I cut some stuff on the hnc tonight! a 1/2-13 thread and the lathe pawn demo program
the best was parting - using CSS+FPR I got one long ribbon for the entire part.
setting lathe tools is much harder than setting mill tools... hope I never need more than 8 ever
yes, don't try to use pid for spindle speed control (ddt on the encoder counts) and then try to cut a thread
it goes nuts when the count resets
so, forget that, no big deal
congrats - very nice
hmmm - that's with the mesa encoder counter (with no vel output)"
yeah I wish it had the nice vel output
once the bugs are worked out of the hostmot2 - are you going to try it>
later with hm2...
that is what I say. ;)
stustev: how is your geometric comp?
played with it all day - I have the error down to < .015 tir moving A and B full motion simutaneous
It should be better and I can get it better but I think it still won'
t be quite right
do you think the comp math is right?
I could not understand it at a glance...
I was thinking I would start from scratch and see if our answers agree - which is why I drew the model
the application is funky - it seems to correct 'correctly' only when I have moved the linear axes to the exact numbers I have in the XYZ error parameters
It is hard for me to understand at a glance. I had to study it for a while to remind myself what it should be
that doesn't seem right...
so if you mount your indicating fixture somewhere else on the table, the rotary errors are no longer corrected?
I like the idea of you writing it from scratch - that way my logic doesn't lead you astray
I will try it Monday, but it seems as if I am not exactly on the numbers the correction is not right.
in my diagram it seems like E1 and A interact to give a significant error, also E3/B, but I don't see how E2 matters much
I have it < .005 tir when moving only the A or the B axis.
that seems pretty good to me
is the latest code on your website?
I believe they each will have the same effect - if they have the same error magnitude
the latest cincikins.c?
I am not sure. I will check
cradek, in the model, is E2 supposed to be the radial distance from the B center of rotation to the A center of rotation?
ok, thanks - the perspective hasn't quite filtered into my brain yet :)
my cad skills suck
SWPadnos: yes the latest cinckins is there - www.mpm1.com:8080/cinci/cincikins.c
cradek: your CAD drawing is exactly correct and very clear
it makes it clear to me that when B rotates "clockwise" away from zero, the head has to go up a bit first, then down. with E3=0, the head is not at its highest position.
E2 is not the distance from the B center of rotation to the A center of rotation - it is the distance from the B center of rotation to the centerline of the spindle
err, I meant with B=0, the head ...
ah ok - it's a "horizontal" number when everything is at 0
it should have an X and Y component
(it may anyway)
in my diagram E2 is a Z offset - sounds like this is wrong
E2 will only have an X component
yes - we are talking about E3
ok, right. B position doesn't change Y
Is it possible to embed vlc in an Axis control panel?
(relative to E2)
BeIdentityCrisis, not without writing or finding some vlc <-> python library
BeIdentityCrisis: what is vlc?
and then with work
network video player
VideoLanClient, video player
the position of the B axis should not change the Y axis position of the tool
ok, so there are X and Z components to the E2 offset
no only Z
the X component of the point at the top of the line E2 is the value of the E3 line
mechanically or mathematically?
I was wondering - it seemed that E3 and E1 would just be offsets to X ann Y
SWPadnos: no, it's much more complicated than that
yes - that gives the three offsets XYZ
imagine how the head moves up and down when B rotates
I thought it should be - it's mea nd the diagram, not the underlying math problem :)
the A carrier is like a pendulum
at B=0 if E3=0, the head is highest. at B=0 E3 nonzero, it's not
same for E1,A
so the A pivot point would be at the intersection of the E1+E3 lines on that diagram?
the disc shows the plane of rotation, not the origin
everything "past" the disk rotates
the E3 line, the E1 line, the vertical line all rotate
you won't see the E3 line move when you rotate A of course
to make the CAD drawing reflect the plane of rotation correctly you would have to move the disc to the other end of the E3 line
that's what I thought
yes that would be clearer
B is a pendulum also
it's not wrong though - the rotation is still around line E3
too bad we can't load (and chain) multiple kins modules - this is like a 2-joint Puma on a 3-axis trivkins mill :)
I have .006 in the value for E1, -.005 in the E2 and -.005 in E3 to get the accuracy I have now
to test the kins, you can set one offset to something big, like 10, and see what the endpoint does when you change the angles
leave the others at 0
yes good idea
tape a stick to the side of the head 10" away from the spindle
it doesn't take a very big number to see quite a change in the motion error
and use sim/vismach for testing :)
.010 and -.010 will yield a big change
on an indicator
SWPadnos: on your craptop, you're using x64 aren't you?
I don't have a craptop
SWPadnos: Yes, you do. I just have a crappier top =)
I had 7.10 A64, but I just stuck in a 200G HD and installed 8.04 A64
SWPadnos: Did you get flash going on it?
dunno - lemme see
I'm not sure
well, find out and let me know!
WEll, got it?
there may have been some place where it asked "do you want to do all these things that aren't free?" and I said yes
well, WHERE was that - Inquiring minds want to know =)
on the web, I'm sure :)
might as well say in the ocean or space
except google is much better at searching the internet than the ocean or space
[02:45:16] <SWPadnos> http://wiki.ubuntu.com/flashplayer9
is the homepage listed for the package "flashplugin-nonfree"
not exactly in this case.. the x64 is the gotcha here
I may have done it when I got DVD playback and AAC audio playback working
I'm trying sudo apt-get install flashplugin-nonfree now, looks liek it's loading ia320libs
I'm trying sudo apt-get install flashplugin-nonfree now, looks liek it's loading ia32libs
if it works, I don't care
may be - it think it did that her ealso
hmmm. that's a good deal for that laptop
or any laptop for that matter
nah, the OLPC is less than that ;)
and a bunch of the eee-like ones too
that's not a laptop, that's a glorifies speak and spell
that doesn't speak
stustev: do any lathes change tools with T1M6 like emc, or are they all like T0101 ?
and we have flash!!!
wow, my laptop is actually fast enough that the Geckodrive website isn't dog slow
SWPadnos: 1280x800 widescreen, had to enable the nvidia driver. Took me a while to get wifi going (solution: http://ubuntuforums.org/showthread.php?t=816780)
and then just flash =)
bummer - they don't seem to mention the weight of that one
will try a DVD now
that may make a good second laptop for me
can Sandie get one for me? :)
cradek: I am not sure. I will try to find out. I would say yes. Some lathes are built with a tool changer like a mill.
no big deal, just curious
is it only lathes that do that? I thought someone had mentioned a mill that also uses that type of T command
SWPadnos: One-day-only sale
(I think it was 6 digits actually)
JymmmEMC, well, it's still today ...
no idea what you'd use 6 digits for
SWPadnos: not TODAY
you have 4 minutes to get there ;)
T0101 is tool one offset one
T0102 is tool one offset two
and there is no separate prep/change, Txxxx does the change
and applies the offset automatically
hmmm - ideally with no spindle stop/orient/restart
I wish we didn't have to g43 all the time. I think it's silly
SWPadnos: This is till Tuesday... http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=6820473&subid=23186764&type=
a mill turn with a swing arm tool changer (or a carousel) would use m6 (m06) to change the tool - I would think
JymmmEMC, wow! is that one also relatively light?
I like the g43
SWPadnos: compared to your craptop?
no, I don';t have a craptop (I told you that before)
you must be confusing me with someone else
stustev: I forgot a g43 tonight and avoided a crash by pure luck
the machines that automatically use the number one offset for number one tool do not allow you to use more than one offset for a tool
oh hmm, I don't want to give that up
education can be expensive
it seems that defaulting to the offset for the selected tool makes sense, then use G43 to change it if you need to (assuming comp is on)
it's not education, I know what to do, it's perfection that I need
perfection is even more expensive ;)
for everything else there is mastercard
SWPadnos: this is what I have http://www.shopping.hp.com/shopping/data_sheet/kc489ua.html
[03:01:43] <SWPadnos> http://video.google.com/videoplay?docid=-5864097001633806354
I do like the way EMC uses the G43 for the 3 axis and 4 axis and 5 axis tool length comp. Other controls use g43.1 (fanuc) and g143 (haas) for 5 axis tool length comp
I would think the G43.1 K# could be done with G43 also.
G43 on your lathe?
(08:01:12 PM) SWPadnos: http://video.google.com/videoplay?docid=-5864097001633806354
yes g43 on lathe gives x,z offsets
ok - I had always used a G92 for each lathe tool
SWPadnos: I don't believe that because of the "girlfriend who's father has" grammar failure
cradek: where is the video of the lathe making chips?
you caught me. I made the whole story up. no videos.
do you even have a lathe?
I haven't even taken one photo have I
It does seem as if I have seen a picture but that may have been a picture of a machine 'just like mine' (or the one I wish to have) :)
maybe I'll take some when I get the area cleaned up - everything is buried now
you have probably been working on the bus instead of a machine
we hope to take the bus out tomorrow
hope the weather is nice
when you get the bus out of the shop and out of the way of the lathe - take a picture of the lathe :)
I had also used a G50 to set the axis values for the X and Z on a lathe with a Fanuc 5T control. That was in 1979 though.
the hnc originally had +-.79999 inch tool offsets - not enough to use exclusively
the 5T had limited magnitude offsets also
[03:23:02] <cradek> http://www.machinesforsale.com.au/corporatemachinerybrokers/m1/image9.jpg
jii is now known as h2i
is that your machine where you picked it up?
no, I was looking for a picture of a 5T
mine's older than that :-)
that looks like the 5T control - the one I ran was on a Miyano 7BC
I wore out a lot of the buttons
[03:30:19] <cradek> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=320274513641
here's one just like mine
(wow I bet nobody "buys it now")
there is no place to bid except for 'make an offer' - nice compact machine
that's a very dirty and crappy looking version of the one that I should have had for $500 (locally)
wow, you should have bought it if it was local
I did, but while I was trying to figure out how to move it here (only 4 miles, but still a pain), they decided they wanted $200 instead
I hadn't paid for it
the first NC control I saw was a GE Mark Century 100. LOTS of boards and a window air conditioner in the side of the control box for cooling
why didn't you pay the $200 instead of the $500?
err - I forgot the extra zero ;)
heh, window air conditioner - that's great
oh, big difference
I bet they still have it...?
yeah - an order of magnitude
still a good deal, but I didn't like the fact that they had changed it like that
could be, but they had 4 they were selling, and I think I had looked at #3 or #4
even if you had paid for it they may have refunded your money instead of letting you take the lathe
that could be
they seem nice enough, but they realized that they had way underpriced the units
when you see a good deal I guess you can't waste any time. it's hard to remember that while worrying about logistics though
yeah, and add to that that the $500 was a big investment for me at the time
did they actually run? $500 doesn't seem way underpriced to me unless they were still in use
yeah but then you would have a lathe now :)
it was under power when I looked at it
with an extra tooling plate even
it was spotless - it's an aircraft shop, they only used oil, never coolant
tooling plate for a set of tools on the cross slide?
you can swap out the turret top
with a set of tools mounted on it
yeah - the top plate that holds 8 tools
those go for $100 or more on eBay
this HGR one has a chuck. the chucks that fit that nose are expensive/rare (no cheap imports that I found)
I don't recall if this one had the chuck. it did have the pneumatic collet closer and the cutoff tool
the closer is neat. I don't believe in the pneumatic cutoff tool...
I don't know that it was pneumatic, but there was a separate parting device
I may just take mine off
yes it is pneumatic believe it or not
it has a limit switch at the bottom to say "I'm done"
I don't know how you adjust the speed
seems like parting without css/fpr would be as bad as on a manual lathe
hmmm - true
it would be nice to know how close to centerline it is
you adjust the holder on the slide
actually I don't even have the blade holder
but I do have one for the turret
this one's closer, cheaper, and I think it looks better: http://cgi.ebay.com/Hardinge-HNC-Chucker-2-5hp_W0QQitemZ190201838136QQihZ009QQcategoryZ12584QQrdZ1QQssPageNameZWD1VQQcmdZViewItem
though the class is broken on one end (a flat piece, luckily)
it's plastic - broken on purpose so the bar can stick out
mine is completely missing on that end
yep - I figured it had to do with some rigged-up "bar feeder"
the hole looks as if it is cut purposefully with a slot to allow bar feeder loading from the operator side of the lathe. the only break is above the slot. someone stabbed it with a bar of material
yeah, there's a nice-ish oval then a break toward the front of the lathe
that one's only 7 hours from here
heh, sounds familiar
count on 3 days
plus a few extra to go borrow/rent a trailer
yeah, I figured - getting there and back is two, then one extra to actually load and drive partway back
(7 hours without a trailer or a lathe)
amazing how long a 7 hour drive takes when you're going 55ish
I think that lathe plus a trailer sufficient to hold it is about the limit for my Jeep to tow
(which is why I bought the Jeep, incidentally :) )
lathe + control + trailer is probably 6k lb
to haul stuff you can't afford?
yep - 6500 lb tow limit
$2k is less of a struggle now than $500 was a few years ago
does it have the electric trailer brakes?
yes, if the trailer does
it's got a 7-pin socket
(I don't have a trailer)
the control and lathe will separate at the control end, it's all plugs except a very few screw terminals
I wonder if I can get a discount if I don't take the control? ;)
the price goes up
the control is all single phase powered. there is a lot of good stuff in it.
no MPG - bummer
the servo supply transformer if nothing else
yeah - I'd take it if possible
no but you could keep the thumbwheel tool offset setter
I haven't yet because I figure I should finish (or at least start) the mill conversion before buying another machine
(actually it works pretty neat - there is a "store and move" button that applies the new offset instantly)
haven't gotten it yet ...
but a lathe is "easy" - you'll be done in a weekend or two
yeah, a weekend or two of actual working on it ;)
(actually I think my mill will be easy after this...)
lots of stuff on a lathe
hmmm. I think we have a skunk in the neighborhood
* jmkasunich bookmarks the HGR one
I'm sure sanity will set in soon
curse the sanity - it causes all kinds of problems
heh, soon we'll all have them
it's a neat machine.
I got my optic interupt switch wired up right finally tonight. I am still not seeing the input in the halscope.
you still working on the craptop with ubuntu?
No SWPadnos' is the craptop, my $400 one is the crappiertop =)
are you planning on using it to control a machine? or sim mode?
I got nvidia video, flash, wifi, and DVD Playing going on it. Working on VMWare now
DanielFalck: Nah, there is no paraport, serial port, or pcmcia slots on it.
I am using emc in sim mode here for visualizing gcode for another machine
But now that you mention it, I'll toss in live emc and see what latency it has
DanielFalck: I didnt think about sim, but I usually design under M$ so didn't have that requirement
Jym1 is now known as Jymmmmm
who did the python picture of a mill a few of months ago and is it downloadable?
Hey, has anyone tried a PCMCIA paraport on a laptop by chance?
alex_joni has, I think
I don't remember if it worked well
SWPadnos: did you see the link I gave you last night?
yes - the $799 laptop?
yeah... but not that expensive $550
hmmm - ok. there were a couple
it adds 3GB ram, bluetooth, 200Gb instead of 120GB, PCMCIA slot, and firewire
[16:56:23] <JymmmEMC> http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=6820473&subid=23186764&type=
was that the 5.6 pound one?
549.99 * 1.0875 - 50 = $550
I *think* this is the $550 one... http://www.shopping.hp.com/product/FE653UA%2523ABA
oh and webcam too
are those on sale today or tomorrow? (or Tuesday for one, right?)
That $550 is on sale TILL Tuesday
Mind you that's $5550 including tax and the mailin rebate. EXCLUDING any extended warranty
No, to get a 2yr warranty for the $400 one was only $50 more
SWPadnos: I'm still leary of the the trackpad but suspect that's a ubuntu driver issue as i lose scrolling after disabling/enabling it and need to test burning cd/dvd
I've noticed on mine that the cursor sometimes goes wacky when you accidentally touch some other part of the pad
due to the added button on the trackpad
Same here, but i think it's a ubuntu thing, not HW
I know the reason now though - I have a Synaptics touchpad that can register up to 3 touches, but it reports the average position of all of them rather than the individual positions
also, when I reenable it, the help window pops up every time
But i suspect that since these are cheap laptops, that they will become abundant and lots of bug reports will come thru
unlike SOMEONE's craptop =)
synaptic driver tries to handle multi-touch? beware of the thumb touching the pad while typing! (hate that , suddenly typing godknowswhere)
SWPadnos: Hey, why did I start calling yours a craptop anyway? I forgot
I don't know - I was going to ask you that
maybe because the fingerprint scaller or webcam (or "Turbo Memory") have no Linux drivers
SWPadnos: I know it was when you were in the hotel in milpitas and it didn't do something that we needed or it didn't have somethin we needed
no, not needed
the only things that don't work are those 3 items, none of which matter to me
(the webcam could be useful, but not too often)
No, i think you were doing SVN and it didn't have something in it
oh, of course there's no parport :)
THERE YA GO.... craptop!!!!!
well, whatever. it does what I need
Ok, let me try and get vmware server installed and go from there.
Except test EMC
no, I can test fine
you don't actually need a parport to load the parport driver
and you can do a lot of testing using sim
SWPadnos: Move my machine with your craptop biotch!
that's not something I need to do, thank you very much
We did that day!
but I could, if I used it as a bludgeon instead of a controller ;)
in eagle, i could not make a wire connect to a 'port5' ( or any other size) is this because free eagle wont handle multiple sheets ( so off sheet nets are disallowed)?
Yeah, that's about ar craptop is good for... to be used as a big hammer =)
SWPadnos: You know to fret over it.... I'm just gonna raz you over it cause your a emc dev and your primary box can't push a machine =)
Jymmmmm is now known as Jym
heh - as if that laptop were y primary machine - imagine! :)
can I send the ouput from make to a file with | ?
| is used to send it to another program
like "make | less"
also, the > will capture stdout only, error messages probably go to stderr
there is an incantation to capture both to the same file, but it escapes me at the moment
make 2>&1 > foo (or similar)
HEY BIG j YOU EVER WORKED ON A FORD tAURUS>>> 96 ish??
uh no, the only ford I had was a 65 mustang and a 67 mustang...
its overheating.. we swapped out the thermostat but i know theres an airlock somewhere
mine was a 69... and i havent played with a ford since... its a buddies car.. what a pita
water pump shot?
how to tell eh??
grinding noises out of the pump?
blocked radiator, blocked hose
no grinding or leaking... so i ruled it out this morning
or radiator plugged up
also check the radiator fins
with the cap off and the engine warm or thermostat removed you should see good water flow in the radiator
the amount of pressure weve had in it today id expect ant thing plugged to have pooped by now
my 91 S-10 had overheating, the inside was fine, but the small fins had corroded away, so there was less surface area (tubes only)
not if it has been filled with hard water a lot and has built up calcuim deposits
this is a closed system you cant see into the rad direc .. the o/flow tank has the presssuer cap...
the radiator itself has no cap at all?
the flippin t/stat is on the bottom of the lower rad hose fcs
you can hardly get a glimps of the rad
tstat or fan temp switch?
t/stat is usually on top, on the exit of hot flow
is the radiator hose from the rad to the pump collapsing inside?
btw is it plastic housing rad?
and are the input and output hoses on the same side?
alot of this it metal
and side of in and out? same side?
there's a trick with same side hose radiators, they fail and it's not noticable
they're hot, but their dissipation is low
can anyone guess what?
dmess: electric fan?
dmess: It loks like it's electric. Does it ever turn on?
dmess: TSB 01-11-6 http://www.alldata.com/TSB/19/96191215.html
[18:13:34] <JymmmEMC> http://www.fordforum.com/ford/make_FORD/model_TAURUS/year_1996/id_01116/TSB.htm
hit the dealership and ask them to print out the TSB for you
[18:14:52] <JymmmEMC> http://autorepair.about.com/library/a/1g/bl602g.htm
btw how would they get coroded coolant if they used nonconductive antifreeze coolant liquid?
high mineral content in the water they used maybe?
well here we use distilled water+alcohol antifreeze
dushantch: 99.999$ of the ppl here use tap water in the radiator
distilled inthe battery though
hmm.. distilled water isnt used here
only in the batteries
dushantch: like I said =)
but then it's conductive, and it makes layers of calcium on the inside of head :)
* anonimasu_ sighs
That's why I asked yesterday why not 100% coolant and no water
dunno, here's like 5l of distilled water for 1.5$ :)
dushantch: HA, they charge $1 for a 16oz bottle of drinking water =)
but a gallon of distilled is about $3
(because there isn't high demand for it)
well, it's not common practice in this part of the world..
btw doesn't mineral content in water negates coolant liquid anticorosive properties?
we have here like coolant liquid that's 40% alcohol, 60% distilled water
I'd much rather spend the extra money on 100% coolant than a 50/50 coolant/water (even distilled) but nobody have been able to give me a good answer as to "why not?"
JymmmEMC: I dont think coolant has the same thermal properties as water
aggresive, usually starts to leak somewhere when you use it
well it has higher boiling point
the only corrosion I've seen alot of is when you've used too little coolant
yep, but when you start to use 100% it eats that corrosion and then starts to leak
jymmmemc: the 100% stuff is just premixed?
russian cars come with 100% antifreeze
oh wow you just mean straight antifeeze?
renesis I'm not tlaking pre-mixed
that's what I figure..
the 40% stuff is premixed
100% glycol or whatever it's called
[18:25:25] <anonimasu_> http://www.jtbaker.com/msds/englishhtml/e5125.htm
if you use plain premixed that would be OK
[18:26:06] <JymmmEMC> http://en.wikipedia.org/wiki/Antifreeze
Flash point: 111C (232F) CC
this is funky data: Melting Point:
WHY IS THAT WEIRD
dushantch: as a solid maybe
ER SRY CAPSLOCK
dude its antifreeze
-13 is low
cars are warm
should be like -40
its half water =\
That's also 90-100% pure.
but blocks srack when parked, not in use
oh is that numbers for the chemical?
but im pretty sure a car starting up would melt block coolant quick
I got a working opto interupt mounted yesterday. I hav'nt been able to see the index or port change state yet in hal scope.
tho yeh cracking blocks and shit, heh
btw. same side hose radiators suffer from breaking the membrane between upper and lower part so they don't cool no more as fluid doesn't go through cells
but goes directly from upper hose to down hose
dmess has never said if the electric fan is turning on either.
Could be something as simple as a fuse or relay
or temp sensor to turn on the fan
well, this is of no real issue, if you find it a problem buy premixed stuff..
well after a while in idle he should see that
dushantch: He should, but did he is the question =)
or just connect the wires on thermoswitch
I was driving from AZ to CAlif and there was a car broke down on the side of the road. They had to stop every 40min, after driving 20min. They saw three mechanics that could fix the problem. I was able to find out that the electric fan wasn't turning on and was a bad relay. Bypassed the relay with a hair barret and made it all the way into SoCal without stopping.
yep, I always have some rope, wire and duct tape in car :)
wouldn't have worked this time, needed something blade-like
used a rope from my jacket hood instead gas cable which broke :)
You dont think I would have stopped in the middle of the desert at night for dudes do ya? =)
well you can put wire in relay connector so to override it :)
nope, wouldn't have fit
well sometimes I stop, sometimes don't
Not in the middle of the desert =)
Well, if dmess stays more than 30seconds ask him to chekc the electric fan; it's relay; and the temp sensor
btw, iI can't think of a relay that couldn't be overriden with a wire, are standards different over the pond?
it's a .187" blade
kinda like this... http://www.freeengineinfo.com/wp-includes/images/relay.jpg
similar as here, I usualy wind the wire around blades and return it
fitting into a fixed socket
and no wire nor tools
How's desert driving? Stresses a car a lot? Sand everywhere?
seems like the people that go desert driving seems to break suspension..
that's offroad, this is hwy
No not a lot of sand damage unless in Palm Springs Area, then you can get snadblasted paint job
Jymm: nope.. I were thinking of desert driving like in the real desert..
Death valley aint a real desert?
heh.. I have no idea..
not desert enough :p
sand in diffs, trans etc. engine?
if you get sand into your diff you have bigger problems.
or into your transmission
btw. anyoone messed aroung with polymer concrete for machines?
or epoxy granite as some call it
I got my opto interupt wired to pin 10 and hal scope is showing it true. even if it is not attached to the opto unit.
true means high, which is the normal state of unconnected TTL or CMOS inputs
if you pull pin 10 low using a resistor, does halscope (or halmeter or halcmd) show it going false?
I looked at the other pins and they were high so I tested the pins and sure nuff 5v so a 100 resistor and it showed false, so wired it through the interupt and bam we got a great signal. so not only does it work but it isolates the pc from the drive logic supply.
thanks for reply SWPandnos
dushantch: fenn had done some concrete with rebar lathe bed , not sure of any additive http://fennetic.net/gingery_machines/index.php?cmd=read&page=lathe_foot_modifications&word=concrete
hm. cutting .125 deep on drill rod makes some heat
cradek: got coolant?
just in a bottle - not enough
the 'chip' eventually turned red, and touched the nozzle, and melted through it
does the machine have a sump, and pump, and all that?
so I stopped
if the chip is red, you are going too fast for coolant anyway
it's impressive to see a red glowing snake though
yeah it wasn't something I'll try again
either cut dry or slow down - coolant could cause thermal shock and chip the insert
its got coolant mixed properly to start with ... but its been topped up with water and glycol intermitently as she'll inhale it... the hoses on the other side of the engine are burping every once in a whole and it inhales more fliud
this apperanntly is a PITA engine once you let the air in...
dmess: checked if the fan starts?
it stars when running but dont run down at ALL
shut down and NO fan
its right cool now .. i have to top it up and restart it
see what happens
some engines are PITA if they get air in the cooling system as they get air pockets in the upper part of head
I don't believe you overheated it and cracked the head so you're losing wather?
the freakin thermostat is in the bottom end of the lower rad hose... i drained most of the engine doing thaT
funny that they put thermostat at the bottom :)
no leakes, but when you heat up compression gets in the water system and presses wather out through water cap
other than the broken plastic tri-fitting that they fixed before he got here.. but i dunno if IT was cause or effect
STUPID would be the opperant WORD
true, as when it's not on the exit water from head it's time to react is pretty high
you got the scenario... you been there
yep, but on other engines, european ones
i cant even tell if the water pump is workin bcz or the air lock
of the air lock
it works, you'd notice the leaks from front seal if it didn't woprk
im almost ready to poke some little hols in strategic hises and then tape em up
i figured... it wouldnt gernade internally without something pisssin
dushantch: gernade = grenade = (slang for) blow up
internal damage, cracks, etc
alex_joni: what's up?
alex_joni: I installed trunk over my current (latest and greatest) emc2. I now get the error: RTAPI: ERROR: version mismatch 533 vs 531. Why?
I should have asked: alex_joni or anyone else.
did you do a rip install?
What does rip mean? I did a build from source.
run in place
No. I didn't I figured I would go for broke.
Lerman: if you already installed from a binary package, it's not wise to also compile and install (that goes for any program, not just EMC)
you get bits and pieces of both versions scattered around your disk
I've seen that before and it has something to do with different versions of rapti
So, should I now uninstall emc2 and then reinstall from source?
EMC has a ./configure option called "run-in--place" for exactly that issue - it compiles but does not install, and you can run from that directory without touching your installed version
I don't know what is the best way to recover
maybe removing the program will remove most files
Lerman: that is what I do with trunk
apt-get remove emc2
personally I don't trust "make install", since it is the least used method of running emc
then install it again
developers use RIP so they can run several different versions by just switching directories, and users use the binary packages
Thanks. It's fixed. did the apt-get remove emc2, then the make install again.
Lerman: were you following the wiki instructions?
One thing that I find "funny" is that the binary install puts emc in /usr/bin, while the make install puts it in /usr/local/bin
like I said, I don't trust make install
Yes. I knew about RIP, but my understanding is that you need to set the environment variables by sourcing a file before running. I didn't want to mess with that or with sourcing the file in my login scripts.
Instead of make install, is there a make package?
I added a couple of bold notes about it on wiki
That would make the package and then it could be installed like any other package.
you have intructions to do that
Lerman: I just modifiy my launcher to be like this or similar to run 2.3 or 2.2.x
what is your distribution ?
Distribution of what? and how can I tell?
sorry it's a french word I search the translation
Lerman: make install does put it in the proper place
but for that to be /usr/bin you need to configure --prefix=/usr
by default autoconf packages chose /usr/local
(this is done automatically by the debian building scripts)
Lerman: I do advise the same thing as jmkasunich
remove the stuff you just installed
I would have thought that the default would be to do the same thing as installing the bin.
and re-run ./configure --enable-run-in-place
then make clean, make && sudo make setuid
I think it's fixed -- with the make install.
emc2 will run from the folder where you checked it out
yeah, but like this it's harder to work on different things
like TRUNK for the latest code
V2_2_BRANCH for the 2.2.x branch
another branch for interpreter changes, etc
I don't want to work on different things (on my shop machine). (Or, I don't think I do.).
distribution is correct in english too ( it's ubuntu, debian, red hat .... )
I think he's trying to install so he can run his machine, not for development work
Lerman: on your shop machine I wouldn't run other things beside the installed binaries
Yes. But distribution of what? Dist of emc, dist of ubuntu.
definately not TRUNK
aren't there some instructions on making .debs in the infrastructure or debian dirs in CVS?
hmmm. I wonder what this little spring goes to
the only command I remember to make debs is "debuild"
The binaries seemed to have a problem. The interp is missing the latest fixes. I figured I'd go with trunk, but NOT keep updating it.
that's the worst idea possible
you'll have a who-knows what version then, which is a nightmare to debug
Because trunk has bugs that won't be fixed.
because it's in no condition atm to run full production
it'll get there after a long while of fixing things
I don't really do full production. I do prototypes.
Trunk will never get there -- sort of by definition.
well.. if you're ok with things going haywire :D
it usually does before branching another stable branch
Do we know when that will be?
or you 'gotta have' some feature in Trunk and can deal with it not being stable all the time
BigJohnT: for that I'd still be more comfortable with a separate checkout of TRUNK
* BigJohnT nods
I wanted the latest interp.
I have made a mistake, the instructions aren't availaible for ubuntu, their are only available for build kernel .deb packages
cr-_-: there are instructions for rebuilding the packages
I want to add a python script that will be executed for files with extension .wiz -- where should I put it in the tree?
but usually apt-get source -b app should be close to that
Lerman: I'd add that as a filter to axis
it outputs g-code after it runs, right?
[21:09:13] <alex_joni> http://www.linuxcnc.org/docs/devel/html/gui_axis.html#r1_11_1
That is the plan. But my .ini file has to tell axis where it is.
What is the search path for such files. I've added wix = wixard2gcode (which is a bash script the calls a python program). Where do I put the script, where do I put the python program? And how does axis find them?
I have searched them but I can't find them :/, I think I have soon read them but I can't rebember where there are
wix -> wiz in two places?
if you put it in PATH then it finds it
for example in /usr/bin
or you can put it in /home/lerman/whatever/ but you need to add that location to PATH
But I want to make this script part of our installation. -- They are part of the wizard system I'm building.
Files with the wiz extension will automatically be prepended with the wizards so that they can be invoked. Gwiz will generate wiz files with invocations of the wizard subroutines.
Originally, I wanted the interp to just find the files dynamically. It can do that, but axis doesn't know about them.
Instead, wizard2gcode will take the wizards and put them at the beginning of the file.
The same mechanism can be used for user defined gcodes/mcodes.
The general question is where do developers (of emc) add scripts containing new features.
This is probably the wrong channel -- s/b emc-devel
I'm going there now.
alex_joni: SWPadnos was tlaking chit about you... said you tried using a PCMCIA paraport card on a laptop
hey boys.... Ney orleans is sinkin'... again...
I am using an optic interupt switch as a spindle index or tach. anybody gknow the average optic interupt puse limt?
I seem to be missing pulses on hal scope at around 1500 rpm
antichip , hello, are u using the parport input from the tacho?
yes pin 10 on para port pulled low by interupt switch
do u have a scope u can look at the signal with?
i recently built a spindle tacho for emc
no scope, it's on the list
and since the spindle can do max of about 5k rpm, i made a 4 slot disk for the spindle which clocks at 20KHz up the parport, and it is fine
do u miss pulses from 1500rpm upwards?
I'm usind a hand drill to spin spindle for testing, before I set-up VFD, so it is missing at 1500
acording to the top speed of hand drill
and the opto switch u have, does it have any response time constraints?
i.e. if u are chopping the beam too quick, does it not respond quick enough?
yeah I'm not sure couse it's a off my parts shelf unit
and the photo detector side of it, do u have to bias that yourself with a R in series with the supply?
and no it looks like on hal scope the pulses are just gone or double or a mixture, if that makes sence.
I may need to make fewer slits
how many do u have currently?
so i bet your parport cant keep up
yes I have a resistor in series with the photo emmitter?
that is a lot
yay well I got it that way?
36 sounds fine to me
1500 rpm times 36 slots = 54000 slots/min divided by 60 = 900 slots per second
if you are looking at it in the servo thread (1mS) then yeah, you got problems
I was hoping for a nice resolution, and emc reads encodes at many higher pulses so I am thinking the opto unit can't or is'nt responding fast enough
really u probably need to look at the o/p of the photo transistor with a scope just to make sure its ok
but the base thread should be able to handle 10x that
I agree with pjm - this is an analog issue - you need to make sure you have the proper LED current, the proper pullup resistor for the phototransistor, and a good clean signal at the parport pin
that is a job for real instruments, not software ones
u could try lowering the value of the R on the phototransistor side
by 10% or so
this would be a nice excuse to get a o scope
10% isn't going to make much difference
what value you do you have now?
u could probably use a soundcard o'scope tool
yay I could I am only drawing 10 ma and they are rated to 50
I was just thinking that pj
antichip: do you have a datasheet for the opto device?
then how do you know its good for 50mA?
umm is what I was told?
a guy on cnczone forums who helped me get them working
does he have a datasheet?
sorta working :)
what is the part number?
step 1: make sure you have the proper LED current - that is based on the datasheet and ohm's law
step 2: make sure you have the proper bias resistor - that is based on the datasheet and ohm's law
pullup on the collector of the phototransistor
heh - datasheet was the first hit on google
abs max LED current 50mA
they specifiy the on state collector current with LED currents of 5, 20, and 30mA
I'd go with 30mA - the higher the better as long as you don't over do it
what is your supply voltage?
ok, Vf is 1.7V, subtract that from 5V and you get 3.3 across the resistor
3.3V / .030A = 110 ohms
yay I got a 330
got 3 of them?
three in parallel will give you 110
No sorry I'm saying I am using a 330
ok, so you are only running about 10mA thru the LED
Or was I have 100 ohm
you are making no sense
was using 330 ohm , I have the 100 omh resistors , and I was pulling 11.2 ma or so
looking at the on-state collector current spec
for the H21A1 version, at 5mA LED current, collector current is 0.15 mA min, at 20mA it is 1mA min, and at 30mA it is 1.9mA min
if I had to guess, I'd say the collector current at your 10mA LED drive is about 0.3mA
5V / 0.3mA = 16.6K
so a 8 um on time and 50 um off seems plenty fast enough for any of my uses
so, if you stick with the 330 ohm LED resistor, your collector resistor (pullup) can be no lower than 16.6K
if you switch to a 100 to 150 ohm LED resistor, the signal will be stronger
5V / 1mA = 5K pull up resistor
yes I will go swap them now
Hey, um, what are we cheering about anyway???