we have g00 z30 m09, then g91 g28 z0, so go z0 home from the z30 above the job, what its doing is going to work zer0 first
apparently nothing, no answer. Hundreds logged in, nobody reading the mail
where g91 should stop it doing that, and just go z home from the 30 hight
so you're trying to move to no waypoint (G91Z0), then to G28 home?
g28 g91 z0, instruct it to return to home position on z only
you don't need the G91
I may have a clue finally. Inspecting the archive .deb with mc, I find it is in /usr/lib, maybe an lkdconfig is in order?
[00:04:25] <SWPadnos> http://www.linuxcnc.org/docview/2.2/html//gcode_main.html#sub:G28,-G30:-Return
gene, could be
g91 stops it going to abosule (work zero)
in theory :)
g90 it will return to work zero (wiping out my parts)
it did it fine on the lath we did, older version of EMC
you could just do G0 Z#5163
v 2.0 something i think
what is the Z value of the G28 position?
I don't think v2.0 had lathe support
atryed with out g91 still does it, maybe its a graphics bug, as using the simulator here
good night all
will G74 (left hand taping) be done with g84 when sorted? just inverse of g84 basily
can it do left hand rigid taping?
ok, we off to make new jig and tool holders, good night all
;) thx for help all,
back agan, and emc is having the laugh of its life at me
for some unk reason, after I'd hand installed it for the 20th time by taking the archive apart with mc and copying the files to where they went, and running an ldconfig, and 3 reboots, it suddenly started working, without X ever finding either the glx or the GLCore modules to load.
and the ^#()^ overrun error is gone with a 38xxx base_thread
run latency-test for a while..
Now the question is, why did emc spend all that time bitching about glx, when its not loaded right now?
It will give you a good idea of what the machine will do..
heh - um, well. at least it seems to work now :)
emc just wants a GL library. I guess if there's some misconfiguration (like trying to load the nvidia library, but not having the mesa library around), then there will be no GL support
eventually, you probably had something that would have worked, but we were chasing the X errors, not looking at whether EMC2/AXIS would run
[01:48:45] <skunkworks> http://www.electronicsam.com/images/house/kitchenalmost.JPG
[01:48:49] <skunkworks> http://www.electronicsam.com/images/house/kitchenalmost1.JPG
[01:48:52] <skunkworks> http://www.electronicsam.com/images/house/kitchenalmost2.JPG
[01:48:57] <skunkworks> http://www.electronicsam.com/images/house/kitchenalmost3.JPG
hey you have my stove
(except mine's black)
that looks nice. I like the counter.
it's nice for folks to be able to be "near" (without being in the way) when someone is cooking
Yes - that is the main reasonn for the way it is layed out now. We would cook for friends and they would pull chairs into the kitchen to talk :)
Really like the stove..
yeah I like it
the only problem I have is it takes a while to boil water.
Yes - Same here.
maybe that's how gas ranges are. I've not had one before this one.
Same here :)
I know know that the big pot will take around 10-15 min to boil.. Just need to remember that ;)
but for sauteing and such - it is really nice
have not baked anything complicated yet.
One last thing, Steve, I found the /lib/modules/`uname -r`/kernel/drivers was still contaminated, one last reboot to clear that out if I can't rmmod it.
you didn't delete that did you?
sure did, it was the nvidia 169.09 version
what chipset do you have in that machine?
it rmmoded just fine, not in use at all, dunno, let me check dmesg
I've just been used to nvidia-chipset machines lately, there's an nvidia module for thse as well I believe
looks like everything is a VIA
ok, then nvidia-named modules should be safe to remove ;)
I just foolded around for aboutan hour, watching the motion...overruns sitting at 0, with a 28u-s base thread, sweet!
if it's screwed up and won't boot, you should be able to select an older kernel with grub, and reinstall the latest kernel+modules
if you choose, you might be able to use the nv driver. you should only have to change the word "vesa" to "nv" to do that, I think
but after the trouble you've had, I'd say leave it alone if it's working :)
dunno what it is right now, but that sounds about right
I hear that, deafeningly loud!
you got vesa working and your rt errors are gone?
that was a uname -r outrput
whoopee, pass me another suds!
first close the package manager and step away...
then reboot and run! :)
comment on the nv, its screen refresh rate is much slower that the 10fps axis assumes, vesa seems to be ok, even with a 28us base thread
the nv driver could be a little faster, I think it has a few accel features (but not full 3D acceleration)
that's "nv", not nvidia
but i'll double that anyway as its still 50% faster than stepcobf chose, yes, i know the diff
SWPadnos: I've also seen overruns with nv that disappeared with vesa. I think nvidia hardware itself is trouble.
that's entirely possible
I've only used the nvidia proprietary driver lately, on non-RT machines
its quitting time, i may have to swim 20 feet to get to the back porch now!
there it works pretty good
actually, I did try some experiments on a lower-end new nvidia card, but I don't remember the results
they weren't good enough for me to replace the onboard video, which (since the machine was headless) still left me with 200ns latencies
but i don't think its rt friendly at all. this is a gforce2 5200-128, cheap card 2 years back
many thanks for your patience guys, bbl
see you. glad it's working
SWPadnos: what was the name of that usb raid box?
SWPLinux: i scored some Peet's coffee before i left the east.... good stuff :)
Green Mountain Coffee Roasters :)
did you experience any New England Coffee when you were here?
maybe on mass pike
it's actually pretty good stuff
they serve it at the close-by coffee shop (that I go to avery second day or so :) )
anything but starbcuks burnt day old taste
yes, Starbucks is like McDonalds to me. you can go into any one of them and get the same thing. It's not great, but it's pretty consistent
I get used to it when flying, since that's what most of the airports have
starbucks is 'office rental' for 'virtual corporations' or at least that's the word on the intarweb
? meaning people who run thier biz from the wifi at starbucks?
right, well, more like a place to meet and still have free wifi
heh - I was trying to figure out what that was supposed to mean
i guessed first !
oh, you definitely got it well before me. :)
well, i've done it.. not at starbucks though
dpkg-shlibdeps is lame *whine*
Twingy's pkg problem?
more or less the same problem, yes
oo i think i got it
man. every couple of days, I get a new email with even lower cruise prices. I'm waiting to see if they'll pay me to take the next one :)
oh, actually these do net out about free
you could write it off as a business expense with a little wrangling
well, I'm the president and my wife is the secretary of the company. we can have a board meeting ;)
beats starbucks i guess
hmmm. well at least the coffee is free once you're onboard
it's not all that great though
my brother in law went to sweden to install lighting systems on new cruise ships. i didnt like the construction details, sounded like crowded bar architecture ( as many as possible and damn the fire exits )
well, they certainly have the lifeboats for it, but a fire onboard could be very annoying
shouldn't spread too quickly given the steel construction though
i was looking for some animations of gcodes, found these, helpful for those who never saw/used 'em http://www.xmlcreate.com/NCGuide/motion_codes.html
the pix dont show how the params relate to the motion tho
damn. now I want some Indian food.
just saw a listing for an indian restaurant, and now I can smell good food. must be hungry
it's probably all the acid you did in the 70's
hmmm. well, that's one theory
SWPadnos: how do you go about bidding on jobs? like figuring out how much time it will take, and what you're worth
that's a tough question
time is just experience
for simple jobs, I can think of similar things I've done before, then add in some fudge factor
but other crap is consumables, how much one hour of machine time costs, and what you want to make per hour
for more complex stuff, it's usually necessary to do some of the design work, so you can have a list of things that are simple enough to estimate reasonably well
ask local shops for quotes on some prints you give them, then be competitive in price and time
I haven't set up a shop rate, since I haven't retrofitted my shop :)
tomp: that's dangerous
I'm talking about EE and software work here, of course
raw materials and such is easy to figure out
CNC shop rates are anywhere from $50/hour to $200/hour, it depends
yeah, you can get two completely different quotes on the same part, and it's not always pure price anymore
especially not in american manufacturing
for my time, I have a fixed rate of $75/hour (though I'm considering raising it this year)
most of the cost-only stuff went overseas long ago
"value added" is the big question
you have to be able to eat a little time if you quote something you need to learn about while you're doing it
look at quotes from online mfctring sites for rough prices, sometimes you can use online calculators, differnt machines are priced with diferent rates
drill press tie cheap, jig grinding expensive
well, i dont have a jig grinder so that doesnt really bother me
lathe is less than mill
hmmm. the only online quotes I've seen are from e-machineshop, and they're about 3x "normal", from what I've seen
if you don't have a VMC you are going to have to compete on engineering prototype services
or a turning center, for that matter
yes that's what i'm doing, trying to figure out if it's worth it for me to tell someone i can build their prototype turbine design
that should be pretty lucrative, if you have 5-axis capability
unless the blades don't overlap or something lucky
not that kind of turbine, and it's not lucrative (he doesnt have much money really)
and if you can measure things that come off a 5-axis machine
oh, then I'd do it if you're interested, but not ootherwise
if you really want to start making money i suggest you take it at cost
one thing I've found to be true (you'll see it everywhere) - if you do favors for someone, they expect you to keep doing them favors
i dont think there's any money in the machining business, honestly
that's not what job shops buying new HMC cells think
so you can tell them you're giving him a discount because you think it's a neat project, or because you're just starting out, or whatever, but don't just quote a low price
but i dont want to get taken advantage of just because i'm learning about it
there's plenty of money to be made from machining, but you have to find the right market
i don't think you are going to be able to not get taken advantage of here one way or the other
if you want to be the guy that inventors can come to for help getting their designs out of their heads and into metal, then you could do well
making pencils won't make you much money unless you have a very good lathe :)
open a place that has all those cheap tools and rent it out to ppl
ugh no way jymmm
dealing with stupidity is not my idea of fun
it seems to work in the SJC area - who knows about IN
no, you charge them for it!
fenn: i don't think you have the tooling to quote "fairly" on a pure machining job.
damn this array is taking FOREVER to build =(
if you were engineering, i think you could do it.
JymmmEMC: it's a cool idea but there's not really a market here
malloc (N*M*sizeof(int)) - what's so hard? ;)
and that's for a 2D array even!
heh - so that's why you wantede to know what the drobo was called :)
you can't beat it for power, I think
hey - they have a CNC plasma cutter now
SWPadnos: Actually... it has a seperate power supply - fsckers, I have evenough dman pwr cubes under my desk already
and the NAS accessory is $200 - oh wth
who has a plasma cutter?
at least, they now have a safety class for one
and a laser too
and a tank full of sharks!
ONE MILLION DOLLARS!! muahahahahaha
i wonder how much debian-maintenance skills were won through much gnashing of teeth
i mean, do people read about this stuff for fun?
if i look at youtube for gcode examples, i learn that g-code is ghetto-speak for living by the gun/living like a gangster
we are clearly gangsters
1500 ipm 4 lyfe
hmm, stumbled on Mike Lynch's 'Parametric Programming for CNC Machine Tools, & Touch Probes'... looks like handy examples, will see if used at Powells
not likely, its an SME publication
hooray for closed information
your tax dollars at work
wait, the amazon page has a very friendly button 'find at a library' and there's a few that have it!
[04:56:49] <tomp> http://worldcat.org/wcpa/oclc/37450727
2 in Ft Wayne!
fort wayne has amazing libraries, yo
um- that was supposed to be "we are clearly gangsters, yo"
so it's a real book, that's something at least
none near chicago, but one in cleveland, and some in michigan, so maybe when i'm on the road..., even New Britain CT ( new brih-en as they say it )
heh - yeah, that's the closest one to me that I don't have to cross the border for
you can request inter-library loan
i think so
hell, worth a shot, good idea. if it works, thats a whole new resource
i got a Krall book on APT from Champaign to the local Elgin lib that way ( still all Illinois )
libraries tend to be much more generous with dead trees than bits, for some reason
yeh mine says free ebooks, but wont let you just download it
you gotta pick up a disk
a few other cnc books from the SME may be of interest http://www.cncci.com/pdf/ss-books.pdf
wow, wanna be pissed about pay-for-information?, look at a place where you can rent this 90$ book for only 35$ per month! http://www.bookrenter.com/products/details/9780872634817/Parametric_Programming_for_Computer_Numerical_Control_Machine_Tools_and_Touch_Probes_CNCs_Best_Kept_Secret/M_Lynch/?utm_source=googlebase_purchase&utm_medium=cpc
90 UKP is not quite 90 USD
eh, nevermind :)
it's not $35 per month, it's $35 for one month
$49.74 ffor 125 days ...
alex had some comment about your question if you scroll back
way way back
seems I do not have backlog that large
log is at http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-03-20.txt,
sorry, seems to be no info really, just a comment
dimas: actually from previous day "<alex_joni> there was a guy in here a while ago, that wanted to build a stepper driver from a lot of resisters and transisters"
maybe he was thinking of Bo^dick
or BoDick^ or whatever
if i've been here for more than a day, i gotta go
heh - see you
alex_joni: yes, it was me, what would you suggest? a lot of resisters is not mandatory
better to put complexity in software than hardware - or lacking software, in electronics rather than mechanics
so, circuits with lots of resistors have their place
critical parts for me - pcb's ans special microchips
if alex could suggest another design - i'll check both for available parts
and some pile of resistors I have inhouse already
dimas: no, not suggesting anything else
was just wondering what happened to your project ;)
alex_joni: it's not yet started :)
number one cause of failure
mmm soggy spinach pizza
idea: would it be interesting to have an event log in emc2?
things the user did, error messages, etc
(most of the machines I worked with all have that)
it would if you could find it from the menu or someplace that is intutive...
that would help newbees I'm sure
BigJohnT: I'm more concearned about integrators that need the feature
to see what operators did/borked
if your a bork and are trying to build your machine it might help that group too...
except for Gamma-X LOL
TGIF is all I have to say
* archivist thinks of a few others with the "need"
* BigJohnT agrees with archivist
unit41 and kev from last night
the log should have LARGE RED ERROR messages
archivist: that's part of the log-viewer
for us old farts that need help seeing smallprint
the log should have [err:], [warn:], [info:] or whatever
what's a log-viewer?
and be in plaintext
probably in /var/log/emc2/configname.loh somewhere
BigJohnT: the stuff you use instead of a regular text viewer, if you want to look at LARGE RED ERROR messages
do I need to know grep to find it?
that is, if anyone is motivated enough to write one
hmm plaintext may not be plain enough for some, with added cluestick
I'm thinking a shock collar hooked to it
* alex_joni was beeing serious
alex_joni: write what?
something in python?
I am interested in putting the other part in place some day
emc2 writing things to /var/log/emc2/foo
for now this probably will go to the features one day might make sense
users can't write to /var/log. you could use syslog, but you can't ask it to create new files.
then maybe creating the files during postinst ?
or sudo make install?
just brainstorming here..
you can't just create a log file where you want it?
BigJohnT: you can probably create one in the config dir
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-03-20.txt
would you just overwrite it on boot up so it won't grow forever?
BigJohnT: dunno.. syslog has a very nice way of dealing with these things
logrotate & all
(plus we have some RT printk's which could make it directly to the syslog - maybe)
ok, I'm just thinking out loud...
now your talking greek lol
alex_joni: can you back up and say what your goal is please. seems like you started with a solution [write some stuff to /var/log] and I don't see how you got there
15:20 < alex_joni> idea: would it be interesting to have an event log in emc2?
a place to log various things about emc2 running, with 2-3 message levels (err, warn, info), etc
cradek: most (if not all) commercial machines I worked with have such a feature
most applications I use don't :-)
I just wonder what problem you are trying to solve by adding this complexity
once we know the problem maybe we can pick the best way to solve it
this is not solving any problem per se
cradek: thanks for the formulas the other day, I'm almost finished writing the Arc Buddy software now...
BigJohnT: you're welcome, but I don't remember what I did
the right angle formulas...
oh right, the trig stuff, you're welcome
right angle trig is one of the few things I still sort of remeber.
cradek: usually an integrator supplies machines.. right?
is there a special language to control scara robot?
acemi: not in emc2
now I controlling a scara with emc2 but I use G/M code
are you using the scara kins?
but for some movements, G/M code is a bit complex
acemi: I agree with you
skunkworks_: no, I use default kinematic
acemi: trivial kins?
I generate path with a python script
yikes. if you made some kins for it - your work area would make sense..
you could program normal x,y,z coordinates.
what is the status of scara kins which is in EMC2?
acemi: the kins work right
you need to adapt the numbers to match your robot though
arm length, etc
there are some issues regarding max vel/accel
there's even a simulation called scara-something.py
(try the scara config in emc2)
scara config crashes X window sometime in my machine
I didn't find the cause
to adapt the scara kins to my system, do I need to edit src/emc/kinematics/scarakins.c ?
or edit the hal file is OK?
* alex_joni looks
hal-file should be enough
setp scarakins.D1 .. scarakins.D6
I have a user interface based on tkemc. I think to use scara kins with this interface
So you can adjust the arm lengths within the hal file?\
this is nifty stuff..
you could thoretically adjust it during operation :D
taking into account thermal expansion of the arm length
acemi: x windows crashes when you run axis + scaragui but not tkemc + scaragui?
fenn: I dind't try scaragui with tkemc yet
axis + scaragui crashes sometime
my system is debian lenny
acemi: i had the same problem but it would crash every time. it also crashed when running two instance of glxgears though
(also debian lenny)
* BigJohnT going fishing
alex_joni: when you guys where getting low jitter numbers - how where you doing that?
jmkasunich, would the list item about two motors also get an accuracy error due to twist in the shaft and possibly an extra resonance mode
skunkworks_: got lucky on a good machine i guess
skunkworks_: but the lowest numbers I saw on SMP machine
for the most part - I have not seen anything consistantly under 10us. I did do an smp kernel but it was like a 600mhz dual machine.
yeah, maybe a more recent one has more chances to work better
around 10us is probably the best it gets
these drives require only 200ns step len and step space
sounds like special HW to me
not parport based step/dir
busy day ?
what's happening here ?
everyone else working ?
ken: you gotta have patience in here..
people have other things going on, but they delurk once in a while
i know...been here..thanks
usally after US working hours :P
maybe you could help me
I can try
i read the manual for g2 and g3
i try to do that and it says...begining and end od arc different
i can get a circle at i and j but no arc
This is as good as this computer gets.. http://www.electronicsam.com/images/pent4.jpg
ken: what values did you try?
arc begins at x .1 y.95 ends at x.35 y 2.2 1/4" rad
probably need a bit more precision
in the math..
no, I think you mean it ends at y1.2
add zeros ?
your endpoint makes no sense
* alex_joni didn't check the math
oops thats 1.95
ken - do you have any cad program that outputs dxf? (that you know how to draw in?)
that arc works fine then
I just tried it
i just get ends are different
x.1y1.95 -> x.35 y2.2, i.25 j0 ?
Then I would use either cradeks relize script and create tool paths directly from acad http://timeguy.com/cradek/autocad
yes but it's still good to know basic gcode
ken: the error message will tell you what it thinks the two radii are. what does it say?
Or output dxf from autocad and use ace converter http://www.dakeng.com/ace.html
tha's what i wrote....i get different ends
Yes it is... (know basic gcode)
i'm new to this and want to learn the codes
Radius to end of arc differs from radius to start: start=(X0.1000,Y0.9500) center=(X0.3500,Y0.9500) end=(X0.3500,Y2.2000) r1=0.2500 r2=1.2500
^^ this is what I got when I used the first numbers you said
you can see r2 is too big by an inch
this is what i wrote for the first arc........
likewise, your error message should help you see what's wrong
g1 x.1 y.1
g2 x.35 y2.2 i.25 j0
no space left on target device
rm -r ~alex
error mess says different ends
ken: what emc version?
ken: it works for me. it would help if you paste the entire message.
2 is kinda vague
ken: we have currently 2.2.4 released as the last version
when you start emc2 you should have a splash wich says v2.2 or something like that
unfortunately, in mini, help/about doesn't show the version number
i got 2.2
mini help is nothing
i tried it in axis same results
ken: don't you get a more explicit error message in AXIS?
i'll tyr again brb
the error message is the same
you really start to see the jitter on the scope at those frequencies
sure you do.. 60 kHz is 0.00001 sec period
is that with double step or not?
yes - double step
ok, so 30kHz effective -> 30 usec period
10 usec latency surely is noticeable there
1/2 division on the scope at 2us
that was wrong
my math is probably borked.. it's 8pm after a long day
the jitter could be as much as 5 divisions on the scope?
well.. it's hard to sync to that signal
you mostly care at distance between two consecutive pulses..
the jitter seems to be less than 1 division while watching it.. But you know
fwiw I tested on mini and it does give the full arc error under "messages".
5 - Radius to end of arc differs from radius to start: start=(X0.0000,Y0.0000) center=(X0.0000,Y1.0000) end=(X1.0000,Y0.0000) r1=1.0000 r2=1.4142
[18:51:44] <SkinnYPupp> http://pastebin.org/24611 http://pastebin.org/24612
I threw together some simple box examples for ken but I gotta run, If someone would pass these along if he returns
thanks for all ya'll help
i tried the first...works fine
i posted what i wrote at http://pastebin.org/24615
i get three different results at three corners
first=worked second=1/2" circle third=different ends
ken: your center points are wrong. the center i,j is specified relative to the beginning point of the arc.
Radius to end of arc differs from radius to start: start=(X1.1500,Y0.3500) center=(X1.4000,Y0.3500) end=(X0.9000,Y0.1000) r1=0.2500 r2=0.5590
the arc on line 9 wants a center below (negative J) the start point but you have it to the right (positive I)
the arc on line 11 is even worse
you cannot just copy i.25 and hope it works. you have to understand i,j and fix them for each arc
Is there a limit to the number of printer ports in emc2?
I should say - 3 should work - correct?
i didn't copy antthing...i wrote each arc
skunkworks_: if you want more than 8 ports, you have to make a trivial modification to the source code.
ah - thanks jepler.
I think there have been reports of people running onboard + extra 2
the third arc....says different ends
that's because the center point (i,j) is wrong
did anyone notice the targeted advertising on those pastebins?
EditCNC software ...
skunkworks_: I used 3 parallel port ( 1 onboard + 1 PCI with two port)
they're all i.25 what do you mean?
I see TOrchmate cnc cutting machinery
ken: I and J refer to the relative location of the circle center
relative to the current location
having all i=.25 can't work for all cases
if they are all i.25, that means you are telling it the center of each arc is .25" to the RIGHT of the start points
that's true for the first arc, but not for the rest of them
let me try
" I and J are the offsets from the current location (in the X and Y directions, respectively) of the center of the circle."
i understand i =x j=y
but then wouldn't i.25 be 1/4" from x1.15 y1.95 ?
[19:18:30] <jepler> http://jmkasunich.dyndns.org/pics/gcode-ref-back.pdf
so ia j are the start and xy the end of the arc
i went look at the pdf
the start of the arc is whereever you are at the end of the prior move. I and J define the center of the arc relative to that point. X and Y define the end point of the arc.
radion 9200 runs glxgears smooooth.
[19:31:14] <cradek> http://blog.modernmechanix.com/2008/03/19/phonograph-disk-directs-automatic-machine/
so g2 x1.15 y1.95 i.9 j1.95 this ?
cradek: neat. can emc do that ;)
i did it ... i had to use negative i or j
Is it me - or are there shavings on the record?
this worked at http://pastebin.org/24622
i now get i and j(i think:) thanks cradex
welcome. glad you got it.
my name is Alessandro and i live in Florence (Italy)
Hello from Wisconsin usa
i trying an emc system whit Mesa m5i20 - 7i33 and two 7i37
i have some questions
is the rigt forum ?
sorry for my bad english but the italian school.....
first question : who load the firmware hal_m5i20 into the board ?
Emc loads the firmware onto the board when it is hal component is used in the hal file. There is a sample config here
[19:56:48] <skunkworks_> http://cvs.linuxcnc.org/cvs/emc2/configs/m5i20/
(also on the livecd
ok second question : i reed in emc manual the pin allocation of P2 P3 P4 connectors where P2 is the encoders and dac connector and on P3-P4 there are other 4 encoder and mask
can i use P3 and P4 as general i/o ?
I think that is the default 4 axis setup. p2 is the 4 axis servo/encoder counter and p3 and p4 is i/o.
(but you can use them as extra encoders if you want)
i need only 3 axis, but i put on p3 or p4 two input for an handwheel
I have seen the plan on anderswallin
Guest145, I don't know if the secondary encoders on P3 are used with the EMC2 driver
for a handwheel you don't need the hardware encoder input
i have an m5i20. i just loaded the stock emc configuration. here's halcmd show's output http://pastebin.ca/950743
( i have no real hdwr hooked up , just the m5i20 in a slot )
may be of use
ok, all 8 encoders are there. thanks
I guess those are inputs by default, and if you don't want to use the encoder feature, just don't connect the output :)
(the hal output m5i20.0.enc-0[4-7]-* inputs, that is)
yeah, dont worry about the names
wow, I think I've lost my ability to make a coherent sentence
you ever had that?
* alex_joni goes to bed
excuse me, if i use all the p3 and p4 input as i/o what happens
nothing bad will happen, you just won't be able to use the extra encoder inputs that are on P3
tomp, were you working on some document that shows the connector pinout <-> HAL pin translation?
yeah, but i forgot where i put :-(
I had considered changing the pin names in HAL to reflect where they are (like parportT), but that would be a pretty big config change, for litttle gain
then there's the idea that it's not much of an Abstraction Layer if you need to change all the higher level configs when low-level stuff changes :)
the old meaning of cpp was the c preprocessor, which allowed name changing
(more correctly, "#define foo bar")
yeh, i used it as a sort of mini language
I attended a class where the instructor mentioned that people actually make programs that run int he C++ template system
found it! hal pin to phys pin http://pastebin.ca/950773
accuracy not guaranteed
heh - cool. thanks
in italy are 9.40pm and is the time to close pc , thanks to all
jepler: pluto step doesn't have an encoder counter?
tomp sai l'italiano ?
non parle italinao, solo un picolo ( -10 spelling :-)
i like the way it sound tho :)
due cappuccini, per favore :)
(had to learn to order coffee in Firenze ;) )
stutta, con grappa
ok per i coffe
half the water
caffe ristretto con grappa
water in the coffe ????????????????????
yes, half the volume, same caffeine
SWPadnos: the universal stepper controller could use the encoder counter for threading - couldn't it?
correto yes but water never
so you can't get to them from hal?
the hardware allows you to ge teither the step count or the encoder feedback per channel, but not both
I was reading - that make sense now.
walkies!, bbl ( io pedo )
darn - so there is really no one hardware step generator solution for a lathe. (with encoder counter) yet
maybe, maybe not
it's easy to use a software encoder for the spindle
bye bye next time i register my name as "albova" good night
not with a USC input
spindles are sloooowwwwww
you'd have to add a secondary input device, like a second parport
unless 1-2 KHz is fast enough to count spindle pulses, which I doubt in many cases
[20:50:40] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=54431
however, if you use a USC channel as the spindle drive, you can just use feedback on that channel rather than the step count
so something like a gecko drive or a frequency-to-voltage converter could be used on that channel
(along with encoder feedback from the spindle, on that channel)
I was going to give him some hardware options.. but there isn't really an easy 1 soluton. I also thought the pluto step had 1 encoder counter.. but it doesn't (as far as I can tell)
I don't think jepler recommends pluto step yet
I don't see where he says he is using steppers
it looks almost as though the present setup has a step/dir drive for the spindle
if that's the case, the USC will work right out of the box
(assuming you can dedicate a channel to the spindle, which should be possible with a lathe)
that one LPT port is not enough to fully control lathe (beside step/dir, home/limit switches, spindle encoder there must be spindle on, turret control, coolant on signals \
when people say step/dir spindle, do they mean pwm+dir for velocity control?
with the right motor, a gecko can drive a 2HP+ motor
other high power drives also have step/dir as an option
I think there are mach interfaces that use the step/dir output to run a spindle.
there's one called the pixie, which I think may be no more
but those aren't positioning devices are they? it's really just pwm?
freq -> voltage converters
to let mach (which only does steps) work with real servos that need analog
oops, I just noticed I stopped caring about this
ok - he does say he'll have a step/dir driver for the spindle, so a USC is perfectly fine, with feedback
and mach only requires 1 index for threading.. ;)
which works perfectly, or so I've heard :)
from someone who heard it from someone
heard it from a friend who... heard it from a friend who.. Heard it from another you've been messin around..
thank you, Mr. Speedwagon
as a friend pernounced it out like rio grande
skunkworks: no encoder counters on pluto step.
I've talked about adding one (for spindle feedback) but there's not room without removing something else .. and anyway, I'm not that interested in developing pluto-step any more
I'd rather wait for seb k. to finish support for all the things that the vastly superior Mesa parport product can do
yay, got the drop-cutter code going again: http://imagebin.org/15087
the plan was to make this fancy search tree to dramatically speed up the algorithm - but I've seem to have slowed it down :)
archivist: regarding screw twist - it will be no worse than with a single motor
i find lotsa web info on how to extend battery life of laptops, but where can i find which ones have a long life ( look for really heavy ones?? )
with one motor, all the torque comes from one end of the screw, and twists it a lot (relatively speaking)
with two motors, some torque comes from both ends, so the torque in any part of the screw is less, and it twists less
oh, is that what the 2 motors on 1 screw is about!
a case for turning the nut and locking the ends down maybe
wow this is really lame. ubuntu ftl: https://bugs.launchpad.net/ubuntu/+source/sudo/+bug/32906