metric is bars?
(101.1e3/14.7) * 16 should be in Pa
I thought it was 101500 Pa per atmosphere ??
what's a few pascals between friends ;)
1.1 is what is on the radiator cap, and it needed to be 16psi equiv
[00:15:51] <jmkasunich_> http://youtube.com/watch?v=O8L49Xq87ms
jmkasunich_ is now known as jmkasunich
speaking of big lathes, do you (jmk) remember where that huge old british lathe photo is
the one that had like 8' swing and could remove a ton a minute?
the one with something like 8 tools and triple back gears
I don't have the url
ok - it was in the metalworking dropbox I think, but I can't remember the name
or the year ;)
03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py:
* new fix for "change current axis with keyboard while mouse jogging"
* fix for a failed import filter with no file loaded
03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/ (debuglevel.py emctop.py lintini.py mdi.py): with emc runscript or emc-environment script or run-installed, the sys.path hack is not necessary; these scripts should always be run in one of these ways
ah - a ton an hour - cheesy
surface speed 6 feet/min
can you imagine what it could have done with carbide tooling?
or even just HSS
yeah, and the lubricans / coolants / coatings today
I'd hate to see that thing going at 28 RPM
with the 10 foot faceplate
I'd hate to crash it
"Immovable Object, I'd like you to meet Irresistable Force"
I'd hate to see what you need to crash it ;)
"pleased ta meetcha"
oops, sorry ;)
ah - my wife is done with her computer - it's now safe for me to upgrade it. bbiab
they are in the process of trashing a lathe like the one JMK linked to at work
it's in parts now
they moved it by hand
I wasn't there to watch though
rollers and such?
dunno, but it has to go down 2 thirty degree ramps
morons built the building on two levels
and then put the loading dock on the lower level
I've moved a few tons up and down those ramps
the place where I used to work had lots of ramps
there was a whole street of small old shops, and they gradually took it over
it drops at least 10'
the buildings were built wall-to-wall, and they'd just knock holes in the walls
but there was a slope
each building was from 6"' to 3' lower than the next one
hey, at least the people who did that had an excuse
the door to the lab at the bottom of the ramps has been wiped out a couple of times :)
I've thought I might do that once or twice
jepler: thanks for the tip on UNITS
I was replacing the radiator in my gf's car. It came with a radiator cap but was marked 0.9 instead of 1.1 and all I new is it needed 16PSI
yay, I have spindle control
SSRs are neat
just on-off though.. you can do so much better.. ;)
but on/off is luxury compared to no control at all
stopping everything when you hit escape is pretty nice.
How is it running?
this concludes my goals for the work to be done before fest
now you can relax!
how many extra in/outs do you have left?
I can't decide whether an RC-servo speed control would be a clever hack or just crap
skunkworks: 6-10 of each, if I wired it right, I think
I'm using only one dout (spindle on) and din (X home/limit)
cradek: RC servo to turn a knob?
I know someone that is hoping to mill a bunch of h_bridges... you do have an extra pwm out on the pluto.
can you do the mechanical parts without it looking like crap? and can you get a suitable range of motion from the servo (I think they are limited to 180 deg, knobs usually do about 270)
skunkworks: the spindle already has closed loop speed control - just have to tap into its controller somehow
jmkasunich: probably not, and you're right about the range of motion
did you once post a schematic of the spindle drive, or was that somebody/something else?
jmkasunich: the hookup couldn't be direct unless I'd want to give up a bit of top speed
that was me
you still have it online
I bet something creative can be done with an optocoupler, a few resistors and caps, and PWM
and duct tape
[02:56:45] <JymmmEMC> http://cgi.ebay.com/TEKTRONIX-2430A-150MHz-DIGITAL-OSCILLOSCOPE-CALD_W0QQitemZ290115167467QQihZ019QQcategoryZ104247QQrdZ1QQssPageNameZWD7VQQcmdZViewItem
not a bad scope
see what it went for?
good price, if it indeed has had all the work and cal done as described
I have a 2440 that I got from work for $200, it hasn't been cal'led for years, but I considered that price a huge win
This looks kinda neat... http://www.owon.com.cn/eng/pds5022.asp
Hesitation is wasting!
from that page - chinglish
oh, heh... owondigital.com
how much they want for those?
cheap for a digital
assuming its not a piece of crap
but is it a chepa digital =)
the manual is really chinglish
[03:05:58] <JymmmEMC> http://cgi.ebay.com/New-OWON-PDS5022S-Portable-Color-Digital-Oscilloscope_W0QQitemZ300112240748QQihZ020QQcategoryZ104247QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
his feedback is 100%
you should read thru the manual and see what you think
and when I asked him about the warranty and no contact info on the mfg, here is his reply:
The manufacturers website is www.owon.com.cn. It is a Chinese company,but if you wait a few seconds for the website to load (it is a Flashsite, kinda annoying), you will see an "English" link on the top rightside of the page. Once you click that, and the website slowly reloads,the contact info is on that main page. If you do ever have to contactthem, they have several people there who speak english. I have soldabout 50 of these now, and there has only bee
Thanks for looking!
I dunno if you've used scope much before - people tend to have strong preferences based on what they are familiar with, and if its different you might not like it
I haven't touched a scope in 15+ years, and it was analog.
they are using actual knobs, thats a good sign
the first couple generations of digiscopes tried to replace knobs with buttons cause they're cheaper
but buttons suck
I like that it has serial, not just usb
I'm really not sure if 25MHz is fast enough for today's stuff though.
you planning on hooking it to a PC?
Only for scrn caps mostly.
25MHz is fast enough for some stuff, and not even close for other stuff, so it depends on what you want to do with it
but I'm fond of serial over usb
25MHz scopes? heh
60MHz and TFT display here http://www.owon.com.cn/eng/pds6062.asp
ds2: new, not used, 100Ms/s two channel digital storage, 6k points/channel, for under $400 is really quite impressive
[03:12:31] <JymmmEMC> http://www.owondigital.com/products.html
assuming it doesn't fall apart in a year, or have usability issues
jmkasunich: but that's a DSO
yeah, and ?
$649 for the 60MHz
so the 100Megasample/sec is not always equiv to a 100MHz analog scope
its never equivalent to a 100MHz scope
jymmm already said its 25MHz
[03:13:47] <JymmmEMC> http://www.owondigital.com/products.html
yes, hence the snide remark about 25MHz
like I said, 25MHz is enough for a lot of stuff, and not even close for other things
I haven't seen ANY digital scopes under $1000 (of course I haven't been looking)
yes.. it is just I find a 100Mhz scope limit
what about the bitscope kits?
I guess Tektronix is wishing they hadn't made their TDS scopes in China
I wasn't counting PC based scopes
I dont want a PC based scope
no fucking drivers to worry about/deal wiht.
I need knobs to twiddle myself
its not the drivers I hate, its the controls
it isn't like that is any different from a Tek scope
jmkasunich: that too.
damn POS uses windows :/
tek scope you mean? or the china one?
they run windows
essential a PC
I bought a high end Tek for work, it uses windows
I swear the tek's must use doze, they're so fscking slow
pain in the butt, since you can't update them
turn the knob three clicks, pause and wait for the scope to catch up
the cheap ones I take it?
we have mostly TDS7000 series at work
and when I say cheap, I mean $5000 or less
they do sync nice though
(I almost never work with repeating signals, everything is single shot)
I'm a mechanical engineer, if it doesn't repeat, kick it until it does
jmkasunich: what do you work on?
I'd think a motor is rather repetious?
when they're working right, sure
not from his company :)
heh I see
when they're blowing up, you only get one chance to capture the event
hence the importance of a DSO
[03:22:33] <jmkasunich> http://www.ab.com/drives/powerflex/700L/
<---- my project for the last couple years
have you ever seen a motor degrade? My boss wants to put variable resistance in series with the windings so he can detect faults
no clue what he's talking about
hehe, that's what I immediately noticed too
Oh man.... another 30+ DVD's to burn *sigh*
I don't either, but I'd rather introduce more natural failure modes than that
usually motor failures are insulation failures (or bearings, but that has nothing to do with the electrical side of things)
I've seen reports of broken rotor bars
yeah, I guess that can happen
but it would either have to be a crap motor, or stressed pretty heavily
we were looking at bearing failures, but even with a hole drilled in the race, our motor is predicted to take 7 years to fail
you make motors?
no we do research on failure detection
in that case, you need to decide what class of failure you want to detect first
broken bars, insulation failures to ground, insulation failures phase-to-phase, shorted turns, etc
all of those things are pretty easy to detect offline, not so easy to detect while the motor is running
(until the smoke comes out, then its easy again)
I wish we could control those failures
control meaning what?
make them happen at will to test your detection?
make them happen in a controlled fashion
what size motors are you working with? fractional HP, a few HP, or big?
and are they running across the line, or on VFDs?
insulation failures are tough to do much about, it works for a long time, then you get gradually increasing leakage current that is undetectably small compared to running current
I'm thinking I should get an isolation transformer before I blow the circuit breaker in the Dean's office again
it gradually increases, but by the time its easily detectable, its no longer increasing gradually
I'm thinking I could come up with a motor modification to simulate that
these are 3 phase motors, right?
insulation breakdown to ground is the easiest to detect
just sum all three currents (assuming a VFD with current sensors, or external sensors)
or run all three phases through one sensor, which is even more sensitive
we're pretty good at small change detection, the real problem is making the failures happen
making failures that behave like real failures
its easy to make fake failures
the literature is full of that
I think I need a slip ring big enough to handle a shorted winding
what kind of motors are you using?
nobody misses them when they're gone
I assumed induction motors, but I don't understand why you'd need slip rings
kind, not brand ;-)
so I can short a winding to ground
the windings are on the stator
good point :)
mechanical engineer here
even a gearhead should be able to tell which parts are the spinny ones and which parts are the sits-still ones
in fact, I thought that was your speciality?
I was just thinking backwards
you made me think of a machining operation I saw at our former (now baldor) motor plant
a motor mounted on a lathe of sorts, with both ends of the shaft supported, and the entire motor turning around the stationary shaft
(this was about a 50HP, not something you'd want to drop on your toes)
they were machining the features on the mounting face of the motor that need to be concentric with the shaft
pretty wild, what were they doing?
like a c-mount?
best way to guarantee concentricity - spin it on its own bearings and shaft, after final assembly
we were looking at our motor's rotor, it looked like a lot of it was cast in place
yeah, they stack the laminations, then inject molten aluminum to make the bars
built up rotor bars and cages don't happen until you get to pretty big/expensive motors
ok, that's what it looked like, then they cast some plastic to space it out from the shaft
looks like it, might not be plastic
I thought shafts were pressed and keyed right into the lamination stack
there's this black stuff that looks like plastic on ours
could be a sleeve, or it could just be rings at either end, and its steel to steel in the middle
I'd be surprised to see plastic in there, because of stress and concentricity issues
I milled into it, it's solid
we wanted a motor with a broken rotor bar
you didn't mill the full length of the rotor did you?
maybe I'm not visualizing what you are... text is horrible for this kind of thing
mini DSO for $249... dual channel, 20Msp
just cut through one of the bars
if the rotor stack is say 6" long, you're saying it looks like they have a 6" plus a bit plastic tube around the shaft, and the rotor stack is around the tube?
it's actually shorter than the rotor bars
so there's a visible gap between rotor stack and shaft?
it looks like they took the laminations, cast aluminum rotor bars making a 5" hollow tube
then they put the shaft down the middle and filled it all up with plastic
[03:49:42] <jmkasunich> http://en.wikipedia.org/wiki/Image:Squirrel-motor.jpg
pretend they filled the white thing up with plastic
cooling should be an issue though...
[03:51:45] <jmkasunich> http://www.phys.unsw.edu.au/hsc/hsc/real_motors/large_images/DSC00073.jpg
[03:51:51] <jmkasunich> http://www.phys.unsw.edu.au/hsc/hsc/real_motors/large_images/DSC00070.jp
found a picture of the rotor, but it's on page 135 of a pdf
thats what I'm used to seeing
that last one that's missing a g at the end of jpg looks solid
oops, sorry bout the g
you're saying that's a lamination?
at the end?
the dark part is a steel lamination
the fact that the bars and shaft stick out a little makes me think one or more lams got grunched up when he cut it in half, so he peel the grunched one(s) off
if you look at the wiki page, the laminations aren't that thick, same concept as ours, just that it's spaced out from the shaft with solid plastic
actually, now that I look at that one, I'm not so sure
another image from the same page:
[03:55:51] <jmkasunich> http://www.phys.unsw.edu.au/hsc/hsc/real_motors/large_images/DSC00070.jp
before it had a date with a saw
you and your g's
same paste, hence the G
[03:56:37] <jmkasunich> http://www.phys.unsw.edu.au/hsc/hsc/real_motors/large_images/DSC00112.jpg
I'll have to look at it again, obviously
I'm going to go play with my mill
also mispeller, i believe
if I was how would I be able to do this ??
* Unition puts his hand in a sock and talks through his hand
better than thru your ass =)
no one likes a butthead
we put those in the crazy wagon that comes around annually
ooops sowwy bout that
numbers fn huge
how did you come up with it
nah, just testing a wireless numpad is all
but for those I use BCMATH,
treats numbers as strings
wireless keypad eh
... wana help me write some drivers for it ?
turn it into a chording one handed keyboard
might be usefull one day :P
PLEASE I NEED TTO CONFIGURATE THE INI FILE FOR ONE
i heave one step motor for 200 pulse/return
and one engine with 10 mm
hoe to configure ot?
your screw moves 10mm per turn?
are you using step and direction control from the parallel port?
lerman_ is now known as lerman
I hnow that its possible to use phase-drives on EMC
Plz anybody can help me to do it?
I want to turn on 2 phases at time, to improve torque
is possible to on steper motor in emc for phase-driver?
-- find the section on stepgen and read about the various types of outputs available.
it's on page 103
thanks for a while!!!!!!!
the figure you're interested in is probably 14.4
jepler: I saw the img you say, but where may i change it?
EMC2 is set up to dir-step heare now. I mean, what do i have to change and where? Thanks a lot!!!
Guest400: you modify the ".hal" files, core_stepper.hal and standard_pinout.hal
you would start by changing the line "loadrt stepgen step_type=0,0,0" to use the step waveform you want from figure 14.4
then you will have to revise a lot of other things: For instance, if you are using a 4-phase waveform, then you do not refer to stepgen.0.step or Xstep anymore
you refer to stepgen.0.phase-A
for each of these pins you will create a signal with "newsig", and then connect it to a physical pin using a series of "link" statements
e.g., if phase A of the X axis motor is on parport 0 pin 1, you would write these 3 lines: newsig XphaseA / linkps stepgen.0.phase-A XphaseA / linksp XphaseA parport.0.pin-02-out
The CALEA compliance deadline remains May 14, 2007... http://en.wikipedia.org/wiki/Communications_Assistance_for_Law_Enforcement_Act
i have one litle problem...when i exit emc2, all the parport pins stay at the same state...i would like some pins to change the state when i exit the program,
for instance, spindle on remains on, even though i close emc
is there a solution for this?
kwaj: what version of emc are you running?
Within AXIS, click on HELP, then on ABOUT
the version number is on the help/about screen
you should update to 2.1.5
then I think the spindle will turn off when you exit
and than it will work fine?
and then it will work fine?
then I think the spindle will turn off when you exit
then I think the spindle will turn off when you exit
Hmmm, never thought of that before. Is thee anything for EMC give give a "heartbeat" ?
JymmmEMC: there are many ways to do a watchdog type circuit
we even have a watchdog component that generates pulses while emc is alive
a pulse + some HW circuit?
ah, charge pump, not watchdog
same idea, different name
that component will generate a square wave while HAL is running and 'enable' is TRUE
what you do with that "outside" emc is something I know less about
"stop if it disappears" is the general idea though
What would I need to connect between the paraport pin and a SSR ?
you need a circuit that outputs HIGH when its input is a square wave and LOW when its output is not a square wave (or vice versa)
do we know the freq ?
JymmmEMC: google for "charge pump" might give you some ideas
it would be whatever frequency you use to run that component
half that, I think
depending on how you define frequency ;)
of the pulse
I'm sure it toggles every time its thread runs
JymmmEMC, the component is so trivial, that it would be very easy to change the frequency if you want
so if you put it on servo-thread you'll get around 500kHz
if you put it on base-thread you'll get a faster rate, e.g., 10kHz for a base_period of 50000ns
It just seems that a pulse -vs- no pulse (stable high or low) isn't all that "safe" in itself.
[18:45:13] <SWPadnos> http://cvs.linuxcnc.org/cvs/emc2/src/hal/components/charge_pump.comp?rev=1.5
hmm charge pump means more than one thing - I didn't find what I wanted right away with google
here are two circuits suggested by mariss. he says they want a 12.5kHz input frequency. http://www.artofcnc.ca/ChargePumpSafety.pdf
JymmmEMC, a charge pump (and some other associated circuitry) is a watchdog: it watches the pulse stream (from software), and if it takes too long to get a change in the pin, then it triggers something (which will usually bean E-Stop or similar)
(no personal experience with these circuits)
a charge pump is simply a way of converting pulses to a constant voltage (or increasing voltage, but we don't need that one)
SWPadnos: Right, I just would rather have something but within a small range than just go/no-go scenario.
the real trick is to make sure that an output stuck high doesn't give you a nice OK signal on the other side
but it's all hardware that detects something - you change a capacitor value to change the frequency cutoff
incidentally - this is the chargepump component: http://cvs.linuxcnc.org/cvs/emc2/src/hal/components/charge_pump.comp?rev=1.5
it's reasonably easy to adda parameter "divider", and only toggle the output every N times through the thread
or you can use a "freqgen" stepping waveform to set the frequency you want
It's more the HW side that I need to know more about. Came across this... http://www.uoguelph.ca/~antoon/gadgets/555/555.html
the 555 is pretty much the opposite of what you want, I think
thatconverts a voltage into a frequency, not the other way around
see Figure 10a
i updated emc to emc2.1.5, but the spindle on still remains on, even when i exit emc
near the bottom
:Basic Missing Pulse detector"
yep - I see it
seems like I'll need to AND the crarge pump and SPINDLE_ON too
ah, ok. they're more or less using the 555 as a one-shot, and as long as you refresh it within the timeout, it's OK
kwaj: when emc is running and you turn "machine off" does the spindle stop?
no, you run the charge-pumo putput through hardware like relays and output enables, so that if the signal is lost, the system is physically shut down
SWPadnos: Yeah, I'm thinking loose cable or boxes crashes.
right, that's an e-stop condition (or should be)
I'd AHTE for the spindle to turn on randomly
the spindle (and anything else) should never be able to turn on without operator intervention
that is a fundamental proviso of good safety design
SWPadnos: so is the charge pump from emc == estop signal?
but that's all separate from the charge pump / missing pulse detector / whatever you use to detect when the controller is disconnected from the machine (or crashed)
no, chargepump is different
there is an enable in the component, so the output could be turned off when EMC goes into estop (though I'm not sure I'd d o that)
so I need to LOGIC AND the estop signal, pulse signal and spindle_on signal to some circuitry
estop is almost always done in hardware as a series connection of normally-open switches (or relays)
if anything breaks the chain, then hardware should make the machine safe
yes, normally open, and closed either manually or when the unit is operating correctly
so if a wire breaks you're fucked?
so that if something loses power, the switch opens, and the machine goes into estop
no, if a wire breaks, that opens the circuit, and you're safe
so it's N/C, and opens of failure.
I say NOP because the relay/switch has to be held closed, and without power will open
during normal operation, the contacts will be held closed, yes
but the relay/switch itself is NO, since it needs power to close ots contacts
SWPadnos: don't scare me like that!
Actually, I think BOTH is a good thing - allows one to check for shorts or opens. in N/O and N/C respectively
well, you'd have to parallel all the opens to be able to see if any is closed ...
and series the closed contacts ...
(wired-or vs. wired-and)
Hmmm, didn't consider the open being parallel.
With three wires (common GND), there shouldn't be both conditions at the same time.
a single set of NC contacts is fine, since a wire break, broken relay, lack of power, ir any number of other things (other than a fused contact ;) ) still causes the machine to shut down
then all the wiring for both could be series (I think)
like cherry switches.
contacts that are expected to be open will cause everything else in that chain to be a "don't care"
that's the opposite of what you want with estop
you only need one set of contacts per device, don't worry about making it "more safe" by adding open contacts, additional chains, etc.
you'll end up seeing that a single chain is the only thing that can really be used, since any single contact opening must stop the machine
I'm thinking SPDT microswitches
(no sense adding more contacts)
Right, but wires NEVER EVER rub againest something and short together.
What if you toss in a resistor per switch... kinda "addressable"
Well, why not damnit! =)
no need, and it complicates the circuit a lot
if it's shorted, then you know something is wrong.
how long is the wire?
it's interesting to think that some kind of shearing of the wire could disable a switch and not stop the machine
how many resistors do you have in series?
yes - that could be a concern
btw, I'm thinking more along the lines of home/limit switches using the resistors
assuming the conductors are run together
it would cause a parallel contact that would cut out the rest of the chain
SWPadnos: My brain hasn't gotten that for on cbale length/loss yet =)
and even with a set of paralleled open contacts, there's no guarantee that you'd be able to detect that, since there's no guarantee that the "accident" would also short that contact pair ;)
all youneed is a TDR
still on 1st cup of coffee
in fact, you can characterize the e-stop chain (resistances of contacts, capacitance of wire, etc), then make sure it doesn't change
SWPadnos: that's why I was saying use BOTH NO and NC
I thought about that for very paranoid secvurity systems
thought about patenting it, but there goes that ;)
no - there's no guarantee that a single event will cause exactly one problem
Time Domain Reflectometer
"cable length tester" ;)
that one I know =)
but you can tell the values of capacitors, resistors, inductors, and wire lengths if you know how to read the TDR screen
a wire length tester is a simplified version of a TDR
Yeah, Like I said, my brain didn't get that far yet =)
ok, back to the random spindle starter
I really should get togethere a patent budget - I'm fairly sure nobody has done that in the security realm yet
SWPadnos: talk to Les
I need to figure out why I'm typoing so well today
It's ok, I speak typo
apparently I just learned it pretty well
[19:25:42] <JymmmEMC> http://www.elecdesign.com/Articles/Index.cfm?AD=1&ArticleID=6419
I googled for "missing pulse detector"
[19:32:23] <JymmmEMC> http://patft.uspto.gov/netacgi/nph-Parser?u=%2Fnetahtml%2Fsrchnum.htm&Sect1=PTO1&Sect2=HITOFF&p=1&r=1&l=50&f=G&d=PALL&s1=4142159.PN.&OS=PN/4142159&RS=PN/4142159
JymmmEMC: there's a nice board from Pmdx I think
breakout board for geckos + chargepump
[19:48:00] <alex_joni> http://www.pmdx.com/PMDX-131/index.html
the better boards are from build yer own
great prices and moderate shipping
trying to install EMC2 on a machine with xubuntu already installed.
I have the EMC2 live cd
the computer does not have an internet connection
I have tried apt-cdrom add
apt-get install emc2
but it cannot find the package
the ubuntu CD is not an apt repository, so you can't use apt-cdrom
anyway to use emc2 from the live cd?
not other than just booting it
thats why I have xubuntu installed on that machine
you cannot use the CD to install the packages on an existing system
you could get the packages themselves though
there are quite a few dependencies that you will need too, especially the realtime kernel
is there an easy way to generate list of dependancies
* skunkworks calles the power of the wiki
* skunkworks channels the power to cradek
[20:20:58] <cradek> http://pastebin.ca/495535
this is the output of 'apt-cache show emc2' on a working system
yes there is information on the wiki but I'm not sure if it's up to date
yeah... Iwas looking at the wiki which got me started on the the apt-cdrom
is your xubuntu 6.06?
I just removed that information from the wiki, because it is outdated and no longer helpful
is there a way to "apt download" to a directory
yes cradek 6.06
I'm not sure how to do that, sorry
apt always puts the files in the cache under /var or something like that
if you use synaptic, you can tell it to download only
Ziegler: can you just carry that machine to a net connection? might be easiest
petv I dont have an internet connection
an any machine?
and there's no easy way to temporarily put this machine on the internet? I notice you are connected to the internet now.
yeah... ya know I may need to just install a eto
I would get the packages and use a USB stick before I had internet on the machine
the CD is for this very problem - it's too bad you can't use it - is it because you don't have enough ram or some other reason?
yes low ram... so thats my next quesion... will emc2 even run with 64 MB ram?
I don't know anybody running emc2 on a system with that little RAM.
some of the oldest guis like xemc will probably run
but you'll be missing out on a lot of the new features in the more modern guis
yeah... I would like to use axis
I gotta take a crack at a few things
ill be back to let you all know whats going on
Ziegler is now known as Ziegler[out]
it seems like this would be an APT FAQ, but there doesn't seem to be a good answer... http://jidanni.org/comp/debian/apt-offline/index_en.html
here is one thing I found, though
(and even "apt-offline" seems to assume you have some slow access from the debian machine)
"He is completing work on a new computer operating system that could destroy the Microsoft monopoly, create a new Internet safe enough for children and the elderly, and provide Associates with a new way to find work and sell products and services."
* cradek checks the date
I found a new use for EMC :)
2 axis machine
X and A axes
Ziegler[out]_ is now known as Ziegler[out]
err you can add stuff with dpkg -i
even without using apt..
anonimasu, yes you can
[22:11:08] <jepler> http://pastebin.ca/495733
atmega8 microcontroller with firmware implementing the USB HID protocol, connected to an encoder for use as a jog wheel
it was much easier than I expected, since these folks now offer their USB HID stack under the GPL: http://www.obdev.at/products/avrusb/hidkeys.html
you where worried about update spped - seems to work?
Guest733: not sure yet -- I can't get very many counts/second with the encoder wheel I have
jepler: nice work.
Guest733 is now known as skunkworks_
skunkworks_ is now known as skunkworks
I don't have any firm data to back this up, but I corresponded with another person working on AVR USB firmwares ("usbtiny") and his guess was that at most 300 uS at a time was spent in the USB code, which would make the quadrature polling rate be around 3kHz -- OK for a jogwheel
that's still fast..
jepler: if you need more speed I guess you can stuff _more_ stuff in packets at a higher jograte..
no, the problem is not the amount of data to transfer (until you get to more than 127 counts per 10ms), but how long the microcontroller is stuck "doing" the USB protocol, unable to see the pulses from the encoder
127/10ms = 12.7kHz so it is not the limiting factor
jepler: isnt that where you need a quadrature decoder?R
so you can service that each timer interupt and throw data off the usb port..
in this case, the one microcontroller is doing both the quadrature decoding and the USB protocol communication
Is the live CD kept up to date?
still has the same name as I downloaded back in feb
I update it once in a while but not for every release
I don't think I've made a new CD since the beginning of the 2.1 release cycle
So I may as well install from the version I have and update?
didn't I see something about the update manager nuking the RTAI kernel?
it's probably time to do another one, since there are many ubuntu updates (maybe 200MB)
only when changing major releases (breezy to dapper)
the live CD has always been dapper, there's no problem
updates work great
Anyone here know how to change the default file handler for .rm files from totem to real player ? I didn't find anyone on the Ubuntu chanel who knew .. .
threaten it with your milling machine?
A flycutter would scare it... wonder if it would change its mind..... realy it's gotta be something simple I cannot find
[23:03:15] <cradek> http://linuxfud.wordpress.com/2006/09/03/ubuntu-linux-file-associations/
isn't that in your mime type config or have they left that behind?
[23:03:38] <eric_u> http://linuxfud.wordpress.com/2006/09/03/ubuntu-linux-file-associations/
two people with stfw skills
[23:03:57] <eric_u> http://www.google.com/search?source=ig&hl=en&q=file+associations+ubuntu&btnG=Google+Search
eric_u: don't give away the secret!
complaint of the day: 2.6 sucks
nobody knows about that site, I keep forgetting that
it seems a lot of people can't find the way to edit filetype associations - it must be a gui bug
I don't like the way it works
I expect rightclick -> open with -> type app name, click checkbox that says "always use this app" -> click ok
gumble .. I can easily twist this 48 edge/rev encoder fast enough to lose pulses. now, is it because of a bug, the polling rate, or the mechanical switches in the encoder itself (noise or limited speed)?
Here's my prob, I'm clicking on a vid link in firefox, it wants to launch external app, happens to be %totem is what it knows about. Selecting ok launches totem... it doesn't know what an .rm file is.. pressing cancel the file appears in the downloads box, but file progress never happens... maybe I just need to get another .rm file to be able to right click on and select always open with ...
jepler: my bet is on the mechanics.
doesn't firefox keep track of its own associations?
hello my genius programmer friends!
Edit -> Preferences
Downloads, View & Edit Actions
jepler: have another encoder?
skunkworks is probably right though, those things are pretty crappy
[23:12:33] <eric_u> http://ubuntuforums.org/archive/index.php/t-145334.html
that was an interesting thing
cradek: yes, I have the etch servo encoders which are optical rather than mechanical
and a 4-pin header I could hook one to (changing my program because it would be different IO pins, of course)
ds2, in firefox edit, pref window has a content tab, and a file types , clicking the manage button a download actions window appears, there is a file type spl for shockwave.... and you can change action but no provision for adding action...
Berry got his first 'city' mouse yesterday.
OK this mechanical encoder is rated to have a "bounce" of 2ms and "chattering" of 3ms; I dunno exactly what that means, but I would not be surprised to learn I was going more than one edge per 3ms when I snapped the shaft and lost pulses...
* jepler goes to dinner