dgarr: interp_error: Named parameter #<zincr> not defined
oh I just added an _
dgarr: reproduced here, great bug report as usual
behavior can change with #<y> value too
is the naive CAM detector forgetting to look at the midpoint of the circles?
forgetting to look at the perpendicular component is my theory
it was actually skipping both arcs, wasn't it. that makes me nervous that I don't understand the problem.
(uh, even though I fixed it)
nope I don't understand it
what is implied when the display in axis shows the arcs, but the execution ignores them?
that pretty much means the problem is not in the interpreter
it is in the motion output afterward
the problem is pretty surely in the naive cam detector since a G64 P setting changes the behavior
in fact it's probably in chord_deviation
yeah chord_deviation is returning 0 which is wrong
these dang full circles
Perhaps we need a worldy-wise cam detector?
it's supposed to be "(naive cam) detector"
but sometimes an argument could be made for "naive ((naive cam) detector)"
I want to get into the problem solving rather than problem making aspects of this project, by the way. But I am on the steep part of the learning curve,
don't worry, the curve is relatively linear.
and doesn't really get any easier.
fwiw, the curve is currently biting my ass
I do do coding, but I have never used C in anger
I used to know Z80 hex codes by heart though
uh. chord_deviation calculated mx and my, but never uses them
I only know that JMP is the two switches on each end up (C3)
(I started young, when you have the time for that sort of obsession)
maybe the return value should be max(hypot(midpoint to startpoint), dev)
SWPadnos: the problem is it finds the included angle of the arc is zero
also used to know z80 hex codes by hart many years back
I bet the mx,my was a false-start way of calculating the deviation (replaced by the smarter formula)
shouldn't the if (rotation<0) ... construct add (or subtract) 2*M_PI?
Rather than 360?
so (th2+th1)/2 = M_PI
I have never found a neat way to do 4-quadrant angles, you always seem to get bogged down in conditionals
hmmm. nevermind. dev should be ==r in my mind
do know a guy who could read papertape , as if it was written text
In prinicile no harder than braille, but you have to wonder how he got the practice
braille is harder
when you read g-code every single day it starts to look like text
even if you're reading dot patterns by looking at them, the encoding is not as simple
I confess that g-code still looks like gibberish
so does braille ;)
he used to fix the pdp 8 and 11 minicomputers
I mentioned, I think, that we retired the last PDP11 at work abut 5 months ago.
back in the z80 days when you used to have to hand compile a few k's worth of code you soon pick them up ,, these days we have assemblers so you dont get to be as farmiliar with them
They ran engine dynos. I have written a fairly big Excel Macro (10,000+ lines of code) to perform automated dyno testing by communicating data values between the dyno controller and the ECU driver. I so wanted to make that work with the PDP cells too, just for the aesthetic anachronism
in soviet russia, engine dynos run YOU
oh ty again andypugh for that link ,, should be able to afford to swith to 20mm rails for less than the cost of the 16mm ones from the other place
OK Z80 fans: http://www.worldofspectrum.org/infoseekplay.cgi?title=Johnny+Reb&pub=MC+Lothlorien+Ltd&year=1983&id=0002635&game=/games/j/JohnnyReb.tzx.zip&emu=3
A game from when user interfaces hadn't been invented
[01:02:18] <cradek> http://timeguy.com/cradek-files/emc/0001-fix-corner-case-where-naive-cam-detector-skips-full.patch
>>AchiestDragon>unistrut beeing galvanised dose not realy allow for welding
A good trick I have learned is to get a bucket of muriatic acid and put it outside - and put the ends of the parts you want to weld in the bucket. The acid will eat the plating off the unistrut or conduit and you can weld without excessive fumes...
jepler: ^^ (?)
Ah! Dave911, I found out how to read current positions into G-code today
(Well, someone on the mailing list told me how)
(Though I also saw hints of a long-term plan to make #_x return the current X in the named parameters section, which sounds better)
Dave911: maybe but 3 points ,, 1st ,,, its cheaper to get 40mm box section than unistrut , if your going to weld the holes then dont matter just make it weaker ,, 2 it would need regalvanising after , may as well start with ungalvanised , and 3 its not realy practical for me to store chemicals here
wonder why atan2's range is [-pi,pi] instead of (-pi,pi]
that seems bogus
andypugh: what did you find out??
Are you on the mailing list?
AchiestDragon> Well then it really doesn't make sense to use unistrut anyway. Muriatic acid is used a lot to correct the Ph in swimming pools. I also use it to remove rust sometimes and clean out my water softener. Handy stuff.
that is the conclusion i came to :)
Basically you can add halui to the the ini file, then net the halui axis positions to the analogue input pins in HAL, then M66 will read the axis you asked for into #5399
ugh, that's an abomination
what are you trying to do anyway?
Yes, but it looks like it will work
Get current X, Y, Z into G-code
no, what is the task where it is useful to have this?
Put stock into the lathe, jog to a point just outside the material, press "go" and have it machine to a fixed length and/or diameter.
oh you mean you want a variable number of passes starting "here" until you get to the specified value?
Yep, there are all kinds of things you can do with that .... I think... That really is not too ugly. I've seen much worse work arounds!
Imagine several lumps of steel roughly cut to length that you want squaring up to the same length, without a probe and without cutting air.
I abuse canned cycles to do that
sounds like what you could really use is a good roughing cycle (and I feel your pain)
Yeah, I have real problems with the concept that G-code doesn't know where "Here" is
well ... it does
it just won't tell you directly
We have done this conversation before: I have tool offsets
yeah I know
cradek: thanks for workng on that g64 problem! i was pretty surprised when it happened on my new little mill since i'm just beginning to cut metal.
actually I bet you can abuse probing to find it
andypugh> You mentioned custom Gcodes - how do you do those? I think I have read every EMC2 manual and if that is documented I must have missed it...
dgarr: it's a very corner case, doing a full circle at atan2's discontinuity
andypugh: G91 G38.3 Z.0001
The <named_parameters> documentation says that #<_a> etc are reserved for just this idea, which sound perfect to me
(only slightly lesser abomination)
andypugh> Thanks for hunting down the answer to that current position reading ... I know of another guy who is starting to use EMC2 also and he had the same question
I am looking into cradeks G38.3 abuse now, perhaps that is better
dgarr: what the heck, I'll commit that - jepler can tell me later if it's somehow bogus
andypugh> Wow, I can see how that one got by me ;-) Gcode abuse... there is nothing illegal about that right ;-)
cradek: seems fair, it is the sort of check you would do anyway if you thought hard when coding it, it can't break anything
andypugh> I'll have to print this out and look at this later .. I'm being hailed Thanks!!
andypugh: I missed something - what?
Your "catch Atan2 discontinuity"
I was saying that it can't really be argued against
Where on the circle is the problem?
G38.3 does do what you want, and you can do an arbitrarily small "probe" move
I am going to guess that it is one of the "cardinal points" and they are quite likely places to be.
andypugh: the problem was a circle that starts at angle -pi and goes ccw to +pi (or the other way)
Yes, thanks for that. I had looked at G38 but decided I couldn't spoof the probe signal, not noticing that failure would mean success.
the "don't report error" codes are the key
-pi is west?
both are west
that's the problem
2pi radians to a circle?
Atan doesn't make the distinction?
Ah, yes, it's a bugger that way.
I think Matlab has a special version of atan that does
<google> Ignore me, matlab atan2 has the exact same issue I think.
[01:37:48] <andypugh> http://www.mathworks.com/access/helpdesk/help/techdoc/index.html?/access/helpdesk/help/techdoc/ref/atan2.html&http://www.google.com/search?client=safari&rls=en&q=matlab+atan&ie=UTF-8&oe=UTF-8
yep same problem, it can return all points on the circle plus one more point
cradek: i'm trying to sort out the spindle control. If I understand this, you use only the index pulse, am I right?
Should be pi to -pi+eps
mIreland: you should give motion.spindle-revs as much position resolution as you have, increasing by 1.00 per revolution
motion also wants a spindle speed in, which you can calculate however it's convenient
it functions but is off by a scaling factor, I believe. can I display the angle of the spindle somehow?
make a mark on your spindle, put a halmeter on motion.spindle-revs, turn the spindle forward 10 turns, see if spindle-revs increases by 10.00
okay I'm getting about 5:1. motion.spindle-revs increaes by ten per 2 turns.
ok, fix your encoder scaling until it's right
that's AXIS_6 in ini file, rite?
encoder_scale in the HAL?
Or am I missing the point?
AXIS_6 is some old experiment - you should delete it
the spindle encoder is in hal
there's a couple of those... lube timer never connected, rite?
yeah I don't think that goes anywhere
setp hm2_5i20.0.encoder.03.scale -400
why did you scale by a negative number?
I guess because it was counting backward
easier than changing wiring
now i finish the PCB!!!!!!!!!!
Does it interface with EMC?
alright... i have a sensible reading and i know how to tune it in exactly. thanks cradek
bet it's 2000 or 2048...
I am going to make a 511 line encoder just to annoy people.
2k i think testing...
andypugh: I've run across 508 line
That's quite random
mm vs inch
25.4 * 5 * 4
Hmm, encoder version of the 127 tooth gear?
hahaha perfect to the decimals. i love it.
put it on a 5 turn/inch leadscrew and you get .01 mm
So you fit that and tell the system it is a 20 pulse encoder, and instant metric lathe (or 200 pulse and work in cm)
fortunately with emc we can have whatever scale we want, so we don't ever need special particular encoder resolutions
I want to work in milli-cubits :-)
millifurlongs are more popular
okay, now reported rpm in the control panel is outlandishly low, at least a factor of two.
OK, time I retired for the night, but my favourite unit, and one that the google conversions facility understands, is the Megaparsec.Barn. A unit of volume with the area of a hydrogen nucleus and a length to the next galaxy
Which, bizarrely, comes out to be about a teaspoonful.
[02:22:27] <andypugh> http://en.wikipedia.org/wiki/List_of_humorous_units_of_measurement#Barn-megaparsec
I've always thought barns was a thoughtful unit for target cross section. all science should be so comprehensible.
this barn is considerably harder to hit though
AchiestDragon, i finish the electronic for the encoder
OK, time to go
panel code says halpin="spindle-actual" I dont see that pin available under HalMeter.
someone what photo of the encoder?
show us, i-pink!
i connect the camera to the PC
[02:36:47] <i-pink> http://www.speedyshare.com/672755871.html
the encoder ^
wow, nice work. that motor is smaller than I had thought.
Okay, my speed issue is solved. pretty sure you are running closer to the top of your speedchanger screw
is about 3 on 12 cm
sorry, talking to 2 persons at once
the speed is 3-4 RPM
appropriate for a 50 foot vertical boring mill
I'm looking for scaling factor which relates VFD command to rpms... I read correct RPMs but do not get what is commanded
atspeed.scale looks likely. has a neighbor setting atspeed.difference Are these the parameters I am looking for?
hmm not it..
i think i've found them, one scale for each gear as inputs to mux2
alright! the map matches the territory at last! speed commanded matches speed measured.
I'm wishing the speed +- buttons bumped a little less than 100 in low gear
mIreland: you can always M3 S123
i-pink, very interesting encoder!
mIreland: sounds like you are sure getting the hang of hal now
getting there... still haven't quite learned to connect every line when I read the files, but I am learning to.
thanks for telling me about spurious axis in ini file, I'd have walked around that superstitiously forever. Now it's gone.
at one time I was trying to map/test my ballscrew with a linear scale. that let me display the value on the AXIS dro.
but the ballscrew was MUCH better than the scale... haha
still have the plot
red is the resolver, the jaggy cyan is the scale
Say, I have one axis of a Mitutoyo DRO, can it be made to interface easily?
wait, red: scale, jags:ballscrew, rite? I see what you mean
toolchanger still reads one off of true position. No clue why; all the switches add up to correct spot per engraved numbers, axis just reports one off.
I'm waiting for parts before sorting that one out, though.
mIreland: that's just the way it is... if I was slightly more bold I'd just say the markings are wrong
scale reads in jumps, i get it i think. so the smoother line is the resolver out
yeah the scale was way too low resolution to be useful for that task
I'm saying the wieghted sum of the turret position switches corresponds to the engraved number. Somehow axis increments it by one.
oh, hmm, maybe yours is different somehow
maybe the Chucker of the beest is special? did you have to fix it?
heck I guess I don't remember now
I did have it all apart and had to fiddle with those sensors
only an observation, I'm sure I'll sort it out
I marked numbers on the encoder(compartment right behind turret). I bet yours is one step off.
and you fixed it with HAL
I'll find it. this is a great way to learn.
So, Mitutoyo DRO, I have a head and the long scale. what does it take to read it into HAL?
I could use the heck out of it...
if it outputs quadrature, just hook it up to a spare encoder counter on your 7i33
I had to take the soldering iron to mine to get the quadrature out, but it was hidden in the box somewhere just waiting to be found
are you serious? hahaha i cant believe what i have laying around
Say, what stops Halscope from reading a probe and serving as a general purpose oscilliscope?
it's not super fast, and you probably don't have analog inputs
I see. some kind of soundcard trick would be needed i suppose
cradek, my machine works like a dream. I owe you one.
And I must say, having a chatline to a global network of experts is a luxury so extravagant it could only be free! I'll be making chips within the week.
that's great news, congratulations
I'm wondering what would be the best way to make a readout of the angle of the spindle in the control panel. It would be a matter of throwing away the count and keeping the decimal, then showing it in degrees.
better would be a way to make the head index itself to a position. I have determined that I can make a good mechanism for clamping the spindle by air.
the ubuntu hardware compatibility list. useful http://www.ubuntuhcl.org
[05:19:54] <mIreland> http://imagebin.org/66977
nice brake, but weird airplane ;)
not super fast. but it is a good match for the tachometer i was using
If I can stop the spindle on a target position, I can clamp the spindle for live tooling. I'll just have to make a disk
[#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/,
wiki at http://wiki.linuxcnc.org/"
SWPadnos: how many cameras would you need for this? http://www.youtube.com/v/2YFkcwtpGZo
oops, um. http://www.youtube.com/watch?v=2YFkcwtpGZo
Philips Carousel for Cinema 21:9 TV - How they did it http://www.youtube.com/watch?v=teoSDTJDjF4
the question is, many meters of baling wire was used?
and gaffer's tape, undoubtedly
wow thats cool
fenn, one. they didn't actually use an array - they had people stand still, and the post production took 5 weeks
right.. i figured that out eventually
otherwise, it would be 140 seconds at 30 frames/second, so it would take 4200 still shots
which could be split up into several scenes
actually, it would be nearly impossible with an array, due to the types of camera moves they do
for instance, you can't dolly forward (or backward) with an array, because the cameras would all be looking straight into each other
right, you have to look sideways
i would not be posible to do with a multicamara single timed shot dew to the path of the movment
they did paint out the dolly track on the long dollies down the hall (for instance), so it's more or less possible
you could have a really fast moving high speed camera :)
or Viper might be another one
hey, has anyone here ever used a stamp roller on a lathe?
seems i have one.. and I just discovered that the company still exists. I'm wondering how pricey/finicky they are. New Method
lettering dies are the kind of thing that sparks neat discussions in here sometimes... wonders how theyr made.
I recall some discussion of calling files on the mailing list, was that ever resolved? I am having trouble making it work
i only found info on the Mcodes 100 thru 199m which uses external files, and are not gcode.
on other controls, i do use external files called by a main program. iirc the syntax is G99 Pxxxxxx, and the file is searched for along the system path.
i refer to subprograms as external files, and subroutines as code included inside the main.
[18:05:02] <andypugh> http://linuxcnc.org/docs/2.3/html/gcode_main.html#r3_7
Says that you just use O<filename.ngc> CALL
great! gotta try that! thx
Please do, I can't make it work :-/
well need to find a better way to get my 3d solids to nc
[18:17:55] <AchiestDragon> http://www.whipy.demon.co.uk/front-pannel-Model2.jpg
like 30mins to draw in autocad ,, saved in acad2000 dwg ,,, importing it to mastercam has taken 1 hour so far
Tom3p I have made some progress by turning logging to full. If it can't find a sub in the program it looks for a file of the same name by appending .ngc. I was thus having it find myfile.ngc.ngc
Doesn't look like it should take that long to CAM, unless it is trying to do an awful lot more than you wanted it to (like start from a foot-thick block)
No, wait, you are talking only about the import process?
andypugh: i agree the call is oXXX fred, and the file search for is fred.ngc, but i dont get the contents of fred.ngc to exec
its at face 344 of 586 atm
it is going up but slowly
tom3p: It is working for me now, but I don't know why
voodoo, dead cats
Did you add the sub/endsub lines (without the .ngc)
The trouble is that there are no errors reported from subs, it just quietly doesn't do anything.
yep... i used the sub/end sub. and o tried stripping the filename at begin of file, and single stepped it. all 'run' but all ignore the code in the subprogram
gotta ask lerhman, he wrote the stuff
changed named files to numbered files, same result
It has stopped working for me now :-)
cool, ghosts in the machine
Well, I did change one of the files.
next time it works, pastebin them
wait, regarding your question 'did you use sub endsub without the .ngc'... did you use names on the 'o<_____> sub (or endsub) line? ( did you say "o fred sub" or did you say "o 123 sub" ?)
Try quitting and restarting Axis, that seems to fix it...
The actual line (pasted) is:
seems to have speeded up must of been some dificult faces ,, but inported fine , so know that works now anyway
andypugh: yes it works now but the umpteen edits left the o word same as the file name, same as the call. i'm glad it can work, but gotta try different names to see the ruleset
It strips spaces and changes to lower case.
and yea ,,, like a bit of a mix up on units ,,, it was trying to make that 700mm by 140mm by 10mm pannel from a 1000"*1000"*1" thick plate
That would be expensive
if you guys figure out the rules it would be great if you would write them down - I've always meant to try using that feature but never have.
realy will have to figure out mastercam now i know i can import with it ,, still not shure how to get mastercam to export as nc though
andypugh: ok, the initial call oxxx call somename... the xxx doesnt need to be in the file, the 'somename' file does not want the file name on first line, the 'somename' file can have any oxxx sub/oxxx endsub pair ( i used numbers )
cradek: will put onto wiki
Yes, that example that has the file name as a single line is bogus
Any file name used should be lower case only and have no spaces
(I am in a VM so can't paste examples)
But O<This File Does Not Exist> CALL shows up in the log as a search for thisfiledoesnotexist.ngc
andypugh: the called file can use named lables eg oethel sub | ... | oethel endsub yes 'oethel' dont know about oodin but i usually dont stare at the monkeys/tease bears
andypugh: i can use internet from a vm. the host os needs the access, the vm just uses whats available to host
(not right now, but have done so inpast often
My problem is that the EMC machine is a VM and the IRC client is on the host Mac, and I can't get copy and paste to work between them
Very bizarrely, if a program with called programs has a problem, reloading the master program isn't enough. You have to quit and restart Axis. It seems that the "Reload" button only reloads the master file and does not re-parse the sub-files.
i'm not getting nested calls ( inside the called program is another call )
i do get multiple calls ok, ( call drill, call cbore, call csink , that sequential stuff works )
i'm believing the plot, if the initial plot doesnt have the calls, then the interp didnt see it my way
Put a line in the ini file, [RS274NGC] section saying LOG_LEVEL = 999 and you should be able to see what it is doing
(Output goes to emc.log in the config directory)
cradek: I accidentally found another way to read the current coordinates. G92 X0 Y0 Z0 then immediate G92.2
andypugh: i got subprograms calling subprograms, but i fooled myself before, the same dir has files o f same contect with a nake and with a number, so the proof is faulty. i've now isolated all named files and test with only numbers. ( will try all names later )
I am pretty sure that named global parameters work between programs. (It worked for me anyway)
Just as an extra point of info
How badly recursive can you get? If A calls B and B calls A?
that's an infinite loop
unless you have a bailout that is always taken eventually
Though it need not be if the calls were nested in conditional structures
recursion is usually only for tree type data structures
which are unlikely in cnc afaics
Yeah, I was just curious if it gets caught.
I would guess probably not. few other languages would
that's the programmer's job, still :)
Could be handy for machining fractals
g-code is a poor excuse for a programming language
i recommend not trying to do anything complicated with it
it's really a data format
Agreed, similar to postscript
except postscript was better thought out
Does PS have loops and conditions? I seem to recall it does
here's an example of postscript code (look at it with a text editor) http://fennetic.net/irc/2inch-45degree.ps
andypugh: no luck mixing names and numbers files, ok if all numbers (eg o111 call 111, and in file 111 ... o111 sub ... o111 endsub )
postscript, see don lancasters website
author of 'the opamp cookbook' a wiz at postscript
How odd, that you can't mix conventions
A friend of mine used to do all his illustrations in raw postscript
>i< cant, i believe its possible
It might depend on which you use first.
haha dont try to organize by hiding stuff in subdirs, thats against the rules
You can't put a path in the filename then?
It does seem a little clunky, doesn't it?
didnt try '/' in the filename, figured the parser wouldnt care for that
i'm having a little trouble with my toochanger. warns that it will violate limits if it runs. I've zeroed out the tool table, no better. where is the setup move for tool switches configured?
does it work?
mm yes, but u dont know how to connect it to the emc2
i have l298 is can help me?
someone can help me with that?
I don't even know what an I298 is
? are you still building servodriver? i have never done such a thing
L298 is H bridge
OK, that will come after you have got the encoder working.
You need to add an encoder to your .hal file and get it set up. Simplest way would be in Stepconf.
the encoder work very good
I mean in the sense of providing position information to emc
i get TTL levels from the 7404 if the encoder is 00 01 11 10
Excellent, so connect those to your parallel port, and add an encoder to your software setup with stepconf
micges1 is now known as micges
Or add it manually if you prefer to edit your own stepconf files.
In your case you will have to edit the .hal file anyway (I think) so that encoder A and encoder Z are the same channel
I'll be off for a few moments - i go to the computer of the EMC, OK?
mmm try to understand what to do
if i go to Applications > CNC > EMC2
Do you have anything working with emc yet?
and i select the ethc-servo
is give me error
And have you used the etch servo config before?
(I have no idea where you are in the process)
trying to sort out file conversion methods atm ,, doing blender >dfx>autocad bit of processing and scailing then doing dwg to >mastercam to nc with a 5 axis machine for stuff like http://www.whipy.demon.co.uk/snapshot5.png
Yes, what is your question?
How much of your machine is already working, or is the answer "none of it yet"?
well a 3d solid is a 3d solid ,, and other than tools sizes a 5 axis should cope with that if made in some sections
hence the processing in autocad to split it into managable sections
blender is far easyer to use to manupulate the surfaces than autocad ,, but blender tends to work to a relative scail rather than an absolute dimentions
this is loook very nice (but my cnc not work :( )
i-pink: What is the error message? did etch-servo work before? Is you hardware connected up?
andypugh: got all named and all numbered working nested, no mixing yet, and no luck with paths in names ( there is a rule about files being in a specific dir , so not pursuing that )
the emc2 send me to this page http://emergent.unpy.net/projects/01142347802
i have the L298 but how he make the encoders???
hii is i-pink from the cnc computer
andypugh: a main with name and with numbered subprogs is ok now, but need to test label<name> call nnnn
btw the angle brackets are neccesary for named files
i-pink: His encoders are commercially bought ones that were probably factory fitted to the motors
tom3p: Putting a path in the filename doesn't work
i-pink: Has your CNC ever worked? Or are you starting from the beginning?
only the x and y axis work (is steppers)
the z axis is DC motor with encoder
Presumably the encoder is mounted on the motor and the H-bridge and encoder are all connected to the computer?
the encoders look like u.s.digital brand ( like the h.p. units ) http://usdigital.com/products/encoders/incremental/rotary/kit/e8p/
is the gary encoder
i-pink's encoder is home-made from mouse parts and cardboard.
But then so is mine.
no.. my encoder is this
i-pink-cnc: yes a gray scale encoder, but jepler has a commercial unit
hewlett packqard sold off its tiny encoder .. now its Avago http://www.avagotech.com/pages/en/motion_control_encoder_products/
They are quite cheap too.
the size of the encode is not importent, i wont to unterstend how to work with this
Is the encoder wired to the computer?
Can you see the pin values changing in halscope when you turn the motor by hand?
no.. i s not connect
OK, wiring it up would be a good start
i dont know to what pin to connect it
Did you create your current configuration with Stepconf?
this is my encoder
[21:58:25] <i-pink> http://www.speedyshare.com/543818401.html
i heve configuration but is work only with the stepper
Yes, but did you create it using the stepconf application?
stepconf would never work with an encoder
stepper and encoder are not (commonly) used together
It won't drive a servo, but it will set up an encoder.
It will also (if that is how the original config was made) show what pins are available.
It will create an encoder and make the hal connections for a spindle speed enoder which would be a good start
i have 2 stepper motor to the X and Y axis, and the Z axis need to be DC motor with the encoder
what to do?
Yes, but my question is if your current config was created with StepConf Wizard or by hand, or is a standard config re-used
by editing the hal
In that case you will have to do it again.
i-pink-cnc: where did that hal come from? from step-config or supplied with emc?
i dont know
now all is not work :(
look at this error
[22:06:39] <i-pink-cnc> http://www.speedyshare.com/953958631.html
the red line box is your workspace, you're at top, if you go up you will exceed the work space.
You need to re-home your Z probably
i think..., begin with a step config, and ignore the output for Z, respect what it says for x & y, edit your own Z into the file. begin with just 2 stepper axis and build on that.
i gotta write down the named subprogram stuff and get back to work, byebye
it give me that error now
[22:09:49] <i-pink-cnc> http://www.speedyshare.com/901120112.html
That's the same error. Move the Z down in Axis (it doesn't matter that it won't actually move) then touch-off to zero
you're still at top of red box
i try to put it on 0
it not work
The physical axis won't move, but you should be able to jog to a lower Z in EMC2 then touch-off to zero there
the emc not work!!
i try to do it
but is not work
You can't run the program, you should be able to jov
jog, I mean
i make the confiroration from zero
now the virtual CNC is work
You need to find 2 spare input pins on your parallel port for the encoder inputs
Look at what is there now, and the wikipedia parallel port page
i use now only in 4 output pins
Use pins 12 and 13 for your encoder
the parallel port get TTL 5V?
Yes, that is fine
from i sent it from 7404
is take several min to connect it
now the configuration is
net Xup => parport.0.pin-03-out
net Xdown => parport.0.pin-02-out
net Yup => parport.0.pin-05-out
net Ydown => parport.0.pin-04-out
net Zup => parport.0.pin-07-out
net Zdown => parport.0.pin-06-out
OK, the easiest way to do this, if that is all you have is to go CNC-stepconf wizard and create a new config. (you can have as many as you want on the same computer)
So exit emc, start stepconf wizard, and create a new machine configuration
what to chose?
create or modify
Fill in the first page, probably 2000 in both pulse boxes, 5000 in the direction boxes, then whatever the latency check says
no, i mean in the first window
put on other
Is there a latency test button at the bottom?
depends on your drivers.
is work good on low frec
But 2000/2000/5000/5000 should be safe
my step driver get up and down (in 2 wires)
I have suggested 5000 twise
A bunch of you got booted
what is it?
looks like one of the IRC nodes crashed
it's called a netsplit. it happens a fair bit on freenode
ok what to do now?
Do a latency test, you should have a button on the stepconf screen
Assuming you are using emc 2.3.0 or newer
how i check?
You press the button marked "Latency test"
ok i see i need to config it Manuel
i cant see this button
how i can set it buy editing the files
OK, it actually says "Test Base Period Jitter"
I was working from memory
the x and Y work good now, i am only need to configur the EMC to send UP/DOWN to the DC motor and config the EMC to read the encoder
all the configuration for the X and Y is good
Indeed. And you can do that by editing the HAL file if you want, but you will have to read the manual and sort out how to do it yourself.
Just press that "test base period jitter" button
where it the "test base period jitter" button
Underneath the "Base Period Maximum Jitter" box
i am in this window
[22:54:54] <i-pink-cnc> http://www.speedyshare.com/197208720.html
You closed stepconf then?
So you are basically demanding my advice, then ignoring it?
Applications -> CNC -> Latency Test does the same thing
i close it because now the X and Y work gooood!
i need only to configur the Z axis
And because you have decided that you don't care about the Z any more?
So configure it then
Edit the HAL files
prices like those wouldv'e stopped me buying shelf brackets
oops scrolled sorry :) I meant those supported rails
the wizard give me only step/dir option, and i need up/down (step for the X Y) so i make it Manuel
MattyMatt: Ah, I see what you mean now.
andypugh, in the HAL file, how i set the encoder?
How are the rack slides working?
not in use yet. I haven't started cutting
i-pink-cnc: I have no idea, read the manual.
I can show you what stepconf set up for me, but you will have to edit it yourself
But it would be far easier to use stepconf in a new configuration, just create an encoder with two channels on the pins you need, and to paste the HAL that creates into your existing one
I've seen a Solsylva design v similar to mine but obviously much stiffer in the table, with thick beams either side of the screw. it's a better design but the shelf slides don't fit in very well
andypugh, i see an option to add quadrature encoder to control spindle speed
and I've decided I need 2 pairs for the table, so the economics aren't fantastic anymore either
MattyMatt, you can help me with the configuration of the HAL file?
Yes, that was the bit you need, but then you need to edit the hal to connect encoder position to z-axis position
i-pink-cnc: Try adding these lines to your hal file http://www.pastebin.ca/1611015
but the spindle is the cutting motor?
Yes, you can chance the word "spindle" to ZAxis if you want
But for the time being, try that and see if you can see encoder.0.position changing in a halmeter as you turn the motor
You should be able to get it to go up and down depending on which way the motor turns
and how i configur the legs in the LPT?
I assume you have 2 TTL signals and an earth?
You might not need to connect the earth, depending on the rest of your wiring
The first question is, if you add those lines to your HAL file, does EMC still start?
i cant understend how i cunfigurd the HAL
the file standard_pinout.hal
Is that the file you are using>
you need to edit emc2/configs/EMC-HAL-STEP-MM/EMC-HAL-STEP-MM.hal
But make a backup copy first
re: subprograms someone might look at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?SubProgamFiles
this is the files
autosave.halscope README stepper.tbl
core_stepper.hal Sherline3Axis_inch.ini stepper.var
core_stepper.hal.bkap Sherline3Axis_mm.ini stepper.var.bak
You are looking in the wrong place. look where I said
my-mill/ Sherline3Axis (copy)/
you can upload your HAL file?
tom3p, it says "the angle brackets are neccesary for filename" immediatly after examples which don't use them :p The later example do use them tho
Those examples are labels not filenames, I think
i-pink-cnc My HAL file won't work for you
It looks like you are using a modified standard config, I have no idea where your actual configs are.
i know, i want to look at the file and look if the standard_pinout.hal is the hal file
And I have other things I really need to be concentrating on right now
or maybe is the core_stepper.hal
tom3p, until I've tried emc, I can't tell you anything, but I know that wiki page needs clarification
I should build my controller and learn emc while I'm waiting to win a drill press on ebay :) I missed a good one today
the next good one has a burnt out motor, but looks 3x the weight of a normal bench press
gbp100 delivery charge will put off gazumpers from further away, I could collect it
I got weird Problem with Screen, after installing fresh ubuntu hardy i installed EMC2 when in generic kernel UBUNTU WORKS FINE..
but whenever i try to use RTAI than SCREEN SPLITS VERTICALLY 8-10 STRIPES :(
/etc/xorg.conf is where the screen modes your screen can handle are stored
i mean i am not able to RUN UBUNTU WITH REAL TIME KERNEL..i am sorry if this is not proper place to ask this question..
I dunno why it should be different with different kernels tho
do you have another screen you could try?
a higher res one?
its laptop DELL 1440
intel GM45 chipset
is this related to KERNEL? coz using generic kernel everything works fine..
chipset isn't the problem, it's xorg or X11 trying to set a mode your screen can't handle
yaa thats true.. how can i change a MODE using Command prompt
coz i am not able to work in that STRIPE MODE
i tried it with nano. ... nano /etc/X11/xorg.conf
I would seek advice from Dell experts about what to change
They are giving support only for the Microsoft OS.. i think
coz this comes with VISTA HOME
and I don't know why it is different with the default kernel. maybe it's Gnome setting the wrong mode (assuming ubuntu uses it)
not Dell, Dell experts ;)
hey mattymatt .. should i try this? i read somewhere resetting GNOME .. rm -rf .gnome .gnome2 .gconf .gconfd .metacity
if that works, then so should logging on as a new user, so try that first
it seems less drastic to me
can I please pick your brains for a moment, is there anyone one here who has or knows any thing about microbot alpha II's or teachmover robot arms ?
hey its not GNOME problem i think.. coz even login screen looks SPLIT 8 times..