jmkasunich: good news, the pump has a glaringly obvious problem
one of the pistons is stuck
gummed up stuck, or broken/rusted/damaged stuck?
not broken as far as I can tell. certainly not rusted.
I put penetrating oil on it. maybe patience is a virtue in this case.
they are pushed up by the cam, right? and down by a spring? hopefully its stuck in the up position, no cam damage
yes all of that is true
there are typical lifters on a typical camshaft
they look just like regular valve bodies except the spring retainer is different
unfortunately I don't think I can get the spring retainer off, since the spring is already pretty much compressed all the way
but maybe it'll just free up with some oil and patience.
ISTR that worked for you a while back
yeah, a little less precise though
for example, I beat on the valves with a hammer
there are no rocker arms, are there?
so its not really "just like regular valves"
lifters push right on the base of the valve
what I think is the lash adjustment is a screw on top of the lifter
did you pull the camshaft, or are you just looking at the gap between stuck piston and lifter (or lifter and cam)
there is nice side access to the pistons/valves
I have not pulled the cam. doubt it would help anything.
with valves, you remove the valve from the head side...
no idea with these
with these things, do you remove the piston from the "head" side?
I really don't know what's in there
I was assuming that they pump into blind holes
with any luck I won't have to touch any of that stuff I don't know how to adjust
I think there is a "sleeve" that the piston slides into. there is a hole in the side of the sleeve which admits the fuel
ah, and the sleeve rotates to vary the admission (at leasts, some work that way)
yes I think so
there is some rotating thingy in there that I didn't mess with
(what I think is) the throttle equivalent is locked in place. there is only one correct speed for it to run.
I was wondering if the other cylinders could give you any clues to the construction of the stuck one
I wonder why it doesn't fire on the other two cylinders. there may be more than one problem.
that seems odd - there is only one correct speed, but there is a wide range of torque, and thus a wide range of "fuel quantity per squirt"
yes I expected some kind of governor thing, but I don't see it
it could have parts missing. hard to say.
hmm, wonder if the thing I think is the shutoff is actually the throttle
I have only seen it fully on and fully off, but when running maybe it does something else.
diesels are FM
well, not the engines, but the fuel pumps
I'm sure its not FM if you look close enough
but few people get that chance
an exploded drawing and a theory of operation would be fascinating
I'd like that too
for a much more practical reason ;-)
one diesel that I find interesting is the Fairbanks Morse engines that were used in WW2 era submarines
no heads or valves
um no valves?
looking for a pic
cool, found the doc I was looking for
[01:14:52] <jmkasunich> http://www.maritime.org/fleetsub/diesel/chap1.htm
gotta scroll down quite a bit
page 15 or so, fig 1-7
wow, opposing crankshafts and one injector running two pistons (1-13)
that's very clever
phase shift between the crankshaft to control "valve" timing
I've thought that if I ever got bit by the "build working scale model engines" bug, thats the one I'd want to build
unfortunately I don't think diesels scale down well
no, I bet not
its also neat how they start them - compressed air, injected into the cylinders
[01:25:27] <jmkasunich> http://www.maritime.org/fleetsub/diesel/chap5.htm#5F
wonder how much the FM injector pump has in common with Yanmar of 50 years later?
could be very similar...
if it works, don't fix it
anyone draw up something in 3d then output gcode for the part?
some apt examples are 3d http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC
tom1: you did that?
That's not exactly what I had in mind.
Say you have a 3d model of a circuit breaker. Then you need to cutout a faceplate to mount it in. Instead of copy and pasting the dimensions of the CB, just have it translate that over, then output the gcode.
and machien the faceplate that is
accounting for tolerances, etc.
I think that kind of process always needs a human directly involved, no matter how much software you have
in apt you might project those details onto a plane, then select a path thru those lines, then the post would give you the gcode
should that little hole be contoured or drilled? == human decision
a 3d model wouldn't tell you things like how thick the faceplate is, what size the thru holes should be, how it mounts to your panel, and so on
cradek: well, the basics anyway, then manually do any post processing converstion
just a single face, not the whole thing.
I guess I don't see what those basics would be
hmmmm, maybe I need to play with google sketch some more.
I think we all wish cnc could replace the requirement for the kind of knowledge a manual machinist has, but it doesn't
all it does is turn the cranks for him very patiently and precisely :-)
not replace, just make it simpler. not the tendous task of copy and pasting dimension.
(then forget one dimension =)
or a gotcha liek I almost had thinking the hole dimensions wee symetrical
you would think something like solidworks could use a 3d object to do a cut extrude on another object
not sure it can though
of course, that would be an interference hole, somewhat problematic
I'm not even sure how to get a 2d face out of SW
and also solidworks doesn't pay attention to collisions
so you can do interference fits
who have *.po for zh_TW
there is no .po for zh_TW
look at http://cvs.linuxcnc.org/cvs/emc2/src/po/README
if you want to make one
jast make zh_CN , http://bp0.blogger.com/_jQr0_XXsOWk/R_Xng4BqIFI/AAAAAAAACbU/F8ZzHtlQ5fs/s1600-h/1.png
translation of the word difficult
would you like to share the .po files with us?
Only a little bit of translation
Translation finished, repeat
I seem to remember reading somewhere that EMC will open another .ngc file with and O word... or am I dreaming?
this was discussed recently on the mail list
if o123 isn't defined, emc will look in o123.ngc
and named subroutines in trunk
I knew I read it somewhere
ok the file has to start with o
[11:37:02] <fenn> http://www.mail-archive.com/emc-users%40lists.sourceforge.net/msg06250.html
thanks fenn, that will make it easy to to my torch touch off at the start of each cut
hmmm, I must be doing something wrong, it doesn't execute the called file...
BigJohnT: it has to be in the nc_files directory (or whatever PROGRAM_PREFIX is
[12:28:34] <skunkworks_> http://www.cnczone.com/forums/showpost.php?p=434567&postcount=78
the reason you dont do current limiting in emc is because it's a safety feature
and the pc is not fast enough
nfc what he's saying about slow/fast response signals
fenn: "fast" and "slow" are decay modes for h-bridges.
"In fast decay mode, the load is connected to the power supplies in the reverse direction of drive during the decay portion of the off time. This has the effect of quickly dumping the inductive flyback current into the supplies during the off period. The slow mode shorts out the load during the off period, resulting in slow current decay." --
[12:35:33] <jepler> http://www.mesanet.com/pdf/motion/7i30man.pdf
my h-bridge shorts out the load.
with from what I understand also give a more linear in-out of the amp.
which is what I have seen
pete had a cool explaination on the list also. Someone was having a problem where it seemed like he had to input quite a high signal to the amp to get the motor to start turning..
He figured the amp was setup to fast decay.
[12:43:21] <skunkworks_> http://www.cnczone.com/forums/showpost.php?p=390997&postcount=2
fenn: that is where it is
I recall reading that certain stepper motor controllers will select fast or slow decay at different points in the microstepping cycle, but at the time I read it the motivation to do so went over my head
I tried o101 call and o<test> call with a files called o101.ngc and text.ngc in the nc_files folder
I also checked the PROGRAM_PREFIX and it points to the nc_files folder
* BigJohnT is off to work
gnix_oag: ni hao, looks good! can you submit zh_TW.po? thanks for your work.
gnix_oag: are you at university? ( your message lloks fine, but i speak yi dian dian :)
tom1: how can i put
ah, wait please
jepler: fenn: how can gnix_oag submit a po file?
[13:36:49] <gnix_oag> http://linuxcnc.blogspot.com/
gnix_oag: 請等待, 其他人將幫助您。謝謝。
[13:40:04] <BigJohnT> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=220220526875&refid=store
anyone... use alta vista to get a rough translation of gnix_oag's work .. use this tool http://babelfish.altavista.com/babelfish/tr
to translate this page http://linuxcnc.blogspot.com/
BigJohnT: "0101 call" command call 101.ngc, not 0101.ngc
acemi: "o101 call" is what I did
I tried "o101 call" and this command open the file 101.ngc
and I tried both o101.ngc and 101.ngc for file names...
acemi: did you have to have anything in the called file besides the g codes
I get an error message too but related with filename
I get an error message too but not related with filename
is this in 2.2.4?
ok that must be it
tom1: send me an e-mail with the .po file as an attachment. I'll add it to CVS, and it will get in emc 2.3. handle revisions of the .po files in the same way -- e-mail me the new file
if you have a file for emc 2.2, send that too -- be sure to clearly mark which is which
hopefully my spam filter lets the message through, most messages with chinese letters are not legitimate messages ..
acemi: I guess I'll have to figure out how to get the trunk or wait for 2.3... thanks for the help
there is a howto in wiki to compile from source
I'll try it this weekend, thanks
jepler: I asked him to send to me, and i will fix up to you, ok, i found an email address for you.
gotta go now must shut down getting keyboard cleaned...
tom1: email@example.com or firstname.lastname@example.org
got it. thx\
gnix_oag is a teacher 老師 :)
looks like he's dealing with the same problem i was (eps illustrations)
the problem is not that dxf or qcad can't handle dashed or bold lines, its just jmk's cad software exports it wrong
bbl ( really)
tom1: your mail
bbl ( really really , now out the door )
gnix_oag: thank you for the translation
tom1: thanks for helping
gnix_oag: 我有文件看起來好! 我現在將送到開發商。:-)
jepler: .po sent to you now
tomp2: your subject line says "zh_TW" but the filename says "zh_CN". Which is correct?
hmmm, looked traditional to me... gotta ask
CN and TW refer to different geographical regions?
traditional = ZH, new style = CN, gotta ask
traditional = lotsa lines newstyle = fewer
ditto for CN vs TW ( TW and some others use old sytle withlotsa lines, the so-called Big5 )
my reference say that "zh" is the language code for Chinese, TW is a geographic code for Taiwan, and CN is a geographic code for China
still gotta ask ;)
my system displays both properly, hard to tell
It could also be set for just "zh"; in that case, systems that are configured for zh_CN and zh_TW would both get this translation -- if people in both geographies would be able to read and understand the text, it's probably better than showing the English (but would they?)
there is also a 'fall back' mechanism to allow prioritizing which localization to pick first if the primary isnt available
it's a different unicode for CN vs TW, i dont think one file handles both
my tools identify the non-ASCII characters as being called CJK UNIFIED IDEOGRAPH-####
but I don't even know what that means
I bet it's intended to be zh_CN, not zh_TW
i've been reading and writing to him in 'traditional' and thats TW not CN, still composing msg
I've added it as zh_CN, let me know if I should change it
gnix_oag: 文件名是' zh_CN.axis.po ' 是正確的? ' CN ' 和不是' TW '???
each character is a word, harder to get message overflow :)
holy crap i can read that
gnix_oag: 謝謝我們將保留文件是' CN '
jepler: looks like CN is correct
tomp2: OK, good
gnix_oag: thank you for your contribution
gnix_oag: may I use the screenshot from http://linuxcnc.blogspot.com/2008/04/test.html
... on http://axis.unpy.net/translations
where would one buy an USB stick online in the US?
a 16GB preferably
[19:04:36] <skunkworks_> http://www.newegg.com/Product/Product.aspx?Item=N82E16820211280
(first thing I found_
I honestly wouldn't ooh
that would go great in my eeepc
hey Alex, how goes ?
spring there yet ?
yeah, 20C and sunny
alex_joni: nice, it's still chilly up here but at least most of the snow is gone
I've been working a bit with Collada for the last weeks, been thinking that there might be a way to use collada to revolutionize the cnc world
Collada is an xml based 3d interchange format with physics and animation support
pretty much all the 3d modeling software can export to collada today, it's emerging as an open standard for 3d models
my reasoning is that using a sequence of machine specific instructions (gcode) to represent a 3d object is really inefficient
a better (future) way is to provide a generic 3d representation of the model with the basic data so that cutting paths can be generated suited to the machine that produces the item
I think they tried to do that using STEP
alex_joni: hmm interesting, I don't really know anything about step
* Dallur digs into google
oh, it's quite a sh*tload you need to read :)
alex_joni: any OSS implementations that you know of ?
Dallur: if somebody tells you that "all the 3d modeling software can export to X", they're lying.
In my day-job industry, they first lied that everyone was using vrml. then they lied and said that everyone was using a step-based format. then it was u3d. somewhere in there, x3d went by too.
* archivist notes solidworks missing from the colada list
but why would solidworks or catia want to interact with anything else
every major cam system supports sldprt files!
and catia, well, if you can afford catia you don't need to concern yourself too much with the particulars of it
jepler: yeah, there have been some pushes, but as far as interchange formats go today it's either fbx or collada and collada seems to be gaining momentum
Dallur: there are some things around NIST
for converting STEP stuff
but the format is closed, so I doubt there are good OSS implementations
oh, and STEP referrs to a lot of things
[20:12:42] <alex_joni> http://www.steptools.com/library/standard/step_1.html
jepler: fbx is closed source (sdk available) while collada is totally open and I think that will make a big difference
(it never caught as they expected it to.. imo)
Dallur: that depends what you need
jepler: it's u3d now?
Dallur: I mostly use ACIS (sat) or IGES when handling solids..
alex_joni: the first issue with it (as you pointed out) is that it's closed
but that doesn't contain anything about toolpaths
alex_joni: here at my day job it's one of the formats we've dabbled with in the past 2 years
alex_joni: I don't know how much use it's getting by our users in the field
jepler: I was very interested in using it
especially the part that you can embed it into pdf's
yes that's what we were doing as well
and have easy 3D views available to lots of people
using OSS I found Meshlab (at sourceforge)
it imports most types of meshes (obj, stl, etc)
and can in theory convert to U3D ..
I hated the time I spent working on that project. the good side was that there is an ECMA document describing the U3D standard
there are some U3D reference implementations
also around sourceforge
the bad thing is that the standard was shitty quality
e.g., a certain detail of the compression algorithm differed between the normative text and the sample code in an annex, but the adobe viewer decoded according to the sample code
how do you do the embedding?
latex + <forgot the name>
presently we generate u3d files with 100% in-house code, and use the 'haru' pdf library to embed it into a .pdf document
"The great thing about standards is that there are so many to choose from"
can has cnc?
one small question. I build a CNC and I'll use EMC2 for milling... how could I do a reference move (move on X/Y/Z-Axis from zero to 100%)??
sorry for my bad englisch im german^^.... I searched for 2 days on google
I'm not sure what you mean by 0-100%
0-100% moving speed?
it works in real distance
0-300m fremd startposition to endposition
thats the word thx
yippee... emc2 sim built from sources on new Ubuntu 8.04 box http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Simulator_Build_On_Ubuntu_8_04_Hardy_Heron
ravennb: kannst du mal auf deutsch fragen?
ok... also ich will ne referenzfahrt machen
also dass er alle achsen einmal abfährt und ausmisst
ne das erste mal
dazu musst du home druecken (oder halt home-all)
das geht aus g-code heraus nicht
ja aber woher weiß emc2 wo mein 0 punkt ist (endlage1 von der X-achse)?
das haengt ab wie es eingestellt ist
meistens gibt es ein taster (home-switch)
thanks for helping, alex
sorry for speaking german ;)
not at all, it's great
oder speichert emc gar nicht wie weit es in XYZ richtung verfahren kann/darf?
ravennb: also das haengt von deiner maschine ab
hab vordere und hintere endlage öffner in reihe geschaltet
meistens ist das so: auf einer achse gibt es eine 0-position
dann gibt es ein soft-limit (min und max)
und es gibt noch min und max endschalter
die sollte man eigentlich nie erreichen (waehrend der normalen arbeit wenigstens..)
ja normal halt
ravennb: genau, jetz..
wenn du die maschine startest
kann es sein dass die irgendwo ist
und man muss zuerst eine referenzfahrt machen (home)
die stellt man in der ini ein (in welche richtung er fahren soll, mit welcher geschwindigkeit, usw)
und dann faehrt er die Achse bis ein Eingang aktiviert ist, und sagt.. ok hier ist 0
dann ab da sind die grenzwerte berechnet
ahhhh also in der ini... gut jetzt weiß ich wo ich zu suchen hab thx... und die grenzwerter muss ich per hand ausmessen und in die ini/hal? eintragen?
wart mal kurz
[23:25:27] <alex_joni> http://www.linuxcnc.org/docview/html//config_ini_homing.html
das ist die komplette doku fuer referenzeinstellungen
man kann verschiedene Arten zu referieren einstellen
thx ich bin zu blöd zum finden^^
ravennb: wenn du dich mit dem englischen nicht zurechtfindest.. einfach mal vorbei schauen und nochmal fragen
obwohl ich jetzt zu bett gehe (ich bin ne stunde spaeter dran als .de)
joa also lesen kann ichs
naja, mit dem kapieren ist es interessanter :D
nur schreiben wird schwer jedenfalls bei technischen sachen
joa also meine hal hab ich auch richtig eingestellt mit boost und sleep etc
joa wünsch dir ne gute nacht
ich nehm mal an du kennst cncecke?
ja schon bin aber nicht sehr aktiv eher nur lesen
da gibts auch ein emc forum, und gute beitraege auch
werd die maschine wenn sie fertig ist auch vorstellen
WE HAVE BEEN INVADED
dauert noch etwas (Mechatroniker abschlussprojekt) bin noch mitten am bauen
ravennb: auch in hier wenn's geht :)
toastydeath: you would think so..
#POLAND: FORMERLY #EMC
toastydeath: it's german :P
i was making reference to germany invading poland
not actually implying you were speaking polish
hmm.. they didn't :P
it was more like a polite border crossing :D
HOW FOOLISH OF ME.
ravennb: ich verschwind zu bett.. gute nacht, und viel glueck
anyway sorry, continue.
thx gute nacht
toastydeath: I'm off to bed (hopefully for good this time.. as it's close to 3am soon)
actually I saw a result from some research in germany lately
and they managed to identify a gene which separates people in different cathegories based on how they like to sleep
ones waking up early in the morning, or ones (like me) that like to stay up late, and get up late
and they proved it's genetic.. not a bad habit
clearly a victory for us
ravennb: wo in deutschland bist du?
toastydeath: they also thought it might get changed with pills in the future
ich bin oefters mal neben siegen
ok sagt mir jetzt nichts^^
das ist schon NRW
aber ganz an der grenze :)
[23:35:34] <alex_joni> http://de.wikipedia.org/wiki/Siegen
alex_joni: if i wanted to change it with pills i'd just go back to usign speed
toastydeath: maybe after getting back to speed you'll get it right :P
ravennb: kennst du Lahn-Dill-Kreis ?
[23:37:36] <alex_joni> http://de.wikipedia.org/wiki/Lahn-Dill-Kreis
also wo der ist ja...
ich arbeite mit einer Firma in Haiger
okay ist ja nicht soo weit weg
aber eigentlich egal ;)
stimmt auch wieder
na dann.. viel glueck mit der maschine
für die hilfe
so ich werd jetzt auch ins bett verschwinden gute nacht