#emc | Logs for 2007-10-19

[00:22:44] <SkinnyPuppy34> knock knock...
[00:24:23] <SkinnyPuppy34> Trying to configure my stepper ini. My mill has .2 inch lead on the xy screws.
[00:26:37] <SkinnyPuppy34> 1.8 deg steppers, 360/1.8=200, 200*5=1000 But setting input scale=1000 only gives 1/2 revolution on the motors.
[00:27:29] <SWPadnos> G203V?
[00:27:56] <SkinnyPuppy34> g203v yes, I set to 10k and it looks like 5 rev but thought I'd ask
[00:28:13] <SWPadnos> ok, 10000 - you wrote 1000
[00:28:54] <SWPadnos> 10k is cirrect - you forgot that the G203V is a 10microstep drive
[00:28:57] <SWPadnos> correct
[00:29:01] <SkinnyPuppy34> Yeah I didn't quite get 2 rev at 4000, I plugged 1000 from my calc, but 10000 is what gave 5 rev
[00:29:13] <SWPadnos> *10 microsteps/step
[00:29:28] <SkinnyPuppy34> Ok, wasn't thinking of the microstepping since it was advertised automatic
[00:29:42] <SWPadnos> yes, but it still does it ;)
[00:30:04] <SkinnyPuppy34> I see, thank you SWPadnos
[00:30:10] <SWPadnos> sure
[00:30:26] <SWPadnos> the general formula is microsteps * steps/rev * drive ratio * screw pitch
[00:43:52] <SkinnyPuppy34> Thats a releif, step calculator looks like I can still get a decent rapid.
[02:13:58] <Skullworks-PGAB> http://www.youtube.com/watch?v=mivM7vrQKmc
[02:25:19] <LawrenceG> I hate it when the OS crashes at the "right moment"
[02:25:53] <LawrenceG> JymmmEMC: you about?
[02:30:09] <Skullworks-PGAB> http://www.youtube.com/watch?v=eNqPTOb31S8
[02:32:27] <SkinnyPuppy34> Ha.. del win.* /s
[02:32:54] <SkinnyPuppy34> grrr if it weren't for mastercam...
[02:38:47] <Skullworks-PGAB> You mean deltree C:\windows /y
[02:42:03] <SkinnyPuppy34> yep
[02:43:01] <SkinnyPuppy34> This is too funny from one of the youtube sidebars
[02:44:11] <SkinnyPuppy34> http://www.youtube.com/watch?v=JahdnOQ9XCA
[02:48:31] <Skullworks-PGAB> that one needed another 15 seconds...
[04:11:14] <JymmmEMC> SWPadnos: ping?
[04:12:49] <fenn> blert?
[04:13:13] <SWPadnos> JymmmEMC, can it wait a few days?
[04:13:25] <JymmmEMC> SWPadnos: ?
[04:13:33] <SWPadnos> whatever it is you're pinging me about ;)
[04:13:57] <JymmmEMC> I could have wanted to see how your are doing, but sure, I guess so.
[04:14:16] <fenn> tell us how you are doing in 3 days :)
[04:14:18] <SWPadnos> heh - if that's the case. I'm doing tired and stressed
[04:14:42] <JymmmEMC> SWPadnos: I got a bottle of cognac
[04:14:53] <SWPadnos> I may make more tea
[04:17:13] <JymmmEMC> SWPadnos: But, you can answer the question I had in a few days.... I want to add a subnet /29 to a host, without using an alias (eth0:30), is there a way to do it?
[04:19:02] <SWPadnos> so you want a /29 on the same interface as something other than a /29?
[04:19:17] <JymmmEMC> No, just the /29
[04:19:25] <JymmmEMC> But I've been told the only way to
[04:19:36] <JymmmEMC> add more than one ip address is to use aliases
[04:19:51] <ds2> 'add' a subnet?
[04:20:00] <ds2> or you want to be able to access it?
[04:20:08] <JymmmEMC> ds2: a subnet in this sense ia a range of ip addresses
[04:20:23] <fenn> what's wrong with using an alias?
[04:20:44] <JymmmEMC> fenn: I don't want to add 40 or 50 aliases.
[04:20:47] <ds2> JymmmEMC: trying to figure out what you mean by add it... add it as in access it w/o burning an address on there or ???
[04:21:00] <SWPadnos> are you trying to add a route, or a net?
[04:21:11] <SWPadnos> "subnet /29 to a host" doesn't make sense to me
[04:21:29] <JymmmEMC> SWPadnos: a range of ip address so they all respond from the host
[04:21:53] <ds2> sure
[04:22:00] <ds2> there are a bunch of ways
[04:22:08] <SWPadnos> ok, and you said "add a /29" - that makes me think you already have a net on that interface
[04:22:28] <JymmmEMC> SWPadnos: I just got assigned a /29 range
[04:22:28] <SWPadnos> ie, it has an address and netmask already
[04:22:33] <SWPadnos> ok
[04:22:55] <ds2> you can assume them to dummy interfaces
[04:23:35] <ds2> or you can apply the foreign listen patch and use the foreign listen interface; the patch was written for squid to do transparent proxying
[04:23:58] <JymmmEMC> SWPadnos: I just can't find the doc on how to add a range of ips to a host is all
[04:24:17] <JymmmEMC> SWPadnos: or even multiple non-concurrent subnet ranges
[04:26:58] <fenn> o
[04:27:16] <fenn> fingers in home position; resume typing
[04:27:41] <fenn> JymmmEMC: i've only ever seen it done with aliases; you can use lo:* if you dont want it tied to a particular ethernet card
[04:29:14] <SWPadnos> for i in `seq 1 6` ; do ifconfig (however you add) 192.168.1.$i -netmask ; done
[04:32:59] <JymmmEMC> SWPadnos: so basically , an alias per ip
[04:33:29] <SWPadnos> I don't know if there's another way
[04:33:42] <toast> or is there.
[04:34:26] <JymmmEMC> SWPadnos: ok, fair enough. I just thought that most admins are way too lazy to add one ip per line and would have a way to say like: add
[04:34:48] <fenn> JymmmEMC: http://www.debian-administration.org/users/lykwydchykyn/weblog/5
[04:34:56] <SWPadnos> they'd probably be lazy enough to automate the process, so it wouldn't matter how "hard" it is ;)
[04:35:09] <fenn> the important bit is 'up ip addr add dev $IFACE a.b.c.d/nm'
[04:35:52] <JymmmEMC> fenn: now THAT looks more like it
[04:35:53] <fenn> if you get the netmask right and use the broadcast address, does it respond for the entire range?
[04:36:06] <SWPadnos> I think that specifies the netmask for that IP address, not that it should add all addresses in that range
[04:36:11] <SWPadnos> no
[04:36:36] <JymmmEMC> fenn: I wouldn't know just yet.... I'm trying to configure the box then take it and plug it in at work
[04:36:45] <SWPadnos> I don't think stations are supposed to respond to broadcast packets, except in special circumstances (like dhcp requests, for example)
[04:37:20] <JymmmEMC> SWPadnos: well, think like a virtual host with 255 ip address
[04:37:45] <JymmmEMC> with each ip bound to a virtual domain
[04:38:01] <JymmmEMC> I think I said that last one wrong =)
[04:38:53] <fenn> JymmmEMC: by virtual domain do you mean an apache virtual server with its own dns entry?
[04:39:47] <JymmmEMC> fenn: For example... to have https, it has to have it's own ip address per domain.
[04:40:25] <JymmmEMC> so the host has to respond to more-than-one ip address, in this case, a range of ip's like 100-127
[04:40:39] <fenn> the way we do it is with aliases in ifconfig.. it works for 8000+ websites, so meh
[04:41:09] <JymmmEMC> fenn: so if you have 8000 ip addresses, you are going to add one alias per ip address?
[04:41:31] <fenn> yep, dont type ifconfig without some kind of piped command
[05:14:49] <fenn> i was talking to a guy today who works with pico-amp current sensors
[05:15:04] <fenn> there's a special room he goes into to test them
[05:16:00] <fenn> i bet you could pick up all kinds of weird stuff with one of those
[05:17:08] <wholepair> all kinds of small stuff - is the room small?
[05:20:57] <wholepair> what are pico-amp current sensors used for? Neurons?
[05:21:11] <fenn> dunno, he works at a cyclotron
[05:23:20] <wholepair> particle accelerator - probably part of some kind of a detector that looks for "events"
[05:23:40] <fenn> yeah thats what i was thinking
[05:24:31] <fenn> you can record the current along a bundle of wires and deduce the path of a particle from the shape of the signal
[05:25:09] <fenn> assuming its traveling fast
[05:26:01] <wholepair> I worked with a professor on a scintilator detector that used fiber optics for look for cherenkov radiation given off in the fiber when particles pass through it. - similar -
[05:26:21] <wholepair> I have some pistures on my flickr account - not that exciting -
[05:26:30] <fenn> i'm trying to figure out how to build neutron counters
[05:26:51] <wholepair> why
[05:27:03] <fenn> uh, because :)
[05:27:14] <fenn> to develop a fusion reactor
[05:28:41] <fenn> seems the usual way is to catch the neutron with a nucleus that will become unstable - in real world terms this means a piece of indium(?) foil
[05:29:07] <fenn> then you can watch for the nucleus decaying
[05:29:09] <wholepair> indium - thats expensive - and smooshy
[05:29:23] <fenn> smooshy?
[05:29:35] <wholepair> soft - like lead -
[05:29:41] <fenn> oh, that's fine
[05:30:01] <wholepair> it makes a good gasket in ultra high vaccum systems
[05:31:06] <fenn> hey this is always nice: Indium is a wonderful metal because unlike so many soft, silvery, low-melting-point metals it's neither dangerously reactive nor toxic.
[05:32:28] <wholepair> why are you developing a fusion reactor again -
[05:32:49] <fenn> to launch tons of crap into orbit
[05:32:56] <fenn> and other stuff
[05:33:45] <wholepair> have I talked to you before? - im not on here often - i just got Pidgin on my laptop and was trying out the irc chat functionality on the only chat room i have been to before -
[05:34:13] <fenn> i remember you were making some overlapping wave patterns using lost foam and bronze?
[05:34:58] <wholepair> yeah that was me - you where doing the lost foam I think - I have never done that - how is emc doing - any new and cool projects -
[05:35:34] <fenn> there's tons of new stuff in 2.2 (or will be)
[05:36:16] <fenn> i just made a hexapod simulator: http://fenn.dyndns.org/pub/irc/hexapod_sim1.png
[05:36:33] <fenn> probably not too interesting unless you're making hexapods :)
[05:36:55] <fenn> the stepper driver now does double the step rate
[05:37:07] <fenn> * fenn is editing the changelog right now actually
[05:38:02] <fenn> lathe threading, rigid tapping, UVW axes, wait-for-input g-codes
[05:40:11] <CIA-8> 03fenn 07TRUNK * 10emc2/debian/changelog: hope i did this right
[05:40:27] <wholepair> I know what a hexapod is - but that picture is bizzare - is that something that would show live movement in AXIS
[05:40:33] <fenn> yeah
[05:40:42] <wholepair> do you have a working machine?
[05:40:42] <fenn> hang on, will take a better pic
[05:40:52] <fenn> no, not yet
[05:41:03] <fenn> <insert lame excuse here>
[05:41:34] <wholepair> <I was given a labatomy in my sleep by aliens>
[05:43:59] <wholepair> ha - did you read accelerando?
[05:44:07] <fenn> huh. now i forgot how to run a program in tkemc
[05:44:16] <wholepair> I read tha a couple months ago -
[05:44:20] <fenn> wholepair: yep, some of it was pretty good
[05:44:36] <wholepair> matrioshka brains -
[05:45:04] <fenn> i'll let the machines figure that one out
[05:45:26] <fenn> i dont think a shell is orbitally possible, actually
[05:46:05] <fenn> the heavier side would tend to go towards the sun and then it would be closer so it would have more gravity pulling on it
[05:46:36] <fenn> you would have to balance it perfecly and then constantly nudge it back on center
[05:47:05] <wholepair> unstable? works off the waste heat of the lower radius shells - good point probably defies some basic law of thermodynamics - that book somehow got me playing around with TiddlyWiki's - that was a waste of time
[05:47:18] <fenn> its a silly idea anyway, why not just gather helium 3 from the sun and burn it in your own fusion reactors
[05:48:40] <fenn> i think we'll be seeing a lot more of wiki's in the near future
[05:49:01] <fenn> when it filters down into the mainstream consciousness
[05:49:25] <wholepair> splendid - get me a straw long enough to get to the sun and I will start sucking hellium 3 for your reactor - I gotta run fenn
[05:49:48] <wholepair> have a nice day - or night - cant remember where you are -
[05:50:07] <fenn> heh night
[05:50:22] <fenn> oh right, screenshot: http://fenn.dyndns.org/pub/irc/hexapod_sim2.png
[05:51:01] <wholepair> oh my - this is funny - http://fenn.dyndns.org/sketches/borg-cow.jpg
[05:51:14] <fenn> thanks
[05:51:26] <wholepair> that looks much better - see ya
[06:02:01] <fenn> oh i see.. the "go" button is off the screen because there's 6 axes displayed
[06:18:59] <JymmmEMC> well, that's stupid
[06:21:12] <fenn> i would have figured it out but the window gets resized to match the screen resolution
[06:28:15] <JymmmEMC> =)
[07:42:37] <alex_joni> morning
[07:45:12] <SWPadnos> so it is
[07:45:24] <alex_joni> SWPadnos: oh-oh
[07:45:31] <alex_joni> sounds like late-nit hacking on drivers
[07:45:34] <SWPadnos> yeah - you said it
[07:45:37] <alex_joni> late-nite
[07:45:54] <SWPadnos> actually, the driver writing is going fine. it's just that the FPGA config doesn't work now
[07:46:07] <SWPadnos> after I made no changes to it
[07:46:11] <alex_joni> :/
[07:49:27] <alex_joni> fenn: yay, the hexapod does get simulated
[07:49:41] <alex_joni> only issue I see is when I drag or rotate, the joints really dance around :)
[07:56:52] <alex_joni> fenn: you placed the tooltip based on world position, and the struts based on joints?
[07:57:10] <alex_joni> that would explain why it does the funky stuff it does..
[07:57:37] <alex_joni> (ideally in the simulation you only use the joints, and compute the platform position/orientation from that..)
[07:58:07] <alex_joni> maybe even colour it red based on constraints (if one of the joints isn't where it's supposed to be - wrong kins or such)
[08:07:36] <SWPadnos> yay. another stupid problem solved
[08:08:02] <SWPadnos> I think I've grown to dislike the Xilinx FPGA tools somewhat
[08:10:07] <alex_joni> at least it's not Labview
[08:10:58] <SWPadnos> that's true
[08:11:15] <SWPadnos> ant it's even possible that the Xilinx tools were part of the problem with LabView FPGA
[08:11:19] <SWPadnos> and
[08:11:32] <SWPadnos> though I think they added the incremental compilation stuff recently
[08:14:37] <alex_joni> * alex_joni yawns
[08:34:11] <anonimasu> :)
[12:19:43] <anonimasu> hm, if you see awallin tell him im searching for him
[12:20:05] <alex_joni> he's hiding
[12:20:13] <alex_joni> oops.. he said I shouldn't tell you
[12:20:15] <alex_joni> :P
[12:21:58] <alex_joni> anonimasu: j/k
[12:22:51] <CIA-8> 03alex_joni 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: copy/paste error. thanks to mxxxxxx for spotting it
[13:55:57] <CIA-8> 03cradek 07TRUNK * 10emc2/debian/changelog: note this bug that alex fixed
[17:55:55] <SkinnyPuppy34> zzzz....