#emc-devel | Logs for 2006-07-03

[01:12:10] <fenn> is there any reason why we need readline version 5 in particular?
[01:12:21] <jmkasunich> not sure
[01:18:22] <fenn> i gotta say there is some weird stuff i dont understand in the test for -lreadline in configure
[01:31:20] <jmkasunich> jepler is the local expert on that
[01:40:56] <jmkasunich> hi
[01:41:01] <zwisk> hello...
[01:41:12] <jmkasunich> you could stay in #emc too, never know when it might drift back on topic ;-)
[01:41:27] <fenn> #emc-devel is pretty boring usually
[01:41:42] <zwisk> True... My irc skills are rusty. I've forgotten how to go into expert mode and be in both and all.
[01:41:49] <jmkasunich> well, zwisk said he was thinking
[01:41:55] <jmkasunich> what IRC client?
[01:41:57] <zwisk> it was dangerous :)
[01:42:15] <zwisk> ircii it seems...
[01:42:31] <jmkasunich> oh, that might be single channel, (or require a true expert)
[01:42:47] <jmkasunich> I've gotten used to gui clients, they use tabs for multiple channels
[01:42:54] <zwisk> I used to do it all the time... like /novice n or something...
[01:43:07] <jmkasunich> just keep /joining and the tabs keep poppin up
[01:43:19] <fenn> * fenn has 10 channels open :x
[01:43:25] <zwisk> new fangled gui's... bah.. :)
[01:43:40] <Mess> or old ones...
[01:43:45] <zwisk> anyway, I was thinking about refactoring again. I should probably stop thinking about that.
[01:43:58] <jmkasunich> not neccessarily
[01:44:05] <jmkasunich> what did you think of?
[01:44:18] <zwisk> The trick would be to refactor while maintaining a high level of compatibility with what's being used now.
[01:44:27] <jmkasunich> yeah
[01:45:02] <zwisk> Cradek was suggesting that the previous work I put into the last refactor might be salageable. I've forgotten enough of what was or wasn't accomplished with that effort that I've deluded myself into thinking he might be right...
[01:45:07] <jmkasunich> and as capabilities keep getting added, it gets harder and harder to do that
[01:45:34] <jmkasunich> (that = keep compatibility)
[01:45:59] <zwisk> yep.
[01:46:21] <jmkasunich> one big goal of the refactor was to move metadata out of shared memory
[01:46:33] <zwisk> Well, the 'brute force' solutoin to that would be to 'convert' all the existing clients.
[01:46:56] <jmkasunich> clients = configs that use hal? hal modules? or?
[01:46:58] <zwisk> Yes, the work I did disposed of shared memory and put all of the metadata in kernel memory. It used ioctl's to access kernel memory in a 'standard' way.
[01:47:37] <jmkasunich> how did you handle commands that result in variable sized output?
[01:47:44] <jmkasunich> (such as "list all pins")
[01:47:48] <zwisk> all of the above :) Whatever would need to be changed, be in configs, modules, or ... (Modules is primarily what I had in mind. Configs shouldn't need much change... I would hope...)
[01:48:39] <zwisk> I used an iterator (if I recall correctly) that would return a record for 1 pin, and allow you to walk the list of pins. I'm not remembering how I locked the list at the moment though.
[01:49:06] <Mess> keep dreamin'
[01:49:20] <jmkasunich> is there a lot of overhead when using an ioctl? (wondering how heavy an operation iterating thru a couple hundred pins would be)
[01:49:20] <Mess> sounds good
[01:49:46] <zwisk> Nope, not a lot. It's not nothing, and it is a copy from kernel to user space, but you don't do it very often.
[01:50:05] <jmkasunich> we no longer iterate for just "show pin" and halscope/halmeter probe assignment
[01:50:16] <jmkasunich> now we iterate for readline completion lists and such...
[01:50:59] <zwisk> I guess some investigation would be in order. I also had a front end on it which used /proc, so you could ls /proc/ and get a listing of all the pins and stuff.
[01:51:16] <zwisk> (It was probably /proc/hal/ )
[01:51:39] <zwisk> As I recall, I quite playing with it when emc2 (and I) moved from the 2.4 to 2.6 kernel...
[01:51:44] <zwisk> So, it's been a while.
[01:51:51] <jmkasunich> yeah, I seem to recall some ideas like a /proc/hal node for each object
[01:52:01] <fenn> a proc interface would be very nice
[01:52:25] <zwisk> Maybe I'll dig that code back up and see about moving it to 2.6 ... then we can play with it and see if it's useful.
[01:52:34] <jmkasunich> fenn: the more I thought about it, the less I liked it
[01:52:56] <zwisk> As I recall, I didn't complete the refactor... Pins couldn't yet be connected or something.
[01:53:04] <jmkasunich> think about the halcmd language... and try to map that onto /proc entries
[01:53:20] <jmkasunich> it gets non-pretty after a while
[01:53:46] <jmkasunich> ioctrl's on the other hand, seem like a more lightweight way to do things
[01:54:09] <jmkasunich> halcmd invokes ioctrls, the kernel space code is probably less than building a massive proc tree
[01:54:16] <fenn> i dont mean for actually doing stuff inside hal, but more for interfacing with userspace applications
[01:54:30] <jmkasunich> fenn: I dunno what you mean
[01:54:39] <jmkasunich> halcmd is used to configure things
[01:54:49] <jmkasunich> pins are used to interface
[01:54:57] <jmkasunich> we already have a clean API for pins
[01:55:09] <fenn> just random crap like, getting the name of a signal connected to a pin (for tooltips in classicladder for instance)
[01:55:27] <jmkasunich> that is a halcmd-like function
[01:56:14] <zwisk> I remember paul being exta excited about ioctl and not about proc, so I did both.
[01:56:51] <jmkasunich> I think paul raised a legitimate concern that the kernel folks are moving stuff out of proc and into sysfs
[01:57:00] <jmkasunich> I sure as heck don't want to support both
[01:57:15] <zwisk> Anyway, the 2 big key improvements in my mind ar the ability to instantiate and tear down stuff without removing or reloading modles, and moving stuff out of shared memory and into kernel memory.
[01:57:25] <jmkasunich> agreed
[01:57:35] <zwisk> nah... support would purely be through halcmd. /proc/ is just kinda showy :)
[01:57:36] <jmkasunich> the two may or may not be tied together
[01:57:56] <zwisk> right. They are different things.
[01:58:07] <jmkasunich> although each requires enough hacking on the internals that it makes sense to do them both at the same time if possible
[01:58:15] <zwisk> exactly so.
[01:58:30] <jmkasunich> way back when, there was just the hal.h API
[01:58:57] <zwisk> Are folks creating hal objects that compile both in user and kernel space these days?
[01:58:57] <jmkasunich> now the halcmd language is as important, if not more so, as a standard to be compiled with (or carefully modified/extended)
[01:59:04] <jmkasunich> no
[01:59:16] <jmkasunich> in fact I'm very tempted to remove the user space code from parport
[01:59:30] <zwisk> I seem to recall that being one of the wishes, but I don't think I've ever seen the user space stuff go.
[02:00:06] <jmkasunich> dual mode components is something I thought interesting at one time, but in hindsight...
[02:00:18] <zwisk> haha ... 20/20 hindsight.
[02:00:27] <jmkasunich> otoh, user space only comps are a usefull thing, and are finally starting to appear
[02:00:41] <jmkasunich> halvcp, halui, the joystick driver...
[02:01:14] <zwisk> I actually think it could be useul. The comments recently on the lists about "simulator" mode, while not quite correct in the way they were written, I think are valid -- Some folks would like to play with these things without modifying their kernel, even if that means *not* having realtime motion.
[02:01:34] <jmkasunich> agreed, but I would not address that with dual mode components
[02:01:47] <jmkasunich> the comps should still be in kernel space, they just don't run in realtime
[02:02:03] <zwisk> perhaps so.
[02:02:22] <cradek> I'm really interested in having a non-rt simulator for development purposes, but I think we're foolish if we let people hook that to hardware because they *will*
[02:02:25] <jmkasunich> as long as we can get kernel/user shared memory with a vanilla kernel, we don't need the RTOS
[02:02:29] <zwisk> Debugging some things is easier in userspace, but I'll concede the point. Having hal work in your jiffies mode seems at least interesting.
[02:03:06] <jmkasunich> if it was straightforward to make every component a dual mode component, I'd be in favor of that
[02:03:30] <jmkasunich> but its not... it would take messy makefile foo just to build them, and messy conditional stuff inside each component
[02:03:31] <cradek> (I do most of my development on a machine that will never be hooked to a mill, and I would rather not run a realtime kernel to do it)
[02:03:41] <zwisk> Ok... showing my forgetfulness and pure ignorance, why is it not easy? I recall we talked some about trying to abstract the hal interface to make that possible once...
[02:04:09] <jmkasunich> I want to keep components clean, so people who are average programmers can make a hal component
[02:04:22] <zwisk> right...
[02:04:51] <zwisk> Dispense with the notion of conditionals inside the components... :)
[02:05:10] <jmkasunich> one extreme idea that I actually expored for a while was to make everything run in user space
[02:05:22] <jmkasunich> use LXRT from RTAI to get hard realtime in user space
[02:05:58] <jmkasunich> but you still need wind up doing a trap and change from ring 3 to ring 0 every time you do an inb or outb
[02:06:07] <jmkasunich> I just can't bring myself to do that in the RT code
[02:06:37] <jmkasunich> its one thing to do a bunch of ioctl system calls for a halcmd operation, its another thing entirely to do that in a driver
[02:06:41] <zwisk> hmm... yeah, I was going pretty much the other way-- almost everything in the kernel.
[02:06:56] <jmkasunich> thats my current approach too
[02:07:56] <zwisk> I should dust off what I was doing before and get back up to speed so I can talk about this more intelligently....
[02:08:13] <zwisk> (in all my 'free' time :) )
[02:08:48] <jmkasunich> ok
[02:09:20] <zwisk> On a slightly related note... I have RPMs for RTAI on FC3 that I'd be happy to put someplace if others are interested. Do we have a good place for those?
[02:09:44] <zwisk> (I asked about that last night, but lost any potential answers in my scrollback buffer :) )
[02:11:02] <fenn> oo oo
[02:11:12] <fenn> i'll be happy to whine when it doesnt work
[02:11:16] <cradek> alex_joni hosts the ubuntu packages, maybe you could ask him
[02:11:18] <zwisk> Haha.. :)
[02:11:41] <zwisk> Well, fenn... these have been running on my primary machine for over a year. They do break with vnc, but that's the only problem I've had with them :)
[02:12:07] <cradek> vnc breaks rtai?
[02:12:07] <zwisk> I do hope to spin a more recent version, though, and those will require testing, undoubtedly.
[02:12:27] <zwisk> Doh ... sorry... my bad... not vnc. vmware.
[02:12:56] <zwisk> The vmware kernel modules apparently do something un-rtai friendly.
[02:13:33] <cradek> no big surprise
[02:13:46] <cradek> vnc would be a big surprise :-)
[02:14:06] <zwisk> yeah. It's one of the big reasons I started using qemu instead of vmware.
[02:15:29] <cradek> I'm glad to know that, I use qemu for dos autocad on my mill's machine and was considering switching to vmware for speed - but now I won't
[02:16:26] <zwisk> If you do get it to work, I'd be interested to hear about it... as vmware is a bit quicker than qemu at the moment.
[02:16:51] <cradek> I'm not going to mess with it, qemu is fast enough for me
[02:17:04] <zwisk> It's completely possible that my problems were a specific interaction with my version of rtai and the kenrel.
[02:17:05] <cradek> I'd hate to screw up a part if I forgot to close autocad
[02:17:07] <jmkasunich> did you have problems running ordinary user space stuff with a RTAI patched kernel in vmware? or running RT stuff?
[02:17:55] <zwisk> I stopped doing this about a year ago, so I'm a bit foggy, but as I recall, the machine screeched to a halt and pretty much locked up as soon as vmware tried installing it's modules.
[02:18:31] <jmkasunich> is this vmware on a machine patched for RTAI, or RTAI on a vmware virtual machine?
[02:18:40] <jmkasunich> gawd, it gets confusing doesn't it?
[02:18:52] <zwisk> I've never tried running rtai inside vmware... always rtai on a linux box on real hardware running vmware.
[02:19:01] <jmkasunich> ok
[02:19:27] <zwisk> I don't imagine an emulated rtai inside a non-real-time vmware instance would be of much use.
[02:19:31] <jmkasunich> I've been considering vmware on a box running vanilla linux, with multiple vm's running various distros with various RTOS patches
[02:19:45] <jmkasunich> its usefull as a compile farm
[02:19:57] <zwisk> That's probably a good use.
[02:20:08] <jmkasunich> build but don't run, so actual RT performance doesn't amtter
[02:20:11] <jmkasunich> matter
[02:20:21] <jmkasunich> but it needs to have the patched kernel and all the right headers, etc
[02:20:46] <zwisk> I don't know what the percentage speed difference between qemu and vmware is, but if I were doing what you just described, I'd use qemu... It's easy to script :)
[02:20:48] <jmkasunich> right now I run individual machines for the compile farm, but a single box would be nicer to maintain
[02:21:01] <jmkasunich> I tried qemu once, too slow
[02:21:28] <zwisk> Did you use the non free-source accelerator stuff? Makes a big difference. And it's of course getting better all the time.
[02:21:31] <jmkasunich> the existing compile farm is a rack of 200MHz Pentiums with 128M each - not speed demons
[02:21:43] <zwisk> But I'll concede that it's probably still slower.
[02:22:02] <zwisk> oh .. yeah.. I have some of those around :)
[02:22:05] <jmkasunich> but when I tried with qemu on a dual CPU 600MHz P3 box with 768M of ram, it was much slower than the existing farm
[02:22:39] <jmkasunich> took over 24 hours just to install ubuntu on the emulated box
[02:22:44] <zwisk> I guess I've been running it on >1.5ghz boxes. Of course, I'm usually running something like word, because I'm forced to.
[02:22:55] <zwisk> eek. That's harsh.
[02:23:20] <jmkasunich> I only have one 1GHz + box, this one (AMD Sempron 2800+, I think thats about 1.6GHz)
[02:24:11] <zwisk> I've come to like the mini-itx intel mobile processors... they've become much more affordable, are low power, and run most things pretty quick.
[02:24:33] <zwisk> hmm... I mixed motherboards and processors. It's a package deal you know :)
[02:25:03] <jmkasunich> for the compile farm, cases are a pain
[02:25:12] <jmkasunich> (unless I have one fast box and emulate the rest)
[02:25:19] <zwisk> Anyways... I'm off to play with the mill some. Thanks for the chat on hal, jmkasunich...
[02:25:25] <jmkasunich> the existing farm is a blade system, 8 in one box
[02:25:30] <zwisk> I'd totally get 1 fast box and run 20 emulators inside of it! :)
[02:25:31] <jmkasunich> but old and thus slow
[02:26:02] <jmkasunich> 1 fast box = 1 cash outlay
[02:26:05] <zwisk> Think of the savings to your electric bill alone! :)
[02:26:05] <jmkasunich> * jmkasunich is cheap
[02:26:20] <jmkasunich> the existing rack is fairly efficient
[02:26:25] <zwisk> Yeah, me too. :) (cheap)
[02:26:30] <jmkasunich> its running on a single 375W ps right now
[02:26:46] <zwisk> Almost all my computers are hand-me-downs in the 'free' category, on their way to the dump..
[02:26:53] <jmkasunich> ditto
[02:27:00] <jmkasunich> but you don't get "fast boxes" that way
[02:27:17] <zwisk> ell, 1ghz isn't 'fast' anymore...
[02:27:32] <zwisk> The 1.1ghz laptop I'm using right now was acquired that way.
[02:27:34] <jmkasunich> I wonder how the generic $249 computer would run if asked to go 24/7
[02:27:55] <jmkasunich> my best trashpicked boxes are dual 600MHz P3
[02:28:01] <cradek> for about a year (the fans would all quit)
[02:28:06] <jmkasunich> never seen a 1G+ in the dumpster yet
[02:28:29] <jmkasunich> cradek: what does it take to get more reliability?
[02:28:32] <zwisk> These are mostly from friends. Being in silicon valley does have it's advantages on the 'junk' side of things...
[02:28:41] <jmkasunich> I suspect a more expensive consumer grade computer isn't the answer
[02:28:43] <cradek> jmkasunich: I'm not sure you can buy more reliability.
[02:28:45] <jmkasunich> I want server grade
[02:29:00] <cradek> jmkasunich: just have a stock of good quality fans etc.
[02:29:09] <cradek> (and make backups)
[02:29:57] <cradek> old server stuff can be really good (and dumpstered because it's large/slow)
[02:30:09] <jmkasunich> thats the best of what I have now
[02:30:13] <cradek> (haha, you saw mine)
[02:30:30] <jmkasunich> but then the tech lag comes into play, not gonna find a fast one in the dumpster
[02:30:40] <jmkasunich> and I bet the new prices are shocking
[02:30:50] <cradek> fast - good - cheap: pick two
[02:31:10] <jmkasunich> well, good and free is what I've gotten so far
[02:31:14] <Mess> thats MY lines...LOL
[02:31:23] <cradek> I have both fast-cheap and good-cheap
[02:31:52] <Mess> but what about fast good and cheap...
[02:32:00] <jmkasunich> I could get fast-cheap, $249 at microcenter
[02:32:02] <cradek> * cradek helps Mess count to two
[02:32:29] <Mess> but there are 3 options...
[02:32:29] <jmkasunich> its infected with a virus tho (that can be fixed, but I hate the idea that even $0.10 of my money is buying a license instead of hardware)
[02:32:56] <cradek> jmkasunich: me too. I haven't bought a prebuilt machine to this day because of that.
[02:33:05] <jmkasunich> Mess: but you only get to pick 2 - thats the point, everything is a tradeoff
[02:33:27] <jmkasunich> "barebones" machines don't have that issue, but they actually cost more
[02:33:52] <Mess> ive been using that line for 15 yrs... dont tell me you got all 3 i'll not believe... ; )
[02:34:15] <cradek> I still buy at most motherboard/processor/ram and scrounge everything else
[02:34:32] <Mess> yup///
[02:39:53] <jmkasunich> well, back to machining... gotta take the collet chuck off and put the 3-jaw on... pain
[02:40:01] <jmkasunich> * jmkasunich is lazy
[02:40:33] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Oword#Sample_4_a_subroutine_for_helical_hole_milling_using_radius_comp
[02:40:53] <cradek> this is cool, I wrote it for doing my motor mounts
[02:41:14] <jmkasunich> helical hole milling?
[02:41:18] <cradek> yes
[02:41:18] <jmkasunich> wazzat?
[02:41:28] <cradek> cutting a hole with an end mill
[02:41:38] <jmkasunich> one that can't plunge?
[02:41:43] <cradek> right
[02:41:49] <cradek> or doesn't plunge well (all of them)
[02:42:12] <cradek> it uses radius comp to calculate the actual helix radius
[02:42:18] <jmkasunich> heh, I usually ramp down while moving in a straight line somewhere in the waste area, then move to the perimiter and work my way around
[02:42:38] <cradek> not much waste area in these holes
[02:45:33] <jmkasunich> damn, how'd it get so late
[02:46:08] <cradek> sundays are like that aren't they
[02:47:09] <jepler> fenn: I'm not very well-versed in configure. Mostly I copied from the existing stanza for 'ncurses'. maybe I got something wrong...
[02:48:52] <fenn> i have libreadline.so.4 installed, but it doesnt seem to find it (it does find all the header files however)
[02:49:45] <fenn> i'm trying to figure out how to include gcc 3.4 standard header files right now
[02:50:47] <fenn> i dont understand why it doesn't see them - i've got symlinks to it everywhere and it's in /usr/local/include
[02:51:25] <jepler> do you have libreadline.so? That's the name that "-lreadline" will search for
[02:51:37] <jepler> lrwxrwxrwx 1 root root 21 2006-06-25 16:02 /usr/lib/libreadline.so -> /lib/libreadline.so.5
[02:51:54] <jepler> I'm pretty sure that's all AC_CHECK_LIB does
[02:52:28] <fenn> how does ac_check_lib work? is that a command or a macro or what?
[02:52:48] <jepler> it's an autoconf ... thing
[02:53:29] <jepler> I think it's an m4 macro which expands into lines of shell script
[02:54:10] <jepler> configure:6204: checking for readline in -lreadline
[02:54:10] <jepler> configure:6234: gcc-3.4 -o conftest -g -O2 conftest.c -lreadline -lXaw >&5
[02:54:22] <jepler> basically it just tests that this link command succeeds
[02:56:10] <fenn> well it finds it now, maybe i changed something
[02:57:28] <fenn> ah it magically apeared: file /usr/lib/libreadline.so.5 is not owned by any package
[02:57:45] <fenn> ok i'm going to bed before i screw things up any further
[05:47:22] <alex_jon1> alex_jon1 is now known as alex_joni
[05:47:44] <alex_joni> zwisk: hi, glad you're back
[06:02:03] <zwisk> actually, I guess alex_joni said hi over hear... :)
[06:02:05] <zwisk> here... )
[06:02:10] <zwisk> hi alex_joni...
[06:38:02] <alex_joni> hi zwisk
[06:38:08] <alex_joni> was away on another terminal
[06:38:10] <alex_joni> how goes it?
[06:42:02] <zwisk> it goes well.... except I strained my back playing with the mill today. Doh!
[06:50:40] <alex_joni> that's not nice
[06:51:52] <alex_joni> zwisk: like I said, I'll be away the next few days, but I'll put the rpm's on linuxcnc.org now
[06:54:24] <alex_joni> zwisk: http://www.linuxcnc.org/rpm/ (I am downloading right now, eventually they will be all there..)
[06:54:36] <alex_joni> you can go on, and announce it on the emc list
[06:55:00] <alex_joni> if you need assistance around linuxcnc.org either jmkasunich or SWPadnos, and probably a few others can help
[06:55:05] <alex_joni> or me ;)
[07:13:15] <zwisk> thanks alex_joni... sorry I was in another window mysefl! :)
[07:20:41] <alex_joni> it's IRC.. pretty common ;)
[07:21:19] <alex_joni> I was meaning to ask you.. I see there is a SMP package
[07:21:26] <alex_joni> how does that work? did you try it?
[07:21:47] <zwisk> I didn't have an SMP box to try it on, so I don't know... but it builds as part of the standard set of things...
[07:22:04] <alex_joni> ok, I was probably expecting that ;)
[07:22:07] <zwisk> Another reason to get more people playing with 'em to see what works and what doesn't...
[07:22:43] <alex_joni> check /msg ;)
[09:13:51] <alex_joni> ok, packages are all on www.linuxcnc.org/rpm/
[09:14:07] <zwisk> cool... thanks much!
[09:26:33] <alex_joni> nothing to thank me for ;)
[09:26:40] <alex_joni> thank you for making them
[09:26:52] <alex_joni> zwisk: I trust you'll send an email to the lists?
[09:27:06] <alex_joni> oh.. just got it ;)
[11:19:06] <cradek_> cradek_ is now known as cradek
[13:28:52] <jmkasunich> hi guys
[13:29:08] <cradek> hi jmk! day off?
[13:29:15] <jmkasunich> yep, 4 day weekend
[13:29:19] <cradek> fun
[13:29:35] <cradek> I'm at work - there was little competition for parking places today
[13:29:43] <jmkasunich> gonna mow the grass (if its dry enough), then maybe go to HGR and the range
[13:30:00] <cradek> I thought you mowed yesterday or so
[13:30:11] <cradek> what's HGR?
[13:30:14] <jmkasunich> I was going to, but it was raining
[13:30:18] <jmkasunich> HGR industrial surplus
[13:31:15] <cradek> oh cool, I was there
[13:31:19] <cradek> get some good stuff for me
[13:31:42] <jmkasunich> "good stuff" is a little vague ;-)
[13:38:56] <jmkasunich> how bout some cutters for your mill: http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4352254&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[13:44:15] <jmkasunich> better yet, a nice mill... http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4352091&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[13:44:22] <jmkasunich> (like mine)
[13:48:08] <cradek> Description: MISCELLANEOUS TOOLING Price: $1.99
[13:48:10] <cradek> haha
[13:48:28] <cradek> ooh, that's a cool mill
[13:48:56] <jmkasunich> the tooling is $1.99 per lb, pick out what you want. I've gotten good stuff that way
[13:49:19] <cradek> dammit I want to go to HGR again
[13:49:31] <cradek> that was the coolest place, but it's x000 miles away
[13:49:33] <jmkasunich> small stuff like 6-32 taps are pretty cheap at $1.99/lb ;-)
[13:49:38] <jmkasunich> yeah
[13:49:45] <jmkasunich> you have surplus center, I have HGR
[13:50:00] <cradek> surplus center is hardly any good. no machinery.
[13:50:00] <jmkasunich> where's the star trek transporter when you need it!
[13:51:28] <cradek> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4365716&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[13:51:42] <cradek> what's with the extra table behind the head?
[13:52:04] <jmkasunich> I've seen those with two heads mounted (each offset about a foot from center)
[13:52:14] <jmkasunich> you can mill two parts at once
[13:52:18] <cradek> wow
[13:52:35] <jmkasunich> also, they made tracer mills, where one head was a sensor that followed a part while the other one milled
[13:53:29] <cradek> neat
[13:53:44] <cradek> I got my very nice Tek 466 at HGR
[13:54:08] <jmkasunich> nice test equipment at HGR? you gotta be kidding?
[13:54:23] <jmkasunich> most of the electronics I've seen there is beat to death and overpriced
[13:54:30] <cradek> not your experience? it might not be nice by modern standards but I like it
[13:54:36] <cradek> it was $100 iirc
[13:54:49] <jmkasunich> I mean nice as in condition
[13:54:53] <cradek> I plugged it in there and it was fully functional
[13:55:00] <jmkasunich> you lucked out
[13:55:04] <cradek> I cleaned some gunk off of it, but it's in good condition
[13:55:21] <cradek> there were a lot of scopes that weren't so good...
[13:55:28] <jmkasunich> I've seen scope with knobs broken off (and the shafts bent, not a simple relace the knob thing) priced at 399
[13:55:56] <jmkasunich> beefy: http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4362952&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[13:56:14] <cradek> Distance: 858 mi (about 14 hours 50 mins)
[13:56:33] <jmkasunich> ouch
[13:56:37] <jmkasunich> not a casual visit
[13:56:40] <cradek> no.
[13:57:14] <jmkasunich> more beef (CNC beef) http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4358026&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[13:57:42] <cradek> nice
[13:57:52] <cradek> but no tool changer
[14:01:24] <jmkasunich> need some taps? 174 lbs at 3.99 a lb
[14:01:31] <jmkasunich> maybe they have a 5-40 for you
[14:02:26] <cradek> haha
[14:02:40] <jmkasunich> they have 94 pieces of 1/4-19 taps...
[14:02:55] <cradek> wtf?
[14:03:02] <jmkasunich> somebody must have made a mistake setting up the tap grinder ;-)
[14:03:03] <SWPadnos> now that's a common size
[14:03:17] <SWPadnos> or it's a typo, and they're 1/4-18 pipe taps
[14:03:55] <jmkasunich> wonder what size these are? http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4361181&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:04:05] <jmkasunich> and is it 9.99/tap or per box
[14:06:02] <cradek> Searching Inventory by Keyword = sherline No Results Found. Please Try Again.
[14:06:48] <jmkasunich> well of course
[14:06:59] <jmkasunich> they are an _industrial_ surplus place
[14:07:00] <jmkasunich> no toys
[14:07:10] <jmkasunich> ;-)
[14:09:04] <jmkasunich> just the thing for high speed milling of PCBs: 50000RPM electronic spindle
[14:09:05] <jmkasunich> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4352315&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:09:28] <cradek> wow
[14:09:40] <jmkasunich> of course, no electronics to go with it
[14:10:49] <cradek> I thought you had to use air powered stuff to get speeds that high
[14:11:14] <SWPadnos> dental grinders are ~100kRPM, I think (and electric)
[14:12:04] <SWPadnos> hmmm - maybe not so high a speed
[14:14:25] <jmkasunich> I wonder if a die grinder could be used as a high speed spindle
[14:14:26] <jmkasunich> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4356870&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=1
[14:14:52] <cradek> sure looks promising
[14:15:52] <jmkasunich> dunno what the bearing quality and runout is
[14:16:13] <jmkasunich> a hand tool probably doesn't demand the same standard as a rigidly held tool
[14:20:11] <jmkasunich> _old_ drill press, kinda cute http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4359634&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:33:04] <jmkasunich> well damn... the range is closed today
[14:33:15] <jmkasunich> at least I called before I drove 35 miles
[14:36:54] <cradek> here's your james g. biddle partial discharge detector: http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4353279&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=3
[14:37:10] <jmkasunich> ohh, just what I wanted
[14:37:32] <jmkasunich> actually, we have gotten involved in partial discharge testing at work ;-)
[14:38:18] <cradek> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4353305&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=3
[14:38:24] <cradek> nice stepper power supply?
[14:38:40] <jmkasunich> bet its AC
[14:38:50] <cradek> oh
[14:38:52] <jmkasunich> 0-140V at 10A sounds like a standard variac
[14:38:57] <cradek> just a variac
[14:38:58] <cradek> yeah
[14:40:47] <jmkasunich> found a 100KV power supply
[14:41:41] <cradek> just what you need I'm sure
[14:42:14] <jmkasunich> 5V @ 50A ?
[14:42:42] <jmkasunich> 10V @ 100A?
[14:43:45] <jmkasunich> heh, a 640V 160A "power supply"... actually its a DC motor drive from the look of it
[14:44:33] <jmkasunich> ohh, a 200 watt CO2 laser!
[14:44:46] <jmkasunich> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4356274&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:45:11] <jepler> what is a "KVAR"? http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4351605&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:45:28] <jmkasunich> kilo volt amp reactive
[14:45:46] <jmkasunich> that is just set of three big ac rated caps
[14:46:12] <jmkasunich> if you have a factory full of induction motors, they tend to draw a lot of reactive current (inductive, current lags voltage)
[14:46:31] <jmkasunich> the power company doesn't like that, the reactive current heats up the wires but doesn't transfer usefull power
[14:46:37] <jepler> I see
[14:46:43] <jmkasunich> so you put caps on the line to draw leading reactive current, cancel it out
[14:47:01] <jmkasunich> (more than you wanted to know... ;-)
[14:50:20] <jmkasunich> hmm, you gotta take their descriptions with a grain of salt
[14:50:52] <jmkasunich> manufacturer VIOLUX, descr: power supply..... its actually a pallet stacked with chairs
[14:51:45] <jmkasunich> heh, theres one that will run a few geckos: 48V 1750A power supply
[14:51:46] <cradek> you'd probably notice that discrepancy while loading the pallet into the truck...?
[14:53:01] <jmkasunich> dunno, maybe not until you try to hook it up
[14:53:18] <SWPadnos> if you on't notice, you deserve the chairs ;)
[14:53:23] <SWPadnos> don't
[14:53:24] <jmkasunich> sits down in chair, mumbling "where's the damned line cord"
[14:56:23] <jmkasunich> cradek: tired of using double sided tape? http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4350219&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:56:46] <jmkasunich> or this one http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4362013&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[14:57:24] <cradek> nice
[14:57:54] <cradek> I may have something like that somewhere...
[14:58:17] <jmkasunich> you must have more crap than I do
[14:58:26] <cradek> nah, just different crap
[14:58:29] <jmkasunich> things that big I have a pretty good idea what I've got
[14:58:53] <cradek> you don't have a barn then
[14:59:08] <jmkasunich> no
[14:59:18] <jmkasunich> if you have a barn full of crap, then you have more than me
[14:59:53] <SWPadnos> "a barn populated with things", please ;)
[15:00:08] <cradek> haha
[15:00:10] <cradek> thank you
[15:00:13] <jmkasunich> no, I'm quite honest about what it is
[15:00:18] <cradek> and it's NOT full
[15:00:29] <SWPadnos> or possibly "a barn densely populated with things"
[15:00:29] <jmkasunich> yet
[15:00:31] <cradek> I know, because I occasionally fit more things in it
[15:01:18] <jmkasunich> do you ever take things out?
[15:01:34] <cradek> I took my motorcycle out this spring
[15:01:53] <jmkasunich> not what I meant
[15:01:57] <jmkasunich> "crap" things
[15:02:37] <cradek> ummm
[15:02:47] <cradek> I'll admit crap, er, things go in more often than they come out
[15:02:49] <jmkasunich> the whold point of collecting crap is so you have it when you need it
[15:03:01] <cradek> definitely
[15:03:06] <cradek> so more crap, er, things is always better, right?
[15:04:01] <jmkasunich> of course if you're gonna have a decent chance of having the particular variety of widget you need, you gotta have about 50 assorted widgets in the crap pile
[15:04:16] <jmkasunich> multiply by all the various classes of widget, and you get a pretty big pile
[15:05:02] <cradek> and not a 5-40x1.5" screw to be found
[15:05:08] <jmkasunich> see
[15:05:21] <cradek> need another barn
[15:05:56] <jmkasunich> or at least more screws in the one you have
[15:06:20] <cradek> oh I don't look there for screws, I look in the garage
[15:06:26] <SWPadnos> heh
[15:06:37] <cradek> can't find a damn thing in the barn
[15:06:40] <jmkasunich> well thats the problem... the 5-40 x 1-1/2 screws are in the barn
[15:07:14] <jmkasunich> bottom of the 4th mason jar on the 3rd shelf in the back corner, under the spare feezlewops
[15:07:30] <SWPadnos> for every screw, turn, turn, turn
[15:07:31] <SWPadnos> you have a lathe, turn turn turn
[15:07:33] <SWPadnos> and the time for this purpose
[15:07:34] <SWPadnos> ...
[15:07:59] <jmkasunich> need to the screw to fix the lathe
[15:08:13] <cradek> that's what the other lathe is for
[15:08:46] <SWPadnos> see
[15:09:02] <jmkasunich> SWP: he'd need an oven too... 5-40 socket head cap screw needs heat-treat after threading
[15:09:10] <jmkasunich> and its a bitch to make the socket...
[15:09:33] <cradek> oh mine would definitely be a slotted head
[15:09:38] <SWPadnos> just make it a sloted screw and be done with it
[15:09:56] <SWPadnos> so basically a minimum home shop needs 2 lathes, 2 milling machines, 2 bandsaws, and a host of accessories
[15:09:56] <jmkasunich> no class
[15:09:59] <jmkasunich> no class at all
[15:10:06] <SWPadnos> I skipped that class
[15:11:37] <SWPadnos> weird - I have a Bacon number of 2
[15:12:03] <jmkasunich> what's your sausage number?
[15:12:13] <SWPadnos> much lower :)
[15:12:20] <SWPadnos> http://en.wikipedia.org/wiki/Bacon_number
[15:12:35] <jmkasunich> I think I know what you are talking about
[15:12:51] <SWPadnos> so you have a bacon number of 3 - scary, huh?
[15:13:38] <jmkasunich> I thought that only applied to "films appeared in", not "persons acquainted with"
[15:13:51] <jmkasunich> since my films appeared in is zero, my bacon number is 1/0
[15:14:02] <SWPadnos> I thikn it's more generalized to "worked on projects with"
[15:15:18] <SWPadnos> assuming it's relazed to include crew, as the page notes, I'm still at 2, but you'd be inf/NaN (along with most of the world, thankfully)
[15:15:22] <SWPadnos> relaxed
[15:18:04] <cradek> haha http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4360136&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=8
[15:18:35] <jmkasunich> $399! they must be outa their freaking mind
[15:19:16] <jmkasunich> what is it, unobtainium core with gold foil shielding?
[15:19:57] <cradek> it doesn't have ends on it or anything, I wonder what they think is so special about it
[15:20:05] <jmkasunich> dunno
[15:20:10] <jmkasunich> maybe another screwup?
[15:20:23] <jmkasunich> $3.99 maybe
[15:20:34] <SWPadnos> maybe it's high quality optical fiber? ;)
[15:21:22] <jmkasunich> yeah, a 500meter piece thats been cut into 20 meter chunks to make it more usefull (note the number of visible ends)
[15:21:43] <SWPadnos> reminds me of the Monty Python skit "String"
[15:21:47] <cradek> me too
[15:21:51] <cradek> due to poor planning...
[15:22:01] <SWPadnos> it's been cut into 3 inch lengths ...
[15:22:03] <cradek> haha
[15:22:16] <cradek> it's cablettes!
[15:22:21] <SWPadnos> "have you ever seen a hospital without string?"
[15:22:40] <SWPadnos> Emperor Calble-ettes, as used in Hospitals
[15:22:46] <cradek> maybe it's waterproof cabling
[15:22:54] <SWPadnos> maybe it was
[15:22:56] <cradek> or water-absorbent
[15:23:02] <SWPadnos> he
[15:23:04] <SWPadnos> h
[15:23:15] <SWPadnos> it's Super Absorbent
[15:23:54] <SWPadnos> holy crap
[15:23:58] <SWPadnos> http://www.mattwalsh.com/twiki/bin/view/Main/JokeEntryMontyPythonStringSketch
[15:24:44] <cradek> Ah, but there's a snag, you see. Due to bad planning, the hundred and twenty-two thousand miles is in three inch lengths.
[15:24:59] <cradek> hahaha
[15:26:02] <SWPadnos> that's the epitome of marketing :)
[15:29:21] <jmkasunich> time to mow the grass, the dew should be burned off by now
[15:29:34] <jmkasunich> if anybody finds something (portable) they want from HGR, lemme know
[15:30:59] <cradek> will do, thanks
[16:10:01] <fenn_> jmkasunich: if you see any small knurling or parting tools in the bulk bins i'd be grateful if you got them for me
[16:10:12] <fenn_> fenn_ is now known as fenn
[16:11:17] <fenn> he's probably gone already..
[16:28:48] <jmkasunich> still mowing (came in to take a break, re-hydrate)
[16:29:01] <jmkasunich> its only medium hot, but very humid
[16:40:24] <SWPadnos> jmkasunich, can you take a look at the selection of 5C collets/ collet chucks?
[16:40:44] <jmkasunich> theres a lot of 5C collets
[16:40:51] <jmkasunich> dunno about chucks
[16:41:03] <SWPadnos> I saw a bin of 5C collets for $4.99 (each, I assume)
[16:41:09] <jmkasunich> collet quality varies, do you have any particular sizes you are looking for?
[16:41:18] <SWPadnos> "standard sizes"
[16:41:57] <SWPadnos> there's a batch of Hardinge collets for $6.99 each as well
[16:42:10] <jmkasunich> what sizes do you already have?
[16:42:17] <SWPadnos> well - let's see
[16:42:19] <SWPadnos> none ;)
[16:42:28] <jmkasunich> what do you have to stick them in?
[16:42:36] <jmkasunich> (lathe spindle, etc?)
[16:42:40] <SWPadnos> I should be getting a lathe Real Soon Now
[16:42:51] <jmkasunich> you know that it takes 5C?
[16:43:11] <SWPadnos> it will be either 5C or 16C, and I can get an adapter for 16C -> 5C
[16:43:22] <SWPadnos> not sure how well those work thoug
[16:43:23] <SWPadnos> gh
[16:43:25] <jmkasunich> ok - so you have a specific lathe in mind
[16:43:25] <SWPadnos> h
[16:43:27] <SWPadnos> yep
[16:44:23] <jmkasunich> wonder what size collets this takes? http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4352870&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[16:44:32] <jmkasunich> and what spindle nose thread that is
[16:44:49] <jmkasunich> I have one for 5C, they are very nice
[16:44:50] <SWPadnos> it looks big
[16:45:01] <jmkasunich> the thread? yeah
[16:45:20] <jmkasunich> overall diameter is probalby 9-10" or so
[16:45:48] <jmkasunich> the spindle thread is probalby on a removable backplate - mine was, I removed it and made a backplate to fit the shoptask spindle
[16:46:49] <SWPadnos> that looked like something I've seen advertised as "step chucks"
[16:47:09] <jmkasunich> they pictured it from the back, dunno why
[16:47:24] <SWPadnos> to shouw the mounting scheme?
[16:47:25] <jmkasunich> the front takes a collet, and you spin the ring (that looks like a steering wheel) to tighten
[16:47:28] <jmkasunich> maybe
[16:48:40] <jmkasunich> SWPadnos: you can get new enco brand collets for 3.99 it seems
[16:48:56] <SWPadnos> ok, maybe the used ones aren't a good idea then ;)
[16:49:20] <jmkasunich> hardinge beats the heck out of enco, IF they're not worn out, sprung, or otherwise busted
[16:50:15] <jmkasunich> they have Lyndex (not hardinge, but not no-name either) for $14
[16:50:46] <SWPadnos> ok. I've seen a lot on eBay as well (I love the term "emergency collet" ;) )
[16:51:15] <jmkasunich> complete set of enco brand 1/16 to 1-1/8 by sixteenths is 64.95
[16:51:29] <jmkasunich> I have a set, they've served pretty well
[16:51:42] <jmkasunich> you can tell they're chinese (finish mostly) but they work fine
[16:52:23] <SWPadnos> I'm sure that would be good enough for me, and if it's not, I can buy the one or two sizes I need from a really good brand
[16:52:35] <jmkasunich> that was my thought when I bought the set
[16:52:44] <jmkasunich> haven't bought any more since ;-)
[16:52:48] <SWPadnos> heh
[16:53:02] <jmkasunich> http://www.use-enco.com/CGI/INPDFF?PMPAGE=285&PARTPG=INLMK32
[16:53:09] <jmkasunich> you have a bport, don't you?
[16:53:13] <SWPadnos> yep
[16:53:29] <jmkasunich> any R8 sizes you need?
[16:53:34] <SWPadnos> nope
[16:53:44] <SWPadnos> I sold a bunch at the workshop, in fact :)
[16:53:49] <jmkasunich> ah
[16:54:22] <jmkasunich> do you do any woodworking?
[16:55:10] <SWPadnos> yep
[16:55:19] <jmkasunich> collets need a rack/box
[16:55:35] <jmkasunich> I made a couple collet boxes I'm pretty pleased with
[16:55:45] <SWPadnos> yep - I was just looking at those on the enco page
[16:55:56] <jmkasunich> take a piece of 1xX, where X is the height of a collet + 1" or so
[16:56:13] <jmkasunich> put a slot to fit 1/8 masonite 1/4" from the top and bottom
[16:56:41] <SWPadnos> masonite == some concrete thing, or the plastic countertop stuff?
[16:56:48] <jmkasunich> put a slot to fit 1/8" alum down from the top, by the length of the tapered part of the collet + 3/8"
[16:57:06] <jmkasunich> make two masonite rectangles and one alum one
[16:57:23] <jmkasunich> make holes in the alum one that the collets will fit in (like the enco rack)
[16:57:36] <jmkasunich> cut the 1xX into 4 pieces, front, back, two sides
[16:57:50] <jmkasunich> glue up with masonite and alum in the slots to make a box
[16:57:57] <jmkasunich> then cut the box open with table say, add hinges
[16:58:11] <jmkasunich> (cut between alum and top masonite, of course)
[16:58:40] <jmkasunich> keeps em clean, and you can stack stuff on top, unlike a regular rack
[16:58:44] <SWPadnos> ah - ok, I was wondering where the aluminum went :)
[16:58:57] <SWPadnos> hmmm - rack == table. that could be bad
[16:59:29] <jmkasunich> when I got my set, it actually included the 12.95 enco rack on that catalog page
[16:59:46] <jmkasunich> I made the box using the top part of the rack, threw away (well, actually, saved) the sides
[16:59:59] <SWPadnos> heh
[17:00:07] <jmkasunich> I made another box for some smaller collets, using aluminum for the plate
[17:00:37] <SWPadnos> I'd probably end up doing something closer to the Mahogany R8 collet stand they have
[17:00:38] <jmkasunich> (I got some swiss screw machine collets on ebay, very nice, 1/16 (I think) to 1/2, by 64ths
[17:01:06] <jmkasunich> pah... they get dirty, can't stack stuff on it, hard to move to a shelf, etc
[17:01:15] <jmkasunich> I should take a pic of my boxes
[17:01:18] <SWPadnos> can't stack stuff on it is a feature ;)
[17:01:36] <jmkasunich> not if the entire stack is on a shelf, out of the way
[17:02:31] <SWPadnos> interesting - MSC only carries Lyndex collets in the 16C size
[17:06:53] <jmkasunich> http://home.att.net/~jmkasunich/Pics/collets.jpg
[17:07:28] <SWPadnos> ok - cool
[17:07:53] <SWPadnos> basically like the standard wooden box that some measurement tools come in
[17:07:57] <jmkasunich> yeah
[17:08:27] <jmkasunich> but the construction is easy... three slots (table saw), then cut four pieces to length and glue up
[17:08:45] <SWPadnos> yep
[17:09:03] <jmkasunich> (glue up as a closed box, then cut open... make sure to mark which end is the top ;-)
[17:09:19] <SWPadnos> or 3 slots (router), then cut to length (slide compound miter saw), glue ...
[17:09:41] <jmkasunich> yeah, if you have a narrow router bit, and a router table with a fence
[17:09:50] <SWPadnos> check, check :)
[17:10:40] <jmkasunich> something else to consider if you are getting 5C collets
[17:10:40] <jmkasunich> http://www.use-enco.com/CGI/INSRIT?PMAKA=235-7050&PMPXNO=950317&PARTPG=INLMK3
[17:10:41] <SWPadnos> my little table saw is such a pain for dadoing, that I almost gave up on it
[17:11:25] <SWPadnos> err - what is that set, compared to "normal" collets?
[17:11:27] <jmkasunich> I've see collet block sets for $20 on sale (thats how I got mine)
[17:11:40] <jmkasunich> it isn't collets (the capacity note is confusing)
[17:11:44] <jmkasunich> it holds collets
[17:11:51] <SWPadnos> it's not immediately obvious how to use those pictured items
[17:11:53] <SWPadnos> ah
[17:12:13] <jmkasunich> stick a collet in either the square or hex block, using the ring shaped nut to tighten
[17:12:27] <SWPadnos> so you can "easily" make a non-round collet holder, with the closer on it
[17:12:30] <jmkasunich> (or the lever thing, quick release, but I usually use the round nuts)
[17:12:45] <SWPadnos> ok, for holding collets in other things - thanks
[17:12:49] <jmkasunich> then you can do things like mill hex heads on homemade bolts
[17:13:02] <jmkasunich> clamp things in a vise to cross drill on the mill
[17:13:40] <jmkasunich> I just used mine to cut a 1" long piece of 3/4 round rod in half on the cheapo bandsaw
[17:13:51] <jmkasunich> too short to clamp in the vise normally
[17:14:23] <SWPadnos> ah - now that's a sitaution I've run into
[17:14:41] <SWPadnos> that's the problem with really big tools - little stuff is hard to hold
[17:14:48] <jmkasunich> its one of those tools that you don't really think you need, but once you have it, it can be very handy
[17:17:34] <jmkasunich> well, I should go finish the grass
[17:17:52] <SWPadnos> ok. I should figure out why these 41 AVR chips on my desk don't work :(
[17:18:07] <jmkasunich> you want any collets? or are you just gonna buy a set?
[17:18:23] <SWPadnos> I think I'll just buy a set later. thanks
[17:18:26] <jmkasunich> ok
[20:37:00] <alex_joni> alex_joni is now known as _aj
[20:37:19] <SWPadnos> SWPadnos is now known as __
[20:37:55] <__> __ is now known as SWPadnos
[20:37:59] <_aj> _aj is now known as alex_joni
[20:38:19] <SWPadnos> had to change, __ might have come back and ghosted me ;)
[20:45:51] <jepler> SWPadnos: compared to 16 seconds
[20:46:03] <SWPadnos> ok - a doubling is nice
[20:46:14] <SWPadnos> or halving, depending on your perspective
[20:46:45] <cradek> slick
[20:47:06] <jepler> yeah, unfortunately I've now picked all the low-hanging fruit (though I've said that before)
[20:47:19] <cradek> that's ok, if the result is that it's fast now
[20:47:33] <cradek> last night I was wondering if I could pull off named variables somehow
[20:47:45] <SWPadnos> do you have profiler runs that you can post somewhere?
[20:47:56] <cradek> even of the form #[a-z]+
[20:48:04] <SWPadnos> so others who are boneheaded (liek me) about profiling can see where to apply effort
[20:48:18] <cradek> you'd have a map of strings to numbers
[20:48:27] <cradek> not sure how you'd "name" the variables though
[20:48:46] <SWPadnos> are those vars saved between runs?
[20:48:58] <cradek> the numbered variables are, I think
[20:49:06] <SWPadnos> so #37 would be?
[20:49:07] <cradek> I think the name-number mapping would be in the gcode
[20:49:22] <cradek> yes I'm pretty sure #37 is saved forever
[20:49:44] <cradek> the trick is pulling off #argh meaning #37
[20:49:48] <SWPadnos> hmmm, and the parameter file is loaded into an array of doubles?
[20:49:52] <cradek> yes
[20:50:04] <SWPadnos> either that or just have a variable file that can have names in it
[20:50:31] <cradek> that doesn't help subroutines whose vars are passed in #1, #2, #3
[20:51:03] <jepler> SWPadnos: you can use 'kcachegrind' to visualize this profile: http://emergent.unpy.net/files/sandbox/callgrind.out.17093.bz2
[20:51:13] <SWPadnos> what I mean is that the variables would be in an associative array, rather than a linearly indexed one
[20:51:18] <SWPadnos> ok - thanks
[20:51:26] <cradek> I was picturing specifying the mapping in the beginning of the subroutine (#xpos,#ypos,#zpos,#depth) = $_
[20:51:50] <SWPadnos> so if a subroutine wants a "Zstep" variable, it uses #ZStep instead of #37
[20:52:03] <SWPadnos> and the array is indexeed with "ZStep" instead of the number 37
[20:52:23] <SWPadnos> that would also be cool
[20:52:25] <cradek> right, I understand, but that doesn't solve my particular problem
[20:52:39] <cradek> where my subr is passed 7 things and I want them to have names
[20:52:46] <SWPadnos> you could do sub (#xpos=10,#ypos=11 ...)
[20:52:49] <cradek> I guess I could do #name1 = #1 etc at the beginning
[20:52:58] <SWPadnos> sure could ;)
[20:53:08] <jepler> all your attempts to make actual programming in gcode not suck will fail
[20:53:19] <cradek> naysayer!!
[20:53:22] <jmkasunich> jepler: right!
[20:53:46] <jepler> give up now and find a nice way to let me write programs in the language of my choice, with calls to the canon interface
[20:53:51] <jmkasunich> if you want to do actual programming, use an actual programming language, and have it spit out g-code
[20:53:54] <SWPadnos> I'd say that the task of defining anything 3D in ASCII is going to suck
[20:53:59] <jmkasunich> jepler: canterp?
[20:54:12] <jmkasunich> hey, anybody need a tapping head?
[20:54:13] <cradek> canterp already does that, but I have nfc how to run it
[20:54:24] <SWPadnos> sold one at CNC Workshop ;)
[20:54:35] <SWPadnos> (got $300 or $350 for it, too :) )
[20:54:43] <jmkasunich> you sold a huge one didn't you?
[20:54:51] <SWPadnos> Tapmatic 90 TC/DC
[20:54:52] <cradek> I'd love one that would work on a smallish drill press
[20:54:58] <jmkasunich> http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4358454&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[20:55:00] <SWPadnos> 30 TC/DC
[20:55:10] <jmkasunich> thats not the exact one... (it has a jacobs chuck on it)
[20:55:16] <jmkasunich> the ones I saw aren't on the web
[20:55:19] <jmkasunich> but the same basic head
[20:55:30] <jmkasunich> morse 2 taper, fits many 12-16" drill presses
[20:55:52] <cradek> eek $200
[20:55:55] <jmkasunich> probably suitable for #4 thru maybe 1/4"
[20:56:04] <jmkasunich> actually, the ones I saw were 89.99
[20:56:14] <jmkasunich> a box of em, maybe 6-8
[20:56:39] <jmkasunich> dunno why the aren't on the web (I'm dissapointed, all I jotted down was the number, figured I could get more info later)
[20:56:47] <SWPadnos> http://cgi.ebay.com/260002483869
[20:56:54] <SWPadnos> oops
[20:57:10] <SWPadnos> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=260002483869
[20:57:51] <SWPadnos> the 30x, 30TC/DC and SPD-3 generally go for ~$100 on eBay
[20:58:42] <cradek> * cradek is pretty good at tapping by hand
[20:59:07] <jmkasunich> my purchase: http://www.hgrindustrialsurplus.com/search-products/product-detail.aspx?id=4354082&searchtable=1&sortExpression=&SortASC=&pageSize=50&currentPageIndex=0
[20:59:08] <SWPadnos> tapmatics are among the slickest things you can find
[20:59:35] <cradek> jmkasunich: what's it for?
[20:59:37] <jmkasunich> 1/4HP 10,000 RPM motor, and high speed spindle, for $150
[21:00:13] <jmkasunich> grinding stuff ;-)
[21:00:46] <jmkasunich> although I have thoughts about making a high speed (for instance PCB) millin spindle with it
[21:01:23] <cradek> how fast can the spindle go?
[21:01:34] <cradek> you could easily make 1:3 pulleys
[21:02:06] <jmkasunich> its designed to do 10,000 and 22,500
[21:02:27] <cradek> cool, 22k would be perfect
[21:02:51] <jmkasunich> * jmkasunich takes apart the motor
[21:03:03] <cradek> bet it's loud
[21:03:24] <jmkasunich> well poo
[21:03:43] <jmkasunich> I knew it needed bearings when I bought it (motor, not spindle)
[21:03:53] <jmkasunich> thought they were ball, just drop in new ones
[21:03:57] <jmkasunich> turns out its bronze
[21:04:14] <cradek> hmm
[21:04:40] <jmkasunich> musta run it without oil (the oil holes are kinda clogged up)
[21:05:29] <anonimasu> just bushings?
[21:05:47] <jmkasunich> yeah
[21:06:23] <anonimasu> hm, why not turn new ones?
[21:07:03] <jmkasunich> not out of the question...
[21:07:07] <jmkasunich> measuring things now
[21:07:26] <jmkasunich> shaft is 0.250 nominal, worn to 0.246-0.248 in the bearing area
[21:08:25] <jmkasunich> bearing is worn oval (due to belt side load), 0.251 x 0.258
[21:08:39] <jmkasunich> pretty easy to press out the old bearing, press in a new one
[21:08:47] <jmkasunich> not so easy to fix the shaft
[21:08:52] <cradek> 258-246 = a lot
[21:09:05] <jmkasunich> yeah, like I said I knew it needed bearings
[21:09:29] <jmkasunich> thats the pulley end
[21:09:39] <jmkasunich> brush end, shaft is 0.2500
[21:10:15] <cradek> turn the bushing 180 degrees, oil it up and call it good?
[21:10:21] <jmkasunich> fsck no
[21:10:31] <cradek> heh
[21:10:35] <jmkasunich> replacing the bearing is trivia
[21:10:38] <jmkasunich> l
[21:10:50] <anonimasu> jmkasunich: heat it with the tig and add metal ;)
[21:10:50] <jmkasunich> but that will still mean 2 to 4 thou slop
[21:10:54] <anonimasu> then file it down ;)
[21:11:18] <jmkasunich> anonimasu: I'd like the shaft to remain straight please... no heat
[21:11:35] <anonimasu> jmkasunich: ah guess you are out of luck
[21:11:35] <cradek> does it have dents for turning between centers on both ends of the shaft? if you make a custom bushing you could just make it a bit smaller.
[21:11:45] <jmkasunich> dents? ;-)
[21:11:51] <cradek> yeah you know
[21:11:56] <jmkasunich> yes, its center drilled
[21:11:59] <cradek> yeah that
[21:12:03] <jmkasunich> lol
[21:12:07] <cradek> so make it smaller?
[21:12:44] <cradek> I'd be surprised if you can't buy undersized bushings for that reason
[21:12:45] <jmkasunich> the ideal tool for that would be.... a toolpost grinder!
[21:12:59] <jmkasunich> I can always make undersize bushings, just bore em
[21:13:03] <jmkasunich> the problem is the shaft
[21:13:25] <jmkasunich> its threaded 1/4-<something fine> on the end for a pulley
[21:13:30] <cradek> it has to be smoother than lathe work?
[21:13:49] <jmkasunich> the bearing part should be a mirror
[21:14:00] <anonimasu> hm, 2 to 4 thou is lots :S
[21:14:05] <jmkasunich> yeah
[21:14:07] <jmkasunich> :-(
[21:14:17] <anonimasu> something like 0.02
[21:14:35] <cradek> mirrors are easy, flat mirrors not so much
[21:14:42] <anonimasu> or 0.058mm
[21:14:51] <jmkasunich> cylinderical mirros
[21:14:54] <anonimasu> what kind of lathe do you have?
[21:15:05] <cradek> yeah, I meant flat in ... polar coordinates
[21:15:05] <jmkasunich> a crappy one - chinese 3-in-1
[21:15:12] <jmkasunich> but thats not a major issue
[21:15:20] <anonimasu> hm ok
[21:15:26] <anonimasu> I'd die if my lathe did that
[21:15:27] <anonimasu> :)
[21:15:36] <jmkasunich> chuck the brush end, run the pulley end on a dead center
[21:15:46] <jmkasunich> runout = 0.0000 with a dead center
[21:16:04] <anonimasu> I'm down to 0,01mm, with a long workpiece
[21:16:06] <jmkasunich> anonimasu: its not really a bad lathe, just not a hardinge or anything
[21:16:07] <cradek> sure but is your cylinder a cylinder?
[21:16:14] <jmkasunich> not this one
[21:16:28] <anonimasu> I hate micro machines :)
[21:16:56] <jmkasunich> tapered from 0.246 at the outboard end to 0.248 at the inboard end (of a 3/4" long bearing area)
[21:17:41] <cradek> jepler: if you make a new interpreter that's not gcode I'll be the first to try it
[21:19:16] <jmkasunich> hmm, maybe put a ball bearing on it, using this:
[21:19:41] <jmkasunich> trick will be to have the bearing centered while the stuff dries
[21:19:53] <cradek> you're going to glue a bearing to the shaft?
[21:20:19] <jmkasunich> "fills gaps to 0.008"
[21:20:25] <jmkasunich> thats exactly what its intended for
[21:20:32] <cradek> wild
[21:20:36] <jmkasunich>
[21:20:39] <jmkasunich> see the pic?
[21:21:22] <cradek> huh
[21:21:47] <anonimasu> hm