nothing to worry about except for the IRC network admins
found a wikipedia article on it: http://en.wikipedia.org/wiki/Netsplit
have to do crawl back under my car...
jepler: set polygon_ratsnest off rats and such worked great
I know have .05 thermals from the pads.
toastyde1th is now known as toastydeath
hi all. I'd like to add a wiki page but the 'Basic' login doesn't seem to be working.
Is the wiki still public-editable?
Perhaps some things have changed in your personal info.
I checked and cookies are on; I tried to admin login with 'emc'
I've never tried before
That should do it.
After you do that go on to the sandbox and try there.
You should see the page become editable.
(Password does not match administrative password(s).)
Let me try with another box.
I just tried it, and it works here
you are using the password of lowercase emc without the quotes right?
you are using the password of lowercase emc without the quotes right?
let me try this other machine
putting it in the administrative field after removing the * that starts out there?
I got in, Firefox wouldn't let me highlight it.
anyhow, I'll do it from here :)
Machining graphite is similar to mining coal: http://www.panix.com/~dgarrett/phpv1/?img=20-graphite%2Fp1040203.jpg
Going to describe my 0-10V spindle control
great. thanks GNieport
cradek: you are correct about the A axis display in AXIS. It doesn't follow the tool. I was looking at the machine and watching the perspective display and thought it was following.
rats - I thought it was working
darn. think you can fix it?
HUH?!? - I can try but do I think I can fix it? NAH :)
I bet you can get the cone tipped. I'm not familiar with how the preview is generated so I don't know how easy that would be to fix
did you get the spindle working again?
I think I can get the cone tipped - yes - but the preview is another matter
yes the spindle is working
The table is cleaned off - I will put a vise on the table tomorrow and cut some metal
I tried to get a video camera for today but no dice. I will try to find one tomorrow. If I find one I will post some video tomorrow.
cool, I look forward to seeing that
I will work on the preview tomorrow - just to see if it makes any sense to me at all.
in trunk, emcmodule.cc contains the code that accounts for BCUVW coordinates for preview and backplot
the relevant function is called "vertex9"
thanks - noted
exactly what I needed
in the current non-configurable implementation, it first takes the offset specified by UWV, rotates it around the Y axis by B, then around the Z axis by C, then translates by XYZ
you'll have to write rotate_x to rotate A around X, then decide on the proper order to do the operations
that would be different for each type of machine I think
yes -- ultimately it'll have to be configurable somehow, but I haven't worked that out yet
that can get complicated
If I can figure it out to make my machine work then I think anyone can figure it out for their machine (with a little direction like tonight)
hm, the cone orientation for backplot is still done in axis.py, around line 969 in my copy. it rotates around z then y
969 glRotatef(rz, 0, 0, 1);
I saw that earlier today - cradek guided me there
I wasn't very active on IRC today
I just might start looking tonight - I have nothing better to do other than sleep
archivist_ub: by the way, aren't you the guy who wanted "A" axis preview? you should read this conversation if it hasn't scrolled away by the time you come back
stustev: ah sleep sounds pretty good to me actually
I'm not familiar with that
It happens on occasion
jepler: thanks for your help tonight. I'm going to bed too.
cradek: you're welcome
see you later
jepler: there are sure a lot of vertex9 references in emcmodule.cc - this will be educating
hi guys... who is Ben Lipkowitz?... he wrote cxf2cnc.py... I was wondering if he hangs his hat in here
thanks.... I am playing with his code.... very promising..... making an engraving widget
my python skills are limited... but am learning by example!
the only way! :)
best way to learn is to have a mission and figure out how to doit
I like python, but find some of it a bit confusing... its a big package to learn from scratch
fenn, are you awake?
SWPadnos, the cool thing abouts fenns code is that it uses qcad font files which are better suited to engraving than true typ fonts. They range from simple stick fonts to slightly more "complex fonts"
ah - I haven't looked at that piece of code yet
I'm scrutinizing axis.py and emcmodule.cc at the moment (catching up on some earlier conversations)
I am basically wrapping a gui around it so it is an imput filter for axis
fenn did some cool stuff like making the fonts cursive.... (connected like handwriting)
about 10 years ago I thought it would be neat to have fonts with embedded code, so you could have a cursive font that actually looked a bit different each time, and which would connect letters together more "naturally"
never got around to it though
fenn, must be across the pond somewhere... his code is all in mm
nope, he's just a smart USian :)
technically, metric is the national standard measurement system here
but nobody cares
same here, but if i tell my wife I have 154.2mm for her, she doesnt know what I'm talking about
all the better
jepler? If you build HM2 I should send you the latest source, the source in CVS is
outdated and has lots of inexplicable booleans for setting the configuration options
the latest source determines what module and how many from the pin descriptions
and moduleID records
it may make sense to send that to Seb
jepler was talking about making changes to hostmot, not hostmot2 (I think)
I did but he has not committed it yet
Also the UCF files for HM2 are different than the ones for HostMot
(all I/O pins are just enumerated (IO0..) so the same UCF file works
regardless for the special functions
ok - but it shouldn't be a problem anyway, since there's a 5i20 UCF file for HM2
I just mention this because the old source make it pretty painful to change configurations...
The new source also includes a buffered SPI interface (16 deep x36 xmit FIFO 16 deep X32 recv FIFO)
that I did to support some of our new SPI interfaced I/O (7i64 and 7I65 so far)
The upper 4 bits in the xmit FIFO select one of 16 channel descriptors that contain
the bit rate, bit count, CPOL CPHA, chip select bits, chip select to frame setup/hold
and whether the recv data should be pushed on the recv FIFO, so each channel that
shares one SPI interface can be quite different. I still need to add recv data deskew
for getting best performance out of isolated interfaces that have turnaround delay
in the received data
[05:09:05] <GNieport1> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Analog_Spindle_Speed_Control
Guest559 is now known as topls
topls is now known as linuxgrrl
linuxgrrl is now known as linuxgrl
topls is now known as sup
sup is now known as suppy
cool! I have my spindle speed display working a treat
btw is there a way to average a value like spindle speed so the values dont flicker around too quickly
add a filter hal component to your spindle speed output
anonimasu, could i pick your brain a sec about the filtering?
i have put;
net spindle-spd-1ppr lowpass.1.in
setp lowpass.1.gain 10
net spindle-spd-1ppr-filt lowpass.1.out
which i think should be ok, but the output is zero all the time
did you add lowpass to a thread so it gets run?
ahh! let me check but i think not
wicked!! thanks for that, its working a treat
that is the answer that keeps giving. it is a very common mistake.
ok i will remember to check that in future
me too - I have done it too :-)
ah yes just fiddling with the lowpass gain, the reading is now pretty stable with it set at 0.01
so adding a realtime rpm spindle speed indication is actually fairly easy (with appropriate pointers)
this is excellent, i can now add other stuff like spindle load etc
i tried a 1 ghz athlon mb and chip just now and I'm getting a latency of 749180!
it looks good until I start firefox...
can a network card cause latency problems?
I don't guess the network card is the problem, tried another one and got jitter as high as 11.4 million... (mozroute is my router computer)
hi alex, I went through all that and disabled all the relevant stuff in the bios.
the first mb I set everything up on was looking good except I get random realtime errors when running emc
this one is far worse...
what kind of graphics card?
first I had a PCI nvidia geforce 2, but then I put in a matrox g400
seems like I saw something about dma on harddrives somewhere but I can't find it.
hmm.. the matrox should be fine
it seems to really act up when I access the hard drive
the hd is the same one as with the other motherboard
SATA? or maybe SCSI?
no, old maxtor 6.8 gig ide
hmm.. that should be fine
mozmck: did you install 6.06 or 8.04?
maybe try the 6.06 on the older hardware
I might do that, or I might try and find a better board somewhere inexpensive.
mozmck: are you sure it's not a SMI issue?
what mobo is it?
if you just leave the latency test running (without touching anything), what are the numbers?
do they change after a while?
it has a via chipset and an amd athlon xp processor
mozmck, Did you turn off lagacy usb in the bios?
heh - nice typo :)
mozmck, I had a problem with latency spikes that were caused by kjournald (the journalling daemon used by ext3)
running the RTAI latency test would show good numbers with spikes every 5 seconds
but the latencies were never that bad
then again, I was using a solid state disk, so there was no seek time
hi again, I'll be in and out today, but I'll be reading. I didn't turn off any usb stuff. let me try that
what could I do about kjournald?
I would think RTAI would pre-empt kjournald?
tower maintenance today bbl.
mozmck, if you use ext2 instead of ext3, then kjournald doesn't run
I don't know why it's non-interruptible - it may be insuring that a fully consistent set of disk writes is done each time
hmmm, I'd have to reinstall to do that wouldn't I?
no, you can run an ext3 hdd as ext2
I don't know how that works when you want to use it as ext3 again
SWPadnos: it shouldn't matter I think
when you reboot in single mode for fixing problems it mounts it ext2 by default
after that it works as ext3 again
hm - ok
so I think the journal gets created each time on boot
well, the idea is that a crash can be recovered later, so you don't want to destroy the journal until you see that you don't need it
but I don't know what the code actually looks like
yeah, but on unmount it writes the whole journal and deletes it I think
except that a crash may result in no unmount being done
right, then on the next reboot it rereads the journal
I disabled all usb in the bios and the jitter is still bad. Maybe not quite as bad...
but if it were mounted as ext2 there would be no journal
I'll try ext2
mozmck: remember to save any important data first
mozmck: and have a rescue disk around
or a hammer :)
:) It won't be too bad to reinstall everything. I only had one important file on the computer and it's saved elsewhere.
S05vbesave S10acpid S10powernowd.early S10xserver-xorg-input-wacom S99acpi-support S99laptop-mode
I removed the above links from my rc2.d, rebooted, and now I'm only seeing a jitter of 12540 so far
hmmm, it just jumped to 15601. I think I'll leave it for a while and see what happens. I haven't gone to ext2 yet...
now the jitter is 680665. don't guess that helped
I think I need to build a small PCB mill
since I have these: https://tech.thk.com/en/products/thk_cat_main_five.php?kk=ah15s39hl&id=3872&st=SFIgOTE4&ss=Mg==&tt=MA==
and this: http://jmkasunich.com/pics/bosch-48k-spindle-2559.jpg
and a xylotex 3 axis board, and some old nema23 steppers
jmkasunich_: sounds like a plan
my problem is that I have too many plans
I don't have the rails to go with the slides, so that makes a good excuse to delay this project
(although it seems they would work fine on a simple piece with /\ edges (45 degrees)
lower load rating I'm sure, but enough for PCBs
load rating on proper rails is 350 lbs
if I stayed away from HGR I wouldn't have so many projects ;-)
I got 12 pairs of those little slides for $18
Heh, I was cleaning out junk at work yesterday... Got a brand new sealed-in-the-box external v92 USR usb-powered modem. =)
heh - I just fixed someones ME computer with an external usb modem. (wasn't usr though)
I wouldn't have kept it except it's totally USB powered, no external PS needed
same with this one/\.
I could not get any of my old pci modems running in it. Gave up
I have (now) three modems... a USR Vioce modem that's serial/usb, a USR Courier, and now this one
wait not courier, the REAL one used for servers and such
it came with one of our machines from italy (took a bit to find the driver)
Yeah, USR was notorious for not placing the model# on the deivce but on the box to locate the right driver, so I unsealed the box, and cut off the end panel that had the SN and UPS barcode and tossed the modem, cables, manual, cd all in a zip bag
tomp: you need a factor of pi in the watt to Nm/s conversion (because SI screwed up and decided angles aren't a unit)
P = T * w, where w is in radians/s
I tried ext2 but that didn't help on my jitter problem. Its as bad as ever.
sadly sometimes you just have to try a different computer
some are just no good.
seems odd though. this is the second motherboard/cpu/memory combo I've tried.
that does seem odd. I have really good luck but all my hardware is old.
what all is plugged into it?
The first is a 1.8 ghz amd and it was looking good on latency-test, but gives random realtime error's when running emc
that could mean too tight of a base period
you mean for power?
no I mean cards
by random I mean it will do it after running for different amounts of time. sometimes it takes a while and other times it did it right away.
I used stepconf to set it up, and entered the worst latency I was getting from latency-test
In this configuration I just have a matrox g400 and a 3com 905tx card in it
I have a g400 in my lathe. have not seen any problem.
in the previous mb that did better I had the g400... oh, I also have a netmos parallel port card in both configurations
yeah, I don't think its the graphics card
the mb has an amd chipset, 1ghz athlon cpu, and 512M mem
the first mb i think is an ecs with via chipset, 1.8 ghz amd cpu, and 512M mem
running hardy 8.04
same hard drive with both motherboards
this guy has some really neat stuff: http://www.panix.com/~dgarrett/phpv1/index.php
i wonder what the white pieces are made from
that's what some of the white stuff on the page are
dgarr: nice electrodes
fenn: panic! i ordered drives and amps yesterday =-O thx for the info
the engraving by panix is sweet, ah, when he says he has an ornamental turning emulator, he means he has a real machine that emulates a rose engine thru 'cnc gearing' cool
panix is just an ISP-thingy
maybe the 'white pieces' were... "This "faux ivory" material has an appearance similar to ivory and excellent machining properties that are important for the patterns and the fine-pitch screw threads used on boxes."
i thought maybe it was soapstone
what is faux ivory? acetate?
ok, Dewey Garret did some cool stuff there
'faux ivory' a quote from his site
bottom of http://www.panix.com/~dgarrett/phpv1/ot.php
his link engine is borqd for me, so copy the link and strip his php gibbereish off the front end
content management systems make web development easy (unless you're the reader)
look at his bowl on top of this page http://www.sqart.org/exh/scapes/p11.htm
says it's a castable resin, i think you can see some of the as-cast surface on some of the "cities" pieces
[19:12:16] <fenn> http://www.panix.com/~dgarrett/phpv1/index.php?img=03-Cities%2Fp1030338.jpg
whew - that was close
for the last day or two, I've been working on this part that needs a 68mm bore
68mm is 2.6772, but somehow I lost the 6, and was thinking it was 2.772
fortunately I only roughed it out to 2.6-and-a-tiny-bit, so I still have finishing room
but I have about 0.1" less than I thought I did
my measuring tools (and the boring head) are all in inches
for final boring, the only thing the CNC is doing is providing the axial feed - the boring head sets the diameter
I love using the digital guessing stick for conversions, stops nearly making that mistake
bore a correct hole in junk, make a pass in junk (and hope it doesnt cut anything) before trusting the boring head 'dial'. boring head graduations are crap ;)
right now I'm fighting chatter
wish I had an insert with a finer nose radius
don't use an insert then?
* anonimasu_ is looking at tooling
it seems like all theese high speed spindle use very high helix angles on the cutters
probably to throw swarf clear
the faster you make swarf, the faster you gotta get rid of it (fast edm is based on faster cleaning, gun drill based on high pressure flush, milling graphite based on high helix...)
I thought it had to do with cutting forces too
to get more force vertical onto the spindle bearings
hmm, like helix angle to cutting plane determines push/pull on spindle, never thought of that
ha - used a diamond sharpening stick to put a better point on the insert, no chatter now
excellent, norbrite to the rescue
jmkasunich_: what are you cutting?
making a bearing seat for a ABEC 7 bearing, so I want it as good as I can get it
cool, I don't get to cut much cast iron at the shop, I get to weld some from time to time
welding cast iron
you need to preheat it
it's not bad, just time consuming
and beat it
spreadsheets are handy things
I have one where I put in the boring head setting and measurements from the last pass
it calculates amount remaining, divides it into equal passes (so spring, etc will be the same for all passes including final) and tells me the setting for the next pass
'user programmable' boring head ( we program the user ;)
good night all
tomp: I think that's how the highspeed spindles manage to last with high feeds and deep cuts
tomp: like the datron machines
for countouring they use 45deg helix angle
(I'm looking at the engraving tools they seem to be of good value
anonimasu_: the velocitraptor, cool!
tomp: the spindles look kindof _weak_
that's why I were pondering
they say "the clear choice for micro drilling operations', i read... the spindle is 2Kw 60krpm, i think... not built for milling.
remember the fanuc mill-drills? where was that edge? lotsa buyers were disappointed or beat the sh(*& outta thier machines
hey, its way nicer than anything in my basement ;)
[22:23:51] <anonimasu_> http://188.8.131.52/images/89/8956891066.jpg
actually I'd like one
I want a mid
mid sized machine that I could fit somewhere at home
like one of thosoe raptor machines ;)
tomp: did you see their aluminium milling videos?
i'm loading method's fanuc drillmate bids now ( cnxn is slow today)
[22:30:50] <tomp> http://www.youtube.com/watch?v=UgtBd_6ujGs
it sounds nice
[22:35:03] <anonimasu_> http://www.youtube.com/watch?v=E82_AZWmeNI&feature=related
that kind of ui makes me love heidenhain.
that impresses me
* anonimasu_ stops youtubing videos
this is no good.
tomp: I cant stop once I start looking at cool stuff ;)
i found it.. google the haas 'office mill' , i saw them at TimTos 2005, it looked like they'd fit thru most home doorways ;)
somebody in taichung must make 'em
your right, back to work!
just look at that ;)
ah, i know a place in Tulsa with 2 integrex, and 2 story high palette changer... there's still money to be made in machining
* anonimasu_ nods
well, over here sending stuff to china is getting too expensive again :)
that integrex is amazing, i just watched it hand the part off from head stock to the tail stock. those mixer parts usually are castings, this is all done from solid.
tomp: I wonder when we'll see emc make a impeller for show .9
i bet cradek can do one now
how fast ;)
woof, now its handing an engine block off from tail stock to spcl jawed headstock
tomp imagines a public integrex with a credit card swiper and usb stick... walk up, mount your stock on the pallette, swipe your card, plug in your program, get a text msg on your phone when the job is done.