#emc | Logs for 2008-07-05

[00:25:45] <Sweeper> I gotta say, converting an atx power supply to a bench supply should be like the first electronics project anyone ever does. screw led blinkeys!
[00:28:58] <archivist> pc power supplies are designed for a base load, bench supplies can work down to 0 current
[00:30:34] <Sweeper> yea, that's why you put in a nice big resistor :)
[00:31:30] <SWPadnos> bench heater
[00:33:45] <archivist> although switch mode psu design has improved, some require some balance to the output loads
[01:00:30] <stustev> anybody home?
[01:01:26] <Sweeper> nobody but us chickens
[01:01:41] <stustev> cluck cluck
[01:02:18] <archivist> Im at work
[01:02:29] <stustev> everybody must be lighting fireworks
[01:02:42] <SWPadnos> someone around here is
[01:02:47] <SWPadnos> I thought the house was falling apart
[01:03:17] <stustev> I haven't heard much this year
[01:03:35] <SWPadnos> me either, until about 30 minutes ago
[01:04:04] <stustev> the 'police' have been patrolling extra - it is against the law to light fireworks in my little town
[01:04:25] <SWPadnos> interesting. it was just legalized in Vermont a few years ago
[01:04:50] <stustev> out here in the wild west it is being restricted more and more
[01:05:55] <stustev> the Kansas legislature just lifted a LOT of restriction on firearm - pistols, rifles, shotguns and placed more restrictions on fireworks
[01:06:21] <SWPadnos> weird
[01:06:53] <SWPadnos> there's no fireworks equivalent to the NRA I guess :)
[01:07:03] <Sweeper> oh there is
[01:07:12] <stustev> true gun control is the ability to - hit what you shoot at -
[01:07:58] <SWPadnos> heh
[01:08:02] <stustev> we can now own fully automatic weapons and short shotguns and rifles - licensed of course - but now legal
[01:08:29] <SWPadnos> so people can shoot hte evil fireworks before they explode
[01:09:32] <stustev> LOL true - i want more guns!!!!
[01:10:26] <stustev> is there any news on joint constraints in addition to axis constraints?
[01:11:23] <SWPadnos> I don't think so
[01:12:33] <stustev> I talked with cradek at the workshop about it. We just had one short conversation. He said he thought some work was being done.
[01:13:35] <SWPadnos> there was some work done, but I think we didn't get as far as we had wanted
[01:14:36] <stustev> I finally have the cinci ready. The gear change component is working very good. The spindle is working very good. I started collection data today for lead screw compensation. I will be working on that in the morning. I will be able to comp A and B but I will have to wait until next week to comp X Y and Z
[01:15:48] <stustev> are the constraints something that can be worked on a little at a time or is that something that needs focused attention?
[01:16:55] <SWPadnos> I think it's a big change that has to be done at once, then more can be added later
[01:17:15] <SWPadnos> ie, there's infrastructure that has to change, then maybe features can be added or tweaked
[01:19:13] <stustev> watching the tool cone move in five axis motion in the axis interface is pretty cool
[01:20:26] <jepler> At Fest I told everyone I knew the solution to the constraints problem, but then I discovered my math wasn't up to it
[01:20:26] <stustev> did I read somewhere that EMC2 could use information in a gcode comment line?
[01:20:57] <SWPadnos> it's possible, and done for several things
[01:20:59] <stustev> somebody is lurking - good evening jepler
[01:21:01] <jepler> hi stustev
[01:21:10] <jepler> how's things in wichita? they are exploding here.
[01:21:13] <SWPadnos> but you'd need to add more things if what you want isn't already there
[01:21:43] <stustev> real quiet here - i think my wife and I will go downtown around 9:30 to see the show
[01:22:12] <stustev> swpadnos - duh
[01:22:22] <SWPadnos> heh
[01:23:07] <SWPadnos> I learned today that there are a few AXIS directives, and I knew about MSG, LOG, DEBUG (?), and maybe PRINT
[01:23:12] <jepler> SWPadnos: what I was talking about earlier are some comments that AXIS treats specially when generating a preview: (AXIS,stop) to stop generating the preview, and (AXIS,hide) and (AXIS,show) to skip part of the program in the preview.
[01:23:16] <jepler> er stustev ^^
[01:23:36] <SWPadnos> out of curiosity, what's the intent with hide?
[01:23:39] <jepler> (AXIS,stop) came up because someone wanted an infinite-loop gcode
[01:23:40] <SWPadnos> it seems dangerous
[01:24:41] <jepler> SWPadnos: for chris's tool length probe, he has to put certain moves in the gcode program. but they're mostly useless in the preview plot once you've got the code debugged, and they confuse the limits calculation. so you can put (AXIS,hide) at the stop and (AXIS,show) at the end to pretend they don't exist.
[01:25:04] <SWPadnos> ok - that makes sense
[01:25:20] <SWPadnos> putting it around a rapid somewhere could be surprising though
[01:26:10] <jepler> there's no real guarantee that the preview is like what the machine will do
[01:26:25] <SWPadnos> that's true, but probably unexpected
[01:26:25] <jepler> O100 if [previewing inside axis] do one thing O100 else do something entirely different O100 endif
[01:26:35] <stustev> I added a wish list item about pivot length and tool length. I want use the posted pivot and tool lengths in addition to the dynamic pivot and tool lengths to run programs from other machine controls.
[01:26:54] <jepler> I am not sure how you write the "previewing inside axis" conditional statement, but I'm sure it exists
[01:26:54] <SWPadnos> what is a dynamic pivot length?
[01:26:58] <SWPadnos> heh
[01:27:40] <stustev> the distance from the gage point of the tool adapter to the pivot point on the rotary head
[01:28:11] <SWPadnos> I guess I'm wondering how that would be dynamic, for a given machine
[01:28:38] <stustev> maybe the term dynamic should be defined
[01:29:07] <SWPadnos> heh - ok with me :)
[01:29:53] <stustev> it is a static number for a given machine just as the tool length is a static number for a given tool. The control uses the static numbers to do dynamic compensation for the 5 axis tool length compensation. clear huh?
[01:30:07] <SWPadnos> ok
[01:30:43] <SWPadnos> so the post will take some length (written to the file as a note) into acocunt when generating G-code?
[01:30:59] <stustev> yes
[01:31:44] <stustev> this then needs compensated out of the gcode program to be able to use the program in EMC2
[01:31:55] <stustev> There are several ways to do it.
[01:31:57] <SWPadnos> if anything other than W changes from the post as a result of that value then it may be a hard problem
[01:32:20] <stustev> it is just a W change
[01:32:47] <SWPadnos> ok, then it may be an easy problem :)
[01:33:00] <stustev> cradek suggested using G54 W and this will work
[01:33:17] <SWPadnos> is it possible for the post to write the value to a parameter instead of (or in addition to) a comment?
[01:33:26] <stustev> I could also remove the static numbers for pivot length and tool length and then run the program
[01:34:16] <stustev> I could do that in my post for files going forward but I have a 'plethora' of previous files.
[01:34:42] <stustev> I could also put the entire compensation calculation in the tool length
[01:34:50] <SWPadnos> an awk script could probably add that to existing files
[01:35:10] <stustev> very possible - perl too
[01:35:21] <SWPadnos> find (PIVOT LENGTH = xxx) and write that out plus another line #1999=xxx
[01:35:24] <SWPadnos> or whatever
[01:35:46] <stustev> wow
[01:36:00] <SWPadnos> ?
[01:36:41] <stustev> adding that to existing files is an easy answer
[01:37:04] <SWPadnos> seems that way. writing the awk is left as an exercise for the reader :)
[01:37:40] <stustev> I didn't want to be changing the configuration in EMC and I didn't want to lie to the tool length - both error prone operations
[01:38:20] <SWPadnos> the other thing you'd need is a variable that has the machine's actual pivot length
[01:38:36] <SWPadnos> that could probably be put in the ini file in RS274NGC_STARTUP_CODES
[01:38:52] <stustev> yes and be able to sum them - and the tool lengths also
[01:39:38] <SWPadnos> oh - you need to correct the tool table also?
[01:40:16] <stustev> yes - as the previous posted programs have a pivot length and a tool length calculated into the gcode positions.
[01:41:11] <SWPadnos> oh, so the tool comp is partly calculated by the post, given a nominal tool table and the pivot lenght
[01:41:14] <SWPadnos> lengh
[01:41:17] <SWPadnos> +t
[01:42:07] <stustev> no the tool length is posted into the positions and the tool lengths MUST then match the numbers used.
[01:42:25] <SWPadnos> hmmm
[01:42:48] <SWPadnos> ok, the post has already done tool length comp?
[01:42:53] <SWPadnos> (in W)
[01:43:29] <stustev> yes - in W - the X Y Z positions in the gcode program are compensated for the pivot length and tool length
[01:44:03] <SWPadnos> ok. I don't know if that will work with just a G54 W offset (cradek may know, he certainly knows more about that than I do)
[01:45:26] <stustev> I am sure it will. I would need to determine the difference between the two pivot/tool length totals and put that amount in the G54 W
[01:45:48] <SWPadnos> the actual tools will probably have different lengths also though
[01:45:53] <SWPadnos> than the post thought
[01:46:28] <stustev> yes - that will be taken care of by the current compensation - I just need to remove the previous compensation
[01:46:40] <stustev> dynamically
[01:46:43] <SWPadnos> so you need tool length comp (per tool) in addition to the pivot length (per posted length/machine pair)
[01:46:49] <stustev> eys
[01:46:51] <stustev> yes
[01:47:18] <stustev> it sounds more complicated than it is
[01:47:44] <SWPadnos> well, since tool comp can be done relative to a "standard" tool, you may be able to do it all with just the tool table
[01:47:56] <SWPadnos> make the standard tool the pivot length difference
[01:48:06] <stustev> it could all be done with the tool table
[01:48:32] <SWPadnos> then offset each other tool by the actual difference in tool length of that tool vs. the posted length
[01:49:09] <SWPadnos> though that probably wouldn't work for pivots that get shorter :)
[01:49:16] <SWPadnos> (no negative length for the standard tool)
[01:49:39] <stustev> why would negative lengths not work
[01:50:05] <SWPadnos> I think that only works for relative lengths. I don't know if you can set the "standard" tool to a negative length
[01:50:17] <SWPadnos> if you can, then it's a whole lot easier
[01:50:38] <stustev> I will try it to see if it handles it. I am sure it will.
[01:50:46] <SWPadnos> cool
[01:50:55] <SWPadnos> then we can check off that feature request ;)
[01:53:19] <stustev> yes it can be but I will still work on getting the numbers into the gcode program. Every time we have human intervention we add another opportunity for error. If I can run with the normal pivot length and tool length and have the control handle the "extra" pivot length and tool length from the previous gcode program we eliminate a lot of human intervention.
[01:53:47] <SWPadnos> what are the programs named?
[01:53:52] <SWPadnos> (ie, the extension)
[01:54:17] <stustev> .nc and .pu1 and .txt
[01:55:26] <SWPadnos> ok, if you can write a filter to add/change whatever needs to be added/changed, then you can make it un automatically when the files are loaded into AXIS
[01:55:33] <SWPadnos> s/un/run/
[01:55:34] <stustev> .nc is from the mastercam system - .pu1 is from ncl and .txt is usually a hand coded program. we could standardize the extension but this gives us an idea who and what did the program
[01:55:48] <stustev> yes
[01:57:30] <stustev> in the previous programs the pivot length and tool lengths are in comments. It will be a simple matter of editing the programs to add a parameter line. most programs have less than 30 tools.
[01:58:51] <SWPadnos> the tool table changes should be automated, since I'd expect the tool lengths to change from time to time
[01:59:01] <SWPadnos> or do you preset them to a specific length?
[02:00:11] <stustev> for machines with 5 axis tool length compensation the only requirement is to get the tool long enough to work. for machines without 5 axis tool length compensation we set the tool to a specific length.
[02:00:44] <SWPadnos> ok, so the table values may (will?) change whenever a tool wears out
[02:01:09] <stustev> good night - must go watch fireworks - thanks
[02:01:23] <SWPadnos> sure, see you. enjoy the - uh - fireworks :)
[02:01:59] <stustev> yes the values change as the tool wears but if there is much wear then the tool will be changed
[02:02:31] <SWPadnos> right, so that has to be done with the tool table, and at least every time the program is loaded
[02:02:35] <SWPadnos> if not more often
[02:03:18] <SWPadnos> a "filter" could write a new tool table, and using gcodemodule, it could also force a tool table reload
[02:05:48] <jmkasunich> is scripts/latency-test usable from the command line, or is it a component of stepconf?
[02:07:46] <SWPadnos> it's usable outside stepconf
[03:13:32] <SkullWorks_PGA1> logger_emc: bookmark
[03:13:32] <SkullWorks_PGA1> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-07-05.txt
[08:22:38] <chr0n1c> so, i got the midibox circut board code down to 18 minutes from 183 minutes! AND figured out how to get 1 pass trace isolation out of pcb_gcode AND figure dout how to get .NC files to .DXF files so i can mastercam them ;)
[08:23:55] <chr0n1c> no more crazy wacked out all over the place cutting.. it'll all be in a neat orderly fashion
[10:00:50] <anonimasu> hm I just bought a surface grinder
[10:00:54] <anonimasu> :)
[10:01:32] <anonimasu> for 250eur
[10:26:05] <anonimasu> brb
[12:22:34] <jepler> chr0n1c: yeah, optimizing rapids by changing program order can make a big difference in milling circuit boards
[13:05:32] <Roguish> good morning all.
[13:07:45] <Roguish> jepler: recall my AXIS realtime delay problem? I think it's related to my jogwheel/classicladder hal file.
[13:08:10] <jepler> Roguish: that's possible; running too much stuff in a realtime thread with too tight a deadline will cause problems.
[13:09:08] <Roguish> please look at: http://pastebin.com/d1aaa6f38
[13:10:35] <Roguish> there are some lines here that i lifted from awallins, but i am not sure exactly what they are or what they do. in particular the '....tmax' items.
[13:11:13] <jepler> they mean nothing and can be removed from your hal file
[13:11:33] <Roguish> i've been rtfm'ing and have not found them in the docs.
[13:11:38] <Roguish> that's what i thought.
[13:11:42] <renesis> 08:22 < chr0n1c> so, i got the midibox circut board code down to 18 minutes from 183 minutes! AND figured out how to get 1 pass trace isolation out of
[13:11:42] <jepler> hal constantly records the time for a given realtime function to execute, and the maximum time. They are available in HAL parameters .time and .tmax.
[13:11:49] <renesis> why would you want 1 cut isolation
[13:11:53] <renesis> that sounds horrible
[13:12:12] <jepler> if someone uses "halcmd save param", these items will be written even though loading them into the system the next time doesn't do anything useful
[13:12:28] <renesis> also pcb gcode ignored Z feed parameters
[13:12:40] <renesis> unless youre using a new version and they listened to me bitching
[13:13:50] <Roguish> i also think i can change the thread timing on the jog wheel. move it to a different, slower thread.
[13:15:16] <jepler> Roguish: moving stuff out of base-thread (if you have stuff in it), getting rid of anything that's not needed, and increasing the times of all threads seem like steps you might try now
[13:15:50] <Roguish> i will give it a try. thanks.
[14:31:28] <stuste1> anybody awake this morning
[14:35:39] <cradek> hi stuart. barely.
[14:35:47] <Vq^> it's 16:35 here
[14:35:55] <SWPadnos> slightly
[14:36:08] <stuste1> maybe morning is not the correct term?
[14:36:27] <stuste1> I have a question about lead screw compensation.
[14:36:46] <stuste1> should I put the information in the file starting with the most negative position?
[14:37:16] <SWPadnos> probably
[14:37:31] <SWPadnos> lowest to highest positions makes sense to me
[14:37:50] <stuste1> one direction or the other makes sense to me
[14:37:52] <cradek> http://cvs.linuxcnc.org/cvs/emc2/configs/sim/screwcompX.dat?rev=1.1
[14:38:06] <stuste1> thanks - I will read it
[14:38:25] <stuste1> I had looked for something but was unable to find it
[14:38:28] <stuste1> bbl
[14:38:33] <SWPadnos> see ya
[14:38:36] <cradek> not sure about the docs
[14:38:50] <cradek> but I found a little example
[14:39:41] <stuste1> this looks as I expected
[14:40:05] <stuste1> the cinci repeats very very accurate
[14:41:56] <stuste1> I dialed in a 16 inch diameter lathe face plate using the XY axes. Then I adjusted AB home position to swing an indicator on the face of the plate.
[14:43:29] <stuste1> I can shut the machine down, restart, home the axes and the AB swing repeats within .0005 tir on the face of the plate. about a 15.5 diameter circle.
[14:43:42] <stuste1> this is much better than the MDSI was able to do.
[14:43:51] <cradek> wow
[14:44:04] <cradek> why is emc better at that? did mdsi not home to index right?
[14:44:05] <stuste1> yes - I am impressed
[14:44:20] <stuste1> I don't know.
[14:44:43] <SWPadnos> maybe there's a FERROR setting that was bigger
[14:44:47] <cradek> maybe the machine is not as sloppy as you thought - that would be nice.
[14:44:48] <stuste1> my guess is it didn't handle the index as accurate.
[14:44:57] <stuste1> that is probably the case.
[14:45:25] <stuste1> the recommended following error for all axes was .055 @ 100 ipm.
[14:45:38] <cradek> woo
[14:45:58] <stuste1> It was still able to cut parts 'relatively
[14:46:00] <stuste1> accurate
[14:46:27] <stuste1> we could cut +/- .01 parts with it
[14:46:54] <stuste1> for most 5 axis aircraft work this is sufficient
[14:47:26] <stuste1> the blend of tool motion was pretty good too
[14:47:47] <stuste1> This should allow the blend and accuracy to be better overall
[14:49:19] <stuste1> well - I am going out in the shop to play with the comp_file for the A axis.
[14:49:24] <cradek> can you measure the rotaries well enough to use comp?
[14:49:30] <SWPadnos> enjoy
[14:49:46] <SWPadnos> A has a limited range of movement, right?
[14:51:52] <stuste1> oh yea - I move the x and z along the tool axis and calculate the angle using ATAN. the W axis is good for that also. This will not be the final set up because I don't have X Y and Z comped with the laser yet. I will do that next week then do the A and B using the comped XYZ.
[14:53:13] <stuste1> forgot - I have a tool in the spindle with 10 inches of straight shank. this gives me a lot of space to measure the angle.
[14:53:20] <cradek> very clever. I can see where doing XYZ first is needed, and then you can build on that
[14:53:36] <stuste1> gets very accurate
[14:53:44] <stuste1> time to play
[14:53:52] <cradek> have fun
[15:43:23] <mung> anyone got time to answer some idiot newbie questions?
[15:43:41] <mung> Well I'll ask them anyway
[15:43:49] <Vq^> wise choice
[15:43:55] <cradek> that's the spirit
[15:43:55] <BigJohnT> that's the spirit
[15:43:58] <cradek> haha
[15:44:00] <Vq^> :)
[15:44:10] <BigJohnT> great minds think alike I see
[15:44:26] <Vq^> we lesser minds doesn't i see :)
[15:44:35] <cradek> I wonder if mediocre minds think alike too (not that we'd know)
[15:44:41] <BigJohnT> :)
[15:45:09] <BigJohnT> mung: you have used up one of your questions so far :)
[15:45:13] <rayh> Um from experience I'd say mediocre minds think alike.
[15:45:45] <mung> I've just got the ubunto live dist, and setup the steppers, and then run the emc2 from menu, and seen there is a demo file autoloaded, and then saved that to a gcode file...... now how do I run the gcode file to get the stepper to move???
[15:46:26] <mung> I could RTFM, but I am too lazy!
[15:46:36] <SWPadnos> in AXIS, you can go to the help menu, and show the AXIS command reference
[15:47:05] <cradek> mung: ouch
[15:47:59] <cradek> http://www.linuxcnc.org/docview/html//gui_axis.html#r1_2_1
[15:52:46] <mung> Unfortunatly I got no net access in workshop, and the machine by the lathe really sux for trying to read docs (tiny screen, and too up high). I get the feeling I am going to have trouble going back and forth between my browsing PC and the workshop to get info.
[15:53:15] <BigJohnT> print out the page and take it with you
[15:53:34] <jmkasunich> get a long spool of cat5 cable
[15:53:37] <mung> thanks for the link cradek, will check it out may be back later
[15:53:44] <jepler> the pdf docs are installed on your system
[15:53:48] <jepler> and contain the same information
[15:54:32] <mung> yeah, there are so many ways to do things...... all of them tooo much trouble
[15:54:33] <jepler> in this case, you want the EMC User Manual available in the Application menu
[15:55:10] <cradek> mung: it's very bad style to ask for help while saying it's too much trouble to figure anything out on your own.
[15:55:20] <mung> I'm just going to checkout cradeks links
[15:55:38] <mung> I am a master of bad style
[15:55:38] <cradek> many people here are experts - save the hard questions for them
[15:56:05] <mung> is there a quick start manual
[15:56:08] <mung> ??
[15:56:13] <cradek> just a little friendly advice.
[15:56:26] <cradek> that link I sent is very close to the quickest start possible.
[15:58:01] <mung> OK thanks cradek, I must stop typing and go and read the info. thanks to all, probably be back later....
[15:58:13] <cradek> hope you get it going, good luck
[17:29:25] <stuste1> the lead screw comp files are set up and working - I worked with the B axis to check the functionality - I will collect data with the laser I hope Tuesday
[17:30:43] <SWPadnos> oh right - A is the "nod" on the head
[17:30:55] <SWPadnos> so it is limited
[17:31:06] <stuste1> yes - both A and B are limited
[17:31:19] <SWPadnos> (I had wondered how you would apply correction to a rotary table-type device, which can run multiple revolutions)
[17:31:20] <stuste1> A move +30 -20 and B moves -30 +20
[17:31:55] <stuste1> for multiple revolutions you must make sure the comps equal zero for every revolution
[17:32:29] <SWPadnos> a single move that does 10 turns would run out of comp data though, unless the repeating is accounted for in the comp code
[17:32:47] <SWPadnos> those comps aren't cumulative, so they don't have to sum to 0 (though they should :) )
[17:32:47] <stuste1> on a Fanuc control the comps are incremental and additive
[17:33:40] <stuste1> why should the comps run out on 10 revs
[17:34:23] <stuste1> maybe that is a case for a rotary scale - 0 to 359.999 degrees
[17:35:14] <SWPadnos> yeah - it's the modulo axis thing - I don't know if it's implemented that way
[17:35:40] <stuste1> I will keep trying until someone implements it or I learn how to :) :)
[17:35:44] <SWPadnos> heh
[17:36:19] <stuste1> I won't put EMC2 on the viper until I have the rotary scale
[17:37:08] <SWPadnos> I'm pretty sure it won't work for a continuous motion rotary as is
[17:37:29] <stuste1> that's ok - for now
[17:37:36] <SWPadnos> comp is being done in HAL, so the HAL component doesn't know anything about what the position command actually means
[17:38:08] <stuste1> oh I thought you were talking about the rotary scale
[17:38:09] <SWPadnos> I think that's the way it is now anyway - I haven't looked (so don't take my word for it)
[17:38:13] <SWPadnos> eithert
[17:38:15] <SWPadnos> -t
[17:38:48] <SWPadnos> Jeff had done something for modulo axes, but he took it out - I guess there were some issues with it
[17:39:05] <stuste1> define modulo
[17:39:11] <anonimasu_> iab
[17:39:25] <SWPadnos> resets to zero every time it hits some number (like 360)
[17:39:43] <stuste1> vvvvveeeerrrryyyy gggggooooooooodddd
[17:39:47] <stuste1> I like that
[17:41:01] <rayh> I thought that an angular axis was wrapped linear. If so shouldn't the axis comp file continue across 360.
[17:41:26] <SWPadnos> hmmm. I may be wrong about comp (at the moment) - I don't see anything that looks like screw comp in the src/hal/components directory
[17:42:49] <rayh> I'm way out of date here. Comp used to be a file in the toplevel directory that got applied in user space.
[17:43:19] <SWPadnos> it may still be, but I recall some conversations about moving the information into HAL
[17:43:22] <stuste1> Hi Ray - is there still a 240 something limit to the comp table for each axis? If so it is not possible to comp a full revolution if you comp every degree so passing the 360 mark or going multiple revolutions is a moot point
[17:43:42] <SWPadnos> that's not a position limit, it's a "number of data points" limit
[17:44:10] <stuste1> yes I understand that but if you comp every degree you run out quick
[17:45:45] <anonimasu_> isnt position interpolated between comp points too?
[17:45:56] <stuste1> yes it is
[17:46:58] <SWPadnos> screw comp is in control.c, which is part of the realtime motion controller
[17:47:37] <rayh> I can imagine a worm that might have some repeatable error. I can also imagine a worm wheel that might.
[17:48:04] <rayh> I can't imagine a hardware error that would require fixing for each degree.
[17:48:48] <rayh> Back then the comp was applied as a linear adjustment between pairs of points.
[17:49:35] <stuste1> that is what I have used on the cinci. It may be overkill but I didn't want to go too far between points.
[17:49:50] <chr0n1c> renesis, 1 pass, not 1 cut...
[17:51:09] <renesis> that doesnt make a ton of sense to me
[17:51:24] <renesis> because pcb will usually have islands to cut
[17:53:22] <renesis> like, youre doing one isolation pass, or what
[17:53:29] <stuste1> on the cinci the motor has a small gear driving a large gear on the end of the worm shaft. The worm drive a ring gear with a smaller pinion gear on it. The pinion runs on a rack. The pinion is a split gear to allow adjustment with the rack. a lot of gears and I thought a lot of comp would be in order.
[17:53:50] <stuste1> comp points
[17:54:39] <stuste1> rayh; did you make it home ok with the load?
[17:54:56] <renesis> anyway, i use 60deg conicals with .005 tips, like .006" deep, .008 entered as diamater, and i do .02 isolation and .004 stepover
[17:55:33] <renesis> much less stepover, and itll sometime push copper to the side, into the path of the old cut, instead of cutting it
[17:57:20] <chr0n1c> ... i'm using 20 degree gutters, i can get a nice .003"- .005" line with them
[17:57:30] <renesis> much less than .015 isolation and its a bitch to solder. someone is going nuts right now over ghetto vias on the .008 space/trace QFP board i did for him, i cant imagine how much more difficult it would all be without the 5 passes of isolation cutting
[17:57:32] <chr0n1c> and they are carbide and VERY sharp ;)
[17:57:44] <renesis> all the more reason to use more passes
[17:57:58] <renesis> oh i said 60 i meant 30 =\
[17:58:28] <chr0n1c> well, i haven't vut it yet.. i may need more passes, but it looks very good on paper.. i had it running last night when i was sleeping with an ink pen
[17:58:36] <chr0n1c> cut it yet*
[17:58:51] <renesis> but i dont know what you mean by pass, if you mean youre only doing on isolation cut, then your shit will be like one copper chip from failure, always
[17:59:10] <renesis> also remember that you cut has a triangular profile, a V cut
[17:59:26] <chr0n1c> if i drop the bal point pen .005 it writes huge lines if i only drop it onto the paper .003 it writes awesome thin lines ;)
[17:59:34] <renesis> so at the bottom, your isolation is under .003
[18:00:19] <chr0n1c> good advice, i'm not trying argue with you ;) just explaining myself..
[18:00:30] <renesis> like, i seen high speed PCB CNC cut boards that are like on pass isolated, its like a waste of a fucking machine
[18:00:51] <renesis> yeah i dunno, maybe you machine and board is diff
[18:01:12] <chr0n1c> i did the one pass isolation in pcb_gcode so i could get a god dxf from the NC file and import it into mastercam and do my own toolpaths so i was like hmmm... lets try one pass
[18:01:23] <renesis> but ive thru ghetto board hell and back enough times, more isolation will make the board better
[18:01:23] <Sweeper> after you isolation cut, you use pliers to pull off the excess copper ;)
[18:01:35] <renesis> well then thats really alot of work
[18:01:39] <anonimasu_> there are theese stones for cleaning pcb's
[18:01:42] <renesis> because you can do 1 pass in pcb-gcode
[18:02:06] <chr0n1c> yes, stones... sandpaper a mill bastard fine tooth file...
[18:02:13] <anonimasu_> yep
[18:02:22] <renesis> you just set the initial isolation to the final isolation, or whatever those parameters are called
[18:02:22] <anonimasu_> that should clean out the chips
[18:02:40] <renesis> im pretty sure that works, and stepover is basically ignored
[18:03:03] <renesis> yeah see when i isolation cut, there is no excess copper
[18:03:28] <renesis> if you do over 50% overlap, the copper likes to be pushed to the side, instead of being chopped up
[18:03:59] <renesis> depending on feed and cutter edge, but like with 50% or less, those dont matter very much
[18:04:19] <chr0n1c> these cutters actually chop up the copper i was impressed
[18:04:29] <renesis> well of course, single pass
[18:04:39] <chr0n1c> there is fine copper dust (not good i know) all over the machine now from the last board
[18:04:49] <renesis> what im talking about happens when you go in to cut passes after the first
[18:05:06] <renesis> haha my machine prob has FR4 and copper in every crack
[18:05:24] <renesis> working fine, i gotta break it down and clean/retune it, tho
[18:05:39] <renesis> been a year and a half or so since i put it together
[18:06:20] <renesis> because yeah, maybe that cool in a machinist sense, .005 gaps
[18:06:57] <renesis> but to actually work with, like assembly and reliability, that shit on an unmasked board can be hellish
[18:08:01] <renesis> i know someone that actually uses fill cutting for all his boards, ive used it on a few, theyre much easier to work with, cost is forever to machine, tho
[18:08:18] <renesis> heh, pcb-gcode docs says they dont actually think anyone uses the fill cut files
[18:08:35] <anonimasu_> lol
[18:08:44] <renesis> 180 minutes isnt long
[18:09:13] <renesis> the time youre savging in art project bullshit and chemical safety and failure at aligning the toner transfers or whatever
[18:09:28] <anonimasu_> 180 minutes is forever
[18:09:34] <renesis> then electrically testing the boards after an etch process, because jezus fuck that fab tech needs it
[18:09:41] <renesis> yeah youre not really taking much more time
[18:10:01] <renesis> whatever, if youre doing electronic project, you prob have some other shit to do or research while its cutting
[18:10:07] <chr0n1c> lol... 180 isn't long... i wrote a 7 hour g-code program by hand once ;) (i got paid for it)
[18:10:16] <anonimasu_> I'd bet you chemical and uv is faster.
[18:10:23] <renesis> jezus fuck i hope you used subroutines
[18:10:35] <renesis> no youre not understanding what im saying
[18:10:46] <renesis> the process itself might be faster
[18:10:56] <renesis> but setup and testing make it take about the same time
[18:11:02] <renesis> and you actually doing shit most of the time
[18:11:09] <anonimasu_> toner transfer dosent really work well except in special cases either
[18:11:16] <renesis> 180min cnc cycle is 180min to handle more project work
[18:11:20] <renesis> or have lunch
[18:11:22] <renesis> or something
[18:11:29] <chr0n1c> i'll take slower over not having to breathe chemical fumes that don't grow on trees
[18:11:41] <renesis> i only have to electrically test do to ghetto vias
[18:11:49] <renesis> its not like etch shit where its a learning experience
[18:12:07] <renesis> when i do boards, theyre done, what i end up with is what i did in the pcb editor
[18:12:21] <anonimasu_> uh.. it's not really a learning experience with uv..
[18:12:30] <renesis> etching after CNC drill instead of just machining traces was a fucked up waste of time
[18:12:44] <anonimasu_> yeah
[18:12:46] <renesis> and considering lead time from prototype fab houses, 180min is not so long
[18:13:02] <anonimasu_> I'd rather have the cnc do some cutting of steel..
[18:13:03] <anonimasu_> :P
[18:13:07] <anonimasu_> and have lunch meanwhile
[18:13:16] <anonimasu_> though pcb probably pays better.
[18:13:26] <renesis> yeah i dont see how thats relevant
[18:13:37] <anonimasu_> whatever that works
[18:13:50] <chr0n1c> renesis, http://imagebin.org/21998
[18:13:51] <renesis> yeah thats what im saying, this shit when i do it always works
[18:13:55] <chr0n1c> there is the traces in axis
[18:14:37] <renesis> okay yeah, thats single isolation pass, with your cutters, i would not want to be the guy assembling that board\
[18:15:02] <chr0n1c> .. it'll be me
[18:15:08] <renesis> i do not want to be you\
[18:15:35] <chr0n1c> i hear the advice on soldering with barely any isolation... i may add another pass before i cut the actual board
[18:15:52] <renesis> not even just soldering
[18:16:24] <renesis> copper chip, fragment of pin after you clip, some random piece of domestic debris
[18:17:13] <renesis> anything can bridge .005, doing .008 shit is bad enough, im pretty much at the point wher id call that doable but impractical
[18:17:25] <Sweeper> * Sweeper sticks a big sheet of styro on his wall to hold salvaged chips and other assorted through-hole components
[18:17:30] <renesis> really, with the cost of pcb, and the cost of time,...
[18:18:07] <renesis> the time it takes to assemble a double sided board without thru hole plating isnt worth it
[18:18:19] <anonimasu_> styro = static
[18:18:20] <anonimasu_> -_-
[18:18:37] <renesis> you can get anti-static foam
[18:18:42] <Sweeper> not here! :P
[18:18:46] <renesis> like they ship samples in, black or pink, feer
[18:18:54] <renesis> just order shit from maxim =)
[18:19:00] <anonimasu_> oh that's other foam..
[18:19:00] <anonimasu_> :P
[18:19:05] <Sweeper> besides, these are salvaged.
[18:19:06] <renesis> maxim = jewelry boxes and black foam, i think
[18:19:16] <Sweeper> if they're sensitive, they're already dead
[18:19:24] <renesis> not true
[18:19:43] <renesis> static death doesnt happen nearly as often as people make out
[18:20:10] <renesis> ive had an nvidia card running in my box for a couple years after it had spent month kicked around a carpet
[18:20:26] <Sweeper> then I don't risk much by storing chips on a styro sheet :D
[18:20:40] <renesis> and i pic microcontroller chips out of my carpet and into a devboard with no issues pretty often
[18:20:50] <renesis> well, doing that is asking for it
[18:21:25] <anonimasu_> yeah but it does happen
[18:21:26] <Sweeper> :)
[18:21:43] <anonimasu_> though with $1 components it's ok..
[18:21:50] <anonimasu_> it'd be work with a 250$ fpga.
[18:21:52] <anonimasu_> worse..
[18:22:12] <Sweeper> if I had antistatic foam, I'd use it
[18:22:20] <Sweeper> I asked around at shops here, nobody has it
[18:22:30] <anonimasu_> well, use styrofoam then :)
[18:22:49] <Sweeper> stuff I buy new sits in its anti-static foam
[18:22:52] <Sweeper> er
[18:22:54] <Sweeper> tube
[18:22:55] <anonimasu_> I think it's worse handling the foam by hand then touching the chips..
[18:23:04] <anonimasu_> not the chips building static where they sit..
[18:23:11] <anonimasu_> (just what I think it might be wrong)
[18:23:22] <Sweeper> yea...rubbing the foam and then grabbing the chip = not so good
[18:24:38] <anonimasu_> :)
[18:26:06] <Sweeper> it's a very nice visual inventory tho
[18:26:44] <Sweeper> if the silkscreening on all the chips was white, I could read the numbers from where I sit
[18:32:04] <chr0n1c> i gotta do take some mustang dash inserts to get silkscreened monday
[18:32:10] <chr0n1c> go take*
[18:36:37] <chr0n1c> oh i lied they are camaro dash parts... the mustang ones are already done
[18:39:58] <anonimasu_> :)
[18:41:39] <dmess> what yr mustang??
[18:42:27] <chr0n1c> we do parts for 65-68 (dash kits, door panels, stainless rocker moldings sometime soon)
[18:42:56] <chr0n1c> and camaros we do 71-83
[18:43:35] <chr0n1c> oh we do upper and lower console restoration kits as well for the 65-68 mustangs
[18:44:00] <chr0n1c> and stainless vent slides...
[18:44:11] <chr0n1c> for the fastbacks..
[18:46:05] <chr0n1c> Dallur, when can i get some of those throwing stars :)
[18:46:39] <chr0n1c> i just saw your nick on imagebin and was being nosey ;)
[18:50:46] <dmess> nice i had a '69 notchback in the 80's that had a perfect interior... sold it to a guy who butchered it into a fast back.. it looked like shit....
[18:51:29] <dmess> sold it the morning after my STAG.. i dont think it was a fair sale..
[18:51:58] <dmess> i was drunk for 16 hrs after that
[18:52:25] <chr0n1c> i never had a mustang... i did have an 84 monte ss though ;)
[18:53:03] <chr0n1c> which.. i cried after i had to sell it
[19:03:56] <chr0n1c> http://imagebin.org/21985 <- offtopic, but F*(&ing funny stuff
[19:07:29] <alex_joni> chr0n1c: lol
[19:11:58] <dmess> dont laugh... thats my MOM's dog... LOL
[19:17:26] <dmess> my 'stang had a 255 cid hi-po with 2.75" single barrel carb....i had it up to 145 mph 1 time..... was supposed to blow up... but it didnt so i decided to keep her...
[19:18:05] <dmess> run her wide open thottle for 11 miles....
[19:22:41] <dmess> had a '69 Boss 302 frame and brakes/suspension with a 6 cylinder in the front...
[19:23:52] <dmess> and no it wasnt the CS edition... ( california special... but had the same engine but not the same carb.
[19:28:09] <chr0n1c> google translate doesn't do latin?
[19:43:04] <alex_joni> good night all
[19:43:21] <chr0n1c> peace out yo!
[20:49:53] <chr0n1c> gtg, later y'all
[20:55:37] <cradek> I would suggest for most cuts a depth of cut around 8% of cutter dia, and for a rotation of the feed handle, just how long it takes to
[20:55:41] <cradek> sing the phrase "Happy birthday to you".
[20:55:53] <cradek> (manual machinists are funny)
[20:57:23] <jmkasunich> heh
[21:21:47] <SkullWorks_PGA1> Sen. John Warner (RINO - VA) suggests lowering national speed limit to conserve fuel
[21:22:30] <SkullWorks_PGA1> this was tried before - it failed then and will fail worse now
[21:24:28] <dmess> i havto breat 110 kph to beak wind resistance on my van
[21:25:37] <dmess> shatt the dilexic in me
[21:26:59] <dmess> cradek... she is sooo cute
[21:27:26] <dmess> we have
[23:02:06] <mung> anyone know how to repair parallel ports???
[23:02:18] <SWPadnos> get a PCI parallel port card
[23:02:46] <SWPadnos> or ISA, depending on the age of the PC
[23:02:59] <mung> recommendations, with opto isolation?
[23:03:37] <SWPadnos> I don't know of any parallel ports with isolation, the parport was never meant to interface to anything "unknown"
[23:03:49] <SWPadnos> there are breakout boards that will provide isolation though
[23:06:10] <micges> hi all
[23:06:20] <SWPadnos> hi there
[23:07:15] <mung> hi
[23:08:16] <mung> what is the recomended way to interface more than 4 axes, as parallel only has 8 outs?
[23:09:09] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
[23:10:48] <SWPadnos> personally, I like the Mesa FPGA cards, because they're quite capable and they're completely reprogrammable
[23:11:11] <SWPadnos> there are other options that may be easier to set up for a mill or lathe
[23:13:47] <mung> I am afraid I'm doing a hobby diy conversionon a really low budget, so want something less than £20
[23:15:10] <SWPadnos> oh :)
[23:15:21] <SWPadnos> isolation will be difficult at that price point
[23:15:28] <mung> I have been trying to save cash so did not bother with opto isolation and realised the error when i dropped a screw driver and shorted something
[23:15:35] <SWPadnos> unless you'd like to build something yourself
[23:15:40] <SWPadnos> oops
[23:16:23] <mung> I still have 4 outputs but 4 blew
[23:17:15] <SWPadnos> a PCI parallel card is probably the cheapest solution - dual cards are within your budget, and if you're careful you won't have to replace it
[23:17:21] <SWPadnos> even if you do, it's still cheap
[23:17:23] <mung> I may just go with a standard card and be more carful next time ;)
[23:17:27] <SWPadnos> yep
[23:17:41] <SWPadnos> you can use up to 8 parports with EMC (more if you want to recompile)
[23:18:38] <mung> its screwed my weekend as I was hoping to get three axes installed and tested :(
[23:19:00] <SWPadnos> well, if you have hardware e-stop, you may still be able to do some testing
[23:19:14] <SWPadnos> the parallel port can have 12 outputs and 5 inputs, I think
[23:19:45] <SWPadnos> the control and status ports are also usable, so you could get 8 outputs (barely), and still have a couple of
[23:19:47] <SWPadnos> inputs
[23:20:08] <mung> I looked at the epp specs and it shows only 8 data lines
[23:20:17] <SWPadnos> 8 fata, 5 status, 4 control
[23:20:19] <SWPadnos> data
[23:20:40] <mung> hows does the hardware e-stop work?
[23:20:51] <SWPadnos> externally
[23:21:26] <SWPadnos> is, you wire relays and whatnot to shut down the machine. you can optionally rell EMC that the machine is off, and optionally allow EMC to shut the machine off
[23:21:30] <SWPadnos> s/is/ie/
[23:21:51] <SWPadnos> pinouts here: http://www.linuxcnc.org/docview/html//hal_drivers.html#sec:Parport
[23:21:59] <SWPadnos> (for the parallel port under EMC2)
[23:23:00] <mung> just use emc hardware wizard to setup output to control pins?
[23:23:07] <SWPadnos> yep
[23:23:15] <SWPadnos> if you're talking about stepconf
[23:23:27] <mung> yes stepconf
[23:24:13] <mung> I really am newbie only messed around with the emc live distro for a couple of hours so far.
[23:24:28] <SWPadnos> have you installed it to a hard disk?
[23:24:42] <mung> no running live cd
[23:25:12] <SWPadnos> ok, in that case any changes you make will be lost the next time you reboot, unless you're specifically saving them to a flash drive or something
[23:25:50] <mung> its just a hacked together box of bits at the moment, as i am learning and botching together a system.
[23:26:39] <mung> I realise all data is lost but i don't have a spare harddisk at present.
[23:26:53] <SWPadnos> no problem - just thought you'd want to know :)
[23:29:11] <mung> I am documenting the project at cnczone, http://www.cnczone.com/forums/showthread.php?t=60541
[23:31:28] <mung> Thanks for info, its late here so I'm go to get some sleep, will probably return tomorrow for more help.
[23:31:37] <SWPadnos> night
[23:32:12] <SWPadnos> where in the UK are you?
[23:32:57] <mung> I'm from the south, IOW (Isle of Wight)
[23:33:12] <SWPadnos> ok. never been there
[23:33:18] <mung> where you at?
[23:33:48] <SWPadnos> I'm in the US, in Vermont, but I've been to the UK a number of times
[23:34:01] <mung> your a emc devel?
[23:34:07] <SWPadnos> sometimes :)
[23:34:56] <mung> yeah, if I get interested I may contribute, but I got soo many projects and soo little time
[23:35:52] <SWPadnos> I know the problem
[23:36:36] <SWPadnos> ok - I guess the closest I've been to IOW is Amesbury or Bath
[23:36:46] <mung> I wrote some CAD for aerofoil analysis for linux about 10 years ago, with rendering and dxf import etc... but have not done much since then
[23:36:53] <SWPadnos> cool stuff
[23:37:59] <mung> anyways, as I said its late. I really must get some sleep. Cheers for info
[23:38:07] <SWPadnos> sure. good nigh
[23:38:09] <SWPadnos> t
[23:38:22] <mung> nigh
[23:38:24] <mung> t