does rtai-info help?
is a text file of all the modules with their paths
Thanks, but no.
The paths are local ones, the .conf file contains lots of lines like:
MODPATH = /usr/realtime-$(uname-r)/modules/rtai_smi.ko
So it is explicitly looking in a realtime directory named after your kernel version.
But, even having fixed that to where the files really are, I get little further
Is there a way to find when a file was last looked at?
It would be enormously helpful to know if EMC was actually looking at rtapi.conf at all
andypugh do you want me to figure out a way to make it look for the modules in a different directory? i can check...
Hang on a bit, I have found yet another rtapi.conf
Aha! /usr/local/etc/emc2/rtai.conf it is!
andypugh: the -ltu options of ls show last access time (but fstab mount options like noatime can disable updating of access time)
I think I found the active file by changing its name
Now to edit it...
Progress, if not success. ' echo "yes this is the file" ' has at least shown which is the correct filr
And now, sucess!
Which is a cut-price version of success
<heads off to edit the wiki>
you have wiki access??
We all have
jeeze. i need to get a wiki account too
It's the WWiki-way
hey you emc enthusiasts, what f/oss cam software is available?
wanting to convert cad drawings to g-code
It's like a dummy login to baffle the bots
autocad version 2 i think does it as well
er, let me rephrase, foss linux software :)
ybit : I thing HeeksCNC might
Kcad is open source if thats what you mean
its all in KDE which is under GNU.... I'm confused
Neo_The_User: oh, i was looking at http://www.kellyware.com/download/index.htm
forget everything i just said
yeah, looks like kcad is a screensaver
yeah i got mixed up
conversion is often cad specific, like you need one for older acads, one for newer, (eg: >r13..) , one for mastercad, one for solidworks...
and there's little foss, just a few scripts to peek into layers of older acads
What is the tag for "a bit bigger and bolder" in wiki?
usually [b] [/b]
or replace b with bold
in most cases..
<b> in this case
= is big == i smaller ( rules on wiki are NOT html, look at 1st page )
So like html
* frallzor if feeling burned out
building a cnc-mill is pure therapy :P
I suspect it is better than hacking kernels and kernel modules
only 1 down
it costs :P
hacking kernels isn't _TOO_ bad.
just adding new features are.
Flushed with partial success. How would one go about installing 2.4-pre rather than having it run-in-place
leaving off the --enable-run-in-place doesn't do it.
cd your/emc/directory && cd src && sudo make uninstall && ./configure --prefix=/usr && make && sudo make install
just leaving out --enable-run-in-place installs to /usr/local
ermm... thats wrong sorry. cd your/emc/directory && cd src && sudo make uninstall && make distclean && make clean && ./configure --prefix=/usr && make && sudo make install
The uninstall presumably uninstalls the RIP version, not the current installed version?
no it uninstalls the installed version
wait no wrong again
sudo make PREFIX=/usr/local uninstall
Even if the installed version doesn't match the source being compiled?
sudo make PREFIX=/where/it/installed/to uninstall removes it from the installation directory hands down no matter what
or am i wrong yet again?
How does it know what to uninstall?
because PREFIX is specified
im not sure really
I am asking purely out of academic interest
I haven't installed and uninstalled emc enough to give you a straight answer.
Hopefully my block of aluminium will arrive tomorrow and my priorities will shift from elegant kernels to pure function
* frallzor wants packages too
but no idea when they'll arrive :(
Yes, darned those postal workers
darn those UK postal workers
they gotz me packages
"lets get these packages going in a haste!!!" "oh wait, TEA TIME"
Better than the US Postal workers having your pkg =)
they are on strike over here
"USPS, can lose an elephant and break an anvil
was the phrase I heard..
archivist royal mail only I assume?
Actually, for the sheer volume USPS has, they are not bad at all.
im getting via ukmail
which I've never heard of
we have a few over here now
know if ukmail.com is any good? :P
What does a 1st class stamp cost in UK ?
about 2 crumpets
37p I think
isn't a crumpet a bakery product?
two of them
30 for second class
so about $0.70 USD to mail a letter?
I just print the postage direct onto the envelope nowadays
no, I think it's ~1.6 UKP/$
andypugh: special software, or any UK citizen can print their own postage?
MarkusBec is now known as MarkusBec_away
you can do it online - I'm on the royal mail website now
Anyone who logs into www.royalmail.com
I wish the USPS would get that far. Only Priority Postage atm.
note that the 39p rate is for a delivery aim of 1 day
second class, which is 30p, is 3 days
not much of a diff.
I'll stick with $0.46 USD =)
so it's about a nickel or so more than in the US for 3 day, and much less for 1 day (attempt)
But then nowhere in the UK is more than a 3 day brisk walk from anywhere else
Isle of Man ?!
I am not sure they are technically in the UK
me neither =)
The Channel Island bailiwicks of Jersey and Guernsey, and the Isle of Man are Crown Dependencies and are often not considered part of the UK, though they are treated as part of it for many purposes including nationality.
SWPadnos, I can't find that .ini file for EMC to make it open up in the nc_files directory. Can you help?
err - if it's a quick thing, yes
are you trying to set the directory that EMC looks at for nc files?
or set the directory where EMC starts up
is emc2 for windows out yet?
JMS-KY: Yes, Widows 3.11
Isn't that just in the general .ini file as "Program prefix"
sure is, even comes with a free version of windows 8
i already have that version :O
I saw Windows 7 today for the first time when a chap in the computer shop decided to show me how they have put the Mac dock at the top.
And I might have to spend a lot of money on a new 27" iMac, 'cos they are lovely.
i miss dos :(
I don't. It's hopeless
andypugh: Why an iMac over a MacMini + LCD ?
Wires. I hate wires
SWPadnos, I mean the actual directory the .ini file is in
But it is mainly the new mouse
I see no *.ini in the emc directory in my home folder
andypugh: Eh, ok. i guess.
err - you mean the command-line option to the emc script?
no, it would be in ~/emc2/configs/something
oh ok :) thanks
actually, ~/emc2/configs/something/something.ini, if you follow the pattern
I was actually genuinely doing accounting in my head like "well, it is £1349, but I will be buying the mouse no matter what, and that is £55, which makes the iMac £1295, which is only just over 1k...
Hi i am having trouble with hal_input - No input devices could be opened......... I read the section 'PERMISSIONS AND UDEV' and added a file to the rules.d directory called 50-keyboard.rules. file contains one line: SUBSYSTEM=="input", MODE="0660", GROUP="plugdev" any ideas on what is wrong?
did you restart udev or reboot?
It would be unusual to need rules for a keyboard
an extra keyboard, maybe
this is the only keyboard
which is presumably working outside HAL?
will you be using a touchscreen?
yes it works fine for everything else
well, it'll be awfully hard to type on it if you're using any keys for HAL functions too
but you may be able to use the lights for HAL output
(I'm not sure if that'll really work actually, but it could be fun to watch)
well what I really wanted is to use some keys as inputs for hal
I don't think hal_input can do that. you can select classes of functions to use (keys, lights, analog, and something else), but not a subset of one class
you can of course just not connect anything you don't want to use
but, those keys still work in X as well, so pressing the Z key will attempt to type a Z (and therefore select that axis), as well as activating whatever you have connected to the key_z pin in HAL
ok how do I use the F13 key for example as an intput
Scroll lock and the windows key have no other uses...
oh - I don't recall what the key will be called
there will be a bit output in HAL, you can see them all with halcmd show pin input
SHIFT+F1 == F13
SHIFT+F2 == F14, etc
ok This sounds easier that I was trying to make it.
You should be able to adapt the instructions here:
[01:47:38] <andypugh> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Simple_Remote_Pendant
To get the keyboard as an input device.
Just use an identifying string from the keyboard name
SWP: Do you know if there is an elegant way round the fact that my installed EMC (2.3.4) won't run with the parport module installed, and the run-in-place 2.4-pre which I want to use for threading won't run without the parport installed?
Obviously I can insmod and rmmod to switch configs, but that seems clunky
no, not an elegant way around
it will be necessary for that to happen though
why use 2.3.4 instead of 2.4~pre all the time?
Yes, why? Are there any reasons at all?
I don't know :)
if 2.4~pre works for you, then you may as well use it
The main one is that I only seem to be able to do run-in-place from a git pull
having 2.3.x around so you can check if something is a bug or regression may be useful, but you can always rmmod as necessary in that case
I kind of had the impression that 2.4 was a flaky beta
err, you mean you haven't been able to install a 2.4~pre version (via make install or by building packages)
But then as a hobbyist I probably should be using the flaky betas to see how flaky they are
you don't have to update and compile every day you know ;)
Yes, that is correct. I tried a make install and it said I had configured for run in place, and at that point my 3 brain cells went on strike.
ok, here's the answer to that one
run in place and installed are mutually exclusive, and the make system is smart enough to know what you chose at configure time
I can scroll back to Neo's answer, but I am curious to see if you have the same one
so it refuses to install for RIP, and it refuses to setuid for non-RIP
./configure, and leave out the --enable-run-in-place option
I tried that.
then make clean and make
then sudo make install
it seems that --enable-run-in-place is sticky
delete config.cache and try again
Now, that I didn't try
( I am not fond of command lines, you can't explore)
sure you can
tab is your friend
especially if you have a very large screen, if you're just looking for a list of all available commands
(there are probably around 3000 of them)
GUI's tend to have a context too. eg I wouldn't just find that "delete" existed, it would probably offer me a list of deletable things..
But that is a fully-argued point by now
sure - type rm<space> and then hit tab twice, and it will list all the files in the directory
if you know it starts with a c, rm c<tab><tab> will list the files that start with c
I think we are talking about different levels of abstraction
there sure aren't little pictures at the command line
Indeed. delete config.cache would probably be a checkbox marked "start from scratch" somewhere. That is sort-of my point
oh, a GUI for the make system, not for the comand-line and file management
In this case, yes,
I am not saying there should be one, just that noobs like me would be able to guess their way through one more easily.
And yes, I can see the arguments why the likes of me shouldn't be let loose with a compiler
Completely random question prompted by the fact that I tried to VNC to the CNC box and it is shut down..
Does RTAI interfere with software shutdown and WoL
don't know about WoL, but it does interfere with shutdown on some systems
I have a feeling that if I boot the CNC box and don't run RTAI then software power-off works, and WoL might. But generally I have to shut down, wait for the shutdown screen, then hold the power button down for 30s. And to restart I actually have to walk round the house to the garage, unlock the door and turn it on.
that doesn't seem like such a bad thing, if it's expected to be running a machine
since you'd have to walk out there to put the workpiece on the table, turn on the driver power supply, etc.
Any workarounds on shutdown? It is a server board and getting shutdown without an immediate restart from the switch is tricky some days
someone had posted a kernel compile option or kernel load option to the user list, I think
but I don't recall who it was, what they said, or if it works
Which bit seems like a good thing? Booted by itself the PC is impotent. The drivers and power supplies are in a seperate bay with their own switch
which is not to say I haven't thought about the possibilities of WoL, webcams and remote switches
sure, but turning on the PC manually seems like the least of your worries when you already have to go out to the machine to actually do anything with it
Hi I am back. Reboot was all I needed after adding the rules file Thanks for your help.
there should be a way of restarting udev without rebooting (like /etc/init.d/udev restart or something), but it's always good to be sure it'll work after a reboot anyway
now it would be nice if there was a pyvcp widget that was a numeric entry device.
SWPadnos: A fair point, but the button is a bit flaky, and has to be held in a long time.
I'd suggest getting a new button
RS should have some that would work
2-way switch soldered to the motherboard and mounted in the front panel
And I am not convinced it is hardware-flaky
The box can be configured to respond to the switch in different ways, or not at all. (1U server)
ok - it may be set to wait for 4 or 5 seconds before powering up/down
But if RTAI is known to meddle with shutdown, that's my question answered. I can live with it, and will stop looking for something I have mis-configured
it is 30 seconds
the problem is ACPI, parts of which can cause latency spikes
other parts control shutting down the PC
Then sometimes it does an immediate restart, cue another 30s
so it's not clear which parts can be enabled without scrweing up latency, while still allowing a soft powerdown
usually, when a system is booting, the power button will immediately shut it off
but that's not necessarily usual for a server system like that
It was cheap (£100), it mounts in the same box as all the other parts, it manages 2092nS latency. I can live with a flaky switch
If anyone is bizarrely interested, the PC is the leftmost unit here: http://picasaweb.google.com/lh/photo/tpgXEB-7Jatf_3Cks8n2Xg?feat=directlink
andypugh check out make xconfig or menuconfig
also is it a multi processor machine?
if you compile the SMP kernel your latency will probably improve
ahh lol all good then ;->
25,000 nS single CPU, 2092nS SMP with a dedicated RTAI CPU
yeah SMP is made of win
i've been following the thread as much as possible for a couple days, would you wiki up how you got 2.1uS latency? ( and is this headless ?)
Cradek described it best, it is a purely aesthetic thing, but having a CPU dedicated to RTAI just seems right
this the max jitter number isnt it?
No, this is a totally standard Ubuntu installation on an old bit of kit from eBay
And yes, the 2092 is jitter
I'm getting 2458 on mine
I can't run a 2uS base thread
but that is with the nvidia binary driver running
coreboot can handle the power_off without acpi, it might be doable from usesrpace with factory BIOS, have to try sometime
I was getting 2500 max jitter with 2.3.0 using the binaries from here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
yeah thats what I was using, but i'm running 2.3.2 I think
You should be able to just do steps 1 and 6.2 and get an SMP system running 2.3.0 in moments.
it wasn't paticularly difficult
However, you can't upgrade that with apt-get for reasons I think I now understand
yeah the emc in the repos is compiled against non-smp kernels
I wonder if the maintainers would consider perhaps an emc-smp meta package that would pull SMP kernels and EMC's from repos
that would be awesome
we don't have an "official" SMP build
and leave the stock uniprocessor stuff unchanged
so that's not really an option yet
And the binary kernel in that link is compiled without the parport kernel module, which is great for 2.3.x, but is a problem for 2.4.x
could there be an official-unofficial one?
there's an experimental one on the linuxcnc website, but it's not in the package repos, and I don't think we'll put it there until it's "official"
One core to bind them all...
Having a dedicated RTAI core just seems so "right"
but it wont be official because you dont "need" it
when 8.04 came out, I don't think there was an RTAI version that we could all boot and run on our multi-core or SMP systems
is there now? (that would rock)
so since we couldn't all use it, we figured that would carry over to the community at large
I don't know
andypugh means i can nvidia drivers run on my box
hopefully I'll get some time to tinker not long before/after 10.04
I don't think I've heard about problems with the SMP builds in /experimental, so that may be a good starting point
if it's the one I did for my P3, it's superbly old
the one on the wiki?
no, these are the ones ehj did for the Atom
the ones you did are still there too
(and happily running some large power supplies :) )
Anyway, I couldn't build an SMP + RTAI + parport kernel, but Neo_The_User seems to like this sort of thing and came up with a minimal .config and instruction page that let me compile a kernel and RTAI that does both current flavours of EMC2 and as a bonus is even faster (though if I ever want token ring or packet radio, I am in trouble)
SWPadnos: Then let me inform you of a problem with the SMP builds in /experimental... No Parport so no 2.4.x
well, he hasn't updated it ;)
did it run on your machine with 2.3.x?
But for 2.3.o then install straight from the .debs is the way to go
Yes but no
there anything funky in 2.4?
it handles parports differently
(among other things)
I'm using mesa
7i43 or a PCI card?
ok, not a problem then
so any major new coolness?
the experimental debs give you a nice, fast, reliable 2.3.0
2.4 isn't out, it's the development cersion
ehj was nice to contribute those but it gives us (and users) a pain if they aren't kept up to date
yeah I know, I was wondering if there was some major new coolness in 2.4
probably - check the git log :)
I broke it all with an apt-get when I should have realised that custom-kernel == build from source
cradek: I think it is fair that it only troubles the naive and dumb users
I didn't mean to imply that.
Valen: Lathe threading is reportedly a whole lot better in 2.4
cradek: I meant to type "fair to say", ie the experimental SMP builds are what they are, and they work well. But if you do daft stuff like an apt-get update then they break.
coordinate system rotation (which works right), nurbs (which don't work quite right)
cordinate system rotation would be handy
some kinematics stuff, not relevant to anyone except robot folks
Are there many robot folks?
I build robots
but they don't run emc ;->
it's hard to tell how many of anyone there are
Indeed. I have no real idea how many users there are for my dynamometer control software, and that is only released within one company. (albeit quite a big one)
what sort of dyno?
interesting, I was thinking how nice it would be to have a megasquirt work with the dyno so you can do a full map of an engine, rather than just pulls
so you use the dyno to hold the engine at a set RPM while you vary fuel + spark etc to get the optimum power for a paticular map point, then vary the throttle one point keeping the engine the same and repeat the mapping
It doesn't actually control the dyno, it performs structured experiments of calibration parameters and acts as the glue between the dyno control firmware, the commercial ECU cmms softare and the logging system.
Valen: Yes, it can do that, but you lack ambition :-)
so do they run dyno's in the manner I was saying?
how "should" it be done then ;->
The gasoline guys do
We do diesels
they need an ecu now? ;->
12 degrees of freedom and 4 noise factors
(mathematically insoluble in finite time)
they don't just pour the fuel in the top and let it sort it self out?
Yes, modern diesels are far more complex than gasoline
I thought they were typically direct injection and all they really changed was quantity and timing?
with multi sequence injection type stuff for stratified charge burns
There are up to 5 injections per stroke, each of which is controlled for fuel mass and timing. At the same time we control boost pressure (variable vane turbo) and exhaust gas recirculation volume
We also have full control of injection pressure
variable vane turbo, that sounds nasty
they do variable cam timing type things yet or not bothered too much given the RPM range?
Fuel injection is disturbingly open-loop so we have a scheme where tiny amounts of fuel are injected on coast-down one cylinder at a time while watching the rank speed, to see what minumum time it takes to actually inject some fuel
No vvt on diesel yet
I was looking at electric valving on petrol to replace butterfly valves and get really really variable timing
But I am looking forwards to the day we get to control the valves electrically. A boosted 4-stroke diesel can switch to 2-stroke once the turbo is spinning...
just blow through for exhaust?
nah still run it electrically
Context: Current 2.0 litre diesel family car engine 168hp. 1979 Z1 superbike, 1.0 litre, 7ohp.
andypugh, that would get you a decent power gain off the line
andypugh: heh can you retrofit my bus?
cradek: yes You don't inject fuel till you need it, so you can purge with fresh air
I like it because you can get rid of pumping losses which is like 20-30 % on a petrol engine
can you retrofit my bus? you only have to beat 8mpg or so
Nothing new: http://en.wikipedia.org/wiki/Napier_Nomad
(a diesel, uniflow 2-stroke aero engine, with afterburner)
Err, yes. We get 45-50mpg out of that 168hp engine
I think the only thing that would get me to trade in or sell my little vw would be to find the same model with the diesel
seems like we're getting more sporty diesels now, in the US
You are, Fiesta and Kuga are due there next year\
nice chatting ttyl
(Kuga is mine)
can't wait to have one in 6-10 years
[03:26:11] <andypugh> http://www.motorauthority.com/blog/1024128_ford-bringing-kuga-c-max-to-the-u-s-after-focus
man, how long until we get decent looking cars back?
these big-assed no-back-window things are everywhere
baby SUVs I guess
'03 GTI VR6
even the new GTIs are uglier now
You don't like Martin Smith's styking then?
I don't know what that is
isn't he the dude that designed the imac ?
no idea :-)
mine is this model: http://dealerrevs.com/car/441462
and by design I don't mean the circuit, just the case
Follow the link
yeah, I did
All the current euro-fords look so similar that I can't tell some models apart.
gas turbine hybrids are surely the way forward
Though I do like the styling, so that's fine by me
if I ever find a decent starter car I'll make a full electric for my tiny commute
I rather suspect that gas-turbine hybrids will be the EDO-DIMMS of the automotive world, outmoded before they are even developed
yeah I thought of fuel cells straight after
that's the proper way to run an electric on liquid fuel3
I bought a powermac 4400, one of only 2 computers ever to use EDO DIMMS
cradek: How tiny is your commute? http://www.vectrix.com/
I thought my P100 used EDO. the 486 used FPRAM
8mi - short enough for bicycle if it was safe (it's not)
EDO is normal, but were they DIMMS or SIMMS?
hilly and narrow 2 lane highway with no shoulder
if it's flat, bet a bakfiets
I've biked it but it feels stupid the whole time
No, even with offline key-transposition that made no sense
get a bakfiets :) it's a dutch flatbed bike that can take a similar load to a Ford Transit
Pedal faster? At 50mph it should be OK
but utterly useless on the slightest hill
Unless carrying a load of helium or unwrapped vacuum?
the bak (dutch for the front bit) is made of wood, so the vacuum would drip
oh, are those the ones that people carry babies and groceries in?
kind of like trikes made of mostly wood
those are darned strange
overkill for babies & groceries. a dutch momma can get 3 kids & a weeks shopping on a 2 wieler
The dutch are strange. Nice, but strange
one thing I found odd was that nobody had a nice bike (in Amsterdam) - they get stolen too often
so everyone is doing everything on these old POS boinger bikes
it's all you need
On a sunday, out of town, you will see folks in lycra on serious kit
I find it strange how people can't ride a bike here without weraing lycra
there ought to be a licensing program for that
ok, you need 2 gears in britain because of the hills, but nobody needs a mountain bike or suspension for the road
depends on the roads
You need more than 2
and in Vermont, the 21-speed is just barely enough
The time the Milk Race went up Winnatts pass is the only time I have seen professional cyclist pushing.
even with transposition, that makes no sense
yeah that's my 2nd gear too :) If standing on the pedal doesn't do it, get off and push
The classic alpine climbs are long, but not all that steep. In the UK we have roads so short it isn't worth wigling, but 1:3
MattyMatt: You asked if there was a fee to spectate at the enduro last weekend. See if you can see a single spectator: http://www.youtube.com/watch?v=26qfN6s-BKg
"The Milk Marketing Board also sponsored the Football League Cup from 1981-1986, renaming it the Milk Cup."
I'd forgotten about that
Eek! look at the time!
I think 4am is a good time to wake up :)
* MattyMatt lives on internet time
SWPadnos: the Milk Race was the UK equivalent of the Tour de France
but now the TdF does part of the race here usually
that's not my idea of cycling tho. hills & bikes don't mix IMHO
I came back from Holland and the skinny gears on my racer exploded on the first hill I came to
MarkusBec_away is now known as MarkusBec
_Poincare is now known as Poincare
MarkusBec is now known as MarkusBec_away
MarkusBec_away is now known as MarkusBec
MarkusBec is now known as MarkusBec_away
MarkusBec_away is now known as MarkusBec
[12:20:35] <alex_joni> http://networkchallenge.darpa.mil/
i thought it was machine tool related..
I get a timeout...the interweb is stuffed with updates to ubuntu
yeey, my cnc stuff arived at the post office today =)
but not computer to run it :(
does anyone have access to this http://www.informaworld.com/smpp/content~db=all~content=a916388183
I would be interested.
someone on the ETH (zurich) campus kindly emailed me last time I asked for something from this journal I think :)
awallin: looks like you can have it for only $37
awallin didnt find it free but stumbled on this http://vision.ai.uiuc.edu/~wanghc/research/cad_cam/paper.pdf
( iput the title into google w/o quotes and got a load of papers )
the paper begins with 'varying engagement' ( how much meat teh cutter has when turning a corner.. a classic problem in edm orbiting.
teh patsh it creates for pockets are very interesting, almost a childs crayon dwg , the path is so curvelinear
to keep the feed up? it definitly doesn't 'turn corners'
tom3p: thanks, I think I've seen this before(?).
the papers are free if your uni has access.
aint got no uni ;) and library system , even thru local uni's in chicago, dont have it
found it, for free, in National Library of Australia, Canberra, main reading room (sorry)
[14:06:20] <tom3p> http://catalogue.nla.gov.au/Record/1577438
bus ticket is going to cost!
lots of frequent flyer miles though
tom3p you an aussie?
dumdidum .. cnc stuffs in about 40 minutes .. yeey =)
problem is that i need to find some fans for cooling:/
what you cooling?
the driver chips =)
according to where i bought it 24V+ needs cooling
probably needs heatsinks as well if they arent supplied
computer store then
got some PII cpus i wont use anymore with fans on them i guess they should do =)
but will be very fun to open the package, even tho i have no computer atm to drive it with :/
more then an very old laptop that has parport =)
pentium 2's are not trivial as i recall
the slot 1 ones
aye its slot ones ... trivial ?
what do you mean by that? :)
taking the fan of a slot 1 cpu is not easy
(think pain in the ass)
ahh, well i guess i have something more i can dismantle =)
put the cpu cooler for my 140W cpu on it maybe :P
should do ;)
I'm going to play with firewalls for a while back later ;->
that would manage I spose
I'd nip over to a computer store, 80mm fans are < $5
Valen, ye true =)
got some 80mm laying around also :)
or maybe its 120
I have used superglue to attach new fans to some verions of the slot 1 package. :)
oh well, more air :P
that or not bother and see if you actually need it ;->
Valen, well cooling is always good =)
and a 120 should cool the whole damn board so :P
yeah, but its gotta be working first ;->
dont forget the blue LED's
Valen: not from Oz, from Chicago ( close )
blie LED's ?
for pimping it up ? :)
have to put that on BEFORE i try it
Ahh, I'm in sydney (300km from canberra)
or maytbe get one of those neon light fans
MrSunshine, that is what i was suggesting, they are usually a LED
aye true =)
should get myself a computer dedicated to the mill later on
will be run with my quadcore to start with :P
so realy hope the parport protection is good enough so a backfire wont burn the computer :)
make a SMP kernel ;->
actually think about getting a PCI port then
a dedicated one so i can make the whole chassi into the cnc box .. powersupply, computer, driver and all
they are pretty cheap I believe
ive got a pci parport board
2 ports on it
that will probably do for anything up to mains ;->
time to hit the gym soon also =)
its a good day )
have the fun
then to draw up and create the motor mounts ...
I think i might have icecream and waffles
Valen, damn you ... daaamn you
(it is 2:00 AM here)
* frallzor yawns
* frallzor likes milling cibatool
how come so quiet today? :P
everyone is partying at work... (halloween)
skunkworks_: do you know what the control side is like on mariss's 'analog stepper? std step&dir?
I am sure it is going to be step/dir
That is his market :)
yes otherwise how could it work with any existing software? (troll!!)
Hmm, this might not be my lucky day. I just had my milling spindle motor go "pop" and blow the breaker. 3 times. Smoke too....
didn't you just smoke some stepper drives also>
step away from the cnc machine!
Yes, but the drives was my fault.
This was running no-load..
Does anyone know what you would expect to find by way of resistance on the commutator of a permanent magnet motor?
a cordless drill motor has something like 0.5 or 1 ohms
This is a 230V motor.
So I would expect higher. However I can't figure out which segment would be expected to connect to which.
that thing had something like 100 ohms http://celer.oni.biz/wp-uploads/2009/10/dremelkopio-480x360.jpg
How many windings?
they're pretty much supposed to be connected from one side to the opposite... so that when you put woltage to the brushes, it goes through a winding
That was what I thought, but then I wondered if it was more complicated that that.
As just using one winding at a time seems inefficient.
indeed. but i don't know, probably there is a reason :P
Starting with one commutator segment I had 127 ohms to about half the others, and open circuit to the other half. The other half have about 4 ohms between all of them.
I suspect that that is not as it should be, but I am not sure.
at least when there's very few ohms at some point between it and the opposite side, there is a problem
could you hook your scope to the brushes and give it a spin with a drill? I think it should be easy to see a short that way.
connect the meter to the brushes and turn the motor slowly, at least then you'll see what the electricity sees 8)
that's a good idea, too, if you happen to have a scope
I have, but it is rather fuzzy and doesn't trigger reliably.(I found it in a skip, probably for good reason)
what happends if i connect a bipolar stepper A B, D C insted of A B, C, D ? :)
it runs the other way.
turns the other direction that is...
mozmck, only ill effect of it ?
i was thinking as you turn it from being + + to + - .. should mess up the steps
as long as the A and B are the two ends of one coil, and the C and D are the two end of the other coil, swapping C and D or A and B will just change the motor direction.
mozmck, humm ok =)
but don't get one end of each coil together or you'll have problems!
could measure ohms
yep, I do that all the time to make sure I have the right wires together. A lot of times I don't even know what is supposed be A vs B, but I know the two go together!
what do you mean with go together ?
on a 8 wire stepper or what? :)
On an 8-wire you can put the coils in series and back to back. Normal current, zero torque. Very puzzling
I mean they are the two ends of the same coil
Especially when you have done that because the data sheet is wrong..
largest ohms reading
ye realy need to find that god damn multimeter
cant for my life figure out wher eive put it :)
or remember where i used it last
Anyway. My milling head motor clearly has a shorted winding. Is it the worst idea anyone has ever had to cut the wires to that commutator segment?
I don't know about that. I would think it would lose power at least.
andy I shudder to think of it
whered u cut@
You will note a tone of disquiet in my voice, but I was planning on gettting some milling done today
I would cut the wire where it is welded to the commutator segment.
that may work fine
anywhere else wood b spooky
There is a chance that the motor will stop on that segment and not start, but this is only a temporary idea.
I wouldn't think it could make it any worse than it is now!
add a big cap
It would certainly be cheaper in fuses than the current situation.
I wonder what the drive will think of it, though?
what kind of drive? vfd?
may not like the impedance mismatch
never know less u try
before doing anything ruinous I'd take it to a motor shop
those seem to be hard to find these days. used to be one near here that closed years ago.
It is a DC motor, but seems to be a sensorless drive, and I can see that not liking a missing pole
is the wire black and sooty in one spot?
or the whole feild melted?
I can get a new motor from the supplier for £65, so probably not worth trying for a repair.
The windings actually look fine, if slightly blackened but that looks like brush debris
Ah, it smells of burnt out motor, no doubt about that.
look for burn thru in varnish
There was smoke, too.
magic smoke not good
Look fine, basically as-new (it has done very little work)
a black spot somewheres
The fuse in the drive is fine, the breaker in the fuse box has popped 4 times (16A) and the fuse in the plug once (5A)
(I thought at first it was a loose connection in the drive)
5amps likely a short
Yes, sorry. I thought I said? 0.06 ohms in one spot, 3k everywhere else
Think is, as I have a VFD for the lathe, might I be better fitting a 3-phase motor?
if its just layin aroud....
No, I would have to buy the motor. The VFD is already hooked up to the EMC machine in the box. I was thinking about switching between motors with a relay.
hard to believe transistors go up to 200 amps now
I got those uln2003an's (4) and picked up a stepgenie for 6 bucs
theres a development board for 25
all set up for 1 motor
you could use it also for the switch?
or mayby a solenoid
Short-term I could just plug and unplug the motors. They are connected with Speakon loudspeaker connectors...
lol just use a wall shitch
Wall switches are not ideal for 3-phase..
"Attention! Speakon is not to be used as a mains electrical connection"
[20:16:12] <andypugh> http://www.neutrik.com/uk/en/audio/210_612654417/NL4FX_detailtechnical.aspx?title=NL4FX
But as they are a very nice 250V / 50A connector with 4 poles, I am ignoring that advice.
They would be good for steppers too.
Anyway, 230V DC motor to suit the existing (cheap) drive in the milling head (no reverse rotation) is £65. I can get a 500W 3-phase motor from eBay for £35, wire it to the existing VFD and get industrial-strength control and a reverse facility. What would you do?
try a new set of brushes first
but thats just me
maybe even revarnish windings
and cross fingers
Huh? The brushes are fine, the motor makes smoke, blows fuses and has a shorted winding.
can you see sparks from the brush?
I don't get the chance.
you said you saw brush debris
Only a little bit
Black dust, like you always see
Hmm, if it is blowing fuses through the drive, that probably doesn't say good things about the state of the drive either, does it?
hard to see how a winding and an upwind 16 a fuse got blown without tripping the 5
or burning out the brushes
The 16A is an MCB, so rather faster than a fuse.
ahh some power spike from supply?
I am just going to see if the breaker trips without the motor connected to the drive.
motors prolly fine with some solder joint run togeather
doubt the wire is actually burnt thru
but you can sure smell burnt copper if it is
There are no solder joints, all spot-welded commutator
Drive with no motor is fine.
I am going to see how it goes re-connected and not starting at the failed coil, just to check the drive is still OK-ish
you dont have anything else to test with?
so might the starter from your car...hehe
cordless drill motor?
anyone know if these guys http://cgi.ebay.com/3KW-WATER-COOLED-SPINDLE-MOTOR-ENGRAVING-MILLING-GRIND_W0QQitemZ230392795570QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item35a47b11b2
bluntz: It's a 240V DC drive. I don't have anything similar
is the motor acutally 240v?
frallzor: I seem to recall from cnczone that they are, and you need a very high-frequency inverter to run them
(It says 400Hz in the spec, and most inverters won't go that high)
Yes, it says 230V DC on the end cover
how many amps?
It doesn't say that, but it is 550W
Anyway, I have concluded that the motor is dead, and the drive must be too.
it says in the auction that it's 3ph
seems I can get a nice danfoss inverter
Connect motor to drive, power up drive, motor turns a fraction, pop, lights go out. smell of burning. Considering that this is with the drive stopped and so the motor shouldn't be even trying to turn, this is a bad sign.
if my model is the right one It can do 1000Hz =)
dont reconnect it
No, it is not going anywhere but the skip now.
try a fan motor or something
I never liked the drive anyway, it has an odd habit of fighting you when you try to turn the chuck, even when stopped.
It's a 230V DC motor. They don't exist in great numbers in a domestic setting.
What doesn't matter?
you may burn up a less rated motor but
if amps arnt too high it may run
just to test
I only have one other 12V DC motor, and I need that to start my bike (and it costs $600 for a new one)
144 max on that 1
no old cordless drills dead in the garage?
I must have 4
Previous to the failure I could turn on the power and the spindle didn't move.
But you have a point. I do have an old mains powered drill.
And they are DC field-coil motors run on AC,
So that would probably work.
If I don't check back in in an hour call the undertaker (it will be too late for the ambulance)
follow smoke officer
"I wanna die in my sleep like my grandfather" "not screaming like the other 4 people in the car"
I have a question about tuning Stepper systems; I have migrated my control from parport to a 7i43, which is running very nicely. I want to optimise the speeds and accellerations etc of the x/y/z axis, and reduce the following error values, is there a procedure to do this?
OK, that is pretty conclusive. Dead driver too, it runs full-speed as soon as it is powered up.
I think that short might have killed the speed-feedback bits, or something.
what a shame
pjm: Do you have encoders?
no, its open loop
In that case isn't following error meaningles?
not really because its still used internally in emc
i.e. the stepgen_maxaccel value etc
Yes, but I don't think it has a great deal to do with stepper tuning.
i could just set the ferror and min_ferror to high values, but i dont think that is the right thing to do
(I am just keeping the conversation going until someone with a clue appears)
me too ;-)
but i guess i first need to optimise the max speed per axis
which i can do, but its tuning the accelleration and stepgen maxaccel etc
You can do it all in stepconf
ah i'm using a 7i43 with hostmot2
(If stepconf works with a 7i43)
its parport only as far as i know
i guess i need a stepconf for hm2
Yes, it will be.
to play with the speed/accell values etc
i just thought i'd ask whilst picking thru the integrators manual etc
However, the max speed and acell are a function of the PSU and stepper drivers, so that is unlikely to have changed much
i have tweaked my closed loop spindle speed control and it is now excellant, even when rigid tapping it keeps the spindle speed perfect
ah yes but in the old parport days, everything was clocked down due to parport IO speeds
i've got my microsteping drivers set at 51200 steps / rev
The lower "granularity" of the hardware step generation might well allow higher speed, but I think max-accell will be the same
the problem is it was never really optimised in the parport days
so this is what i need to do now the controller is 'proper'
Just max out the axis speeds in the ini file and see how fast you can go then, and then back it off a bit :-)
LOL yes that is possible, i just wondered if there was a proper way to do that!
Without any actual position feedback you can't really do much actual tuning.
i guess then i just need to optimise the speed / accel of each axis
my Z is working very well indeed, on a pecking cycle its acting pretty damn fast
You could try running a program that has a sharp stop and reverse (small radius) at one end, a long travel then a large radius at the other and run that cyclically. Keep increasing the accell till the motors stall at the fast end, or until you see the position creeping as it misses steps.
I found this interesting http://126.96.36.199/search?q=cache:8JGfdydFfJoJ:emergent.unpythonic.net/01188134101+emc2+ic+driver&cd=14&hl=en&ct=clnk&gl=us
yes thats an idea, is it possible to mod the accell value without restarting
I don't know.
You would need to look in the HAL guide and see if axis max accel is a parameter or a pin. If it's a pin you could wire it to a pyvpc custom control.
i need to do some more rtfm'ing
ah yes thats a good idea
i can use the hal meter to check that out easily
andy did you build the drive yourself?
I am not seeing any max-accel parameters in the hal configuration
No, it is the one that comes with the mini-mill
I need to build in some kind of overcurrent into my controller so that doesnt happen to my motors
Mine is (or perhaps was) one of these http://www.kbelectronics.com/data_sheets/kblc_data_sheet.pdf
andypugh, the MAX_ACCELERATION = etc is in the ini file
Yes, I was hoping it was then used to set a parameter visible in HAL, but it seems not.
ah ok, i was just picking through with the hal meter to see if i could see it
been out to help commission a new manual Sieg tonight
looks well made ish, but why dont the lazy bums adjust them properly
It is used to set stepgen.n.maxaccell, so actually it should be available
You can't wire a slider to it (I don't think) as it is a parameter
But in the HAL configuration viewer you can type HAL commands, and "setp stepgen.0.maxaccel 100" works fine
Because proper setup takes longer than building and would double the cost?
pjmcnc: As you are probably not using stepgen, you might have to look elsewhere.
apparently the hostmot2 stepgen uses these params too
There you are then, fiddle with setp in the HAL Configuration viewer.
doing that now, thanks for the pointer!
dont go to close to stepper limits
I think the idea is to find the stepper limits
Though I have heard that stalling a stepper can demagnetise them (rapidly oscillating field is basically a degausser)
I am not sure if there is any truth in that
maybe permanent damage to a PM motor?
that looks like a sweet drive andy
how much was it?
It came with the mill
I think they are about $75
nice 180 vdc
[22:15:42] <andypugh> http://cgi.ebay.com/KB-Power-KBPC-240D-Penta-Drive-DC-Motor-Speed-Control_W0QQitemZ270474074028QQcmdZViewItem?rvr_id=&itemid=270474074028#ht_500wt_780
(I think that the KBPC has superceded the KBLC)
[22:18:03] <EbiDK|AWAY> http://www.eevblog.com/
This might interest the electronics geeks here :)
is there a command to turn of the "tool" when seeing the paths being milled? my tool wont follow anymore
in that little window to the upper right that is, with the paths
[23:42:58] <frallzor> http://www.youtube.com/watch?v=DFM140rju4k
some nightly entertainment