wouldn't that be afternoon?
Lerneaen_Hydra: I just got up.. so I take it as morning
and yeah, it's 4PM
lots of late night coding? ;)
no, actually I went to bed quite early (about 1am)
but a good 15 hour sleep once in a while is really nice :D
03jepler 07HEAD * 10emc2/src/Makefile: fix build error 'cannot find -lemc'
machine levels are from hell.
* anonimasu has just concluded that
cradek: what was that about two CPU's?
Lerneaen_Hydra: it would be nice if the realtime stuff ran on one CPU, while the other stay free for userlevel stuff
might get some good increase in performance
indeed that would be nice
paolo says you can get latencys < 10us that way
ah, that sounds nice
no, but it's hardware issues not software
can't get the second processor to work at all
it's old hardware
so it could work if you had working hardware
it would be nice to test anyway, might be a way to get very fast step rates
if someone buys me a fast pentium D board/processor/ram I promise to test it
shouldn't there be many older 600mhz dual processor workstations available for cheap?
that's what mine is
the rtai author
I got it as one processor and tried to find the additional parts to plug in to get two
cradek: and that particular one happens to be broken or all older ones won't work?
that sounds difficult
I can't say for sure what's wrong with mine, not enough information
'There were errors detected during the LaTeX run.' -- fresh-ish kubuntu 6.06 system
What IS a CNC for?
OK, got it now...
forget I asked
I was just trying to type up an answer for you
alex_joni: it needs the 'tetex-extra' package, apparently
I think cradek ran into this too
I think I have it installed
sounds like a lyx dependency problem
I can easily add tetex-extra to build-depends
but it would be nice if configure would detect the missing file
(I don't know exactly what file it is)
I think apt-get install lyx should pull tetex-extra
and I suspect it's rather a bug in their dependencies
lyx depends on tetex-extra
[15:37:06] <alex_joni> http://packages.ubuntu.com/dapper/editors/lyx
hm, I probably installed lyx-qt, not lyx
try installing both
good morning all. little help? running axis with new head build of this morning. Tuning gives error "error: can't read HALCMD, no such file......"
excuseme. should be "....no such variable....."
hm, it's possible (likely) that I broke it. Do you know about how old your previous checkout or update was?
couple of days ago.
not sure it worked then. can't remember using Tuning.
by the way. the Tuning thing is very nice feature!!!!!!!!!
hm, how are you starting tuning?
from 'machine' and then 'calibration' in axis
is there another way?
that's what I tried, but with the 'sim' configuration it doesn't do anything useful
it gave an error when you pressed 'Test'?
it gave me errors on steppers
where it's not really usefull
but I suspect it might also be a ha l
but I suspect it might also be a hal incompatibility
installed vs. run-in-place
mine is a run-in-place
03jepler 07HEAD * 10emc2/tcl/bin/emccalib.tcl: fix leftover use of $HALCMD
a few weeks(?) ago I made some changes to the way the scripts (including emccalib) find the emc2 installation. I missed some stuff in that file.
try updating and see if it is fixed for you
Jepler: all better now. thanks.
glad I could quickly fix it once I knew about it
PID tuning is in the news again I see
wild numbers in the posted ini file, really high P and almost no I
Tuning is that way.
Trying a make on the current checkout to see how the doc stuff works.
WARNING: Included file `~/emc/emc2/docs/src/config/ini_homing.lyx' has textclass `article' while parent file has textclass `book'.
yeah, I've seen that, but I'm not enough of a lyx guru to know what it means
I think we use book for everything (except one or two chapters)
That must be the case. I'll see if I can find it after a bit.
The lyx stuff does seem to slow down the completion of make.
takes about a minute or so to make the docs here
but just like the code, it will only make the docs if the source (.lyx files) has changed
so when somebody edits docs, make will rebuild the pdfs, the rest of the time, no delay
ah that doesn't sound so bad.
* Lerneaen_Hydra pokes jmkasunich
you don't recognise this mill by any chance?
[18:07:27] <Lerneaen_Hydra> http://www.brannehylte.se/upl_prod/1033.jpg
I have almost zero exposure to commercial small machines
most of what I know about commercial machines is from seeing large ones at the industrial surplus places
someone said you might. I can't seem to find anyone that recognises it, other than that jymmm used one long ago and forgot what it's called
the orange color reminds me of the EMCO cnc lathe that was at the cnc workshop
I tried searching for emco 45 cnc but didn't find anything, so probably not
infact, the blurry name at the bottom left of the control panel _might_ be EMCO
you think so? I found the leading letter to look like a T but it may be a logo/icon or something
the rest looks like it may well be emco
definetly CO at the end though
I guess sometimes the surplus place does have small mills - this is cute:
[18:14:24] <jmkasunich> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4704937&searchtable=1&sortExpression=&SortASC=&pageSize=50¤tPageIndex=0
expensive, but cute
hey, not that bad
sounds far too pricey
then again, from 1995
EMCO cnc lathe: http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4696267&searchtable=1&sortExpression=&SortASC=&pageSize=50¤tPageIndex=1
I don't think the color matches after all
hey, that's the lathe I have
except I don't have a toolchanger
mine were much cheaper though ;)
yeah, prices at the surplus places have been rising lately
I got two of those and some odds and ends for a cake :D
though I gave them the cake afterwards
I got them off a cnc distrubutor
they had been lying around for several years and they didn't know what to do with them
Lerneaen_Hydra is very lucky
indeed I am. though it certainly helped that a teacher from my old school came with me when I went there, as he knew many of the people there from before
still, a good deal nonetheless :D
they were easy to retrofit too
Learneaen_Hydra: try looking for Terco 45
[18:22:44] <jmkasunich> http://www.desktopcnc.com/out_of_prod/old_terco.htm
that looks just about right
from that page: "hard to tell for sure because their site is in Swedish"
you should be able to handle that ;-)
I can't say they seem to have much information on their site
hmm, the mill seems way to small/weak for anything worth buying
thats the trouble with commercial sites - they have no incentive at all to support older products
40mm throw in Z? that's laughable
barely enough to toolchange
fine for 2D engraving, but thats about it
Howdy Good People of Earth, and you guys in #emc too!
I need something more meaty, say 200x100x100 or so, and 500-1000W
is that spindle wattage?
yeah, I would like around 1hp
thats a pretty powerfull spindle for such a small workspace
1HP = 746 watts
guess I've been spoiled by the 750x350x600 and 24kW spindle mill at my old school
24 tool toolchanger, the works
I was working with a motor the other day that would make a nice high speed machining spindle
bit big though
160HP at 12000RPM
18m^3 or so
sounds quite good for cutting alu
bring it on..
the intended application is actually a vehicle
diesel-electric hybrid garbage truck
oh, I see
driving only wheels?
or is that the generator
yes, two driven axles, one motor on each one
diesel engine drives the generator
[18:33:20] <dmess> http://www.eclipseaviation.com/
hmm, seems strange to choose a motor with such high speed (rotational speed)
makes for a big gearbox
not really - high speed improves the power/weight ratio
motor size and wieght depends almost entirely on torque, and very little on speed
jmkasunich: Got any low speed motors over there? Say in the 3" diameter range?
the gearbox is one-speed, and the motor is designed for a very wide speed range
oh, so 12k rpm max?
Jymmm: nothing I work on is that small
not optimum rpm?
I think it can deliver 160HP from about 3000 to 12000 RPM
below that its constant torque
yes... we me and a friend are working in a Hydrostatic drivee from 1 electric engine... 4 motors 1 on each wheel
oh. ok. that makes more sense
jmkasunich: I figured, but hand to ask. Would love to find a direct-fit low-speed spindle so I can use my tapping head.
Jymmm: I've got some motors (at work) where the _shafts_ are 3" diameter....
uh, is " inch or feet?
" is inches, ' is feet
oh, then that's not too bad
I was thinking three feet diameter
makes for a big winding assembly
you really think big!
i seen a 66 foot vertical lathe...
the biggest motors I've seen (in person) were at a steel rolling mill
up to 45-60 foot tall pieces...
shafts were probably about 24 inches diameter
jmkasunich: after working at large bearings SKF one thinks that axles are always big
i built many like them
the largest ones they make are 2m in diameter
ive tined 2 M jounals..
those steel rolling motors had something like 100000 foot-lbs of torque
9000HP at only a few hundred RPM
and then, for the first roll stand, they geared that down even further!
[18:40:41] <Lerneaen_Hydra> http://www.skf.com/cmimages/056952.gif
<-- one of the larger ones
Well, lets see if I got lucky this time... Hit the electronics fleamarket this morning. Got a Dell P3 600 256MB no onboard video or audio, and I sealed PCI parallel card for $18. Got Ubuntu installed, well have to see if emc chokes on it.
jmkasunich: so things could get nasty in that roller?
input: red hot steel bar, about 300x300mm
the are not forgiving... ive seen accidents...
output (after 6 rolls): steel sheet, about 2000mm wide x 1mm thick
oops, I'm confusing two mills
there must be a large size delta between each stage
ours were I beam 36" x 16".... gross loud and VERY aggressive..
that one was about 1000mm x 50mm thick going in, and 2000mm x 1mm going out
the steel probably looks like putty :D
jmkasunich: shorter when coming in?
the other mill was a rod mill - 300x300mm going in, 20mm diameter round going out
the incoming area is 25 times greater though
input is moving at a slow walking speed
ours was continuous casting... came thru 300 or so rllers as gravity let the molten steel thru a copper die
when you squish down the cross section, the length increases
yeah, shorter when coming in
slow walk speed going in, 40 miles/hour or more (60-70km/hour) going out
longer coming out
dmess: my father worked on a continous caster - very impressive stuff
where there surface imperfections so you could see the speed of it?
yes, especially on the input
glowing red, with scale and such on it
to build.. until you seen one.. its hard to imagine the scale...
we sent 1 mill to china... MY bad... ; (
they had armature ammeters (these were DC motors)... the needle would be quivering around zero before the steel entered
as soon as the steel hit the rollers, the needle would bounce up to 15kilo-amps and start quivering around there
GO til the pour was done.. no breaks...
or was it 5KA, on a 15KA full scale....
must have been 15KA full scale
lots of power nonetheless
we didn't even attempt to directly control the armature power
we had 300 amp SCR bridges controlling the field winding of a generator
the generator armature was connected direct to the motor armature
generator spun at a constant speed by a big AC motor
we went in to clean up a re-bar bust out once.. they figure we cut and slashed out 750 KM of rebar in 2 days..
we control both generator and motor fields to control the line speed and maintain the proper tension in the steel
missed a guide... spaghetti EVERYWHERE
dmess: oh no...
oh, at a rolling mill?
jmkasunich: sounds like an effective way of controlling
but it takes a lot of rotating machinery
and with motors that big you nearly have 100% efficiency
there was one (very big) room that had 4 AC motors (22KV, 14,000HP each)
each AC motor drove three generators, total of 12
two generators were paralleled for each roll stand (6)
each stand had three motors
so 4 AC motors, 12 generators, and 18 motors, all in one room
total close to 150,000 HP
IIRC, they had 300HP motors just to run the cooling air fans for the big motors
jmkasunich: now *that* is manly
the electrician we were working with was less than 40 years old and already deaf - he had hearing aids in both ears
that's not very good
interesting place to visit, but I sure wouldn;t want to work there
the bar mill was scarier
not as big HP (but still maybe tenl thousand HP total)
but the rods could miss a guide (like dmess mentioned) and fly everywhere
all the places people might be were armored with heavy guage perforated steel
even catwalks 10 meters above the follr
seeing a video of that happening would be cool though
just like sheet, the rods go faster as they get skinner
at the end they're doing close to 100km/hr
do they have a protective atmosphere?
no, just air
and they don't get much oxidisation?
oh they get lots
ever look at hot rolled steel? its covered with black scale
I haven't worked with that before
(most steel for machine shop type stuff is cold rolled)
we had this bar of uh
I-beams and such are hot-rolled
seghärdningsstål, which is a type of high-tensile strength oil hardend steel
* alex_joni loves hot-rolled steel
alex_joni is strange
so EASY to weld
which was covered in black scale
oh, I forgot, alex_joni is a welder
I guess thats the same thing
jmkasunich: once it's been cleaned of course
a PITA to grind it all away
jmkasunich: I welded on hot-rolled once..
couldn't get it to look bad
try machining it and you'll hate it
it wouldn't matter what parameters I would set up.. it just looked great :D
I won't :P
somebody on rec-crafts-metalworking once described it as trying to get a good finish while machining a pink pencil eraser
heh.. that soft?
soft and "gummy"
like copper but worse
I take it
something like that
sounds like needs for high speed spindles?
does it get stuck in the flutes of mills too?
alex_joni: no, you just don't machine it at all
I did some work in it, and it seemed simple enough to mill
a 16mm endmill
I-beams for building office towers and bridges don't get machined
flame cut to length, punch the bolt holes
oh, this was no I beam
its very ductile, punches great
it was a 20x40mm bar
Lerneaen_Hydra: you said "oil hardened"
thats not the same as hot-rolled
similar surface scale, but the metal is completely different
it was called: "ss-en 10 083-1-c45e"
hot rolled is used mostly for beams and such, and for sheets that will be processed by forming/stamping/punching, not by machining
cold-rolled is used for stuff that needs to be machined
it's apparently meant for "machine parts with higher requirements of strength, for example axles, wheels, and screws. can be flame or induction hardend"
sounds like good stuff
yeah, Re of 450 n/mm^2 when hardend
breaking stress of 800n/mm^2
so it's got lots of give even when hardend
how do you "brake" steel?
* Lerneaen_Hydra resists smart ass response
Jymmm: push/pull/bend it until it breaks
go for it
* jmkasunich didn't resist
cast iron I can see, but not steel.
anything will break if you apply enough force to it
cut yes, but break?!
everything will break from fatigue given enough cycles
tighten a steel guitar string too tight, and what happens?
as long as you bend it past the elastic part (usually even under is enough) enough times it will break
take a paperclip and bend it back and forth many times
Lerneaen_Hydra: thats fatigue
oh, that type of break
you can break things just with pure tensile force too
you have shearing force too
its just that the forces needed to break steel objects are higher than most people can generate at home
0.6 times as strong as tensile
and compression, which is much stronger
you could straighten out that paper clip into a wire and break it just by pulling on the ends
as long as it doesn't buckle
jmkasunich: you'd need fingers of steel +5 to do that though
paper clip 1mm diameter would probably break at somewhere between 100 and 300 lbs tension
yeah, you'd need steel fingers, or a vise
and a 4' crowbar as a level
eeesh... I can't type for crap today! lol
HEY! I think I got a machien that works
I meant the computer I bought this morning... it might actually be able to run emc
at least it didn't choke this time when I started emc
real time, not sim.
Lerneaen_Hydra The laptop I have hates emc, and was the other system I have. That's why I've been running TurboCNC all this time.
Soooooooo, *IF* this machien will run emc, then I'm going to try and run the console remotely on the laptop as there is limited desk real estate for a 17" monitor, kybd, mouse.
Jymmm: haven't you been able to find an older computer as a freebee?
Lerneaen_Hydra: Yeah, at about 0630 this morning for $18 =)
sounds like a deal
Jymmm: I suggest you run the /usr/realtime/testsuite/latency/kern/ tests
can't remember the exact path, but something along those lines ;)
Jymmm: that sounds about right ;)
how fast was it?
some P3 I think Jymmm said
alex_joni right now, I'm running memtest =)
I just tossed in a 256 DIMM that's been "questionable" all these years.
oh.. that's safest for overnight testing ;)
03alex_joni 07HEAD * 10emc2/debian/ (control.in control rules): initial attempt at handling multiple kernels and deb dependencies
Well, we're up to 400K+ errors. Ya think it works?! lol
not really :D
Ya know, I bet if I heat this sucker up, it'll work.
Ok, so what temp do they cure solder paste at?
03alex_joni 07HEAD * 10emc2/docs/src/code/ (5 files): moved emc2_code_notes, older CVS history in the documents/ module
03alex_joni 07HEAD * 10emc2/docs/src/Master_Developer.lyx: initial version. will hold information about the code (for developers)
03alex_joni 07HEAD * 10emc2/docs/src/Submakefile: added EMC2_Developer_Manual.pdf/Master_developer.lyx to the build system
03alex_joni 07HEAD * 10emc2/docs/.cvsignore: something new to ignore
Has anyone tried running AXIS remotely? Like from a laptop, and then run emc on on another machine (hope I said that right)
Jymmm: that should work.. but you need another linux PC for AXIS (hope you know that)
alex_joni: Yeah, just not sure if I need RT on the laptop though.
I just dont hav ethe room to put up a 17: crt
or even another kybd
sure.. no worries
there's no sugar :D
I'm wondring if I should install a 2nd nick and use a crossover cable between the two mahcines for the contol portion
wifi might work too.. but I advise against :)
I got 5000' of CAT5, no need for that
This is the PCI Parallel card I got today.... it's DUAL paraport: http://koutech.com/proddetail.asp?linenumber=48
Jymmm: I run AXIS over an ssh connection all the time, using the X protocol. It works fine.
Jymmm: the only limitation I know of is that you need to have the nc files on both pc's in the same location
I wonder if I should use it instead of the built in one.
alex_joni huh? I need to copy the nc files to BOTH machines?
jepler Cool. Why ssh ?
Jymmm: ssh is the only safe protocol to connect to another machine (rlogin and telnet are kinda obsolete)
Jymmm: it's the way I'm accustomed to running a program on another linux machine
This is across the lan, or maybe a crossover cable.
Jymmm: you need to have them locally so the Open-dialog shows them to you when you try to open a file, and they need to be in the same place, because it will send the absolute path when instructing the remote emc to run it
alex_joni: that's not important when displaying remotely over X -- the Open dialog is running on the same system as emc
I thought I just needed the files on the emc box, so when I open a files from the remote it'll utilize a local emc box path.
I don't need to run X on the emc box, do I ?
Jymmm: you can ssh into the machine (with X forwarding enabled), then when you run emc it will export the display to the local machine (that's jepler's way of running it)
so you'll get the GUI on the local machine, not on the one with emc on it
No, I mean the way that emc does it internally. like keystick on the emc box, axis on the laptop.
I dont want to run X on the emc box if I dont have to.
Jymmm: maybe I'm not expressing myself clear enough
there are 2 ways to do it:
1) remote X terminal
1. do it without any emc help (ssh with X forwarding, that means no X needed on emc's side), you're using a standard protocol to make the GUI appear on the remote side
2. do it with emc's help (run keystick or usrmot locally for no X needed), then run another GUI remotely (be it axis or tkemc, or mini)
the second solution assumes emc is playing along (minor tweaking to the config files), and for it to work OK while loading files, you need to have the files on both PC's
so with #1, you dont need X server running on the emc box?
I've never done remote X anything, so my ignorance is showing here =)
You can run the X server (the part that actually runs the display) on one machine, and the X client (the application itself, such as firefox) on another machine. They use the X protocol to communicate, like web servers and web browsers use HTTP.
so, no, you don't have to run an X server on the mill machine
so, "X Server" is really the GUI part of X, and the client can be nothing more that pure instruction transfer (low cpu reosurces)?
I suspect both methods will behave similar (performance wise)
I'm just trying to take away overhead from the emc box is all.
what processor power do you have?
Just dedicate it to machining alone.
The new emc box will be a P3 600 with 256MB
should be enough even for X
but if you do plan to run X on it, install Xubuntu (leightweight)
The p3 800 I had choked on RT Axis
dead locked up the machine.
I suspect a different issue there
sadly, the P3 800 is 360 mad the p3 600 is slot 1 =(
yeah, the mobo's
slot1 cools better
I got one of these today, it's a DUAL pci paraport card... http://koutech.com/proddetail.asp?linenumber=48
I think cradek mills all the time on a 667 MHz Pentium 3
I wonder if I should even use the onboard one or not
don't suspect it to be faster
jepler Right, but then a P3 800 shouldn't be an issue, but it is =)
alex_joni even it's docs says it's faster than onbaord paraport
alex_joni: "p to three (3) times faster data throughput than comparable ISA bus or motherboard parallel port"
NetMos Nm9715 ?
the chip on it
never heard of them either.
they say they support linux though =)
take a look at the card, and read what's written on the black chip
what do you want to know?
what the name of the chip is
you just said it
ok, so same one?
I was reading off my card :)
oh, you have this same card?
KWE 220N-2 on the top right
written in white
same one.. not from Koutech though
03jepler 07HEAD * 10emc2/src/ (Makefile.inc.in configure configure.in): SITEPY is the location in which to install the files from lib/python for 'make install'
on the chip
that's the serial (the last one) and is expected to vary
Is yours a DUAL paraport too?
and NOT faster than the onboard one :)
at least not regarding RT
it might be faster in ECP when transferring data to a printer using DMA
ah, ok. So between onboard, adn the dual, I'll have a shitload of I/O huh
I'd use all 3 :D
oh I meant to ask which should I be suing ECP or EPP ?
you can build yourself a nifty control board
neither.. the standard mode
why? I thought the others were faster?
Jymmm: not for what you want
you don't need to send lots of info out in bursts (like ECP and EPP do)
you need to change the state very often
and ECP, EPP only make that impossible
oh, so ecp/epp are buffered?
emc2 won't run properly if not in bidir
Jymmm: one of them is, but you have differnet addys you need to write to for both modes
base_address + 400, if I remember it right
ok, I'll take your word for it (this time =)
I dont need all 3 paraports, but it was sitting right there in front of my face, still sealed in the box and was only $3
I was surprised to see it's still being sold and supported (by the appearnce of the plastic wrap)
it's quite a deal
a new card is around 20-30$ over here
Jymmm: if you plan to switch over to cheap way of servos, then you might need all 3
PWM for speed control, inputs for encoders
and of course you can always set up a control panel (with buttons & knobs & switches & leds) for controlling the machine
I was thinking of making a PS2/USB pendant
using one of these: http://www.ultimarc.com/ipac1.html
axis is written in python, wonder if I could run axis under M$ ???
Jymmm: you certainly could
I failed to do it..
you need to compile some parts of emc on doze first
because axis has an emc-interface which needs to be compiled and linked with certain libs
libnml and libemc come to mind
oh, nm then. I'm clueless on the compile part it
did you look at the ipac
yes, interesting.. but if you have the parports.. it might be cheaper to not use that
and you need to code some software for it to work with emc
nope, pure kybd emulation. static
anything you can do from the kybd, you can do from the ipac
and it retains it's config between power off's too
there are lots of things you can't do from the keyboard in emc2
but you could do from a parport
halui has a list of 100+ pins
each controlling something
like what as example?
spindle-speed override :D
that's missing from the gui's
hmm.. maybe not as many as I thought
those could be added to the gui though, couldn't they?
by end-users I mean
if they know how.. sure
you mean "if they know python"
or tcl (if they want tkemc)
axis in tcl ?!