#emc | Logs for 2007-11-14

[02:11:35] <cradek> haha http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=130173579919
[02:11:42] <cradek> now what possible use could that be
[02:12:35] <cradek> "MaxNC to CAT50 spindle adapter"
[02:26:50] <ds2> isn't that like an adapter for a 10000lb class IV hitch for a Geo metro? ;)
[02:27:20] <toastydeath> I HAVE SCHOOLED
[02:27:43] <toastydeath> sometimes you have a boring bar that is cat50
[02:27:52] <toastydeath> or some other equally ridiculously expensive thing
[02:28:04] <toastydeath> you'd probably buy the cat50 adapter than another 7000 dollar boring bar
[02:28:13] <ds2> the MaxNC spindles I have seen have a taig taper and thread (3/4-16) nose
[02:28:20] <toastydeath> or not
[02:28:34] <ds2> now... a 7000 boring bar + a cat 50 adapter... hmmmm
[02:28:52] <toastydeath> the machine interface is usually fixed on expensive stuff
[02:28:55] <ds2> maybe its been modified to thread on via the pull stud ;)
[02:29:00] <toastydeath> loool
[02:39:47] <tomp2> geda netlists might be used for hal files.
[02:39:53] <tomp2> example of simple nets http://imagebin.org/11726.
[02:40:01] <tomp2> automatically created netlist with notes http://pastebin.com/d510092e9.
[02:40:03] <tomp2> gschem source file http://pastebin.com/m797ff10b
[03:21:00] <Jymmm> /joioin #phpc
[03:55:51] <stustev> good evening
[03:56:20] <cradek> hi stuart
[03:56:33] <stustev> I say again - sweet
[03:56:40] <cradek> is it working over there?
[03:56:47] <stustev> oh yeh
[03:57:01] <stustev> looks very good
[03:57:23] <cradek> is it sort of the right shape? I didn't have much to go on.
[03:58:09] <cradek> when you get your kins working, you should be able to plug in the geometry of the simulated machine and see what happens
[03:58:12] <stustev> the shape is very good. it is not exactly like my bridge but the motion is represented almost perfectly
[03:58:38] <stustev> somehow I figured that
[03:59:00] <stustev> I have been MDI'ing the machine around.
[03:59:25] <stustev> the motion is like as normally expected
[04:00:00] <cradek> I'm getting a few 'exceeded soft limit' errors that I don't expect, not sure what that's about
[04:00:55] <stustev> I expanded the soft limits to +120 and -120 for the fourth axis and +720 and -720 for the fifth axis
[04:01:24] <cradek> ok, I didn't know what is normal
[04:01:31] <cradek> seems like maybe it needs some more Z travel too
[04:01:37] <stustev> my machine has +/-120 for the B axis and the C axis has slip rings
[04:02:01] <stustev> I haven't tried much on the linear axes
[04:02:26] <stustev> I was able to run the axis.ngc and that was good enough for me
[04:03:20] <stustev> my machine has 200 inches X travel, 100 inches Y travel and 40 inches Z travel
[04:03:51] <stustev> I haven't checked the limits on the linear axes yet
[04:04:23] <cradek> 1000x1000x300 mm, much smaller than your numbers
[04:06:05] <stustev> if I change the linear limits will that change the size of the components on the screen? I don't suppose so but it's worth a question
[04:06:24] <cradek> nope
[04:06:43] <cradek> but you can enlarge the machine without too much trouble.
[04:07:07] <stustev> in which file?
[04:07:33] <cradek> src/emc/usr_intf/axis/scripts/5axisgui.py
[04:08:01] <cradek> you could make it look just like your machine (with some nontrivial amount of effort)
[04:08:51] <cradek> huh, surprising how much time it took me to write so few lines of code - it looks so simple
[04:09:08] <stustev> I don't need that kind of realism. I just want to run some of my current programs in the sim
[04:09:31] <cradek> then just change the limits and it'll cut through itself, no problem
[04:09:48] <cradek> and share a nice program :-)
[04:09:49] <stustev> don't you know it. when you get done it seems like it should have been 5 minutes work but it takes 5 days
[04:10:25] <cradek> you would have to post with this machine's geometry to make the program look right?
[04:10:42] <stustev> no
[04:11:09] <stustev> the pivot length would be the only requirement
[04:11:37] <cradek> I understand
[04:12:01] <stustev> I will have to make some adjustment to account for the type of rotary motion.
[04:12:12] <stustev> my machine is a little different.
[04:12:47] <stustev> the rotary axes are on a 360 degree scale
[04:13:17] <stustev> this means the axis command never exceeds 359.999 degrees
[04:13:29] <stustev> the next .001 degree move to 0.0
[04:14:08] <stustev> also the command C90.0 and C-90.0 point the tool in the exact same direction
[04:14:29] <stustev> the + moves the axis in one direction and the - moves the axis in the other direction
[04:14:44] <stustev> but the position is the same position
[04:15:39] <cradek> man do people agree on how anything in gcode works?
[04:15:55] <stustev> there is no REAL standard
[04:18:39] <cradek> in the simulated machine, if you call the end of the 'pointy thing' the gauge point, the pivot length is 250 mm
[04:19:14] <stustev> in 1998 I talked with Kurt Zierhut from Haas Automation. He told me Haas thought about changing the axis designations because there is no standard and Haas might want to differentiate themselves from the rest of the world.
[04:19:34] <stustev> I will call that the pivot length
[04:20:20] <stustev> The Okuma machines have unique controls and some unique treatment of g code.
[04:21:07] <stustev> my bridge machine is configured in one of two ways to configure a Fanuc control. The other way is just like the machine you have built.
[04:21:12] <cradek> a nice thing I saw on one of your machines that will be challenging in emc was the plunge and face jogwheel function
[04:21:17] <toastydeath> most controls agree on fanuc 2d stuff
[04:21:27] <toastydeath> but even then.
[04:22:09] <stustev> surprisingly :) I have thought about that a little bit
[04:22:31] <stustev> I don't think it should be too bad.
[04:23:20] <stustev> toastydeath: yes, it has finally converged to a semi-standard
[04:25:29] <cradek> I suspect we're going to extensively test inverse time feed mode soon if you get a program running. I think that part of the interp is not as well exercised as the rest.
[04:25:40] <stustev> cradek: the plunge and face could be implemented just like the single point threading and rigid tapping?
[04:26:11] <stustev> yes we will see extensive use of inverse time feed
[04:26:37] <cradek> we'll have to explore how teleop mode works (step 1). step 2 is doing it with a jogwheel
[04:27:05] <cradek> currently I think the guis let you teleop in axis directions only. we would want to add some new directions.
[04:27:47] <stustev> the plunge direction of my machine is the current tool axis vector
[04:27:58] <cradek> yes
[04:28:18] <stustev> the face direction is orthogonal to the tool axis vector
[04:28:18] <cradek> and that comes from only B,C axis positions right?
[04:28:26] <stustev> yes
[04:29:40] <cradek> of course there are infinitely many of these directions...
[04:29:49] <cradek> how do you pick one?
[04:30:07] <cradek> getting sidetracked - that's down the road
[04:30:38] <stustev> how do you pick a direction?
[04:30:53] <stustev> of the orthogonal?
[04:31:00] <cradek> yeah there's one plunge but infinitely many face dirs
[04:31:53] <stustev> on my machine it is perpendicular to the axis of rotation of the B axis
[04:33:07] <stustev> as you rotate the C axis the B axis rotates.
[04:33:25] <stustev> the machine move perpendicular to the B axis
[04:34:32] <stustev> I find the facing movement is not very useful
[04:34:43] <stustev> The plunge movement is very useful
[04:38:18] <cradek> running out of battery - I need to get going
[04:39:22] <cradek> hope this tweaks some interest in 5 axis...
[04:39:25] <cradek> goodnight all
[04:42:52] <stustev> goodnight also - I need my beauty sleep
[14:54:27] <skunkworks> This is awesome. http://youtube.com/watch?v=JmJoyuUJj2Q
[15:12:01] <cradek_> cradek_ is now known as cradek
[16:08:35] <jepler> oh, I see now -- in classicladder, the "D" output of a "T" block is energized just once when the timer runs down, while the "R" output is energized whenever the timer is running.
[16:13:32] <jepler> no that's not quite right either
[16:52:20] <alSMT> need some pointer on how to start to write ladder logic I click add and a properties box pops up what exactly does it want I type FwdinI0 and select no contact and say ok and nothing appears what am I missing?
[17:01:25] <jepler> on the main window, you click "editor", then "modify" (to change an existing rung) or "add" (to start with a fresh rung)
[17:01:39] <jepler> then click the type of contact you want to add, then click where on the rung you want to add it
[17:02:57] <jepler> then fill out the fields for that contact. For example, after I place a NO Input, I might enter %B0, or %B1, etc (to use a result generated elsewhere in my ladder) or %I0, or %I1, etc (to use a value coming from outside classicladder)
[17:03:20] <jepler> for an output I would enter %B0 (to set a result used elsewhere in ladder) or %Q0 (to set a result that will be used outside classicladder)
[17:03:32] <jepler> you have to type it in uppercase: %B0, not %b0
[17:03:40] <jepler> then click apply, then place the next contact
[17:04:03] <jepler> to change a contact after placing it, choose the "arrow" button instead of a contact type; to remove contacts, click the "eraser" button
[17:04:18] <jepler> the classicladder GUI is a hard program to love
[17:04:54] <jepler> you won't see the rungs updating until after you press "OK" to exit the rung editor
[20:33:34] <jlmjvm> jepler:is there any way the cp can be linked to the F2 on,off button instead of the external estop?
[20:35:00] <jepler> jlmjvm: anything you could do in your old configuration, you can do in your custom.hal file. but stepconf always sets up the charge pump in the same way, if you choose "charge pump" as the function of a parport output pin.
[20:37:19] <jlmjvm> ive got the custom.hal stuff from the other day where its tied to the estop button
[20:38:43] <jlmjvm> does step conf set it up for the F2 button normally?
[20:40:10] <jlmjvm> i think it does and the problem the other day was because of the external button
[20:40:22] <jepler> the intention of stepconf is that when the emc is not in estop, the charge pump output will be active (toggling). the only reason you had to add any lines to your custom.hal to get that behavior is that there was a bug in stepconf's handling of the charge pump enable signal
[20:40:42] <jlmjvm> k,my bad
[20:42:03] <jepler> you have the freedom to do whatever you want: for instance, leave "charge pump" selected as an output, but in your custom.hal file link the enable to something else (e.g., the signal xenable which will be TRUE when in Machine ON and not when in other states). (If you choose to do it that way, "unlinkp charge-pump.enable" first, or you'll get an error when you regenerate your configuration with a later version of stepconf)
[20:42:24] <jepler> or you can choose to do it in the exact same way you did it in 2.1 by copying whatever lines of .hal code set up the watchdog the way you liked it before
[20:42:46] <jepler> (in that case you'd set the pin as 'unused' in stepconf)
[20:44:08] <jlmjvm> i never used it in 2.1 other than using the dallur example and testing for voltage
[20:44:18] <jlmjvm> i like that xenable idea
[20:45:28] <jlmjvm> thats what i was doing to turn the power enable contactor on originally,worked great untill you turned the computer off and on again
[20:45:58] <jlmjvm> then all outputs would go high untill fully booted and in emc
[20:49:14] <jlmjvm> the xenable idea would be the way to go,that way any error would turn off the enable cantactor
[21:04:09] <jlmjvm> jepler:that was too easy,works like a charm
[21:04:53] <jepler> I'm glad to hear it
[21:05:00] <jlmjvm> just changed net estop-ext to xenable in my custom hal
[21:05:03] <jepler> did you try crashing the computer, just to make sure it works as expected?
[21:05:33] <jlmjvm> how should i crash it
[21:06:00] <jepler> well I was joking really
[21:06:27] <jepler> but that's supposed to be one of the strengths of using charge-pump instead of a simple "enable" signal: you lose the toggling signal when the PC crashes
[21:06:39] <jlmjvm> lol,this is working great
[21:07:16] <jlmjvm> now if you toggle the estop it stays off till you hit F2
[21:08:43] <jlmjvm> before it would enable as soon as you hit the emc icon,and stay on sometimes if you turned the program off without the estop button engaged
[21:09:08] <jlmjvm> i like it
[21:10:52] <jlmjvm> thanks a bunch
[21:11:26] <jepler> you're welcome
[21:48:51] <buckie555> in the TOOL_CHANGE_POSITION definition in the ini file is there a wildcard that can be used for some of the fields? - During a toolchange I want to go to a certain z pos but want to stay where I am in x and y
[21:49:17] <skunkworks> I think that is fixed in trunk..
[21:49:36] <skunkworks> Atleast I remember cradek doing something with it.
[21:49:54] <cradek> that is not currently possible, but I'd also like it
[21:50:54] <skunkworks> Ah
[21:50:58] <buckie555> OK so am I correct in thinking if the definition is present the x y and z are all moved and if not present none are moved?
[21:51:29] <cradek> yes
[21:51:50] <buckie555> thanks for the clarification.
[22:02:58] <buckie555> the upgrade to 2.2 instructions talk about 'gpg --keyserver pgpkeys.mit.edu --recv-key 96935D7D' etc. Will this still work if the node does not have network access and if not what should I do instead?
[22:04:56] <cradek> if you skip the gpg steps, you may have to reassure it when it complains that the package is untrusted
[22:05:17] <cradek> I think that is nothing other than typing yes or no or clicking, but I don't remember for sure
[22:06:41] <buckie555> --no-debsig maybe?
[22:06:51] <cradek> not sure, sorry
[22:07:02] <cradek> but clearly you will be able to figure it out
[22:07:21] <buckie555> well as long as dpkg exposes some way around the authentication I should be ok
[22:07:34] <cradek> yes, I think it just asks you
[22:07:50] <buckie555> cool
[22:08:26] <alex_joni> it only issues a warning which you can click away (if you use synaptic)
[22:08:33] <buckie555> I have all my toolchange code working on my dev platform - just need to test it on the live machine now
[22:08:37] <alex_joni> if you install a package with dpkg -i , it doesn't care
[22:08:47] <cradek> buckie555: excellent
[22:08:57] <cradek> we want a video if it works!
[22:09:04] <buckie555> alex_joni - great thanks
[22:09:08] <cradek> heck even if it doesn't, those are sometimes fun to watch too
[22:09:34] <alex_joni> remember to use 1-2" lexane in front of the camera the first time :D
[22:09:35] <cradek> we saw ours drop tools on the table, put tools where there were already tools, etc etc...
[22:10:23] <cradek> bbl
[22:10:24] <buckie555> If I break another set of drawbar push rods my partner will floor me. Needless to say I now have logic in place that prevents any spindle rotation whilst the pnuematic drawbar is actuated
[22:11:06] <buckie555> WHEN it works and I've stopped grinning I'll happily take some video
[22:11:17] <cradek> I missed when ray activated the coolant pump that was full of ten year old coolant, but had only a hose pointing at the ceiling
[22:11:21] <cradek> I only heard about it later
[22:11:31] <cradek> it takes lots of trial and error...
[22:11:43] <cradek> best of luck, and be careful
[22:13:06] <buckie555> yes thanks - I will be. I plan to test as mush as possible with cylinders disconnected and the carousel out of the way. Only when I'm happy with each stage will I let it run free - and even then my hand is going to be right over the big red button
[22:15:59] <fenn> why do drawbar pushrods break if the spindle is turning?
[22:16:06] <fenn> is that supposed to be a safety feature?
[22:17:17] <fenn> if it were me, i'd do any safety logic like that with real relays (in addition to software)
[22:18:22] <buckie555> the rods on the sabre are necked. I can only assume it was to ensure that in such a condition the rods would break and not damage the piston or spindle assembly
[22:19:40] <buckie555> Well I figure that the original siemens control did it all in software so I might as well too. I have retained all the original hardware safety circuits
[22:21:24] <buckie555> When they snapped it was brown trouser time - The 15hp spindle servomotor wasn't going to stop for anything
[22:21:57] <fenn> that's like a small automobile
[22:22:40] <buckie555> yep - no substitute for power
[22:23:00] <buckie555> or rather torque
[22:23:03] <fenn> um, hrm. what about tool rake?
[22:24:19] <buckie555> well and that too
[22:24:56] <buckie555> A slight aside but has anyone here ever come across a Crudginton multispindle cnc lathe?
[23:25:38] <alex_joni> has anyone seen this? http://www.gsimple.eu/
[23:44:30] <user_> user_ is now known as skinnypuppy34
[23:48:09] <skinnypuppy34> Alex looks cool . Have you had a chance to run it?
[23:55:56] <toastydeath> fff
[23:56:50] <dmess> hi all