#emc | Logs for 2008-08-27

[00:03:04] <antichip> ok I pulled off the trigger and finally saw a .5 v across the out put and ground
[00:17:04] <SWPadnos__> SWPadnos__ is now known as SWPadnos
[00:26:36] <jmkasunich> owad: did you figure out your MDF cutting issues>
[00:26:37] <jmkasunich> ?
[00:27:18] <owad> yes, I asked Patrick, who designed the system, and he said I should be doing at least 50 ipm for MDF, whereas I was doing 16.
[00:27:35] <jmkasunich> yep - 16 in/min at 20K RPM means only 0.0004 per tooth (2 tooth cutter)
[00:28:00] <jmkasunich> with that slow of a feed, you are rubbing instead of cutting = heat
[00:28:18] <owad> I'm worried about these cheap 1/8" RotoZip cutters holding up, though.
[00:28:58] <jmkasunich> 20K RPM on a 1/8 cutter is 650 SFPM
[00:28:58] <owad> oh, I took a picture: http://tinyurl.com/67qx7w
[00:29:16] <jmkasunich> dunno much about MDF, but that seems fast for steel - should be fine for carbide
[00:29:31] <jmkasunich> lol
[00:29:35] <jmkasunich> a bit hot
[00:30:17] <jmkasunich> can you change the spindle speed?
[00:30:19] <owad> think that endmill can be salvaged?
[00:30:37] <jmkasunich> not really
[00:31:04] <jmkasunich> almost anything can be salvaged if you want to spend enough time at it
[00:31:16] <owad> any source you'd recommend for long 1/8" endmills?
[00:31:17] <jmkasunich> but you said "cheap" Rotozip cutters - start with a fresh one
[00:31:32] <owad> well, that's not one of the cheap rotozips, in that picture... :(
[00:31:36] <jmkasunich> oh
[00:31:49] <jmkasunich> you didn't answer my question - can you vary the spindle speed?
[00:31:54] <owad> yes
[00:32:00] <jmkasunich> slow it down then
[00:32:19] <jmkasunich> set the speed at say 10K, set the feed for 0.002 per tooth, per rev
[00:32:19] <owad> 10,000 - 23,000 RPM
[00:32:35] <jmkasunich> 10K RPM x 2 teeth = 20,000 teeth/min * 0.002 = 40 ipm
[00:32:38] <jmkasunich> see how that cuts
[00:32:48] <owad> I'll do that. Thanks!
[00:33:13] <owad> tonight I'm going to get the vacuum hooked up. the dust is really filling up the workshop
[00:33:17] <owad> and the smoke. :D
[00:33:23] <jmkasunich> if/when you increase the speed, make sure you increase the feed to match, keep feed per tooth constant
[00:33:56] <stustev> I think you will like a faster feed - the MDF is very abrasive
[00:33:58] <jmkasunich> is the tooling HSS or carbide?
[00:34:12] <jepler> hm, one of these 4096 permutations of 12 pins has to be the right combination to make my 3 resistor DACs work right...
[00:34:14] <jmkasunich> stustev: 0.002 is 5x faster than he started at ;-)
[00:34:29] <stustev> much better
[00:34:31] <owad> it doesn't say, but it certainly looks like steel
[00:34:41] <owad> no carbide
[00:35:09] <jmkasunich> then you probably don't want to take the speed too high - focus on increasing the feed per tooth, until you either run out of machine speed, or things start to get rough (vibration, cut quality)
[00:35:23] <stustev> jepler: just 4091 attempts to go
[00:35:32] <jmkasunich> you want chips, not dust
[00:35:44] <owad> I had one snap, earlier today
[00:35:45] <stustev> chips and NO smoke
[00:35:48] <jmkasunich> also, don't try to go too deep in one pass
[00:35:59] <owad> I've been doing .25" per pass
[00:36:00] <jmkasunich> 1/8" cutter = 1/8" depth per pass
[00:36:15] <owad> ok
[00:36:29] <jmkasunich> two reasons not to go too deep - cutter strength (snap) and chip clearing
[00:36:38] <jepler> my project of the evening is 640x480x4096 colors from an FPGA: http://axis.unpy.net/files/01219715210/img_7499-medium.jpg -- the test pattern is static from an LFSR plus what should have been a grid of color swatches..
[00:36:55] <jepler> (well, it's a grid -- just not in the order I anticipated :-P)
[00:37:04] <stustev> If you create heat you melt the resin onto the tool - then chips (dust) cannot get to the cutting edge or escape
[00:37:08] <owad> say, to you use gnccam, by chance?
[00:37:16] <jmkasunich> not me
[00:37:21] <jmkasunich> I use humancam
[00:37:33] <stustev> not me - NCL
[00:37:53] <jepler> aha, got it right in much fewer than 4096 tries :)
[00:38:13] <jmkasunich> I think the average for random would be N/2.....
[00:38:14] <owad> what's NCL like?
[00:38:25] <stustev> APT like
[00:38:42] <owad> it can't convert dxf?
[00:38:51] <stustev> no
[00:41:51] <skunkworks> jepler: neatg
[00:41:54] <skunkworks> neat
[00:44:24] <stustev> where would I look to change the action of the g43 code for 5 axis work?
[00:45:04] <SWPadnos> uh oh
[00:45:20] <stustev> I would like the 5 axis tool length comp that acts upon the W axis - what do you mean uh oh?
[00:45:26] <SWPadnos> heh
[00:45:34] <jmkasunich> I think cradek already wrote that
[00:45:38] <jmkasunich> its in a branch somewhere
[00:45:47] <stustev> let me repeat and expand on my question
[00:45:50] <SWPadnos> two things: 1) what the heck are you doing and 2) there are a lot of places to look for that answer :)
[00:46:06] <stustev> if you will be patient I will explain
[00:46:11] <SWPadnos> sure
[00:47:01] <stustev> I would like the 5 axis tool length comp that acts upon the W axis to also change the Z by the same amount when the g43 Hxx is used
[00:47:22] <stustev> and also the reverse when g43 h00 is used
[00:47:38] <stustev> no motion would result
[00:47:50] <SWPadnos> is that "always" or only when W is aligned with Z?
[00:47:57] <jepler> so you're aware of the inifile directive [TRAJ]TOOL_LENGTH_IS_ALONG_W but it doesn't do what you want?
[00:48:00] <SWPadnos> no effective motion, for instance on a knee mill
[00:48:31] <stustev> jepler: YES it works great
[00:48:49] <stustev> no effective motion as on my cinci
[00:49:29] <SWPadnos> hmmm - I'd have to look at the mechanics again - I don't remember which kind of motion that one was (too many machines, not enough time :) )
[00:49:44] <jepler> what happens if the tool is not exactly vertical when you turn on/off g43?
[00:49:57] <stustev> I haven't tried that
[00:50:06] <jepler> on the old control, that is
[00:50:47] <stustev> if it accepts it on the new control I would want it to adjust all axes involved.
[00:51:12] <stustev> let me explain what I have to do on the cinci
[00:51:24] <antichip> ok I was excited about having a .5 v on the opto switch, well I do get a full 5v when I shine a light at it? the exact as another opto I tried
[00:53:18] <stustev> with 9 inch tool length in h05 - I use g43 h05 - it changes the W axis by 9 inches - to move from there in motion I have to g91 w9 z-9 - resulting in no motion but the A and B axes then pivot around the tool tip
[00:53:40] <stustev> the cinci has on 14 inches of Z travel
[00:54:14] <stustev> I cannot move the W axis by the full nine inches and then move the Z axis back by the full nine inches
[00:54:57] <stustev> the cinci has ONLY 14 inches of Z travel
[00:55:25] <jepler> ok, well to answer the question you first asked -- where is the code concerning turning g43 on -- it's in the function called Interp::convert_tool_length_offset in src/emc/rs274/interp_convert.cc
[00:56:02] <jmkasunich> ONLY 14", eh?
[00:56:05] <jmkasunich> I'll trade you
[00:56:54] <stustev> well with 120 inches X and 42 Y the 14 seems rather short - not very balanced - they call these machines profilers
[00:57:03] <jepler> to answer the question you didn't ask, emc never produces motion directly as a result of G43/G43.1/G49. Instead, the offset along a given axis is applied in the next move that includes that axis letter. So I think in emc you *can* g43 h05 / g91 w9 z-9
[00:57:17] <jmkasunich> yeah, I was surprized to hear only 14 on such a big machine
[00:57:23] <jepler> the second line will actually "do" the move, and assuming that the tool is exactly vertical it won't produce any motion
[00:57:28] <jmkasunich> but I've only got 3.3" of Z here, and it drives me nuts
[00:57:56] <stustev> Jepler: that is exactly what I have been doing
[00:58:49] <jmkasunich> stustev: then what is the problem? doesn't that work for you?
[00:58:52] <jepler> ok i wasn't 100% sure where you were describing the pre-emc control and where you were describing emc's present behavior ..
[00:59:22] <stustev> the W display changes when the g43 is read - the motion is implemented on the next move
[00:59:40] <jepler> anyway, I *think* that if you change the code near USE_TOOL_LENGTH_OFFSET in that function I name, you can get different behavior, including a move when you G43/G43.1/G49
[01:00:14] <jepler> , including a "move" that ends up with the motors in the same positions but the DRO showing different values on Z and W
[01:00:21] <stustev> jmk: yes I works - in a program - no problem - MDI - much more typing
[01:01:05] <SWPadnos> I wonder if it might be "better" to implement something like the lathe tool tables, but use the two offsets in Z and W
[01:01:37] <stustev> jepler: that move is what I am looking for - no motion at all but the axis positions are changed
[01:01:42] <SWPadnos> (or more generically, have the ability to offset a tool in any axis, maybe with column headers or something)
[01:03:17] <stustev> tomorrow I will try g43 with the AB axes not at 0
[01:05:26] <stustev> right now I have the cinci moving around around a 1 1/2 inch ball on the end of a tool. I get .006 movement on an indicator mounted along the Y axis when I move from A20 to A-20. that is +-.003 - not bad
[01:06:11] <jmkasunich> that represents an error in the kins (pivot length or something)?
[01:07:03] <stustev> I believe this is geometry error. The pivot point of the A axis does not coincide perfectly with the centerline of the spindle
[01:07:30] <stustev> I have a compensation routine in my kins
[01:08:18] <stustev> I can move the indicated amount by changing a parameter in my hal file
[01:08:56] <stustev> I have on parameter for the A axis and two parameters for the B axis
[01:09:02] <stustev> on/one
[01:09:11] <jmkasunich> neat
[01:09:41] <jmkasunich> would that work for machines with a large (intentional) offset between spindle axis and a rotary axis?
[01:09:48] <stustev> yes
[01:10:22] <jmkasunich> dunno if you remember but we were talking about my someday gantry mill at the workshop, and I was considering offsetting the spindle to reduce the amount of Z travel needed to compensate for tilt
[01:10:51] <stustev> yes I remember
[01:11:31] <jmkasunich> as I started to work the details of that head tilt bearings it got less attractive to do that, but its not off the table yet
[01:12:05] <stustev> this is the same kind of compensation as Max has - the rotary table can electronically adjust the XY axis to spin around any arbitrary point on the rotary table face
[01:12:08] <jmkasunich> the cinci ring gear/arc way doesn't crowd the table because it doesn't tilt that far - I want to tilt 90 degrees one way
[01:12:35] <stustev> yes - the cinci has 50 degrees total on A and B
[01:27:19] <stustev> here I am again - reading the code in interp_convert.cc I see the tool length cannot be negative - I need to use negative tool lengths :)
[01:30:37] <stustev> I have a lot of programs for the cinci that have tool lengths and pivot lengths calculated into the g code. If I can use negative tool lengths I can use these programs by subtracting the programmed length (pivot length and tool length) from the actual length (pivot length and tool length) and putting the difference in the tool table - this will usually be a negative number
[01:30:39] <SWPadnos> I don't see that limitation, though I haven't looked at the definition of USE_TOOL_LENGTH_OFFSET
[01:31:29] <stustev> I was just looking and the comments alluded to that - maybe the code will allow it though - I will look farther
[01:32:15] <SWPadnos> are you looking at this line: "The H number in the block (if present) was checked for being a non-negative integer when it was read, so that check does not need to be repeated."
[01:32:45] <stustev> my mistake - it was referring to the H number not the length - yes - it just read it again
[01:32:49] <SWPadnos> right
[01:33:34] <jmkasunich> FWIW - I've used negative tool lengths on my machine (not W offset tho)
[01:34:38] <stustev> I think (heaven forbid) I was wrong :)
[01:34:53] <stustev> :(
[01:35:02] <SWPadnos> heh
[01:35:11] <SWPadnos> happens to the best of us
[01:35:15] <SWPadnos> (or so I'm told :) )
[01:35:20] <stustev> heh
[01:36:02] <jmkasunich> "things happen" - thats why I use negative offsets
[01:36:17] <jmkasunich> instead of using the spindle nose as my gage line for tool length, I use a point 4" out
[01:36:35] <jmkasunich> that way if I ever do a move with length comp turned off, I don't ram the entire tool and holder into the work
[01:36:45] <stustev> good idea - the tools crash less frequently that way
[01:40:28] <stustev> I will try some negative offsets tomorrow
[01:42:26] <stustev> bbl
[01:43:37] <SWPadnos> see you
[02:21:05] <cradek> I also like to consider the longest tool in a setup as the zero length tool, and then have everything else negative
[02:21:35] <cradek> I don't measure tools offline, but if I did, I'd pick a nice number like jmk says
[02:29:37] <CanSir> Is there a way (gcode or halui) to reset the absolute X,Y,Z position? Currently I exit EMC and then start it back up.
[02:32:52] <cradek> homing is how you set the machine coordinate system
[02:33:28] <CanSir> I don't have home or limit switches installed though.
[02:36:24] <jepler> halui.joint.#.home will activate the homing procedure for the given axis
[02:37:33] <jepler> when you configure emc without home switches, the "homing procedure" consists of equating the current motor position with the axis position given in the inifile (HOME_OFFSET) and then moving to another axis position given in the infile (HOME)
[02:40:50] <CanSir> I don't see how that helps. I start up with my machine in an unknown x,y,z so now I manually jog to the lower left corner, exit EMC and then start it back up so it knows where x=0, y=0, z=0.
[02:41:30] <CanSir> Am I going about it the wrong way?
[02:42:51] <jepler> then "homing" without limit switches is exactly what you want. You jog to the reference location, then hit "home". If [AXIS_0]HOME=0 HOME_OFFSET=0 HOME_SEARCH_VEL=0 then that will equate the current position of that motor with the machine coordinate X=0
[02:43:29] <CanSir> Ahhh. I sometimes miss the obvious. :)
[02:43:50] <jepler> sometimes the obvious is hard to state
[02:44:04] <jepler> (if you used stepconf, simply turning off the home switch inputs on the port configuration screen should give you these lines in your inifile)
[02:45:24] <CanSir> I've done too much mucking around to run stepconf on my real config. If I run into trouble though, I'll create a new 'fake' one to compare it to.
[02:45:39] <CanSir> Thanks jepler!
[02:46:04] <jepler> OK good luck
[02:46:11] <jepler> I think those items I named are the important ones for this
[02:47:31] <CanSir> In looking over the documentation, the only other one that looks important for this application would be "Home Sequence" or "Home use Index"
[02:54:55] <stustev> cradek: how are you this evening?
[03:01:19] <stustev> CanSir: if you are homing in the manner jepler described then "Home Sequence" will do nothing for you - you determine the sequence as you home each axis. "Home use index" is for when you have encoders with an index pulse.
[03:24:08] <cradek> stustev: pretty decent. not much work done on the lathe today though. maybe tomorrow.
[03:24:56] <stustev> I hadn't heard anything this evening about the lathe - a break is necessary sometimes
[03:25:10] <stustev> what do you have to do on it?
[03:25:51] <stustev> did you see my question(request) about the tool length functionality?
[03:26:17] <cradek> the two big remaining tasks are to set up control for the spindle vfd, get spindle position feedback somehow
[03:27:43] <cradek> yes I think I understand what you are asking. The only problem I see is this. If there is an ABC offset a move like g91 w9 z-9 will cause motion
[03:27:51] <cradek> because w,z are in different directions
[03:29:28] <stustev> yes - this will only work with ABC zero - if it is implemented as I would like to see the axis positions will be updated without any axis motion - the dro positions will be the only thing to change
[03:30:33] <stustev> I don't know how much work this is. I looked at the file. I don't see a way to do it. This is way beyond me (at this time).
[03:31:55] <stustev> The machine is working well. The control is working well. This is not a gripe or a big need. It is something I saw while running it today. It would be a time (typing) saver when in MDI. A program can handle this without a problem.
[03:32:47] <stustev> I can hardly wait to get the motion adjusted and show a video of the 5 axis motion spinning around a point with indicators in three directions.
[03:32:56] <cradek> there is provision for the interpreter to resync the position. I think you would need to cause it to do this - first at the canon level, shift the coordinates in the way that causes no joint motion [layer violation - it has to know about the machine kins] and then have the interp reread the axis positions
[03:33:59] <cradek> sounds neat - I look forward to seeing that video too!
[03:34:03] <stustev> i will look at that
[03:35:12] <stustev> bbl - my wife just called
[03:35:16] <cradek> if there is a rotary not at zero, be careful to abort with an error so you do not end up causing a joint position jump! very important to handle that case. it could be bad.
[03:47:13] <SWPadnos__> SWPadnos__ is now known as SWPadnos
[03:57:58] <stustev> I want to update the XYZ positions if there is a rotary head not at zero. This is certainly possible. I will be able to give it some time later (after the cinci is running production).
[03:58:36] <SWPadnos> can you change tools with ABC != 0?
[03:58:59] <stustev> yes - I can change tools in any position
[03:59:09] <SWPadnos> bummer :)
[03:59:44] <SWPadnos> you could make the toolchange position have A=B=C=0, so it will always orient correctly before you'd be likely to change the TLO
[03:59:48] <stustev> why is that a bummer? the only thing I have to watch out for is if I have room to remove the tool from the spindle
[04:01:02] <stustev> I think that will be the usual way we do it. There is no real reason to change tools at anything other than AB zero.
[04:02:09] <stustev> We may want to change the offset at other than AB zero positions for blending cuts (mismatch). We use two or three offsets per tool many times.
[04:02:55] <SWPadnos> hmmm - ok then, my solution won't work
[04:05:00] <owad> jmkasunich: 40 ipm looks great when I use a good endmill. The rotozip, though, drags behind. My good 1/8" endmill only has a 1/4" cutting depth, though, and a 1/4" shank. Is it possible to affordably get 1/8" endmills that cut cut 1" deep, or will they all drag at that length?
[04:07:01] <stustev> I am thinking that a 1 inch DOC with a .125 end mill is hoping for a little too much.
[04:07:20] <toastydeath> i concur with the good doctor.
[04:08:11] <toastydeath> a 1/8" endmill with a 1" shank is going to be a) expensive, and b) for delicate tool and die applications with very little metal removal.
[04:09:04] <owad> even if I'm working with wood, not metal?
[04:09:45] <toastydeath> i'd be shocked if you got 1" on 1/8" even in wood
[04:09:49] <toastydeath> i think you might do it in cardboard
[04:09:53] <owad> heh
[04:09:55] <toastydeath> foam, definately
[04:10:05] <stustev> yes - wood is very abrasive - the edge dulls quickly - then the tool pressure builds - then the tool breaks
[04:10:14] <toastydeath> if you needed to seriously do that depth
[04:10:20] <toastydeath> i'd do very small steps
[04:10:26] <stustev> cardboard will be much like MDF
[04:10:27] <toastydeath> .075 - .100
[04:10:37] <toastydeath> or maybe less, depending on how it ran
[04:10:57] <stustev> foam will cut easily as will carbon (for edm)
[04:10:57] <toastydeath> write up the toolpath with a ramp down, and just run it incrementally
[04:11:06] <owad> with the rotozip bit?
[04:11:18] <toastydeath> which are you replying to, owad
[04:11:39] <owad> to the suggestion that I do in in small steps
[04:11:55] <toastydeath> anything you have that's 1" by 1/8"
[04:12:31] <owad> what's the shallowest cut I can take?
[04:12:45] <owad> (without friction being an issue)
[04:12:48] <toastydeath> uh, .001-.002 probably
[04:12:51] <toastydeath> z depth
[04:12:58] <toastydeath> but there's no reason to do that, none whatsoever
[04:13:18] <toastydeath> start at .050" and work toward .1"
[04:14:30] <toastydeath> with a rotozip cutter you're going to be fighting the chip clearance since those things have a friggin million teeth
[04:14:37] <owad> I could do the initial couple passes with the sturdier endmill. Perhaps that would help keep the deeper cuts on track, with the rotozip
[04:14:56] <toastydeath> what happens at the bottom that you can't use that same sturdier endmill
[04:15:06] <owad> it has a 1/4" shank
[04:15:15] <toastydeath> oh, doesn't fit the hole
[04:15:16] <toastydeath> okay
[04:15:29] <toastydeath> i dunno, if you're going to have to do it might as well bite the bullet and just let it run
[04:15:50] <toastydeath> if you were doing production i'd say yes, use sequential tooling, but since you're not making 40 of whatever it is (at least i hope not)
[04:16:01] <toastydeath> just write the toolpath, step it down incrementally, and let it run.
[04:16:14] <owad> in the z direction, can friction be an issue if I go too shallow, like it can with the x & y?
[04:16:19] <toastydeath> yes
[04:16:32] <toastydeath> but i gave you your Z depth, so that won't be an issue.
[04:17:44] <stustev> are you trying to contour a mold?
[04:18:04] <owad> the drag is so bad with 1/8" passes, I'm concerned it will still be apparent with 0.05"
[04:18:07] <owad> wooden gears
[04:18:28] <toastydeath> what kind of machine do you have
[04:18:35] <toastydeath> gantry with acme screws?
[04:18:48] <owad> basically
[04:18:50] <stustev> the gears are laying flat on the table and you are cutting the tooth profile with the side of the cutter?
[04:18:53] <owad> but I don't have acme screws yet...
[04:18:56] <toastydeath> also how fast you go determines the drag
[04:19:03] <toastydeath> if you are trying to push that poor 1/8th cutter at 40 ipm
[04:19:06] <toastydeath> or 30
[04:19:15] <toastydeath> you're itchin' for some hurt on deflection
[04:19:32] <toastydeath> if i had something that long, i'd be trying to run it as slow as possible
[04:19:35] <owad> yes, the gears are laying flat, and I"m cutting the tooth profile
[04:19:44] <owad> well, I tried it at 16 ipm, and it caught fire :)
[04:20:00] <toastydeath> even if i didn't have much control over spindle speed, i'd still run it down near like, 4-5 ipm
[04:20:15] <stustev> the gear is 1 inch thick? This must be a mold for casting a gear.
[04:20:38] <owad> well -- right now I'm just playing with 3/4" MDF, b/c that's what I have
[04:20:57] <stustev> will the finished gear be MDF?
[04:20:57] <toastydeath> your spindle has no speed control?
[04:21:11] <owad> the spindle can go from 10K to 23K
[04:21:20] <toastydeath> and it caught fire at 10k?
[04:21:27] <toastydeath> at 16 ipm?
[04:21:29] <owad> at 20K & 16 ipm, it caught fire
[04:21:43] <owad> and that was with a 1/4" endmill
[04:21:54] <toastydeath> then i guess i'd try plunge roughing it first then
[04:22:03] <owad> what's that?
[04:22:08] <toastydeath> rather than try to get the profile by milling the contour
[04:22:13] <toastydeath> you step over in small amounts
[04:22:18] <toastydeath> and "drill" down to the full depth
[04:22:21] <toastydeath> in pecks
[04:22:35] <toastydeath> the pecks give it time to cool back down
[04:22:40] <owad> that makes sense
[04:22:56] <toastydeath> it's also the most rigid type of cutting
[04:22:56] <stustev> can you drill a hole at the root of each tooth?
[04:23:07] <owad> I don't see why not
[04:23:58] <toastydeath> what we do on this big stainless steel pocket is drill down with an endmill using 1/3rd cutter diameter stepovers
[04:24:07] <toastydeath> and just trace the profile
[04:24:31] <toastydeath> and then, once the whole thing is roughed out, you can take several light passes at full depth and just dust the ridges off
[04:24:38] <toastydeath> .010" and .005" stepover
[04:24:39] <stustev> the plunge roughing is a good idea. You can drill a hole at the root of each tooth and then rough (normally or plunge) with a larger tool and finish with the size tool you need for the root radius.
[04:25:57] <stustev> this is for show? MDF is easy to hand work. Drill the hole rough it close. Hand work it in.
[04:26:26] <owad> it's mostly for learning how to use the router
[04:27:29] <owad> Is there a good source for cheap endmills? There must be something better than $1 rotozips, that is still affordable.
[04:27:43] <toastydeath> dunno what your limit on "affordable" is
[04:28:03] <stustev> ok - plunge rough some of it - climb mill rough some of it - conventional rough some of it
[04:28:04] <owad> I'm thinking under $30 per endmill; ideally around $10
[04:28:06] <toastydeath> cheap endmills tend not to last long and aren't very sharp
[04:28:27] <toastydeath> i don't think you'll find an endmill worth using for 10, unfortunately
[04:28:37] <owad> so what's the cheapest endmill worth buying?
[04:28:38] <toastydeath> someone else here might have a suggestion
[04:29:18] <toastydeath> don't know, my shop only uses carbide
[04:29:20] <stustev> you will find carbide the cheapest - long term - use the rotozips to perfect the motion and then use a carbide cutter to cut the gear. One cutter will do the whole thing
[04:29:46] <owad> what do you mean, 'perfect the motion'?
[04:29:58] <toastydeath> make sure the program cuts where you want to cut
[04:30:06] <toastydeath> learn to jockey the machine around
[04:30:21] <toastydeath> $1 rotozip cutter = very inexpensive disposable learning tool
[04:30:40] <owad> http://www.amazon.com/4Fl-SE-Carbide-End-Mill/dp/B000I6I5NK $5 carbide?
[04:30:53] <toastydeath> give it a shot and see if it works
[04:31:45] <toastydeath> the cheaper the carbide is, the more brittle it is and the duller the starting cutting edge
[04:32:00] <toastydeath> those probably won't be an issue in mdf
[04:32:13] <stustev> you don't want to bang corners and overload the cutter - you want to make simple clean cuts - figure that out with the rotozip cutters. A $5 carbide cutter is also a cheap tool. If your router is not pretty sturdy then you don't want to try the carbide on steel but MDF should be fine.
[04:33:17] <owad> it's not sturdy
[04:33:24] <owad> What about grizzly? http://grizzly.com/products/H3661
[04:33:59] <toastydeath> i'd be more interested in the 5% cutter
[04:34:00] <toastydeath> er
[04:34:02] <toastydeath> 5$
[04:34:21] <toastydeath> because 34 bucks is closer to what you'd spend on like, a Niagara or OSG endmill
[04:34:56] <toastydeath> i don't buy our endmills, lemme see what they run
[04:35:03] <owad> Would a Niagra or OSG be stronger?
[04:35:17] <toastydeath> Niagara and OSG are sharper, and they're tougher grades of carbide that hold a better edge
[04:35:35] <toastydeath> so not only will they take the shocks a non-rigid machine generates, but they'll keep the edge
[04:35:47] <toastydeath> but like everything, they cost money
[04:36:01] <toastydeath> if a 5$ carbide endmill holds you over, i'd be dumb to say don't do it
[04:36:56] <owad> Would a Niagra/OSG bend just as much as the rotozip, at 40 ipm?
[04:37:16] <toastydeath> much less
[04:37:21] <toastydeath> it would still bend, though
[04:37:33] <toastydeath> and carbide is not a fan of bending a whole lot, it's very brittle
[04:37:43] <toastydeath> but again high end carbides are more resistant to that kind of nonsense
[04:37:54] <toastydeath> check this out
[04:37:55] <toastydeath> http://www1.mscdirect.com/CGI/N2DRVSH?PACACHE=000000065940741
[04:38:14] <owad> invalid part
[04:38:46] <toastydeath> one more time
[04:38:48] <toastydeath> http://www1.mscdirect.com/CGI/N2DRVSH?PACACHE=000000065940770
[04:38:51] <toastydeath> if that doesn't work then i dunno
[04:39:01] <toastydeath> but what i'm linking is the MSC page for Niagara endmills
[04:39:05] <toastydeath> 1/8th
[04:39:10] <toastydeath> they're like 20 bucks
[04:39:14] <owad> nope. is there a part # I can look up?
[04:39:22] <toastydeath> maybe not exactly what you're looking for but uh, they'll probably work
[04:39:23] <owad> that's not bad at all
[04:39:53] <toastydeath> on the home page, on the left menu, click third item down "end mills, saws, cutters & tool bits"
[04:40:08] <toastydeath> section #3 is end mills/saws/etc
[04:40:21] <toastydeath> click on "single end mills"
[04:40:25] <toastydeath> then "standard"
[04:40:36] <toastydeath> then it'll give you all their endmills.
[04:40:44] <toastydeath> and you can start filtering by name/size/etc
[04:41:08] <toastydeath> i would suggest NOT filtering for a full 1" depth of cut, because those are going to be like 110-150 bucks in 1/8th size
[04:41:59] <toastydeath> oh wait, there are cheap ones
[04:42:02] <toastydeath> in hss
[04:42:13] <toastydeath> oh wow.
[04:42:15] <toastydeath> a+
[04:42:42] <owad> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=140260792020
[04:42:47] <owad> a+ to which?
[04:43:14] <toastydeath> oh, i was just saying a+ that there are a bunch of 1" flute length, 1/8" diameter endmills
[04:43:24] <toastydeath> for like 20 bucks
[04:44:21] <owad> How do I decide how many flutes?
[04:44:40] <toastydeath> go with 2
[04:45:00] <owad> why?
[04:45:02] <toastydeath> wood, aluminum, and other soft materials all get 2 flutes, or 3 on a rigid machine
[04:45:11] <toastydeath> the primary goal is chip clearance
[04:45:34] <toastydeath> you can take a large bite with a two flute cutter and not really stress the thing out
[04:45:56] <toastydeath> the shear strength of the material is low, so even with the flutes "hammering" it's a fairly low amplutude vibration
[04:46:31] <toastydeath> with something like steel, two flutes will quickly produce a very large amplitude vibration
[04:47:14] <toastydeath> so to get the feed rate back up
[04:47:22] <toastydeath> they start stacking a ton of cutting edges on the cutter
[04:47:35] <toastydeath> so where in aluminum i might be taking a .005-.010" per tooth
[04:47:42] <toastydeath> in steel that might drop to .003
[04:47:54] <toastydeath> but since i have 6 teeth on a 1" cutter in steel...
[04:48:11] <toastydeath> it brings the horsepower of the machine back to bear
[04:48:48] <owad> thanks
[04:48:51] <toastydeath> np
[04:49:14] <toastydeath> 3 is the "ideal" number from a dynamics perspective, but that's on expensive machines with very high feed rates
[04:49:29] <owad> should I worry about the coating?
[04:49:30] <toastydeath> 2 is just a really good choice for general purpose machines
[04:49:35] <toastydeath> no, you shouldn't
[04:49:43] <owad> solid carbide vs carbide?
[04:49:51] <toastydeath> nope
[04:50:11] <owad> what's the difference make for sold vs non, and the coating?
[04:50:22] <toastydeath> that's probably just MSC doing a retarded extra category
[04:50:30] <toastydeath> i'm not aware of very many brands that have non-carbide cores
[04:50:40] <toastydeath> but they add toughness to the tool
[04:50:53] <toastydeath> MSC likes to add extra categories where none should exist
[04:51:09] <SWPadnos> the people who type in the information don't necessarily know what they're talking about
[04:51:09] <toastydeath> the coating is for higher tool speeds in steel
[04:51:27] <toastydeath> some coatings help aluminum, but that's very few since aluminum cuts like butter anywa
[04:51:30] <toastydeath> y
[04:51:35] <SWPadnos> so if a manufacturer says "Solid Carbide!", and another just says "Carbide", it'll be two categories
[04:51:50] <owad> heh
[04:51:57] <SWPadnos> you see it a lot on NewEgg or PriceWatch also :)
[04:52:02] <toastydeath> as the coating heats up, it starts to act like a solid lubricant, helping chip flow and reducing horsepower requirements for the cut
[04:52:21] <toastydeath> it also allows more brittle grades of carbide to be used that are sensitive to thermal shock, and can't handle coolant
[04:52:31] <owad> Are all of MSC's brands decent? Kennametal, Accupro, Hertel, Atrax...?
[04:52:40] <toastydeath> most should be pretty good
[04:52:55] <toastydeath> kennametall is a pretty well known brand, i recall hertel being spoken of highly
[04:53:00] <toastydeath> accupro, atrax? i dunno.
[04:53:04] <toastydeath> could be good, could be not - io
[04:53:06] <toastydeath> i'm not sure.
[04:53:20] <owad> there's so many choices from msc, it's overwhelming
[04:53:21] <toastydeath> my shop is 90% OSG
[04:53:33] <toastydeath> for your purposes, i would not worry overmuch. seriously.
[04:53:48] <owad> it's only like $7 more for osg
[04:53:51] <toastydeath> kennametall is a middle of the road kind of manufacturer, you won't go wrong with them and you also won't get shafted
[04:54:32] <toastydeath> you're cutting mdf - almost any cutter will do that pretty darn well
[04:55:41] <owad> for other sorts of woods, and maybe plastics or foam, will it be an issue? I don't plan to do metal
[04:55:46] <toastydeath> nope
[04:55:57] <toastydeath> two flute carbide with your spindle should be pretty good
[04:56:05] <owad> cool
[04:58:11] <owad> I can get Accupro for $15, Kennametal for $18, and OSG for $22
[04:58:35] <toastydeath> i'd go with kennametall for three bucks.
[04:58:38] <toastydeath> really i'd go with osg
[04:58:46] <toastydeath> but you're cutting mdf.
[04:59:24] <owad> I'm more concerned about bending/breaking than sharpness
[04:59:57] <toastydeath> i'd be suprised if you bent enough to break any of those endmills
[05:00:31] <owad> by bend, I meant drag. You don't think that'll be an issue?
[05:01:08] <toastydeath> what do you mean by drag, exactly
[05:01:41] <toastydeath> from the sound of your word choice, i'd guess that you were pushing the rotozip cutter way too hard and it wasn't clearing chips
[05:02:06] <toastydeath> but i don't want to like, randomly guess
[05:02:10] <owad> the rotozip bit bends, when I cut at 40 ipm. it's not vertical
[05:02:24] <toastydeath> yes, that's a nature of the asymmetric forces of cutting
[05:02:34] <toastydeath> one side of the cutter is trying to dig into the material, the other side wants to push away
[05:02:46] <toastydeath> especially in a full slotting type of deal
[05:02:58] <fenn> logs logs logs.. y'all talk a lot
[05:03:12] <owad> it makes a pretty ugly mess, doing corners
[05:03:17] <fenn> owad: i suggest use-enco.com - it's MSC's discount flavor
[05:03:46] <owad> thanks fenn, I'll take a look right now
[05:03:49] <toastydeath> corners are usually a pain because you have to realize you're going from a situation where you've got some fraction of the cutter diameter engaged int the cut
[05:03:50] <fenn> i like the name of your website btw, mentioned to some people in #hplusroadmap
[05:04:03] <owad> thank you
[05:04:13] <toastydeath> to a situation where the cutter is fully engaged
[05:04:29] <fenn> i've been following the dave gingery path myself
[05:04:47] <toastydeath> so you can either grin and bear it, decelerate into corners, or take light depth, high feed cuts at 90% the cutter diameter or so
[05:05:30] <fenn> but gingery is hopelessly outdated so now i'm basically inventing my own system of post-industrial knowhow
[05:05:34] <toastydeath> also a small skim pass cleans it up most of the time, if it chattered bad
[05:08:57] <owad> toastydeath: last fall, I got to use a shopbot at MIT. we started out with a rotozip bit, and it dragged terribly, and then broke. Then we switched to a really pricey endmill that we weren't supposed to use, and it worked beautifully. So I really do think the rotozips are junk, but I don't know what the endmill was, or how special it was.
[05:09:08] <fenn> owad: why are you using MDF to make gears anyway?
[05:09:21] <toastydeath> owad: it's that they're built for very light cuts
[05:09:26] <toastydeath> it's more of a rasp/file
[05:09:38] <owad> fenn: mdf is what I have at the moment. I'll probably use some nicer-looking wood.
[05:09:46] <owad> (for making a clock)
[05:09:47] <toastydeath> you COULD get a rotozip to cut, but you'd have to really set the machine up well
[05:10:01] <toastydeath> count the teeth, set the feed per tooth very low
[05:10:02] <fenn> yeah, double cut flutes are only good for fibrous materials like circuit boards
[05:10:03] <toastydeath> etc
[05:10:12] <fenn> its also good for turning styrofoam into dust
[05:11:08] <fenn> the endmill will eject chips out of the groove, whereas a double cut rasp like a rotozip will just get clogged
[05:12:08] <owad> Here's the difference I'm getting in cut quality. Good endmill on the left, rotozip on the right: http://sandbox.owad.org/cut.jpg
[05:12:42] <owad> and charred remains at the top :)
[05:13:45] <owad> fenn: how's gingery outdated?
[05:14:04] <cradek> wow - in the future, after the few inches of cut with flames shooting out, you can probably assume it's not going to improve, and stop it :-)
[05:14:22] <fenn> no mention of CNC, china makes small low-quality manual machines for less than is possible to make from scratch
[05:14:38] <toastydeath> grigery?
[05:15:07] <fenn> owad: more reasons here if you're interested: http://fennetic.net/machines/21st_century
[05:16:13] <fenn> toastydeath: dave gingery wrote a series of 'how to' books on aluminum casting, scraping, and building machine tools using self-generating processes
[05:16:18] <toastydeath> oh
[05:16:27] <fenn> you basically bootstrap the industrial revolution in your back yard
[05:16:35] <toastydeath> lol
[05:17:20] <owad> fenn, are these updates of gingery's designs, or totally different?
[05:17:54] <fenn> owad: it started out as updates and corrections to gingery, but then i sort of got distracted and nobody else contributed, and it mutated into something else
[05:18:15] <fenn> i now think that traditional cartesian axis machine design is bad
[05:18:34] <fenn> inefficient use of mass and labor
[05:18:40] <owad> what's better?
[05:19:28] <fenn> it depends on what you're doing, but most of my ideas are based around the octahedral hexapod or some kind of tensegrity structure (cable hexapod, basically)
[05:20:43] <fenn> we're trying to come up with a program/website for designing self replicating systems
[05:20:58] <fenn> reprap just doesn't cut it (no pun intended)
[05:21:53] <owad> the reprap prints are looking very good
[05:22:31] <fenn> if you want low-res warped plastic parts that can't have overhangs more than 45 degrees, then yeah
[05:23:11] <fenn> kind of hard to make rockets and computers and mass spectrometers with that though
[05:23:23] <owad> true
[05:24:31] <toastydeath> lol
[05:25:05] <owad> http://blog.reprap.org/2008/05/quality-development.html
[05:25:18] <owad> that sort of quality will be very useful for a lot of applications, though
[05:25:33] <owad> the one on the far right, that is. :)
[05:25:33] <fenn> i hope they do some ceramics research before going down the pointless fields metal circuitry dead end
[05:26:43] <fenn> hmm. "While most folks are trying to get a Darwin or some other kind of reprap machine together and working, Nophead simply bought a CNC positioning table and attached a Mk II extruder to it."
[05:27:02] <fenn> do they get equivalent quality out of the darwin frame?
[05:27:15] <owad> I have a reprap head I'm hoping to mount on my router
[05:27:51] <owad> I don't now about equivalent, but there are some good-looking prints from Adrian and Vik, and they're using the darwin frame.
[05:28:34] <fenn> it looks like the handles are made of different materials too, like (left to right) PCL, HDPE, ABS
[05:29:24] <owad> Your shaper looks really nice.
[05:29:37] <fenn> not my shaper, some guy in texas i think
[05:29:48] <fenn> i only made the lathe
[05:30:05] <fenn> should have started with the shaper in hindsight (ah hindsight)
[05:30:15] <owad> is there a pic up?
[05:30:26] <toastydeath> it would be cool to have a planer
[05:30:34] <owad> is a shaper all that useful, when you have a cnc vertical mill?
[05:30:39] <toastydeath> a cnc planer
[05:30:41] <fenn> http://fennetic.net/machines/lathe_modifications <- click on links under completed
[05:30:50] <toastydeath> owad: depends on the size of the mill
[05:31:08] <fenn> shaper can do lots of weird things that are impossible on a mill, like gear teeth or keyways
[05:31:19] <toastydeath> with a desktop sized mill they retain a lot of functionality
[05:31:21] <fenn> also it makes smoother surfaces for things like dovetail ways
[05:31:34] <fenn> and the cutters cost much less
[05:31:35] <toastydeath> with a bigger mill, they tend to be relegated to specific tasks
[05:31:38] <toastydeath> like fenn is pointing out
[05:32:22] <fenn> the interesting thing is that if you're designing the machine, it can be reconfigured in different modes, i.e. shaper, horizontal mill, lathe
[05:32:35] <fenn> using mostly the same components
[05:32:52] <owad> there's always some cute Atlas shapers at the cabin fever auctions. Wish I had room for one.
[05:34:58] <toastydeath> i have pictures to upload
[05:35:02] <toastydeath> of a 2" diameter endmill
[05:35:05] <toastydeath> but i am so lazy
[05:39:46] <owad> ooh, Enco has the 1/8" Atrax carbide endmills on sale for $10.39
[05:40:27] <fenn> toastydeath: i feel your pain
[05:43:45] <toastydeath> lol
[05:43:48] <toastydeath> ty
[05:45:25] <owad> I better get to bed. I really appreciate everybody's help.
[05:45:42] <toastydeath> nighty
[05:45:49] <owad> 'later
[06:44:59] <anonimasu_> fenn: nice
[08:36:32] <pjm_> good morning
[09:12:21] <ra3vat> ra3vat is now known as dimas_
[11:42:06] <cradek_> cradek_ is now known as cradek
[21:28:34] <CanSir> How does the "Home" button in AXIS do it's work? I'd like to setup a pyVCP button that homes both X & Y to the current position.
[21:31:55] <CanSir> The home button is AXIS works great, but I have to select the X axis, click Home, select the Y axis, click Home. I'd like to setup a single button to home both.
[21:37:37] <SWPadnos> set "HOME_SEQUENCE" to 0 for both X and Y, set it to -1 (I think) for any other axes
[21:37:56] <SWPadnos> then ctrl-home will home X and Y - that's the "home all" command
[21:44:28] <alex_joni> SWPadnos: if you have HOME_SEQUENCE set up right, AXIS will display a home all button, not the regular home button
[21:51:51] <alex_joni> good night all
[22:00:50] <SWPadnos> alex_joni, the home all button is in addition to the individual home, I think (so you can still home a single axis)
[22:01:12] <anonimasu_> hm.. can you home Z then home X and Y togther?
[22:01:17] <SWPadnos> sure
[22:01:31] <anonimasu_> nice
[22:01:38] <SWPadnos> set HOME_SEQUENCE to whatever sequence you want, starting with 0 (first)
[22:01:53] <SWPadnos> or set to -1 (or maybe don't set ??) to not include that axis in a home-all operation
[22:02:04] <anonimasu_> I mean sequentially
[22:02:10] <SWPadnos> you can have multiple axes with the same sequence number
[22:02:21] <anonimasu_> hm.. ok?
[22:02:22] <SWPadnos> sure, 0=first, 1=second ...
[22:02:43] <SWPadnos> [AXIS_0]HOME_SEQUENCE=0
[22:02:47] <SWPadnos> [AXIS_1]HOME_SEQUENCE=0
[22:02:51] <SWPadnos> [AXIS_2]HOME_SEQUENCE=1
[22:02:56] <anonimasu_> I see
[22:03:00] <SWPadnos> would do X+Y together, then Z
[22:03:04] <anonimasu_> that's a neat feature
[22:03:19] <anonimasu_> (I usually end up homing z first to clear parts and clamps)
[22:03:35] <SWPadnos> yeah - I noticed I had it backwards after I had typed it :)
[22:03:41] <anonimasu_> haha ^_^
[22:03:48] <anonimasu_> *yawns*
[22:04:12] <anonimasu_> thanks for enlightening me
[22:04:28] <SWPadnos> sure
[22:07:10] <rich__> Hello from UK. Scuse if I make errors as a newbie to this
[22:07:18] <cradek> hi rich
[22:07:51] <rich__> Am new to Linux but trying it for a cnc project and needed bit of help
[22:08:21] <rich__> Hi cradek
[22:09:04] <cradek> the only rule in irc is to not ask to ask -- just ask!
[22:09:17] <rich__> Thanks :-)
[22:09:26] <cradek> well there might be more rules LIKE DON'T YELL, haha
[22:09:38] <cradek> what's your project?
[22:10:45] <rich__> just building mini cnc router and saw that Linux cnc was available.. so guess what.. not only take the plunge from Microshaft to linux but cnc too!
[22:11:31] <rich__> Been years since did IRC too I add.. bit rusty!!
[22:14:46] <rich__> Can you tell me.. is it poss to run EMC2 without hardware present? I get errors
[22:15:08] <cradek> yes everything under "sim" is for no-hardware
[22:15:18] <cradek> try sim/axis or sim/axis_mm if you prefer those
[22:15:32] <rich__> Thanks.. thut I'd tried that but maybe not..lol
[22:15:47] <rich__> gimme sec
[22:15:59] <cradek> if you try again and get an error, let's see what it is
[22:17:26] <rich__> still error. ending line is "Module rtai hal does not exist"
[22:17:48] <cradek> how did you install emc?
[22:18:12] <rich__> from the site
[22:18:23] <cradek> the cd from linuxcnc.org?
[22:18:47] <rich__> just sec.. will check
[22:20:03] <rich__> downloaded hardy-install from the home page
[22:20:26] <rich__> had installed Ubuntu first
[22:20:33] <cradek> ok
[22:20:36] <rich__> am new to all this
[22:20:48] <cradek> when you install that way, you end up with a choice of realtime and non-realtime kernels.
[22:21:03] <rich__> now ur bamboozling me..lol
[22:21:15] <cradek> when you boot, there is the opportunity to hit escape early in the boot process. this lets you pick the realtime kernel which will be linux-something-rtai
[22:21:34] <rich__> oh??
[22:21:38] <cradek> the easier way to install a system that's meant for just emc is to use the cd distribution from linuxcnc.org
[22:21:54] <rich__> u mean order the real hard copy?
[22:22:03] <cradek> this is the same as the normal ubuntu, with emc added, and with only a realtime kernel
[22:22:31] <cradek> http://www.linuxcnc.org/content/view/21/4/lang,en/
[22:22:35] <rich__> wish I knew more bout Linux b4 delving..lol
[22:23:04] <cradek> but it sounds like all you have to do is reboot and pick the rtai kernel from the list.
[22:23:58] <rich__> ok.. will persist... but will prob be back here soon I guess! Thanks for your help
[22:24:01] <cradek> I have to run - hope you get it
[22:24:04] <cradek> welcome
[22:24:12] <rich__> thanks.. c u later