70's
but its a good idea
I have a few power R's available
ignore prices http://www.part-number.co.uk/partsubindexE1RES.html
and I may have raided and used since I listed
archivist ok i will have a pick through that list, i need to calc the value first
pjm_, its how I get rid of surplus, about 2 sales a year :(
i should do the same
i have far too much crape here
just none of it useful ;-)
with impending redundancy I need to make it work
a mate of mine found a bag of 200 opto switches at work in the weee bin
so i should be getting those for encoder making etc
"none useful" is what I thought of 100+ video tape motors, but I sold the lot to Poland
i did all the mechanics for a simple 8 ppr + index encoder just to have a play with closed loop spindle speed before I get a mesa card
and he didnt haggle at the price :))
some stuff like the resistors I use myself and dont regard as a sellable item
any way off outside for a bit of yard clearing here
rob btw are u the guy that posted the question on spindle load monitoring to the list?
hi PJM yes i am
pjm_, ill be back around later
gdb is not showing all the variable with 'info locals' - it shows some of them - it gives no explanation why it does not show the others - any ideas?
are you compiling with -O0?
also, good morning
no - will try it - compiling with --enable-run-in-place --enable-simulator
in the Makefile replace any -Os with -O0
(it's not a configure option)
I remember now - you said to do that before - thanks
cradek: that did it!!!! - many many thanks
jepler: re the [DISPLAY]GEOMETRY: i verified that cvs checkout was unmodified, but same misbehavior
on two machines: one RIP using simulator (not RTAI kernel), the other RIP (and RTAI).
then, I did make clean; make and display with liveplot on or off seems ok on both machines.
dgarr: neither jepler nor I can reproduce a problem now. can you give a careful step-by-step for me to follow, please
hmmmm .... pretty quiet!
not that I have anything very exciting.
However I do need to know what branch of the cvs to download to make certain I get the hm2 support
cradek: understood. to clarify: after make clean && make i see no problem. didn't understand that make clean would be needed for this case
sorry for the trouble
ugh, a bug went away with "make clean"? that makes me unhappy :(
dgarr: is your clock set right, and has it always been recently?
clock is right on both machines
it's always been right as far as i know
hello everybody, donno if anyone remeber me (wasnt here that long). im however glad to be back and recovering from shouder surgery. I had promised to compleat the swedish translation and intend to do so.
neat, it's great to have translators. glad you're doing well.
last time around i was about to try useing a single signal per rev spindle sensor and think someone told me i would need to recompile emc with some exsperimental code, yes ?
yes that's not in a released version yet. it will be in 2.3.0 which is due soon.
allright, the code was hosted somewhere temporearly before i think ? where can i get it now.
[16:43:59] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2
"encoder23" is the version of the enhanced encoder component that can be compiled to work with emc2.2 using "comp". http://media.unpythonic.net/emergent-files/sandbox/encoder23.c
Instructions on compiling a component: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#How_to_compile_and_install_a_component
or you can follow cradek's instructions and upgrade to the CVS version of emc, which is getting close to "beta release" status
thanks for the info
I think that the last time, encoder23.c is the route you were trying. Here's the old IRC log, maybe it helps jog your memory too: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-10-19.txt
starting at around 14:37 it looks like
now I am recovered to where I was two weeks ago. we just received a package of work to quote - 1000 part numbers - darn work - I will be BUSY the next two weeks
busy is good
busy is very very good BUT this means no EMC stuff - ugh!
I would rather have it this was as no work also means no EMC stuff
we are preparing a room for the intense quoting - it is good to see people scrambling around with something to do
now we have to WIN some of this work
when is 2.3.0 estimated to be released ?
[17:21:57] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Emc2.3Status
cradek: I've been using your roll the dowel method to set my Z on my mill and like it better than the paper trick.
however my Anilam control is not as good as EMC and I have to do the math in my head for the dowel thickness
after touching off
what's the roll the dowel method?
stuart showed me that trick. it's better to move up than down - you can't break a point off the tool.
I put a 0.500" dowel on the work and lower the tool to just below and push the dowel up to the tool then raise it 0.001 steps until the dowel just rolls under the tool
as you get close to the 0.500 the dowel really moves
yeah that's why I like to use a big size. it's really easy to feel
meh.. crappy keyboard
I have a keyboard like that also. It types the keys I hit instead of the ones I want to hit. :)
stustev: indeed that was the case ;)
how can it not know I want to type /away
(especially since I type that daily)
maybe we need some 'hot' shorthand keys - that won't help - the shorthand will just be wrong :)
I don't want a keyboard to type what I think - that would be very dangerous :)
I saw some nice CNC machine today
some Mori Seiki iirc
[19:00:08] <alex_joni> http://sos.noaa.gov/equipment/
cradek/jepler, i must be doing something wrong, ive added the encoder23 and ive added the bits and peaces i found at http://firstname.lastname@example.org/msg01003.html
im however reading very low values from the hal meter when probeing spindle-speed
Kohlswa: is the read/update function connected to the base thread?
addf encoder.update-counters base-thread
addf encoder.capture-position servo-thread
what is your base-thread period?
ugh, sorry, got disconnected,
[20:21] alex_joni: Kohlswa: is the read/update function connected to the base thread?
[20:21] Kohlswa: addf encoder.update-counters base-thread
[20:21] Kohlswa: addf encoder.capture-position servo-thread
what is your base-thread period?
Period FP Name (time, Max-Time)
99733 NO base-thread ( 5433, 30687)
how about in the ini?
BASE_PERIOD = 100000
that's awfully slow
100.000 usec = 100 msec = 0.1 sec
that means that your BASE thread updates 10 times / second
oh ? ive just used the setup wizard to create the configureation.
so you can count encoders up to 5Hz speed
alex_joni: Kohlswa7578 is trying to use 1PPR threading, so the signal frequency is probably not the problem
Kohlswa7578: the "position-estimated" output of encoder is in revolutions and "velocity" is in revolutions per second (not RPM); is it possible that everything is OK, but you were expecting to see RPM?
5Hz means max 300 RPM..
well under spindle-speed useing the hal meter shouldent that be rpm ?
what is spindle-speed?
and BASE_PERIOD = 100000 means 100us -- are you sure you had your coffee today, alex_joni ?
jepler: I had a long time ago :)
max speed my small spindle should be about 2800rpm.
this is what i added to my hal file.
setp encoder.0.counter-mode true
setp encoder.0.position-scale 1.0
addf encoder.update-counters base-thread
addf encoder.capture-position servo-thread
net spindle.phase-A parport.0.pin-15-in => encoder.0.phase-A encoder.0.phase-Z
net spindle-index-enable motion.spindle-index-enable <=> encoder.0.index-enable
net spindle-speed encoder.0.velocity => motion.spindle-speed-in
net spindle-pos encoder.0.position-interpolated => motion.spindle-revs
spindle-speed is in RPS then
ah, that adds up
Kohlswa7578 is now known as Kohlswa
well im off to bed, thanks again for all the help
SWPadnos_ is now known as SWPadnos
I made a new entry on the wiki for adding SPI support on to hostmot2. It is under "Misc stuff" in the table of contents.