Looking at the LS244 datasheet I am not sure it passes enough current to drive one opto, nor is it clear that the gate pin switches that current.
244 is tristate when output disabled, so the current is switched. and they can handle a normal LED current (20mA?) fo sho
Output disabled = voltage to the G pin?
it was called OE when I were a lad
there are 2 out enable inputs IIRC. one must be high and the other low to enable
[00:06:58] <andypugh> http://media.digikey.com/pdf/Data%20Sheets/Fairchild%20PDFs/DM74LS244.pdf
Shows one enable for each bank of 4
the LS373 has latched inputs too. I'm probably getting confused. that was my other favourite chip
you could put pulldowns on the inputs to the LS244, so you get a clean digital low output when the input is disconnected, if that's what you're after
No, I don't think that matters. As long as 18mA passes through the amp-enable opto until I tell it not to, all is good.
24ma sink from the 244, so that's good. your only problem is getting 5V from the hot side safely
There is a 5V supply for that.
3A of it
I'd be tempted to use a latch, so it couldn't trigger accidentally if the PC reboots etc
It isn't the end of the world if it does, the drives are active until EMC2 has finished booting at the moment and it isn't a real problem
i am just concerned that I am massively over-driving that p-port pin, and while I am sorting that out I might as well get better default behaviour
naah. the centronics spec was based on LS TTL, so all printer ports will give you that 24mA
it's just that when you overdo it, you've got a dead mobo these days
Yes, but I have 4 amplifiers on one pin..
OK, one tube of 18 74HC244s on it's way to me from RS Components. That should last the rest of my life. Don't suppose you want a couple?
amplifiers? you can directly drive opto LED with TTL
By amplifiers I mean stepper drivers
I have my supply :) a pile of old mobos anyway
if it's opto isolated, that doesn't affect the current drawn by the input
Each opto on the drivers takes 18mA @ 5V. I am sinking 4x that current on a single P-Port pin
ah. this would be a job for a bc108, if I still had my babani books
are the drivers on separate PCB? do you have to use separate isolators?
I asume there is no problem with connecting the ouputs of all the channels of the 244 together?
if you tied the corresponding inputs together
The opto-isolators in question are inside the drivers.
if fact that's silly. just tie the inputs together to yoy pport pin, and use 1 output per opto
I was rather planning on connecting the -V of each driver's enable input to a single input pin of the 244, then connecting all the outputs to earth.
So the 244 isn't a current-in = current-out device?
no. shottky inputs and darlington pair outputs
Bugger, so no use at all then?
Or not as I anticipated
it sinks its output from its +ve
I need to sink 18mA per stepper driver to somewhere, anywhere.
But only when the p-port pin is low.
one great use of Digikey.com is to avoid all the 'data sheet scams'. just goto Digikey and lookup any chip, free pdf's and no email plague.
I have an RS account. I check the datasheet and then order. But you can see all the datasheets without an account too, I think.
that's pretty much what you get with a pport pin. when it's low, it will sink 24ma
the 244 would just be insurance
I need to sink 72mA, controlled by one p-port pin, as I have 4 drivers.
use a 7400 for cheapness even
from digikey (via ti ) http://focus.ti.com/lit/ds/symlink/sn74ls241.pdf
is for 240 242 244
7400 is inverters tho, I forget which one is plain buffers
one pport will drive up to 10 TTL inputs
I am not trying to buffer a signal though, or at least I don't think I am.
if you use 4 buffers in parallel, you get in effect a digital current amplifier
Yes, these aren't TTL inputs, they are 18mA opto-isolator LEDs in a sealed drive unit.
This is the driver. I have 4 of them. The input circuit diagram is at the very end.
[00:36:36] <andypugh> http://www.motioncontrolproducts.co.uk/pdf/MSD325_stepper_drive_datasheetV2.0.pdf
so, half a 244 with the 4 inputs connected to one pport, and you're sorted
And that arrangement sinks the opto-current to the 0V line, not the parallel port?
yeah the 244 will pull them low
so current will flow if the anode is connected to the same 5V
as the 244
though a current limiting resistor, like any LED
So, if I pull-down the input to 0V current will flow through the optos even with the p-port off or not connected, and then when the p=port goes high, the current will stop? That sounds perfect.
it doesn't pull it down to 0V, but to around 0.7 to 1.4V iirc
I don't really care as long as current flows and photons happen inside the drives. This isn't logic-level stuff.
And the buffer inputs are high-enough impedance for me to not worry about limiting in that direction?
Or I could just leave the enable terminals disconnected :-)
one pport will drive 10 TTL LS inputs at least, maybe fewer with a pulldown, but it'll work
andypugh: the drive needs the signal to go low with respect to the VCC. so, it wont work if you turn off the VCC and ground the line.
if you turn off VCC, the drive sees nothing, and does not activate PUL DIR or ENA.
you'd need a VCC supply independant of the computer with the pport.
I would wire the enables permanently enabled, as the outputs go hi-z when disabled
My issue is when VCC is present because the control box is powered up, but EMC2 isn't booted yet.
I like to be able to turn off the motors thought, to move stuff by hand, or just so they aren't buzzing away uselessly if I am meddling with HAL files etc.
I wish someone would port the mIRC whiteboard addon to xchat :)
you CAN discuss circuits without diagrams, but I don't like it
ok, nothing will happen until enable goes true, so the task is to not do that until emc2 is up and running.
you can get a lot of step and dir but they dont mean anything until after enable is valid.
That is what I am trying to achieve,
lets work on that. lets see if the group has done that ( i bet they have )
The issue is that it takes 18mA per drive to keep them disabled if they are powered up.
The way it works is that there is a +5V supply to each of the 3 input optos of each drive. The P-Port pins connect to V- of each opto.
So, pin-low = current flow = drive disabled, pin-high = no current flow, = drive enabled.
It all makes sense until you try to enable/disable 4 drives from one P-Port pin.
so use the 244 as a fanout amplifier
the drive pdf says ( i dont know why i can copy and paste from the pdf )... sez that 'usually left disconnected... enabled' and thus the crux of this problem.
permanently enabled, so outputs follow the input(s)
Aye, I was explaining to Tom3p what it was we were dicussiong.
why i cannot...
Because those words are an image on the PDF not texy
Yeah, it is all me being picky about not wanting drives buzzing away making motors hot when I am not even running EMC2
Permanently enabled would work fine, and free up a valuable p-port pin, but I decided I wanted to be able to turn the motors off.
I suppose I might as well buffer all the P-Port lines? I have ordered 144 channels of 74LS244s after all.
why would the drives be buzzing away if you're not trying to print? is your bios or OS trying to do PnP ?
I am not sure, but they do. Chopper frequency I assume.
try diabling pport PnP in bios and OS
Or maybe some level of microstep dither?
no, i can select each letter but thats o/t. the pdf is misleading when it says 'single ended' and 'differential'. it has a way to avoid using differential by only using the '-' . the drive enable is high enabled, so wont do anything until the pport goes active.
otherwise the pins should be in one state from switch on. one random pulse when you power up then silence
The motors don't move, so it isn't spurious pulses, so it must be noise generated in the drives themselves.
i dont think you have a real problem, try a real test
yeah the anodes are commoned in the drive, so you don't have the option of differential drive
Yeah, I am happy enough in the pin-high drive-active configuration, but in pin-low that shared pin is sinking a lot of current.
yeah one per TTL output, so use 4 non inverting gates of any description. 244 are handy to have a tube of :)
Well, the real test was last night, where one drive wouldn't enable at all, and another kept dropping out. I think that was possibly due to a spurious current path through a telltale LED though (one that I fitted this time last year when I knew even less than I know now)
that's 5 outputs then :)
Yes, I guess so.
Eek! 0210 and work tomorrow
I'm sure you can amplify TTL with a couple of cheapest transistors, but 244 is the lazy way to get on-spec I & O
first the neccesary signals: http://linuxcnc.org/docs/2.2/html/config_stepper.html#sub:Adding-enable-signal
2) how to implement that... pport to ls244
I already have an enable signal, the system was working until last week when I blew up some drivers, then started looking at things with a more experienced eye.
[01:13:04] <tom3p> http://linuxcnc.org/docs/2.2/html/config_stepper.html#sub:Adding-enable-signal
upto 500mA per channel ( things that can go BANG have guts )
wrong paste... http://www.st.com/stonline/products/literature/ds/1536/uln2805a.pdf
uln2803, use 'em a lot, 5V in, 24V out, real industrial type control signals
That's old stuff by the way, stepperconf does amp-enable nowadays, and it's all "net" not "addsp" nowadays.
heh. old docs abound
But anyway, I really need to sleep.
500mA would drive my biggest stepper directly
middle of \o/ gnite
yeah, but the absolute max power dissipation for the entire package is only 2.25W
that'll be a heat limit. those chips were invented before ram heatsinks :)
no, I won't try it, not this week
right, lotsa immed power, not much average
55 deg C/W is the thermal reistance Junction to abient
so get the ceramic military spec
I'm just trolling now :)
well all the numbers are there in the data sheet....
the amplifiers uses 18-19mA when active, and 0 mA when inactive. what was this all about? the inactive current draw? ( 24V, 1270 ohms -> 19ma or 5V , 270 ohms --> 18mA )
there is no inactive current draw
looks like the homing sequence succeeded. I'm running out of obstacles.
thats easy to fix
just say something like "what could possibly go wrong"
"Yeah, that looks about right"
works for me
20mA at 5V or 24V no problem, the 500mA he joked about would get the die rather hot
okay, so now i need to zero Z and X. My tool is currently at x0z0. how do i set current position as zero?
how does one zero the dro?
no idea, sorry, i don't use emc
g92 x0 z0
woo victory is mine
yah, i feel extra dumb i knew and forgot
just figured there would be a button. I'll add one
what if you ain't at zero
I usually use the "touch off" button
it's alive! I'm running lathe pawn example...
boy do i have abig grin on my face
flawless. glad there were no m6s in there:)
so i'm pretty sure my following error is hovering around .002 when jogging at 100ipm.
i can now easily send screenshots i got eth drop.
cool. I'd like to see your plot then. you should be easily able to improve that .002 if it is steady-state error
okay. i've been setting it to roll. I'll try to catch clean moves to files.
yeah, just take a screen grab of it
some ff1 ftw
well I'm off to play with the mill
screengrab? im an idiot. is this as simple as pushing printscrn?
yeah or alt-printscreen or something, whatever grabs just the current window
on std ubuntu instal: applications|accesories|take screenshot
thats mitey useful
hah printscreen works too (doh! tho its Fn Ins on this laptop )
sure does. now ive got a file, what's easiest way to get it to you?
[03:31:45] <cradek> http://imagebin.ca
stick it there, then copy the url here, then everyone can see it
here goes. http://imagebin.ca/view/ODKw73S.html
that looks really excellent. you only need to adjust your FF1 and you'll get good following. the error is very clearly proportional to the velocity.
there might be a tiny bit of high frequency oscillation - not sure - if it sounds rough, try fiddling with D (higher or lower - look for a sweet spot)
but the majority of this error can be fixed right up with FF1
good. now am i correct reading 2 thou? as in inch?
it looks like about 2 div at 500 microinch each, so that's .001
do your servos talk at all when at steady state?
I think they make a tiny bit of noise, but don't move perceptibly
mine will at times ring when they stop abruptly. it seems to diminish as i tune, tho.
oh thats better
wow, i'm under half a thou ferror both ways. now i'm fiddling with D trying to make it stop ringing at random stops. changing it certainly has an effect.
excellent. the graph was pretty good to begin with. how big is the machine?
hardinge HNC bar and chucking machine.
it's a tank.
hardinge, bought by bridgeport, now UK based, no more Connecticut in Bridgeport:(
nic emachine, i saw the chute buttons on your i/f
is it new? new ones were built in Taichung at FineTech. same place i built several edms
new = 2005 to 2008
ser#195. it is decidedly not new.
then its better :)
man, this thing refuses to fail. it's the best thing i ever bought hands down.
I seem to have lost both grid zero switches, but there is a chance that they still work.
i peeled off the part chute but i hear they are for keeping.
I'm still tweaking in the config, and i've never run emc before about 45 min ago, so it's a rush all the way. buttons you saw were from cradeks config, he has similar machine.
wondering what I will have to add to index and brake the spindle... it wants live tooling and i think it can be done.
bigger lathes usually just hang another toothed pulley off the spindle
and a rather large servo
it freewheels when it isn't in use
and for braking/clamping, it's usually a literal disk brake and pad
so you get your big, main spindle drive and pulley, your spindle servo and pulley, and a disk brake.
there is an M code they use to alternate between spindle drive, servo drive, and braking.
and obviously, the spindle will need an encoder
you can use a small servo if you are only indexing to clamp
and then it's just like m90 C<angle>
and it will unclamp, index, and reclamp for you
this thing has a kickass brake but leaves a little lash because it is down by the motor. if i could keep it loaded aginst the brake i'd be in business with existing equipment. the trick would be stopping at the index points.
i think it can be done.
the belt will not handle milling
I've just had an idea for a head changer on a table router. the Z-way ends (with a taper) lower than the screw, and when the X is aligned with a rack, the Z ways are continued up in the rack. with a biro-style toggle latch at the top the Z motor could effect the head change
if all you want to do is like, axial drilling it should be fine
good point thankyou. i've run a chucker but this is the first ive owned.
not in a hurry to burn it up
those brakes wear out pretty quick
at least on the hardinge chuckers we have
still, small hi rpm tooling shouldn't load the spindle any more than normal cutting, i'd think.
it's not a spindle issue
it's a holding the thing still issue
the spindle's fine
even a little bit of springyness in the brake system will break an endmill
that's why the clamp is universally on the spindle itself for live tooling machines
either as a disk or as a torquemotor
I see. so the correct answer is a disk brake on the spindle nose, if I really intend to pursue this. I have the pneumatics laying around...
I can fab a serviceable disk on engine lathe. this brings me back around to indexing. what is the best way to index the head?
oh hell yah there is an unused bracket space back there.
I've got a bicycle brake disk doing nothing. no calipers or anything tho
you can hang it off the ass end of the thing, too
I wouldn't want brake dust up by the thrust bearing
but it's up to you
cycle ones are all cable pull, so no good for pneumatics
have part in hand. but the point about the dust gives me pause...
if it were a crappier machine i wouldn't have an objection
but the spindle bearing in that machine is very, very good and it would be a shame to crap it up
an extra oilseal on the outside?
not sure how you'd fit one to a hardinge spindle
objection taken. now I have to resort to clamp not brake.
how about engine braking? is it a brushed DC motor?
mattymatt: the motor is hung from a pulley
that's not acceptable for live tooling, it has to have a firm clamp on the spindle itself
gotcha belts are springy
a brake belt around the chuck?
leather or sth
if you can brake it with the motor, index it, and then just use a disk brake for a clamp
then you'd get no dust
and could hang it (somehow?) off the nose
but you won't be using a chuck if you do that, given that spindle.
a disk fitted to the spindle nose. i have noomatic works for disk brake, just need to ensure no drag. I'm planning for collet work mostly on this machine.
you could have the whole disc in an enclosure, or use a drum brake
i really wish hardinge made a 16c spindle machine
that could keep the dust in
it will function more as a clamp. I will just have to tune it carefully.
gotta add something the machine just worked!nofun.
build another lathe just the same, but with a disk inside the headstock >:)
plan H, use magnetic braking
Honestly, this doesn't look like a half-bad idea right now. there is literally a bracket pad in the correct position. it could have been an option or something.
ali disc, permant magnets in a caliper style bracket that swings in. that would brake but not clamp
I could unbolt it in a minute.
It doesn't hurt that I have the rest of a shop standing idle around me!
I could confine it's use to short runs, and make the disk to be installed after turning.
arg, my headchanger can't run entirely from the Z screw, because there's no way to re-engage the half-nut once it's released. it'll need a solenoid at least
oooh, headchanger? what sort of beastie are you working on?
just a scribble so far. this needs proper V shaped sliding ways and a half nut. all beyond my means
If I try and use a half nut on M8 the M8 will just bend out of the way :)
basically it just moves the Z slide up onto fixed fingers attached to the roof, so each head has a slide
something for the mk3
hey, ever heard of rolamite bearings?
it's a neat trick, using two dowels and an S-loop of spring steel.
are they the ones that run on round rails?
they work by a balance of spring tension, curling and uncurling in unison
They make machine ways of extreme rigidity and precision, provided they can be kept very clean.
my company makes air bearing dovetails for diamond turning machines
for contact lens mould turning
that's my concern with having the carriage part of the head, it can get dirt in it when you put the head down on a bench
and transferring from Z-ways to fingers might cause wear if alignment isn't perfect
air bearing dovetails? what level of precision does that ential? zounds like the land of millionths.
I'm thinking in terms of woodworking machines so far, so dust-free is not an option
.0001" parallelism in 24"
the frictionless part is more important or we'd send the slides to get lapped
toastydeath: does that account for thermal dimensional changes? ;)
air bearings can be 24 inches long?
mireland: sure they can, but i'm talking about the slide
the actual bearings themselves are flat to 20 millionths
the dovetail the bearings go on is usually in the 24" ballpark
toastydeath: so the warm spindle on one end won't violate the parallelism spec?
ds3: it's a diamond turning machine, not a normal lathe
they go in dtms, rather
the whole machine has thermal control
ah I see
spindle chillers, etc
not necessarily, optics doesn't have to be done at 68
as long as you hold the temperature AT something, it will be fine
so many places hold it to 72 because it's easier to do
and less costly
optics is more about finish than raw accuracy, especially for contact lenses
so the temperature isn't as rigidly controlled, but vibration is
sleep now, nite
so nobody knows about rolamites? they were invented at Sandia. apparently they are public domain. they make a near frictionless slide with the side to side rigidity of the spring steel you make them from.
they mite make cheapgood slideways, if an industrious person should try.
[06:01:16] <MattyMatt> http://www.youtube.com/watch?v=O-1Drlfuoi4
an industrious person with a slip roller I presume
springsteel. should be flat when unsprung. these would be dangerous to create or destroy.
but scaleable to very small sizes
with pizzacutter combine rollers they mite even tolerate some dirt
hence the name >:)
linear steppers are the thing to dream abount
huh. i think i will dream about those a while.
[06:21:25] <MattyMatt> http://www.youtube.com/watch?v=AjkT6VgaQLo&feature=related
I'm having trouble w adobe plugin at the moment. im trying tho
I'm writing a user component in python. I want to be able to home the axis. It looks like I use the emc.command.home(0) to home the X axis.
When I run that, my component just dies silently.
I've had similar issues reading statuses, eg emc.stat.position
To get that to work I need to "emc_stat = emc.stat()" first.
And use emc_stat.position
I'm new to python, does anyone know why I need to assign to a variable first?
Actually, I think I just answered my own question - () - its a function call. doh!
ftkalcevic: that happens a lot in python when fetching data from a C lib or app, you can't operate directly on the data you need object getters and setters in the wrapper
is it emc_command().home(0) and emc_stat().position then?
emc.command() and emc.stat() i mean
Yes. That worked for me.
I'm now struggling to execute an MDI command from my python user component. I eventually want to zero the axis, but I can't even get a test rapid working.
This is what I am doing...
emc.command().mode( emc.MODE_MDI )
emc.command( "G0 X0" )
oops, typo. That should be emc.command().mdi("G0 X0")
then a wait complete.
If I don't set the mode to MODE_MDI, axis will complain about not being in MDI mode, so I know commands are getting through. It's just the mdi() call isn't doing anything.
I'm basing the series of commands on what Axis does. Any ideas?
look at mdi.py
[08:48:28] <alex_joni> http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=src/emc/usr_intf/axis/scripts/mdi.py;h=f5b4f23eb9a56deb14f54ac4f45fd2c9438c0ac8;hb=HEAD
Hmmm. It works when I...
c = emc.command()
c.mdi( cmd )
But it doesn't work when I...
emc.command().mdi( cmd )
* alex_joni isn't that familiar with python trickeries ;)
I don't like python. Give me something with a compiler and strong type checking.
ftkalcevic: You must add c = emc.command() at top of your program, and the use c.smth
emc.command() is command channel class and must be created firstly
I ran the StepConfig wizard and input my latency numbers etc and when I look at the .INI file I see that the base thread is set to 100000 which I believe is 100 uS..
That works out to a 10 khz step output I believe. The latency test indicated that the max I could run hz wise was something like 50 khz...
How does StepConfig determine the INI file setting for the base thread? Do I just alter the INI to crank up the max step rate?
SWPadnos: 2TB = $160 http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=8419165&subid=27062592&type=
>>ftkalcevic>I'm now struggling to execute an MDI command from my python user component. I eventually want to zero the axis, but I can't even get a test rapid working
What did you read to be able to get to the point where you could figure out how to do Python user components?
SWPadnos: 1TB = $60 http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=8419165&subid=27062593&type=
NewEgg had 1.5TB on sale for $99 yesterday, but I didn't get any
I should put together my file server with the 1TB drives I already have, then upgrade if I run out of space
(rather than upgrading before I put it together, like I've done before :) )
SWPadnos: Ah. I'm waiting for 2TB to go down
yep, me too
if I can get 2TB drives for under $150, I may do it
but the 1.5's for $99 is a better unit price
I need 8
8x 2TB = 6 RAID5 + 2 hot spare ~= 10TB NAS Goodness
better to use 7 in RAID6, plus a hot spare (or no hot spare and just use 7 drives)
I've found that the 4GB drives are a little small today, I prefer to use 9GB or even bigger on my new machines
I have several U160 SCSI 9.1GB drives, if you want to put together a new machine
I have plenty, thanks :-)
Me too, the make great coasters!
I even have some 18G SCSIs like what's in git.linuxcnc.org
and a bunch of 4.3/4.5G ones too, if you prefer SCSI RAID
I have one of those too
I went all out on that machine
should I send it to you so you can do mirroring?
no really, I have plenty
[13:06:43] <jymm> http://www.sears.com/shc/s/p_10153_12605_00911404000P?keyword=c3
plus I have a rebate form good thru tomorrow for a free cordless tool
of the 18GB - you big spender you
jmk gave me 3? 4? from a dumpster-dive a year or two ago
I really can't imagine having TBs worth of data
me either, though I do work with large numbers of still images from tiem to time
cradek: I have 4TB full JBOD as it is.
but you pirate movies and stuff, so that doesn't count :)
jymm: what is it all?
I think I actually have over 1GB of data that I'd like to keep safe
cradek: MediaJukebox, GIS Data, Hi-Res Imagery, Personal data, backups, etc
a lot of it is datasheets, which theoretically would be downloadable later, but it's easier to leep them here
the one thing I'd like a huge datastore for would be to get the full-res Blue Marble images
but that's definitely a "need", rather than a need
SWPadnos: Did I tell you I have one of these? http://www.cidesign.com/usa/product_detail.php?id=60
oh, I don't think so
I guess you "need" 16 drives then
oh wait, that's the ad, not the product
2U 8 Bay with SATA Backplane hot swap chassis
did it come with the stack of SATA controllers?
FDD and slim DVD burner surprisingly
No, that's what I'm still shopping for.
I can't complain, got it for $50
let me know if you find a system that really supports hoot-swapping under Linux
I haven't looked into it carefully lately, but I think I remember that it isn't trivial to get live swaps working right
That's based on the controller's ability, no?
drive, backplane, controller, and OS all have to work right
Well, the backplane works with 3Ware and A_________, so I thig it's ust a matter of wat the kernel support is
There are headers in the backplane that you connect to the control for the LED's per drive
yep. I've got that on one workstation and on the 4-drive cage I have
the last time I checked, I'm pretty sure there was an issue with Linux recognizing new drives when they're hot-plugged, or stopping the FS (or letting it still run, which is preferable) for a rebuild
A good thng when hot swapping =)
I don't remember the details, it's been a while since I looked at that
a while being?
at least a year
SWPadnos: you just gotta love that a RAID card has it's own web server thogh =)
ayeah, as long as I have an ethernet switch sitting in the machine
SWPadnos: No, it has it's own ehternet port =)
exactly - I'm talking about having an extra port to plug it in to :)
SWPadnos: If you have to worry about not having a spare port, you have bigger issues to worry about =)
SWPadnos: I'll give you a switch !
I probably have 200 ports here, but in my case, I have a server in a closet, and the patch panel isn't in the same closet
so there's a cable run to the server, and I'd need another cable run (or a switch) to connect multiple devices
* jymm shakes his head
the cable run is in the wall, so it's not too easy to run another (though I do have conduit installed, so it's not as hard as it could be)
SWPadnos: Run more strings and tin cans =)
I'll get more pigeons
Now you're thinking =)
steve_stallings is now known as steves_logging
Um, High Tech Redneck???? http://www.frys.com/product/5940674?site=sr:SEARCH:MAIN_RSLT_PG
My IRC browser went down for a little while so I am not sure anyone answered by previous post - if you did thanks.. If not, I figured it out anyway :-)
Dave911: if you ever need to see whats happened, or refer to another day... http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/
or logger_emc bookmark
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-10-09.txt
Thanks ... for the logger link..
I am looking at an example config that comes on the LiveCD the Stepper-Gantry .. There is a Hal component loaded called "gantrykins". I cannot see how it is loaded. I have another config I did with StepConf for my lathe and gantrykins is not loaded - it can't be seen in the Machine-Show Hal configuration Any ideas? I've looked through the ini file and all of the other files I could...
...find that is part of the config ??
gantrykins is used to assign the joints to the axes in another hal file that is loaded in the INI file. Is "gantrykins" part of another module? I did a web search and found some mention of source code on Jeff E's website ( the meat one :-) )
src/emc/kinematics/gantrykins.c i havent found a cfg in 2.3.3 that uses it yet
these files might make a sim system that use XYYZ http://axis.unpy.net/01162326817
(gotta take off a fender right now bbl )
doh try stepper-gantry!
I'm just using the Live CD install and 2.3.0 and it is used in this config. It loads up ok - no errors but I haven't tried to actually run it yet...
Fender - as in bent fender - do you have kids? ;-) I have two fenders that need fixing. Already replaced both mirrors .. Thanks for the link :-)
I'm back, my obit wasn't in the paper. Can someone tell me how, in this modern self configuring X we have now, to force it to use the vesa driver?
Damn, I just found an update has removed that stanza from my milling machines 6.06 install. WTF??????
So according to the log, that machine is bacvk on the nv driver, and that was trouble (latency) before. So I need to convert both.
Back to the xorg list I guess.
if you have a backup of the older xorg.conf file, you should be able to just copy the correct stanzas into the newer config file
I'm surprised that 6.06 would have an update like that (in fact, I thought they ended 6.06 support in April or June of this year, at least for deskyop apps)
That I do, many in fact, all saved with their saved date as xorg.conf-20090308162951 for the newset of those. Can I copy/paste that "Driver" section into this 8.04 installs xorg.conf?
Whoops Section "Device" not driver.
I think so, but you'll need Device, Screen, and maybe ServerLayout (or similar) sections
I don't remember how ServerLayout is used. or was it DesktopLayout or something?
The device by itself isn't enough?
I don't know ;)
I'm not sure if you need to explicitly tell xorg that you have this device, with that monitor, and you want to use it "here!"
the device section may be enough - you could find out and let me know :)
That's what I was afraid of.
I'm going to reboot, so if I don't come back in 5 mins or so, you'll know I fubared it. :)
heh, sounds good :)
hi all, first i'm happy to say you that my CNC machine is working great now ... thanks to all helper ! ;=)
please anybody know a link to find a postprocessor for Camworks ?
someone on room ?
lol hi archivist
google says you (or someone else named Phil) asked the same question in 2006 :)
yes it was me ;)
but no answer
perhaps you need to write it
I found a place that makes custom post processors
or modify one that is close
i have an old version for older camworks ( its not me to do that ) .. but now seem to work, but i have ome trouble now !
i dont know how to do a new one !
you should be able to modify the one you have to fix the "one little problem" :)
SwPadnos ; i found too, but not free ! :(
there should be a way of editing the post in CamWorks
in fact, i dont understand what is wrong .. mays i send a sample of my Gcode ?
you can, but I won't understand it either
and I don't know that anyone here uses CamWorks and EMC2
how can i send a file to try to open in your EMC2 ?
i have do many search with google .. but nobody seem to use camworks & EMC2 !!.. what you use ?
make a small one and put it on http://pastebin.ca
I don't use either at the moment :)
please can you open in EMC2 : http://pastebin.ca/1607485
well had sucess getting emc2 working but not a lot of sucess in finding sutable ways to get my drawing files converted to nc format for use with it
what is the problem you have with the code? (I don't have an EMC2 machine handy at the moment)
i don't know how to translate the message in english ( i have EMC2 in french ) .. some problem with coordonates, origin ..?
using kcam i can inport dfx files (assuming just 2d cutting profiles ) , but even the output nc format from kcam wont work in emc2 format diferences i guess
circle/arc origin, or machine origin?
AchiestDragon : exactly same for me !.. work perfectly .. but hard to generate Gcode .. ( or i'm too bad to do a postprocessor .. )
idealy could do with a nc preporcessing program a bit like mastercam that can do solid model to gcode paths
open my file, the problem is near line
but that in its self i guess is going to be rather large to file as a single feature request :)
i don't know mastercam .. i'm working with solidcam .. and like i have an old postpro for camwork i use camworks ..
mastercam know generate Gcode for EMC2 ?
i have other issues with mastercam though , as it does not support the file types i usualy work with for cad
SWPadnos : near line 9
so realy it would need to support eather dfx format or one of the default brlcad file types
mastercam know open IGS file or maybe solidworks files ?
mastercam does a great job with solidworks files
ty eric, may be i will try it
dfx is more coman within cad drawing applications though
mastercam can generate gcode for EMC2 ?
you need a post
so same problem with mastercam !... i dont understand that there is not a place to share our postpro !..?
I thought there was a mastercam post floating around
and having to say use qcad that saves in dfx , then inporting into another porgram to give you IGS to inport into mastercam to preporcess then save into gcode to inport into emc2 is not an ideal way
dxf is inherently 2d though, correct?
just that qcad is ,, dfx in autocad has full 3d support
Bummer, it is going with the xorg.conf.failsafe 800x600 screen, butt fuggly. There is no samsung entry for this monitor, which is a samsung 22" hidef tv whose tuner has gone south
SWPadnos are you there ?
but I have no machine to test with, so I won't be of much help
do you have an idea about my sample ?
something seem bad in that line : N30 G90 G54 G41 D27 G00 X-245.000 Y4.950
the issue of posts is industry-wide
<SWPadnos> can I just delete all the xorg.conf.* stuff, it seems to be rewriting it to suit itself anyway
and the fact that someone sues anyone that makes a control with programming capability doesn't help things
there is some 1000's users of EMC2 .. nobody share her postpro ?
gene_, yes, newer versions of xorg don't need anything in the conf file, it's all detected at X startup
I just wasn't sure how much of the old file you'd have to recreate to make sure it does what you want
Phil31, post processors are unique to the program that's generating the G-code
so you can't take a mastercam post and use it in camworks
the down side fro EMC2 is that if the only way to preprocess the gcode is by using solidworks and mastercam then sort of puts a damper on the situation as both are windows based , and both are rather priecy to get licences for
SWPadnos : yes i know ! but if mastercam say in native generate gcode for emc2, i can work with it and forget camworks !...
could say the same with autocad other than already have a licence for autocad
you asked if anyone had shared their post - I was only pointing out that 100 people could have shared post processors for other software, and it wouldn't help you with camworks
I made a post (not sure how conmplete it is) for TurboCADCAM, and you can have that if you like
no, i use windwos programs on my personnal PC but EMC2 on the PC of my machine ...
SWPadnos : i think about a place where all body share the postpro he use .. everybody say " here it is a post fro XXX cad software and YYY ( like EMC2 ) machine controller ..
BJT-Work : yes .. something linke this ! ...
i will try to post on this link my old postpro for old camworks version ..
.. can we post to this page ?
i try to post , but something wrong on the link ... ?
done ! ;)
be back later if someone can help me with with Gcode problem ...
see you later guys
N30 G90 G54 G41 D27 G00 X-245.000 Y4.950
I didn't load or run it, but if the problem is on this line (and why doesn't he say what the error is!?) maybe he has no tool 27
it preps tool 07 above on line 8, but never loads it
so there's probably a post problem (no M6) and one or more operator problems (no tool 27)
(I doubt any postprocessor is a substitute for knowing some gcode, and in general, what you're doing)
cradek: he is gone
yeah I know
maybe he'll check the logs or something.
ok, I didn't think of that... your the boss today
then I declare that people will stick around after asking questions
yep that is how we do it here who ever comes up with a good idea is boss for the day
or if one of us says lets go fishing or mushroom picking or something like that
I only need one good idea per day to always be the boss? that's a low bar.
it's got to be a good one
I better ration them then
well we have 2 of the 5 machines done we are building today so I'm taking the rest of the day off :)
cool, everyone should take friday afternoon off.
tool late SWPadnos cradek is all ready the boss
rationing good ideas was his idea anyway
see you guys later
I have two questions; Is there a FeatureCAM postprocessor for EMC floating around anywhere? How do I stop Ubuntu's screen-blanking when idle?
1) no idea, 2) system / preferences / screen saver
also check power management
Y'all talk to much! shut da he11 up alreay!
yeah it was the screensaver setting; I feel dumb now. I was sure I'd unchecked it already. Thanks!
Is support of G12&G13 planned for future releases of EMC?
... or G84? :)
_Poincare is now known as Poincare
hm, I see there are newer releases...scratch that g84 question.
um what are G84, G12 and G13?
g84's rigid tapping, which I see listed in the manual as unsupported. g12 & g13 are circle pocket routines.
we have rigid tapping - see g33.1
we do not have any circle pocket routines
I've only personally seen them on a Haas mill. I used them a lot boring tiebar holes in sideframes. They are nice because you call them off their center.
yeah I'm spoiled by the Haas machines, and expect the same functionality at home as I have at work. EMC's been great for getting close to that, much better'n Mach.
perhaps you could add these to EMC and submit a patch to the developers
or, if you could write up a full spec of how the codes should work, perhaps someone will add them. the wiki would be a good place to put this.
ok, I'll do that; explain the command and just what the controller is doing and the variables it's juggling. I've considered at least writing a routine to emulate it.
yes that is another option
they peck to a depth, spiral out at a width, then finish pass. you can tell it to skip out to a diameter if you are widening an existing bore. HAAS user manual should tell enough to emulate the function.
if you write a spec, be sure that you describe the behavior, not the implementation. I'm not sure what you mean by variables, but that sounds like an implementation issue. You should describe what the gcode says, and what path it causes the machine to take. You should also explain how (if) it interacts with cutter compensation.
I've used the spiral routine that comes with the examples, but with the 'real' routine, you set the starting and ending arcs individually. Nice when you just need to make a finish cut in an existing hole.
I remember wishing it would rapid rather than feed through air on completion. ours will be better than theirs.
it's possible the user manual gives all the information, but remember the developers probably do not have manuals for other controls
if those are freely available online, a link to the manual would be a great thing to put in your spec
if they are not freely available, you can't really usefully refer to them
the user manual just shows how to use it and the path it'll take, not the things it's thinking about to get there.
I'll include real details.
it could be emulated in scripts i bet. most important consideration is the possibility of name collisions with incompatible implementations. these codes may have resisted the standard for a reason.
when you are done, you might send a mail to the emc-developers list saying what you have done, and maybe you can get someone's interest if you can't implement it yourself.
awesome i will certainly use this
yeah , I certainly can't implement it.
but you probably guessed that when I came in here asking how to turn off my screensaver. :)
Is it possible to mill a 3mm dia circle with a 2mm dia cutter using diameter compensation? I can't get a sensible entry move.
It's more that it won't allow the cut, it complains about gouging.
lie about your cutter diameter, half of it is exactly your pocket radius.
I am wondering if I need to toggle compensation on and off in some way. It looks like it will cut the first profile fine, does it get a fixed idea of where "inside" is?
If I lie about the cutter diameter it will surely mill the wrong hole?
just telling it your cutter is 1.99mm should be enough
It would be simple enough to just fix the cutter dia and mill explicitly, but I am curious about how compensation works.
andypugh: turn on compensation on a long move above the work. then as archivist says, just helix down
(if you are not yet using emc2.3 it might be somewhat harder to do this)
It isn't the length of the entry move it hates, it seems to insist on doing wierd semicircles on straight moves, so I am clearly missing the logi
what version are you using?
(and what's the exact error message?)
are you remembering to use G1 on the next linear move after a g2 or g3?
I'll shut up now, sorry.
that is not a requirement in emc - it can compensate arc-arc-arc just fine
I can't remember exactly, it was something like "Can't cut concave corner without gouging:
Bear with me, I will reboot the VM
yeah that's why I shouldn't be offering "advice". I'm a machinist, not a programmer.
vlmarshall: cnc machinists often give good advice, but controls are so different sometimes the advice doesn't match emc exactly
thanks. I just jumped in on something I thought I knew, while my mill is running behind me.
"Arc move in concave corner cannot be reached by tool without gouging"
pastebin your gcode and tool table?
Generic 2mm dia tool. I guess a 2" one would work too
I don't see where you load the tool or otherwise tell it which one to use
I did that in the MDI.
I am just curious if I need to fiddle more, or if it can't be done?
looks like you can do it with extreme care in entry and exit moves. if I change the diameter to less than 1/2 of the hole dia it works as-is
Odd, as it is something you could do without a second thought as a manual machinist.
I dont use compensation when milling, and I admit holes are milled with the rotary turning here
I wasn't sure what cutters I have, so wanted to leave my options open :-)
It's not a problem, it just seemed a good use for compensation
anyone know a cheap supplyer for http://www.axiscnc.com/pdfdocs/tbr.pdf
this type of rails and bearings in the uk ?
seen them on fleabay
yea axis cnc on there ,,, i got what was suposed to be a dremel/proxon spindle mounting from them dosent fit a dremel so not realy willing to use them again
Is £2.90 per 100mm and £7.40 for a carriage too much?
well cheaper than £3.67 plus £8 postage anyway
for same 100mm
The £7.40was the slide, not postage
[21:07:05] <andypugh> http://www.slidesandballscrews.com/tbr16-supported-guided-rail-p-139.html?cPath=52&osCsid=9625642cc78dd89f4a494d0d1c58ca3a
Most of my stuff has come from there, mainly by accident as I thought I was buying from Motion Control Products. Bournemouth seems to be a hotbed of CNC product sales.
Their ballscrews seem very cheap too.
still comes to £127.60 for the total 4.4m but better than 161.48 that axis charge
kool ty for the link andypugh , bookmarked it
was going to use trapizoidal leadscrews
Price up the ballscrews, there may be less in it than you think,
Arguably good leadscrews are better than cheap ballscrews though
the screws cost about the same for eather ,, its the nuts that seem to be more of a cost issue
sorry ebay link but http://cgi.ebay.co.uk/x3pcs-Cylindrical-10x2-Trapezoidal-BRONZE-Nut-LRM10-2D_W0QQitemZ310159999257QQcmdZViewItemQQptZUK_BOI_Metalworking_Milling_Welding_Metalworking_Supplies_ET?hash=item4836fa0119&_trksid=p4634.c0.m14.l1262
remember that un-preloaded ballnuts will require doubling-up to control backlash.
2 nuts to make an antibackash assembly , rather than trying to make a ballnut or £45 for a single ball nut
the trapizoidal ones seem like a better option to me
I have a 6tpi acme tap if you want to borrow it.
i need to reduce costs a lot but i dont want to compormise the machines acuracy too much
the frame is going to be from 40mm square 2mm thick steel box section its only approx £14 for a 7.6m lenth
You have to be careful about where you cut cost though, or you end up with £100k prototype cars that can't be used because the road tax has run out and the re-tax cost isn't being signed off....
(Of course, I made that example up, no huge multinational would be so stupid, obviously)
well welding together a frame is within what i can manage , so are the trapizodial leadscews and nuts ,, those rails are a comparmise as would of liked to use harwin 15mm rails and bearings but would put it out of my budget
The aluminium extruded profile isn't a lot more expensive than that, by the way, and looks dead technical.
would an aluminium profile machine be up to machineing brass parts on though
Dunno, and I was wrong about the cost at that size too
[21:27:56] <andypugh> http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=0223638
i can mig weld the steel but would have to bolt it if it was alli
your box section may not be stiff enough, depends what you want to do
stiffness is not a problem as going to use a latice type frame work assembly ,, lot of welding though
could even go to i beem profils and a lot thicker if needed
but 40mm box would not be a problem with the spindle power i am concidering using
Bolting that Ali profile is very easy though, you only need to drill a hole to get an allen key into. I used the tiny version to make my controller box.
bolting things together i usualy find leads to it vibrating loose during use
I am not saying that welded box won't be stiffer and better. But it won't look as nice :-)
although with welding i would usualy have to design in adjustments to correct any welding deformation
yea i know ,, would look better from aluminium
be lighter also
Easier to reconfigure too.
We used it a lot in a previous job, building assembly and inspection lines. It is very rigid when bolted up.
Maybe not iron-casting rigid, but it fills a lot more space than box, so is torsionally very stiff. You still have the fact that the Young's Modulus of ali is a lot less than steel though
ok only the inner section of the main frame shown on this but that was using 4" by 2" by 1/4" thick aluminum channel for the framework http://www.whipy.demon.co.uk/main-assembly.jpg
there was to be an outer frame providing extra support also
still not toatly shoure about the weight issues on the z axis if i make it with a steel frame , will definatly need countrbalance weights on it
forget the stepper position on the y axis , as like the z going to be a double leadscrew
It might be instructive to do the beam deflection calcs for steel box and ali
The ali section moments of inertia are published, you might have to calcuate the steel one
roark's stresses and strains
have the values for 40mm box someware
0.24 " at 110lb on free end supported 36" away
Simply-supported is probably more relevant than cantilever, but for comparison purposes it is fine.
same as for unistrut
i did try uinistrut some years back , but the bolting method tensd to allow for vibration to losen the bolts and then it tends to drop out of allignment because of the slack
At the correct torque nothing should work loose
unistrut beeing galvanised dose not realy allow for welding
how best to explain the problem i had ,, drill 2 lenths of bar say 500mm long ,, bolt though , tighten to rated touqe of nut/bolt ,,, hold other ends of bar and it losens as if you had a 500mm long bar on a socket ove the nut
Aye, though you could use the corner brackets as well
yea probalbay better that way
but welding is more stable in the long term , although as you say does not lend itself to reconfiguring the layout
unistrut would be a nice option and price wise good , the down side is the fixings they seem to want to sell you them in packs of 100 or so
and for 1 machine that works out quite expencive
as to looks when finished was thinking of masking of the faces where the rails go and other inportant surfaces and getting it poweder coated
theres a place down the road does that ,, if alli would of got it annodized
main aria of use would be to make parts for cavity amps and other uhf rf stuff like http://www.ve1alq.com/cavity1/cavity1.htm
mostly copper then silverplated
but some brass fittings also
I understand they have to be very tightly toleranced, or is that only if you really and truly care about the frequency?
the tolarances go down to a few hundered microns for actual tuning , but the tuning is adjustable and +- .2mm or so is acceptable on the construction
Use a cold-chisel then :-)
I reckon my dad and grandad could use a chisel to that tolerance, then scraper and blue for the next two decimal places.
yea but would have more flexibilaty with a cnc , and end up with parts that look the part and would justify spending the costs to get them silver plated
Yeah, I built a CNC because it is too late to do a machine fitter apprenticeship. :-)
Scrolling back: No need to anodise that rail profile, it is already clear anodised
i was meaning the frame if i made that of alluminum
I meant frame profile
was meaning the rest of the frame parts as i would of been using a diferent section to that
like the channel section sat on the machine in this pic http://www.k7peh.com/imageHiQ/rawparts1.jpg
a local company uses a far ammount of that stuff for walkway gantries arround the hoppers and air vessles they make , qiute a bit of it ends up in the local scrap (when they make a bit the wrong size ) but been told there going bust
so a cheap supply of it is going to dry up rather soon
4" by 2" 1/4" thick , not got funds to buy enough in for what i will need to make the machine from , so looking at another design for the time ,, if theres enough there when i have some cash then will get some and use that
but it would need anodizing once machined
£1 per kg
on a £0 budget atm as its allocated to xmas pressents for kildet and famaly
will have to see though may manage to scrape together enough by end of next month ,, need about 85kg of it though and about another 40kg of 10 and 20mm thick aluminium plate
i finish the encoder (on the matrix) it work very with 2 LDR and 2 LM741 and one 7404 (to get TTL)
now dismantled it to build it on a PCB
how i can cut PCB board
here in night and i cant make noise..
someone name a free program to do 2.5d cam in. got some dxfs but cambam wont do them properly
pen-knife, score then snap
stanely knife ,,, score the board ,,, then break it like you do tile cutting
i try to cut him with japanese knife, but is not work
I know this way, but I need to cut a small square
small drill , about 3mm dia in a pin chuck and turn it by hand ,,, , drill a few holes arround the edge use side cutters then to hack out the center and file it suqare ,, posible to do that at 3am without waking the cat asleep on the desk next to where your working
alhough filing it to a suare may
although / althougth ,,, suare /square
Is it possible to set the spindle speed through a pin? Not spindle override, the specific speed, eg 100rpm.
On my pendant, if I select spindle on the axis-select switch, I want to adjust spindle override when a program is running, otherwise, in manual mode, i want to set the spindle speed.
So I can't do it in a hal file? I need to do it in python and use the mdi?
ftkalcevic: I don't think there is a good way to do that currently
I think in theory halui could have a "counts" style input for spindle speed, but it doesn't currently
it has "increase" and "decrease" pins which are for buttons that increase/decrease by 100 rpm for each press
Actually, Spindle Override looks fine. I didn't realise it applied to the spindle speed in manual mode.
So was it concluded that radius compensation on small holes was too much bother?
i generated a gcode file that gives very different behavior with g61,g64,g64p- http://www.panix.com/~dgarrett/stuff/h3.ngc
demonstrable on simulator (axis.ini) -- maybe i'm doing something wrong -- to check, run it, then change g64p.001 to g64 or g61
now i finish to cut 2 squares of PCB
dgarr: Try the mailing list with that one?