#linuxcnc | Logs for 2013-06-13

Back
[01:08:10] <Loetmichel> morning
[01:42:37] <RyanS> Is this guy getting a little too carried away? http://www.toolsandmods.com/projects/marty-nissen-project-lathe
[01:42:55] <DJ9DJ> moin
[02:41:57] <RyanS> shaping machines are bizarre
[08:41:00] <skunkworks> heh - Windows 'hate'
[08:41:13] <skunkworks> (funny I just heard that...)
[08:48:37] <asdfasd> when I study in university 10 years ago somebody writed on the wall
[08:48:53] <asdfasd> I hate windows. signature: John
[08:48:58] <asdfasd> and below that
[08:49:09] <asdfasd> I hate John. signature: windows
[08:49:10] <asdfasd> :)
[08:53:13] <skunkworks> heh
[08:53:24] <archivist_herron> hey I am working at a windows company today!
[08:53:26] <skunkworks> This is in reference to windows '8'
[08:53:38] <skunkworks> different windows!!
[08:54:05] <archivist_herron> win8 is the biggest bag of sh** I have seen
[08:54:09] <asdfasd> different but people still say the same :)
[08:55:06] <Jymmm> archivist_herron: You forgot Vista.
[08:55:21] <Valen> win8 is worse than vista
[08:55:27] <Jymmm> Valen: Really?
[08:55:33] <archivist_herron> and ME
[08:55:38] <Valen> vista they tried to add new stuff and do fancy things, and sucked at it
[08:55:39] <archivist_herron> combined
[08:55:49] <Jymmm> Eh, ME don't count. That was just fluff.
[08:55:56] <Valen> win8 they took something that worked, and took out working stuff
[08:56:04] <Valen> and added crap in its place
[08:56:26] <archivist_herron> added the worst interface they could think of
[08:56:34] <Jymmm> Heh, I was just talking about that with someone actually.
[08:56:39] <ssi> they run a one on, one off release schedule
[08:56:51] <Valen> they are doing 8.1 soon
[08:56:59] <Jymmm> Thus why I dont' run M$ on bare metal
[08:56:59] <Valen> i hear the start button is making a comeback
[08:57:05] <ssi> 3.1 was good, 95 was crap. 98 was good, me was crap. XP was good, vista was crap. 7 was good, 8 was crap
[08:57:15] <Valen> 2000 was good
[08:57:22] <ssi> yea but that's an NT product
[08:57:23] <Valen> 2000 and xp were both good...
[08:57:24] <Jymmm> ssi: lol, I never realized that
[08:57:25] <ssi> different line :)
[08:57:26] <Valen> yeah
[08:57:29] <archivist_herron> I still have a 2k box
[08:57:37] <ssi> they have to make a crap os every other time so that you're HAPPY to switch to their next os
[08:57:48] <Jymmm> NT 351 and 4.0 were good.
[08:57:56] <Valen> lol, I think the problem is they went IPADS!!!!OMG
[08:58:04] <Valen> so they tried to make an OS for ipad
[08:58:07] <jdh> I still have DOS, w98, nt4, w2k boxes
[08:58:10] <Valen> which in it self is fine
[08:58:28] <ssi> I got off windows back in 98
[08:58:33] <Valen> but when I want to run 16 different spreadsheets and some bigass cad package its not much chop
[08:58:39] <ssi> I ran 98 beta in `97, and then switched completely to linux :P
[08:58:56] <ssi> I still run XP in VMs though
[08:59:00] <ssi> for things like solidworks and ISE
[08:59:02] <ssi> and esx client
[08:59:09] <Jymmm> client?
[08:59:22] <Jymmm> ssi: ESX from VMWARE?
[08:59:43] <ssi> yeah
[08:59:48] <ssi> esxi technically
[08:59:52] <ssi> the management client is windows only
[08:59:57] * Valen just uses the missus "gaming machine" for any of that stuff ;->
[09:00:03] <Jymmm> ssi: oh, that fuscking thing.
[09:00:04] <ssi> I have an esxi server in a datacenter
[09:00:13] <ssi> yeah that fucking thing heheh
[09:00:19] <Valen> I gave up on vmware when kvm got into its stride ;->
[09:00:22] <ssi> if I had to do it over again I might try to go kvm
[09:00:25] <Valen> ganeti + kvm w00t
[09:00:28] <ssi> but it's a lot of work to start over :)
[09:00:30] <ssi> and downtime
[09:00:43] <Valen> have 2 machines?
[09:00:51] <Jymmm> ssi: and the $5000+ to move it all over to
[09:00:53] <Valen> I mean a cluster setup
[09:00:58] <Jymmm> machine^
[09:01:54] <Valen> I cheated for my migration, fire up the new machine, shut down the guest, boot ubuntu livecd on the new guest instance and the old one, the DD | NC the image over ;->
[09:02:02] <ssi> I can't afford the hosting for two machines
[09:02:12] <ssi> hell I can't really afford the hosting for one machine
[09:02:19] <ssi> it doesn't pay me back enough to justify its existence
[09:02:22] <ssi> but it's massively underutilized
[09:02:27] <Valen> I have 2x 1RU's for clients VM's
[09:02:33] <Jymmm> ssi: heh
[09:02:40] <ssi> mine's a 2u, and it's a whitebox build
[09:02:56] <Valen> Dell R210 II's nice and quiet and cheapish
[09:02:59] <Jymmm> ssi: Not your case, but check out the last image here... http://alpinelinux.org/downloads
[09:02:59] <ssi> single i7 consumer proc, 16g ram, 8x2TB on an adaptec raid6
[09:03:11] <archivist_herron> mention of cluster reminds me of....add lustre to your cluster with knacker lacquer
[09:03:15] <Jymmm> ssi: It's xen on a lightweight distro
[09:03:22] <ssi> nice
[09:03:37] <Valen> I am wondering about making a solid block of a server, just buttloads of 1.5Ghz quad core arms, with 2gb ram and 30gb fast flash
[09:03:38] <ssi> we use kvm at work for our "private cloud"
[09:03:57] <ssi> but I don't manage the infrastructure side of it
[09:04:03] <Valen> probably pack a few thousand machines into each RU
[09:04:13] <Jymmm> ssi: I've been trying to work on a 'virtualized desktop', Like ESXi for desktop if you will.
[09:04:18] <ssi> Valen: that's the opposite of how things are done these days :P
[09:04:21] <Valen> i know
[09:04:26] <Valen> but its interesting isnt it
[09:04:26] <ssi> and there's reasons for that
[09:04:37] <Valen> there are, but there's also reasons not to
[09:04:39] <ssi> you have significant overhead in things like power distribution, networking, etc
[09:05:02] <Jymmm> ssi: bandwidth is cheap in a DC compared to power.
[09:05:04] <Valen> yes and no, those things have also become drastically cheaper
[09:05:10] <ssi> Jymmm: yes it is
[09:05:15] <ssi> but I don't mean networking
[09:05:18] <ssi> er
[09:05:18] <Jymmm> =)
[09:05:19] <Valen> I think he means having a thousand port switch might be kinda spendy
[09:05:22] <ssi> I dont' mean bandwidth
[09:05:29] <ssi> I mean cabling, phys, switches
[09:05:34] <Jymmm> ah
[09:05:34] <ssi> PLANNING
[09:05:37] <Valen> Mine was going to be all solid state
[09:05:45] <Valen> switches on board
[09:05:56] <Jymmm> Eh, do what akami does =)
[09:05:56] <ssi> nice thought, but without PHYs, you have MII to MII comms
[09:06:00] <ssi> and those are wide
[09:06:23] <ssi> it's far more practical to just have one big honkin machine, and virtualize
[09:06:28] <ssi> then you have virtual networking
[09:06:29] <ssi> single point power
[09:06:41] <ssi> less hardware infrastructure needs
[09:06:42] <Jymmm> SPOF too =)
[09:06:52] <ssi> sure, but you design around that
[09:07:02] <ssi> you have the same spof if you have a thousand arm cores in one box
[09:07:04] <Jymmm> two $7000 machines =)
[09:07:17] <Valen> you seem to be thinking that there is all this stuff hanging around to do the support
[09:07:36] <Valen> the idea is you get 30 "sheets" of PCB made, totally self contained
[09:07:48] <ssi> yeah I realize that
[09:07:53] <Valen> stack them and power them
[09:07:57] <ssi> but even the networking on that board is inconvenient
[09:08:07] <ssi> I've been working on similar stuff
[09:08:14] <ssi> sheets of pcb that hold dozens of bitcoin mining asics :P
[09:08:19] <ssi> it's a smaller problem to solve
[09:08:21] <ssi> and still a lot of work
[09:08:42] <ssi> there we're just talking about routing SPI around to everyone, but still comms is a hassle
[09:08:45] <Valen> I was thinking of putting phys in, because you could have a run of up to a meter anyway
[09:08:46] <ssi> power distribution is a hassle
[09:09:03] <Valen> those asic's also run stupid hot and use assloads of power too
[09:09:03] <ssi> phys would be a good idea, because you don't want to run MII long distances
[09:09:07] <ssi> yes they do
[09:09:25] <Valen> these arms TDP is usually more like half a watt
[09:09:26] <Jymmm> ssi: ASIC, talk about bitching about power.... you'll need like 80A if you fill a rack, plus the cooling
[09:09:38] <ssi> I'm aware :)
[09:09:52] <ssi> single chip wants 6A or so Vdd
[09:09:58] <Jymmm> OUCH
[09:09:59] <ssi> ~5W
[09:09:59] <Valen> what vcc?
[09:10:02] <ssi> 0.9
[09:10:11] <Valen> thats not too terrible really
[09:10:25] <ssi> no it's really not bad for what it is
[09:10:31] <ssi> but it means fat switchers
[09:10:34] <ssi> and a switcher per few chips
[09:10:38] <ssi> big power planes
[09:10:42] <Valen> I was looking at some of the coinminer rigs they have heatsinks that look more appropreate on high end audio gear
[09:10:49] <ssi> yep
[09:11:02] <Valen> I was thinking by Z stacking I could just feed the power vertically
[09:11:16] <Valen> but some nice big blower fans in to cool the whole thing
[09:11:20] <ssi> sure
[09:11:20] <Valen> turbine style
[09:11:25] <ssi> it's absolutely feasible
[09:11:28] <ssi> I just don't think it's worth doing
[09:11:33] <Jymmm> Valen: forget that... liquid cooling
[09:11:47] <ssi> I think for less money than it'd cost you to build this thing, you could buy a single server from dell that'll virtualize out to the same width and more power
[09:11:52] <ssi> more computational power
[09:11:59] <Valen> I don't think you would tbh
[09:12:08] <ssi> arm SoCs aren't as powerful as they sound :)
[09:12:13] <Valen> The way I look at it is for $50 or so you can get a real hardware that has about the same spec as a minimum spec linode
[09:12:18] <ssi> 1.5GHz of arm isn't the same as 1.5GHz of xeon
[09:12:26] <Valen> totally true
[09:12:27] <Jymmm> ssi: is too
[09:12:53] <Valen> but if you share 8x 3ghz xeon cores amongst 500 users, well you may well be better off with the quad core arm
[09:13:16] <Valen> its not cut and dried either way
[09:13:18] <ssi> 500 $50 arm boards is $25k
[09:13:27] <ssi> before you start adding the plant to connect them all to a network
[09:13:39] <Valen> thats a pretty hefty dell I'll grant you
[09:13:44] <ssi> you can buy a lot of virt servers for $25k
[09:15:07] <ssi> anyway, I gotta run to a meeting and talk about our configuration management for our virtual servers :D
[09:15:10] <ssi> bbiab
[09:15:16] <Valen> nighty night
[09:15:56] <Valen> quick look at dells
[09:16:26] <Valen> $25k seems to get you 4x Intel® Xeon® E7540 2.00GHz with 1Tb of ram
[09:16:54] <Valen> thats 24 cpu cores amongst our 500 theoretical users
[09:17:23] <Valen> Vs 2000 arm cores with no contention
[09:17:38] <Valen> no VM overheads
[09:17:44] <Valen> dedicated flash storage
[10:23:56] <IchGuckLive> hi all B)
[10:24:49] <IchGuckLive> today first time this year we hit 30 degrees Celsius = 86 degrees Fahrenheit in germany
[10:24:53] <IchGuckLive> O.O
[10:59:14] <Tom_itx> IchGuckLive you've got some catching up to do, it's been 95+ F here this week
[10:59:38] <IchGuckLive> we will see this on monday shure
[11:00:08] <IchGuckLive> im off till later By
[11:08:27] <nikola_> peter , in the uper right corhet , there is a hibrid cirquit called p-rs1
[11:08:44] <ssi> has a hunger
[11:26:57] <tjtr33> prepping to drive to Wicjita, replaced rotors and pads, 5 miles later , rotor sheared of hub! http://imagebin.org/261251
[11:27:35] <tjtr33> i think its welded and the only good welds are where it really fractured, rest of 'welds' just sheared
[11:29:14] <archivist> supposed to be solid cast
[11:29:40] <tjtr33> i know, od was ground so i didnt see anything
[11:30:06] <cradek> tjtr33: yikes, hope nobody got hurt
[11:30:21] <cradek> those are some pretty lousy rotors
[11:30:53] <tjtr33> i'm ok, limped home on hand brake
[11:33:41] <CaptHindsight> what auto parts chain did they come from?
[11:34:17] <KimK> Top of picture says Autozone
[11:35:01] <archivist> you want your money back and extra for new underwear
[11:35:36] <knbb> Hmm emc?
[11:35:52] <knbb> Is this the channel for EMC?
[11:36:00] <jdh> not the EMC storage
[11:36:06] <CaptHindsight> either really good high friction brake pads or really low grade alloy rotors
[11:36:25] <knbb> jdh is there such a channel? FOr storage
[11:36:39] <knbb> Or better, for emc specifically?
[11:36:52] <jdh> no idea... sorry.
[11:36:59] <knbb> Np, ty anyway
[11:37:04] <archivist> freenode is about opensource and free so unlikely
[11:37:29] <knbb> Well, there is a Windows channel :)
[11:37:40] <archivist> sad
[11:38:16] <tjtr33> archivist, i really did tear my shorts and the old jeans when angrily ripping the parts off in the rain at night :) i think i said some bad things too [ enuf ot ]
[11:38:45] <archivist> I imagine they will need a number of channels for disgruntled win 8 users
[11:40:24] <jdh> windows doesn't seem as bad to windows users as it does to 'us'
[11:41:38] <cradek> yay for managing expectations
[11:42:02] <jdh> MS, lowering expectations for 25 years
[11:42:37] <Jymmm> jdh: What makes you think they ever RAISED them in the first place?
[11:43:22] <jdh> I didn't imply they did.
[11:43:37] <Jymmm> jdh: After all these decades, M$ STILL doesn't have a bootable recovery media.
[11:43:49] <Jymmm> jdh: They are the only one, everyone else does.
[11:44:02] <jdh> ENOPARSE
[11:44:17] <skunkworks> bootable recovery media?
[11:44:27] <jdh> I started with VMS & unix...
[11:44:55] <Jymmm> skunkworks: Yeah, shit happens... toss in your install media and boot form it, do your repairs/recovery, etc
[11:45:01] <Jymmm> from*
[11:45:11] <jdh> I've done that in windows
[11:45:19] <jdh> for years
[11:45:35] <Jymmm> jdh: Nah, you have the CLI shits.
[11:45:38] <skunkworks> Jymmm, you can boot off the cd to the 'recovery conosle' which is pretty much dos command line..
[11:45:44] <archivist> I too have repaired win systems with the recovery disk ok
[11:45:55] <Jymmm> skunkworks: Yeah, that's not what I'm talking about.
[11:46:20] <skunkworks> do you mean something like a 'live cd' (in linux terms?) then no.
[11:46:23] <Jymmm> OSX/Linux LiveCD sorta thing.
[11:46:43] <jdh> there is a mini-bootable XP
[11:47:07] <Jymmm> jdh: It's a clusterfuck to configure though.
[11:47:28] <jdh> ok. I guess I won't use it anymore then.
[11:47:57] <Jymmm> jdh: Especially things like network drivers so you can image your hdd before it totally fails, etc.
[11:48:47] <skunkworks> I have fixed a lot of microsoft machines using the linux livecd...
[11:48:51] <Jymmm> jdh: OSX is obscenly easy in comparison.
[11:49:08] <Jymmm> skunkworks: Yep, me too, far too many.
[11:49:13] <skunkworks> yep
[11:50:21] <Jymmm> skunkworks: This is the best overall be it M$ or nix http://www.sysresccd.org/
[11:50:40] <Jymmm> skunkworks: even has a PXE server
[11:51:05] <skunkworks> cool
[11:51:06] <knbb> I really need some EMC storage help, any idea to where I could turn?
[11:51:21] <jdh> knbb: call emc support?
[11:51:23] <Jymmm> knbb: http://na.com/
[11:51:45] <knbb> I cannot call them atm
[11:52:00] <CaptHindsight> knbb: have you considered making your own cnc mill?
[11:52:02] <knbb> A shot in the dark, I know this isn't the place to ask. Sry
[11:52:04] <archivist> heh, you have to pay the bill
[11:52:16] <knbb> cnc mill?
[11:52:18] <ReadErro-> emc -> pay money -> they fix
[11:52:18] <knbb> :)
[11:52:29] <Jymmm> knbb: http://www.netapp.com/
[11:52:39] <ReadErro-> netapp = lhame
[11:52:43] <cradek> knbb: freebsd+zfs is an awesome storage solution
[11:52:54] <ReadErro-> VNX or GTFO
[11:52:56] <ReadErro-> ;)
[11:52:59] <Jymmm> cradek: That's FreeNAS btw
[11:53:08] <knbb> using VNX
[11:53:10] <ReadErro-> cradek, thats far from enterprise level storage though
[11:53:26] <ReadErro-> i run that at home though
[11:53:28] <cradek> heh, yeah not buzzword-compliant enough
[11:53:47] <ReadErro-> cradek, well scalability wise
[11:53:47] <Jymmm> ReadErro-: http://www.ixsystems.com/
[11:53:51] <cradek> also not expensive enough and there's not enough lock-in
[11:54:41] <ReadErro-> cradek, its not buzzword or cost, its performance and scalability
[11:54:58] <ReadErro-> when you have hundreds of drives, a simple freebsd box wont cut it
[11:55:05] <archivist> linuxcnc has lock-in....low price and feeeeeeeatures
[11:55:12] <Jymmm> ReadErro-: sure it will.
[11:56:30] <ReadErro-> well i do this on a day to day basis, and i must respectfully disagree
[11:56:49] <Jymmm> ReadErro-: How's a petabyte sound? http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/
[11:57:52] <cradek> that's actually linux and not zfs
[11:58:03] <Jymmm> $117K vs $2.9M
[11:58:04] <ReadErro-> yes, lets trust millions of dollars worth of data on some concept idea..
[11:58:21] <Jymmm> ReadErro-: It's not concept, they actually use that
[11:58:40] <cradek> this is so OT... sorry I started it
[11:58:45] <jdh> heh
[11:59:05] <cradek> or can we blame knbb?
[11:59:15] <Jymmm> cradek: we can
[11:59:24] <cheesepuff> anyone maybe have some ideas on settng up MPG pendant. I have attached to extra pci card
[11:59:29] <knbb> haha
[11:59:58] <cradek> cheesepuff: you can hook an mpg to any set of inputs and use the software encoder module to count it
[11:59:59] <ReadErro-> Jymmm, i dont see any redundancy built in to that at all
[12:00:05] <ReadErro-> if 1 mobo dies you are SOL
[12:00:14] <ReadErro-> EMC has heavy redundancy built in
[12:00:21] <cheesepuff> and when I use hal meter to look at parport.1.pin-04-in pretty sure I should see an change when the switched is selected.
[12:00:24] <cradek> or if you have hardware encoder counters use one of those
[12:01:02] <cradek> cheesepuff: ok sounds like you need to explain more about what you have and what problem you are having
[12:02:29] <cheesepuff> cradek: I'm just using a extra pci card c22 pendant interface.
[12:02:43] <cradek> sorry I have no idea what that is
[12:03:25] <Jymmm> ReadErro-: LOL we spent $1.4M with EMC, brought it to it's knees, they ended up giving us $167K worth of more gear to compensate. it was funny. They thought we were exaggerating are needs =)
[12:03:37] <cradek> OH STOP
[12:03:46] <cradek> GRR
[12:04:00] <cheesepuff> cnc4pc.com C22 card.
[12:04:13] <cradek> aha
[12:04:24] <cradek> have you asked them for help?
[12:05:17] <cradek> or do you have a link to specs for the hardware etc?
[12:06:30] <ReadErro-> you cant throw a bunch of crappy SATA cards in a chassis and call it reliable enterprise storage Jymmm
[12:06:41] <cheesepuff> http://tinyurl.com/3yqd4lv
[12:07:08] <tjtr33> http://cnc4pc.com/Tech_Docs/C22-Pendant_Interface_Board_Rev2.pdf
[12:07:08] <ReadErro-> not to mention that doesnt even include SAS or SSD drives
[12:07:21] <CaptHindsight> tjtr33: did you visit Amerimold 2013 in Rosemont?
[12:07:50] <willburrrr2003> Good day all, Who here is familiar with the modbus functions, I have a question on addressing....
[12:08:05] <CaptHindsight> went yesterday since it overlapped with SUR-FIN and Fastener Tech
[12:08:47] <tjtr33> CaptHindsight, got a dead car right now ( scroll back :( did they have EDM ?
[12:09:20] <cheesepuff> C22 is wired up to pendant just like that...
[12:10:41] <CaptHindsight> tjtr33: a little, talked to Poco Graphite about EDM for micro-edm applications using 2 photon polymerization to make smaller tips
[12:11:05] <tjtr33> cheesepuff, you'll need to make a cable from the c22 to a parallel port ( or other digital io ), using the 'mach3' names as a guideline
[12:12:06] <tjtr33> CaptHindsight, did you read the article on printing graphite electrodes? some japanese uni working on the idea, in irc chat of the last 7 days
[12:12:32] <CaptHindsight> tjtr33: show was pretty small, yeah, that was my post under my other nick :)
[12:12:59] <tjtr33> CaptHindsight, are you working with l84supper? wicker park dlp printing?
[12:13:07] <tjtr33> oh
[12:13:08] <willburrrr2003> When addressing the slave in Modbus setup, do I put the modbus address in the Modbus element block (example Idec PLC M11 is modbus address 001010)
[12:18:48] <CaptHindsight> tjtr33: not sure if we will make it down to Wichita, battling with RTAI, kernels and Linuxcnc issues
[12:18:55] <cheesepuff> tjtr33: did all that.. so far think I just need to figure out confgurtion issues
[12:19:10] <IchGuckLive> hi all B)
[12:19:35] <willburrrr2003> or is that address block for the number of the plc component i am trying to address (ex M11, i would enter 11 in the block)?
[12:19:42] <CaptHindsight> if we get it stable by next week maybe we'll bring a PC down to play with, I'm sure we can find something to hook it up to :)
[12:21:36] <CaptHindsight> there's an RTAI kernel patch now for v3.8
[12:22:05] <IchGuckLive> willburrrr2003: M11 has different funktions on what version are you 2.5.0 git 2.6.pre master
[12:23:33] <CaptHindsight> tjtr33: http://amerimold13.mapyourshow.com/5_0/exhibitor_results.cfm?type=parentcategory&parentcategory=1A
[12:23:43] <CaptHindsight> EDM vendors at show ^^
[12:24:07] <willburrrr2003> M11 is the internal IDEC relay number, like %B in classicladder, not an M code in linuxcnc
[12:24:32] <willburrrr2003> I am just trying to figure out why I am having issues setting up Modbus communications
[12:24:43] <IchGuckLive> ok
[12:25:22] <tjtr33> CaptHindsight, thx
[12:25:27] <IchGuckLive> im in trouble with the usb gamepad as it kicks out fro time to time and then i need to walk long long ways
[12:25:57] <willburrrr2003> when the Modbus " communications error" pops up, it doesn't tell me what generated the error, is there an erro log that I can look at to see what caused it?
[12:30:59] <cheesepuff> how do you tell which I/O ports are running in?
[12:31:29] <IchGuckLive> cheesepuff: HAL
[12:31:34] <cheesepuff> loadrt hal_parport cfg="0x378 0xbc00 in"
[12:31:42] <pcw_home> USB is not a good industrial protocol
[12:32:12] <IchGuckLive> cheesepuff: the halmeter holds a list of the pins
[12:32:18] <IchGuckLive> parport.1
[12:32:33] <IchGuckLive> there you see if its in or out
[12:32:50] <cheesepuff> yeah it's in
[12:33:14] <IchGuckLive> the halmeter also shows True or false
[12:33:45] <cheesepuff> hmm maybe wiring issue I see can see pin-05-in flip true false
[12:34:40] <IchGuckLive> the signal connecting to this pin is in hal-show
[12:35:17] <IchGuckLive> if you put it on the watchlist you see more then one pin at the time
[12:36:16] <IchGuckLive> with your parport config you shoudt get 13more inputs
[12:37:56] <Tom_itx> ok, i've got an issue i'm not sure where to begin to diagnose. When running a program, midway thru a cutter path it will hang and won't continue unless i restart from the beginning. Sometimes during this, linuxcnc will ask me to insert the next tool when it's nowhere near time to do so. Occasionally i will get this error message: "can't do that (EMC_TASK_PLAN_RUN) in manual mode"
[12:38:07] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/cnc/configs/sherline/
[12:38:13] <Tom_itx> are the latest config files for it
[12:38:54] <Tom_itx> i did try increasing the SERVO_PERIOD thinking that might help but it didn't
[12:39:47] <IchGuckLive> tom is there a manuel board hooked up to the mashine
[12:40:06] <IchGuckLive> buttons axis haleffect sensors
[12:40:44] <Tom_itx> 7i43 7i47
[12:41:06] <Tom_itx> all the wiring is shielded including the steppers
[12:41:24] <Tom_itx> manual limit switches and 1 ir sensor for the spindle speed
[12:41:40] <Tom_itx> using debounce on the limits
[12:42:18] <IchGuckLive> (EMC_TASK_PLAN_RUN) im getting this all the time on my usb jojpads
[12:42:18] <Tom_itx> at first i thought it might be because the shop was too hot but it was cool this morning so i thought i'd try it again but no joy
[12:43:10] <Tom_itx> also sometimes my pendant will hang but jogging the axis using the axis screen will allow it to continue working normally
[12:43:27] <IchGuckLive> you got 4 halui in your ini something is trying to trigger this
[12:44:47] <Tom_itx> i might try removing the counters and see if that makes a difference
[12:44:52] <Tom_itx> cycle counters
[12:45:13] <IchGuckLive> i confused also the interpreter by OR cammanding pendant usb and pyvcp
[12:45:44] <Tom_itx> i tried 2 different PCs too
[12:46:00] <Tom_itx> the atom did work a bit better than the old clunker but still got errors
[12:46:37] <Tom_itx> one running 8.04 and the other running 10.04 both on 2.5.2 ver
[12:46:58] <IchGuckLive> i got 25 all the same and only 4-6 are troubeling all the same config the same gamepad
[12:48:25] <Tom_itx> i increased the SERVO_PERIOD from 1000000 to 1500000 and it didn't seem to make any difference
[12:48:47] <Tom_itx> base period not used
[12:50:34] <cheesepuff> ok not wiring
[12:51:53] <IchGuckLive> cheesepuff: you dont need a reset time for the second IN parport
[12:52:17] <IchGuckLive> therfor it might toggle
[12:53:37] <CaptHindsight> PCW: how many 7i65's can the 7i68 support?
[12:53:54] <pcw_home> Tom_itx: bad memory or hard drive?
[12:54:21] <Tom_itx> mostly new equipment
[12:54:24] <Tom_itx> tried 2 pcs
[12:54:26] <pcw_home> 4 (firmware limitations) 6 is hardware max
[12:54:36] <Tom_itx> the atom board is next to new
[12:54:54] <cbjamo> I'm getting an error, I think its related to axis, but I'm not sure. "HAL: ERROR: pin_new(halui.jog.selected.increment) called with already-initialized memory"
[12:55:39] <cbjamo> I'm using ja3, that may be important.
[12:56:05] <CaptHindsight> tjtr33: have you ever used the T-slot from Frame-World http://www.frame-world.com/ ?
[12:56:56] <pcw_home> Tom_itx: Not sure how the interpreter can skip ahead
[12:57:16] <Tom_itx> it's got me baffled a bit
[12:57:39] <CaptHindsight> pcw_home: trying to decide on what combination of boards to use for 7 servo's and lots of IO
[12:57:44] <pcw_home> Did this just start (and if so what changed)
[12:57:48] <Tom_itx> bbl but if you think of something i'll read about it when i get back
[12:58:29] <pcw_home> what kind of I/O?
[12:58:51] <CaptHindsight> tjtr33: their factory is 10 minutes away but they only sell through distis 1hr away
[13:00:28] <IchGuckLive> cbjamo: hi INCREMENTS = in displaysec initalisises this
[13:01:06] <CaptHindsight> pcw_home: 50 pairs LVDS
[13:02:09] <cbjamo> IchGuckLive: My machine is metric, so INCREMENTS has 5, 1, .2, .1, .05, .01, and .005mm
[13:02:15] <IchGuckLive> cbjamo: page 157 in the integreater manual
[13:02:37] <cbjamo> thanks
[13:04:18] <IchGuckLive> cbjamo: what are you trying to do
[13:04:31] <cbjamo> um, are you sure that is the right page? 157 refences pluto-step pinout...
[13:04:55] <IchGuckLive> V2.5
[13:05:26] <IchGuckLive> search for halui.jog.selected.increment
[13:06:18] <cbjamo> Get my hal file set up. It stops giving the error if I set [TRAJ]AXES=6.
[13:08:06] <IchGuckLive> cbjamo: geometry ?
[13:08:28] <IchGuckLive> dident you use joints
[13:08:35] <IchGuckLive> ja3
[13:08:37] <cbjamo> Like a ganrty, but two motors for every axis rather than just y
[13:08:49] <cbjamo> yes, I'm using ja3.
[13:08:56] <IchGuckLive> ja3 needs joints no axes
[13:09:21] <IchGuckLive> 6. is an error
[13:09:27] <IchGuckLive> no dot
[13:09:49] <cbjamo> right, I have [KINS]JOINTS=6
[13:10:03] <cbjamo> that was just a period, I don't have a dot in the file.
[13:13:17] <IchGuckLive> cbjamo: http://linuxcnc.org/docs/html/gui/halui.html
[13:15:01] <cbjamo> hmm, I don't have halui.jog.selected.increment in my hal file.
[13:15:30] <IchGuckLive> as you use joints it might be halui.joint
[13:15:49] <IchGuckLive> http://www.linuxcnc.org/index.php/german/forum/49-basic-configuration/22342-jog-increments-using-a-6-position-selector-switch
[13:16:01] <Jymmm> Bastards... Waterpik.com product "Lifetime warranty" is in reality a one-time-only replacement.
[13:16:55] <IchGuckLive> cbjamo: there is now mux8 available for a selector
[13:17:16] <IchGuckLive> cbjamo: keep in mind the first needs always to be zero
[13:18:54] <cbjamo> I don't have a physical ui, I'm just going to use axis.
[13:19:33] <IchGuckLive> ok im off as its late here by till tomorrow !
[13:26:40] <skunkworks> Tom_itx, I would be looking at your halui connections... maybe disabling them and testing.
[13:28:33] <archivist> Tom_itx, I have a pc with a thermal problem, is it time related after the cpu is worked hard
[13:36:09] <tjtr33> CaptHindsight, (sorry working on car ) re frameworld: nope, but i think they're a fav of JT's , so i'd say it was good stuff.
[13:36:59] <tjtr33> CaptHindsight, (sorry workin on car) i have not used frameworld but iirc JT sez its good stuff & price
[13:42:04] <CaptHindsight> tjtr33: no problem, let me know what route you took and how long the drive is to Wichita, Google estimates 11+ hours
[13:51:19] <Tom_itx> archivist, no this happens even at first startup
[13:51:31] <Tom_itx> on both pc's
[13:51:44] <Tom_itx> i will look at halui and disable a few things
[13:52:11] <Tom_itx> all the recent configs are on the link i posted...
[13:52:37] <Tom_itx> if you happen to see anything unusual
[13:58:03] <Jymmm> Tom_itx: What USB devices do you have connected to it?
[13:58:32] <Jymmm> Tom_itx: And What make/model is the mobo?
[14:03:58] <willburrrr2003> is there an error log I can look at when I get A "Modbus communication failure" message, to see why the failure was/is?
[14:06:25] <FinboySlick> www.youtube.com/watch?v=zMhVPId-pRA Now here's what I call being forced into shape.
[14:13:53] <Jymmm> FinboySlick: The thing that gets me is there was ZERO hesitation/resistance... YOU WILL COMPLY... "SQUISH"
[14:21:58] <CaptHindsight> http://www.youtube.com/watch?NR=1&feature=fvwp&v=2suW1kgiw-0 this was next after that video
[14:22:00] <Tecan> (2suW1kgiw-0) "Explosive panel forming with oxy/acetylene - the first attempt." by "ksv04" is "Autos" - Length: 0:05:11
[14:23:28] <Jymmm> This (the video itself) just seems "dated" to me... http://www.youtube.com/watch?v=44XD5mZoM_0
[14:23:30] <Tecan> (44XD5mZoM_0) "Roll Forming" by "SteelFormingSystems" is "Tech" - Length: 0:06:41
[14:27:17] <CaptHindsight> I should make something cheaper than makerslide using roll formed cold roll
[14:28:10] <CaptHindsight> but then I'd have to make a video to prove that it works :)
[14:36:41] <archivist> its old technology book:- High velocity forming of metals American Society of Tool and Manufacturing Engineers 1964
[14:37:47] <cheesepuff> hmm maybe wiring issue I see can see pin-05-in flip true false
[14:38:36] <cheesepuff> oops
[14:40:08] <archivist> hehe
[14:44:31] <CaptHindsight> http://www.ebay.com/itm/Small-Garage-Sized-Bridgeport-1-2-HP-Milling-Machine-Vise-Veriscope-INV-11246-/230992817697?pt=LH_DefaultDomain_0&hash=item35c83eae21 when did Bridgeport make a 0.5hp 4250rpm head?
[14:48:21] <fragalot_> "sized right for your small garage"
[14:48:24] <fragalot_> I find this offensive.
[14:48:55] <fragalot_> lol, just noticed the massive "NO CANADIAN BASTARDS"
[14:49:53] <Aero-Tec> can one turn on and off the A axis rollover with Gcode?
[14:50:38] <Aero-Tec> I want to cut a circle and have 360 be 0
[14:51:17] <Aero-Tec> now when I go to 0 it goes all the way around back to 0
[14:51:46] <Aero-Tec> I know there is a ini 360 roll over
[14:52:03] <Aero-Tec> but would like to know if I can control it from gcode
[14:52:58] <CaptHindsight> fragalot_: heh, quite a character
[14:56:44] <Aero-Tec> I am not seeing any Gcode for axis roll over
[14:57:15] <Aero-Tec> but there is one I thought that would let you reset 0 with gcode
[14:57:34] <Aero-Tec> is that doable?
[15:01:14] <fragalot_> should be
[15:01:42] <fragalot_> not sure about proper gcode, but on my 3D printer, I often use m92 E0 to reset the E axis
[15:01:50] <fragalot_> sorry, not 92
[15:01:59] <fragalot_> G92
[15:05:10] <Aero-Tec> G10 L2 P1 A0
[15:05:23] <Aero-Tec> looks like that one would be the right way
[15:08:24] <Aero-Tec> G92.2 maybe
[15:09:01] <Aero-Tec> G92 looks to do more then I need
[15:09:40] <Aero-Tec> also says something about moving all the the axis offsets
[15:09:47] <Aero-Tec> not sure how that works
[15:09:58] <Tecan> rinky do
[15:14:15] <cradek> no you can't change the wrapped-rotary setting from gcode
[15:14:36] <cradek> of course you can set an offset that makes the current position become zero
[15:16:45] <Aero-Tec> some thing like this
[15:16:48] <Aero-Tec> G10 L20 P1 A[#<NumberOfDeg>-360]
[15:17:03] <Aero-Tec> I over shoot 0 by 10 deg
[15:17:28] <Aero-Tec> so that should make A axis be at 10 deg
[15:17:33] <cradek> not sure I understand your situation
[15:17:43] <Aero-Tec> so a0 will be a short trip
[15:17:53] <cradek> maybe you should be using wrapped rotary mode
[15:19:15] <Aero-Tec> was hoping to be able to use gcode to sim wrapped rotary
[15:19:34] <cradek> why don't you want wrapped rotary?
[15:19:42] <Aero-Tec> save restarts and zero resettings and such
[15:20:11] <Aero-Tec> I think I have EMC save the setting when shutting down
[15:20:37] <Aero-Tec> as maybe I could do a restart and not loose Zero
[15:21:22] <Aero-Tec> what I am doing not is a one off
[15:21:27] <cradek> I'm not really following you
[15:21:50] <Aero-Tec> did not want to restart software for a quick and simple one off
[15:22:07] <Aero-Tec> I set the axis zero by hand
[15:22:16] <Aero-Tec> edge finder
[15:22:25] <Aero-Tec> no home switches
[15:22:51] <Aero-Tec> so no quick and simple way to home and zero things
[15:23:07] <cradek> I don't understand what any of that has with the choice of using wrapped rotary or not
[15:23:14] <cradek> has to do with
[15:24:09] <Aero-Tec> wrap is not turned on right now, so I would have to shut down and make a new INI with wrap turned on, with start EMC again
[15:24:32] <Aero-Tec> then not with
[15:25:05] <Aero-Tec> or is there a way to load new ini and not restart the program
[15:25:49] <cradek> but you could have edited the ini and restarted a hundred times by now.
[15:25:55] <Aero-Tec> is there a quick and easy way to turn wrap on?
[15:26:08] <cradek> yes, edit the ini and restart linuxcnc
[15:26:20] <Aero-Tec> true, but I am learning
[15:26:53] <cradek> if you move the machine to a known point (g0 g53 x0 y0 z0 a0) before you shut down, you can pretty much keep track of your position
[15:27:00] <Aero-Tec> I thought it needed a restart to turn on axis wrap
[15:27:01] <cradek> just move it to wherever you consider home
[15:27:18] <cradek> brb
[15:27:50] <Aero-Tec> that would work, but what about the backlash?
[15:28:24] <Aero-Tec> will it keep track of the back lash directions for each axis?
[15:29:44] <Aero-Tec> I think I have EMC save things when it quites
[15:30:01] <Aero-Tec> not sure if backlash is one of them
[15:41:49] <tjtr33> CaptHindsight, Thx! new front rotors installed, tests ok, thinking of going Chi->LakeOfTheOzarks->Wichita
[15:41:51] <tjtr33> (stay overnite in http://www.mostateparks.com/lodging-unit/57780/lake-ozarks-state-park-outpost-cabin-1 )
[15:42:04] <andypugh> Is a 1925 lathe going to use UNC or ASME threads?
[15:42:54] <tjtr33> CaptHindsight, crap! " Single-night reservations are not accepted "
[15:44:12] <andypugh> Ring and ask, they might be desperate.
[15:55:56] <Tom_itx> Jymmm, it's the Intel Atom D525 MB and no usb devices connected to it
[15:56:44] <Jymmm> Tom_itx: No kybd/mouse?
[15:56:49] <Tom_itx> i'm gonna strip out the halui stuff and try it
[15:56:56] <roycroft> i thought unc was developed during wwii
[15:56:56] <Tom_itx> both ps2
[15:57:06] <Jymmm> Tom_itx: k
[15:57:38] <roycroft> so the yanks and brits could screw each other as easily as they screw themselves
[16:28:16] <andypugh> roycroft: Yes, that was my thought. So what thread standard did you use before that?
[16:29:06] <DJ9DJ> gn8
[16:32:29] <Tom_itx> ok i went back to a 'stock' linuxcnc. removed the postgui stuff and ran the file. it still stalls midway through a cut.
[16:33:01] <roycroft> we used amse while you used whitworth
[16:33:55] <roycroft> and i think it was after the war when the unified thread standard became widely adopted
[16:34:03] <roycroft> around 1950 or so
[16:57:42] <andypugh> Tom_itx: That's pretty unusual.
[16:57:57] <Tom_itx> i can't seem to figure it out
[16:58:14] <Tom_itx> everything is shielded and rewired
[16:58:22] <Tom_itx> tried 2 PCs
[16:59:37] <Tom_itx> removed the references to postgui.hal and display.xml
[17:00:05] <Tom_itx> i'll try a different parport cable just to see
[17:01:59] <Tom_itx> axis doesn't lock up but the gcode quits executing
[17:02:28] <andypugh> That's not the normal way round.
[17:02:43] <Tom_itx> i'm sure open to suggestions
[17:02:49] <micges> Tom_itx: can you pastebin that gcode?
[17:02:53] <micges> ,
[17:02:55] <andypugh> Does the rt system carry on running (if, for example, you twiddle a limit switch?)
[17:02:59] <Tom_itx> yeah but it's nothing unusual
[17:03:18] <Tom_itx> andypugh, i can check when i return from an errand
[17:03:36] <andypugh> There is no chance that something is randomly setting feed-hold? (maybe it is wired to a floating pin)
[17:03:41] <Tom_itx> no
[17:03:50] <Tom_itx> it does too odd of things
[17:04:03] <Tom_itx> like sometimes it asks for a tool change when it halts
[17:04:09] <Tom_itx> midway thru a cut
[17:04:44] <andypugh> Your computer is haunted. No other explanation.
[17:05:18] <Tom_itx> i honestly don't know where to start debugging
[17:05:32] <Tom_itx> back in a while...
[17:31:18] <Aero-Tec> Tom_itx, what Gcode are you running?
[17:32:12] <Aero-Tec> some one asked for it to be paste bin, sound like a good idea
[17:32:42] <gene78> Hey guys, whats with the 2.6.0pre branch, update-manager shows it to me, but won'tr take the check in the box
[17:33:59] <gene78> The problem is that I need the -dev stuff to get comp, but the versions don't match, so I can't pull the -dev stuffs
[17:34:13] <Aero-Tec> I have not seen that
[17:36:14] <gene78> I fired up synaptic but thats in a libmodbus dependency hell
[17:37:30] <gene78> Din-din time, back later
[17:56:55] <andypugh> gene78: I thought you already had comp working?
[17:59:15] * pfred1 has a comp working
[17:59:55] <pfred1> http://i.imgur.com/OjNG078.png
[18:00:25] <pfred1> I'm probably going to change the Window Manager to Fluxbox though
[18:00:46] <pfred1> I just wanted to see what LinuxCNC was like in Trinity :)
[18:01:30] <pfred1> I still might give Gentoo another shot too now that I know of a few build issues
[18:03:08] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/cnc/cnc_files/Cube.txt
[18:03:15] <Tom_itx> there's the cnc file
[18:03:24] <Tom_itx> doesn't stop in the same place
[18:03:27] <Tom_itx> random
[18:03:41] * pfred1 didn't like The Cube
[18:03:41] <Tom_itx> so far always after the first tool
[18:04:10] <Tom_itx> andypugh, i'll go try the limits this time
[18:12:39] <atom1> andypugh, the limit switches are still active when it stalls
[18:13:03] <andypugh> So, RTAI is still running.
[18:13:18] <atom1> appears so
[18:13:34] <atom1> spindle stays on, just stops moving
[18:13:54] <pfred1> servo spindle?
[18:14:04] <atom1> no, it's a standard sherline
[18:14:18] <atom1> no spindle control except on off
[18:14:35] <pfred1> andypugh you did a servo spindle didn't you?
[18:14:35] <andypugh> pfred1: Sort-of
[18:14:57] <pfred1> your one gear video where the spindle is slaved ot the axis
[18:15:51] * pfred1 needs "stalls" explained
[18:16:00] <andypugh> Yes, I have done that twice now in fact.
[18:16:41] <pfred1> what stalls the program the PC the mill?
[18:17:42] <andypugh> pfred1: The program
[18:17:46] <pfred1> oh
[18:18:01] <pfred1> best thing to do there is install fron git and try that
[18:18:14] <pfred1> hope whatever is wrong got fixed
[18:18:48] <pfred1> that is what I would do at any rate
[18:19:04] <andypugh> I think that it is more likely to find less rather than more reliability in a dev build.
[18:19:38] <pfred1> oh I don't know it probably worked for someone I think I'm running a git build right now
[18:19:56] <andypugh> I think this has to be a local problem, or there would be more complaints. When you swapped PCs I don't suppose you swapped the memory across or anything like that?
[18:20:35] <pfred1> the biggest problem I ever had with the CD image was related to video drivers
[18:20:58] <pfred1> I tried to run the live image with an ATI card in a system and it just froze up
[18:21:38] <micges> atom1: does it stall each time on same gcode?
[18:21:42] <pfred1> swapped it for a nvidia card and it worked fine
[18:21:48] <atom1> no
[18:21:51] <atom1> random
[18:22:05] <pfred1> random sounds like RAM
[18:22:22] <pfred1> RAm errors can be very random
[18:22:34] <pfred1> maybe that is why they call it random access memory?
[18:22:45] <atom1> came back out and a whole screen full of errors with it sitting idle
[18:23:25] <micges> what errors?
[18:23:30] <pfred1> errors in a terminal?
[18:23:49] <pfred1> like text errors?
[18:24:04] <pfred1> like something you could possibly paste into pastebin?
[18:24:26] <atom1> i didn't check, just shut it down and am changing the parport cable
[18:24:44] <pfred1> not all parallel cables are created equally
[18:25:02] <pfred1> when choosing cables go for fatter ones
[18:25:06] <atom1> i've been using both of these though for quite a long time
[18:25:13] <atom1> these are
[18:25:19] <andypugh> I almost guarantee the parport cable has nothing to do with it.
[18:25:21] <pfred1> yeah the skinny ones skimp on grounds
[18:25:47] <atom1> andypugh, i tend to agree but i try to check everything
[18:26:08] <micges> atom1: after linuxcnc start do in terminal: echo "4" > /proc/rtapi/debug
[18:26:39] <atom1> micges, when it errors or anytime?
[18:26:45] <micges> when it stops again please save dmesg output
[18:27:05] <micges> right after start
[18:27:11] * pfred1 likes syslog too
[18:27:39] <andypugh> micges: The problem seems to be in the interp. So turning on the userspace logging might help too.
[18:27:45] <pfred1> if they really had system errors they are probably logged in syslog
[18:27:54] <atom1> how do i copy dmesg to a file?
[18:28:06] <pfred1> dmesg > file.txt
[18:28:18] <atom1> it should still be there shouldn't it?
[18:28:29] <pfred1> well dmesg gets rewritten on boot
[18:28:37] <pfred1> but errors in syslog will still be there
[18:28:46] <atom1> i haven't rebooted since all those errors
[18:29:04] <pfred1> then give it a look
[18:29:14] <micges> atom1: add [EMC]DEBUG = 0x7FFFFFFF to your ini file also
[18:29:14] <atom1> i'll post it
[18:29:40] <pfred1> this system has a nagging error I never got around to fixing that spams my dmesg
[18:29:53] <pfred1> [353790.844023] usb 2-6: reset high speed USB device using ehci_hcd and address 2
[18:30:02] <micges> atom1: and set terminal to infinite lines
[18:30:14] <pfred1> I can echo something to something else in /proc and fix it but I usually don't bother
[18:30:36] <pfred1> you can less dmesg
[18:30:49] <pfred1> dmesg | less
[18:31:42] <pfred1> then you get to page back
[18:31:50] <atom1> http://tom-itx.dyndns.org:81/~webpage/cnc/cnc_files/dmesg.txt
[18:33:30] <micges> atom1: yes, that's it, post here again after rtapi debug enabled and error shows up
[18:33:54] <atom1> where do i add it in the ini?
[18:33:57] <andypugh> What thread is a screw that is 0.147" OD and 40tpi?
[18:34:09] <micges> [EMC] section
[18:34:13] <atom1> 4-40?
[18:34:20] <atom1> no
[18:34:24] <atom1> that's smaller
[18:34:28] <pfred1> 4-40 is tiny
[18:34:33] <andypugh> .112
[18:34:45] <andypugh> (actually less)
[18:35:14] <atom1> 6-40 takes a .113 drill
[18:35:23] <pfred1> well I just measured a 4-40 and got .104
[18:35:36] <andypugh> Need to subtract .00625 for the thread rounding.
[18:35:49] <atom1> that's for a tap drill
[18:36:13] <andypugh> No. That's different.
[18:36:14] <pfred1> andypugh sounds like you have a "special"
[18:36:33] <atom1> then the threads go to 8-36
[18:36:44] <pfred1> 8-32 is common
[18:36:49] <atom1> 6 is probably the largest standard fine thread
[18:36:56] <atom1> 8-36 is fine
[18:37:14] <pfred1> hmm if I hit those they probably end up in a beats me bucket here
[18:37:20] <atom1> ^^ 6 with 40 threads
[18:37:48] <andypugh> Looks like 8-40.
[18:37:53] <ssi> I need to finish up and publish the threadcalc iphone app I wrote a couple years ago
[18:37:57] <micges> 'nite
[18:38:06] <atom1> andypugh, you sure it's not 36?
[18:38:16] <atom1> that's standard fine
[18:38:38] <pfred1> sure you can't just run it throug ha die?
[18:38:44] <andypugh> That's .164 nominal. Knock off .00625 for the thread-flats, and that leaves .158. The fattest of the screws I measured is .157.
[18:38:57] <andypugh> (the others are .147)
[18:38:59] <atom1> sitting idle i get "can't do that (EMC_TASK_PLAN_RUN) in manual mode
[18:39:24] <andypugh> It's definitely 40tpi. I wanted it to be 36 :-)
[18:39:25] <ssi> andypugh: remember the flat callout changes with pitch
[18:39:32] <andypugh> It's 1/8P
[18:39:35] <atom1> heh
[18:40:09] <ssi> ok, as long as you remember ;)
[18:40:10] <pfred1> atom1 sounds like lyrics from an 80s pop song
[18:40:16] <andypugh> atom1: Something is sending random messages to your system.
[18:40:41] <pfred1> no, no, can't do that!
[18:42:27] <atom1> ok i added that to the ini
[18:42:29] <atom1> now what?
[18:42:31] <pfred1> I often wonder if a jounraling filesystem is the best thing to run on an RT system
[18:42:31] <andypugh> I just googled for 8-40 ASME. The first hit was no help to me at all.
[18:42:49] <pfred1> I use EXT2 and just take the risk
[18:43:12] <pfred1> [ 2.532334] EXT4-fs (sda1): mounted filesystem with ordered data mode
[18:43:16] <atom1> i'll try running the file again
[18:43:28] <pfred1> atom1 did you set the debugging flag?
[18:43:39] <atom1> yes
[18:43:44] <pfred1> OK just checking
[18:44:33] <atom1> where does that log to?
[18:44:53] <pfred1> should log to dmesg or perhaps /var/log/syslog or messages
[18:45:07] <pfred1> just do tail -f /var/log/syslog
[18:45:13] <atom1> ok just stopped midway thru the cut and asked for a tool change
[18:45:14] <pfred1> and you can see crap pop up in real time
[18:45:34] <pfred1> it'll like scroll for you
[18:48:48] <atom1> posted new dmesg, same link
[18:50:21] <pfred1> you can copy paste pretty easily in terminals too
[18:50:53] <Roguish> Hey out there.. Can the Mesanet 7i43 be used with stepper motors
[18:51:21] <pfred1> atom1 is your parallel port disabled in your bios?
[18:51:31] <atom1> i don't think so
[18:51:38] <pfred1> atom1 [ 2029.533653] hm2_7i43.0: board at (ioaddr=0x0378, ioaddr_hi=0x0778, epp_wide ON) found
[18:51:45] <atom1> it wouldn't work at all if it were?
[18:51:54] <pfred1> ioaddr=0x0378 is the standard IO port for a parallel port
[18:52:05] <atom1> epp mode?
[18:52:10] <pfred1> any mode
[18:52:23] <atom1> lemme go check
[18:52:25] <pfred1> that is the address
[18:52:26] <atom1> back soon
[18:52:55] <atom1> i haven't reset it so it should be
[18:53:01] <pfred1> I don't know maybe there isn't a conflict
[18:53:14] <pfred1> but it sure is something I'd look at
[18:53:41] <pfred1> long and hard
[18:55:25] <atom1> it's enabled in EPP mode
[18:55:39] <pfred1> atom1 OK so is this board connected to the standard parallel port throug ha cable?
[18:55:47] <atom1> yes
[18:55:50] <pfred1> OK you'
[18:55:52] <atom1> i swapped cables too
[18:55:55] <pfred1> OK you're OK then
[18:56:13] <pfred1> I thought it was maybe an internal board trying to sit on the same IO
[18:56:33] <atom1> no, bare bones atom D525
[18:56:41] <atom1> nothing on the USB either
[18:56:53] <pfred1> PS/2?
[18:57:00] <atom1> kbd and mouse
[18:57:10] <pfred1> well they're round jacks right?
[18:57:15] <atom1> yes
[18:57:31] <pfred1> PS/2 is actually higher performance than USB for keyboards and mice
[18:57:41] <pfred1> no polling
[18:59:55] <atom1> ok, mdi i just told it to move to x0 y0 and it didn't make it all the way
[19:00:06] <atom1> stopped about .2 from the destination
[19:00:20] <pfred1> are you tailing syslog with the debugging flaf set?
[19:00:23] <pfred1> flag even
[19:00:26] <atom1> soon after that the axis jumped a bit
[19:00:40] <atom1> it's not open atm
[19:00:47] <atom1> after rebooting
[19:00:57] <pfred1> well how are you supposed to spot events if you're not looing for them?
[19:01:03] <pfred1> looking even
[19:01:28] <pfred1> something kooky is going on and you have to catch it in the act
[19:01:39] <atom1> tis on now
[19:02:22] <pfred1> if this doesn't work you may have to build LinuxCNC with debugging symbols or something esoteric like that
[19:03:36] <pfred1> which is only fun to a limited segment of the population
[19:03:50] <atom1> not this segment
[19:03:58] <pfred1> yeah I'm not very keen on it myself
[19:04:10] <pfred1> but sometimes we have to go these places
[19:04:39] <pfred1> beats the alternative shich is getting the mushroom treatment
[19:04:57] <pfred1> you know be kept in the dark and fed crap
[19:12:43] <PCW> when it stops can you read some HAL pins?
[19:13:07] <atom1> well it just locked up the whole system there for a bit
[19:13:26] <atom1> i had to close lcnc with the keyboard commands
[19:13:26] <pfred1> supposedly an app called screen can help there
[19:13:28] <atom1> no mouse
[19:13:32] <PCW> Hm bad hard drive?
[19:13:50] <atom1> i'm getting similar errors on 2 different pcs
[19:14:03] <PCW> with all different parts?
[19:14:15] <atom1> yes
[19:14:22] <pfred1> same 7i43 or a different one?
[19:14:49] <atom1> one is running 8.04 and the other is running 10.04 both with 2.5.2 and same mesa cards
[19:15:03] <pfred1> so the mesa card is the constant?
[19:15:25] <atom1> i'm not ready to blame anything just yet
[19:15:37] <pfred1> well you have to blame the one thing the problem follows around
[19:15:39] <atom1> but i have swapped out parport cables and pcs
[19:15:56] <PCW> Unless its shorts out your power, I dont see how the 7I43 can cause this kind of error
[19:16:37] <atom1> i haven't noticed any power issues but don't have a meter on the supply either
[19:16:42] <PCW> even if it scrambles the EPP bus the worst you will get is a EPP timeout
[19:17:06] <atom1> no errors to the log file when it quit this time
[19:17:13] <atom1> it just stopped
[19:18:00] <atom1> i replugged the mouse to get it back
[19:18:12] <pfred1> oh man don't hot swap PS/2
[19:19:02] <pfred1> usually modern hardware tolerates it but there is no guarantee it will
[19:19:47] <PCW> Try the sample 7I43 config (with step drives powered down)
[19:21:20] <atom1> not sure i can do that easily without unplugging a bunch of wires
[19:21:50] <PCW> cant just turn off step motor power?
[19:22:00] <atom1> i'm running a 7i47b bit file
[19:22:12] <atom1> the 5v is wired into the main supply
[19:22:28] <atom1> well it's using a center tap
[19:22:39] <pfred1> a center tap you say?
[19:22:47] <atom1> 24v
[19:22:58] <atom1> 48 is too much for the smps
[19:23:25] * pfred1 is jealous
[19:23:32] <pfred1> I only run 24V
[19:23:39] <pfred1> crappy TB6560s
[19:24:01] <PCW> You could unplug the flat cables from the 7I43 (with power down)
[19:24:29] <atom1> the 50 pin?
[19:24:48] <atom1> part of the wiring is on the 7i43 though
[19:24:59] <atom1> no 50 pin on the 2nd port
[19:25:26] <PCW> but I really doubt its I/O related, nothing low level I/O can do to interp sequence to make it skip ahead
[19:25:39] <atom1> i could unplug the geckos
[19:26:26] <atom1> wanna have a look at it next week? :)
[19:26:48] <PCW> also if you had EPP errors you would expect timeouts and ferrors
[19:27:00] <atom1> i'm not logging any errors yet
[19:27:47] <PCW> yeah really looks like a memory/stack kind of problem
[19:28:15] <atom1> would swapping cards around make a difference?
[19:28:18] <atom1> or unplugging one
[19:28:35] <atom1> i think i can run 1 on this board but i'm not sure
[19:28:44] <atom1> it's got 8G iirc
[19:28:50] <atom1> 4 or 8
[19:29:23] <atom1> how would that explain errors on the other PC?
[19:29:53] <PCW> config file error?
[19:30:20] <PCW> did you change something in the HAL or INI?
[19:30:37] <atom1> http://tom-itx.dyndns.org:81/~webpage/cnc/configs/sherline/
[19:30:51] <atom1> there they are except i removed the halui stuff
[19:30:57] <atom1> postgui i mean
[19:31:18] <atom1> from the ini
[19:31:32] <pfred1> PCW initially RAM errors were my gut guess
[19:32:52] <pfred1> atom1 before I run LinuxCNC on a machine I really like to stress test it to make sure it is 100%
[19:33:26] <pfred1> LinuxCNC can be enough of a pain without other issues cropping up to make it even more interesting than it already is
[19:33:59] <atom1> i'm gonna shut it down for tonight
[19:34:07] <pfred1> no run memtest
[19:34:20] <pfred1> just let it run a bunch of passes
[19:34:38] <pfred1> even memtest doesn't catch all RAM errors though
[19:35:20] <pfred1> but if you can do 3 passes OK it is a good sign
[19:36:22] <atom1> where do i find that?
[19:36:29] <pfred1> it should be on the install CD
[19:36:40] <pfred1> should be an option on the boot menu
[19:36:59] <pfred1> says something like memory check or test or some such
[19:37:15] <pfred1> almost every LInux install has memtest
[19:37:57] <pfred1> or you can download like a flash image of it off the net
[19:38:17] <pfred1> but it should be right in the opening boot screen of the install CD
[19:38:46] <atom1> i'll go look for it
[19:38:46] <atom1> bbl
[19:38:51] <pfred1> it is basically a hacked Linux kernel and all it does is check your RAM
[19:40:36] <pfred1> I had flakey RAM issues with the last PC I setup LinuxCNC on but I got them straightened out
[19:48:43] <Jymmm> pfred1: "Value Ram" ?
[19:49:15] <pfred1> Jymmm no it was pretty good stuff actually
[19:50:32] <Jymmm> pfred1: Ah. I just bought 16GB but paid a lil extra for the good stuff as it's "technically" not supported =)
[19:51:02] <pfred1> all RAM is pretty delicate stuff
[19:51:13] <pfred1> billions of parts and if one goes it is shot
[19:51:26] <Jymmm> pfred1: It can be, but this has better specs than most.
[19:51:32] <pfred1> doesn't matter
[19:51:49] <Jymmm> pfred1: Corsair Vengence
[19:51:53] <pfred1> not like any RAM manufacturer can rewrite the laws of physics to favor their hardware or anything
[19:52:47] <pfred1> far as i know there are only really 3 RAM manufacturers in the world anyways
[19:53:00] <pfred1> least people who actually make the chip dies
[19:58:09] <Tom_itx> this is kingston
[19:58:10] <Tom_itx> ram
[19:58:44] <pfred1> any ram can go bad
[19:59:01] <pfred1> the law of high numbers just works against the stuff
[19:59:03] <Tom_itx> i'm aware of that
[19:59:56] <pfred1> supposedly Linux can map around bad RAM
[20:00:44] <pfred1> I had one module go bad on me and I was in X Window and when I opened up firefox my machine like turned into molassass
[20:00:52] <pfred1> I was like what the heck?
[20:01:11] <pfred1> then I did free -m and the machine reported i only had 22Mb of RAM installed
[20:01:58] <pfred1> up until I tried to run the browser it seemed OK
[20:02:35] <pfred1> another place where bad RAM will show up is during big compiles
[20:02:56] <pfred1> where the machine will create huge data structures in RAM GCC will catch checksum errors
[20:03:16] <Tom_itx> i'll let this memtest run a while
[20:03:21] <pfred1> GCC is probably the best RAM test on the planet
[20:03:37] <pfred1> I've seen it catch RAM that memtest didn't
[20:04:50] <pfred1> Tom_itx if memtest hasn't caught anything in 3 passes it probably won't catch anything at all
[20:05:03] <pfred1> but I've seen bad RAM pass memtest too
[20:05:22] <gene78> Back from dinner, now, what do i do to make linuxcnc-dev install?
[20:05:42] <pfred1> aptitude install linuxcnc-dev ?
[20:06:57] <pfred1> that is assuming the package name is really linuxcnc-dev and you have aptitude installed
[20:06:57] <gene78> I'm on the 2.6.0-pre branch, and becauseof dependencies nothiing in linuxcnc will update
[20:07:18] <gene78> yup, thats all here.
[20:07:39] <pfred1> well a lot of folks say not to update
[20:07:56] <pfred1> but i have without negative consequences
[20:08:08] <pfred1> but I usually build a custom kernel too
[20:08:21] <gene78> libmodbus is the depemdency its yelling about
[20:08:53] <pfred1> it doesn't exist on my Squeeze system
[20:09:22] <gene78> This is 10.04-4 LTS, stock kernel
[20:09:29] <pfred1> you might have to rebuild your dpkg database
[20:09:57] <pfred1> yeah whatever package that is is obsolete now as far as Debian is concerned
[20:10:49] <gene78> Its also saying that the dev package is for a newer build than I have. Too bad I can't copy-paste between machne screens
[20:11:06] <pfred1> start an ssh session
[20:11:42] <gene78> Whats the routine to rebuild the dpkg database?
[20:12:09] <pfred1> that's a good question I've redone it but when I did it the command to do it was right on the screen for me
[20:12:42] <gene78> Let e see if I can do it on this machine. BRB
[20:13:53] <pfred1> http://people.adams.edu/~cdmiller/posts/Ubuntu-dpkg-recovery/
[20:14:18] <pfred1> I didn't have to do all of that it was like a single command as I can recall
[20:14:21] <gene78> linuxcnc-dev:
[20:14:23] <gene78> Depends: linuxcnc (=1:2.6.0~pre0.4496.gd2f30b2) but 1:2.6.0~pre0.4470.g3e81d30 is to be installed
[20:14:44] <pfred1> update your database aptitude update
[20:16:18] <pfred1> that is kind of kooky it looks like the cev package is a bit behind
[20:16:22] <pfred1> dev package even
[20:18:31] <pfred1> man a lot of the ubuntu help seems less than helpful anymore
[20:19:09] <gene78> aptitude update wants to fix it by removing linuxcnc
[20:19:25] <pfred1> this is when I start to hate dependencies
[20:19:41] <Valen> anybody in .au looking for hardware http://www.graysonline.com/sale/5006729/manufacturing/insolvency-sale-unreserved-cnc-engineering-equipment?spr=true
[20:20:03] <gene78> makes 2 of us
[20:21:02] <gene78> does this problem exist if i go back to the 2.5.2 branch?
[20:21:20] <pfred1> I don't know I built LinuxCNC from git
[20:21:47] <gene78> which branch?
[20:22:20] <pfred1> I think 2.6
[20:22:30] <pfred1> it was like a month or so ago
[20:22:38] <gene78> humm, on which distro?
[20:23:00] <pfred1> well first i tried Gentoo and ran into a few problems so I went to Debian Squeeze and got it all to work there
[20:23:11] <pfred1> http://i.imgur.com/FUL7PwG.png
[20:23:25] <pfred1> now that i have some of the config wrinkles ironed out I may try gentoo again
[20:23:42] <gene78> a month ago it was fine, and being updated almost daily. but its been gghosted out for a week or more now.
[20:24:13] <pfred1> I had problems getting RTAI from cvs but there was a work around for that
[20:24:36] <pfred1> seems their cvs server is broken
[20:25:37] <gene78> ah, well, i expect it will get sorted, i was hoping before Wichita though.
[20:25:49] <pfred1> it does seem like you have to wait until the planets are aligned to get this stuff to work sometimes
[20:26:01] <gene78> :)
[20:26:26] <pfred1> building LinuxCNC from source and getting it to work isn't always a trivial feat
[20:26:58] <pfred1> I had to email some guy in italy who writes RTAI to figure out the final thing
[20:27:32] <pfred1> had to do with the fact I was using a multi core CPU
[20:28:24] <pfred1> but he told me one setting to do in the kenel and it all worked after that
[20:28:53] <gene78> which is why I let TPTB do it,I just report what I think is a bug. :) Learning a new language at my age (78) is probably not in the cards. Sigh.
[20:29:12] <pfred1> I'm 48 so I'm kind of pre PC myself
[20:29:21] <gene78> I can recall when that was fun, 40 years ago.,,
[20:29:21] <pfred1> but i try
[20:30:18] <gene78> so I think I'll head back to the house for the night, thanks.
[20:30:22] <pfred1> from what I've seen of the stock kernel config going through the trouble does seem worth it to me
[20:30:34] <pfred1> it is a kitchen sink of sorts
[20:31:38] <pfred1> my system was loading all kinds of modules and drivers it didn't really need
[20:32:14] <Tom_itx> you gotta be able to determine what you need and don't need before attempting that though
[20:32:24] <Tom_itx> i'm not close to that
[20:32:30] <pfred1> Tom_itx it's not that hard
[20:32:50] <Tom_itx> nothing is that hard once you've done it once or twice
[20:32:52] <pfred1> run this lspci -vv | grep driver
[20:32:58] <pfred1> that is half of the battle
[20:33:23] <pfred1> if it ain't there you probably don't really need it
[20:34:03] <pfred1> now do modprobe
[20:34:08] <pfred1> and see how long that list is
[20:34:41] <pfred1> well lsmod actually
[20:35:39] <pfred1> on stock kernels a lot of modules load that are questionable if you really need them
[20:36:44] <pfred1> one thing Gentoo got me into again was getting away from an initramfs
[20:36:56] <pfred1> and building just what I needed monolithically into the kernel
[20:37:10] <pfred1> until I ran Gentoo I didn't even think it was possible to do that anymore
[20:38:36] <pfred1> configuring a kernel well is pretty challenging to do even if you've done it a dozen times
[20:38:51] <pfred1> there are an awful lot of options
[20:39:10] <pfred1> somewhere north of 2,700 the last time I looked
[20:40:00] <pfred1> honestly it takes me a few tries on any machine I do it on sometimes as many as 6
[20:40:25] <pfred1> but by then I get close to dialed in
[20:41:11] <pfred1> I should build a new kernel for this system it has a cockeyed bug in it that annoys me
[20:44:08] <pfred1> if I don't issue a command similar to this my dmesg gets spammed with messages echo "1" > /sys/devices/pci0000:00/0000:00:1d.7/usb3/3-6/avoid_reset_quirk
[20:45:23] <pfred1> but my other fix is to just not look at dmesg :)
[20:46:21] <pfred1> because other than that this PC is all of that and a bag of chips as far as I'm concerned
[20:59:53] <Willburrrr2003> wooohooo , after working on my system the last couple days i finally have modbus communicating with my external PLC !
[21:01:09] <Willburrrr2003> so far i only have clasic ladder turning on a relay output in my PLC, but i feel it is a big step forward!
[21:05:59] <skunkworks> yay!
[21:11:41] <Willburrrr2003> i am not seeing a modbus setting to read internal bits from the PLC, they are the equivilant of the classicladder %B variables...anyone know which setting will let me read internal bits?
[21:17:10] <Willburrrr2003> having communications to my external PLC leave lots of exciting possibilities, like setting up an automatic tool turret which i have been dreaming of for awile now
[21:45:56] <Tom_itx> pfred1, first pass no errors
[21:47:35] <pfred1> Tom_itx keep it up
[21:47:55] <pfred1> it takes a while huh?
[22:00:14] <Tom_itx> i just let it run
[22:00:21] <Tom_itx> and went on to other stuff
[22:04:34] <Willburrrr2003> what are you making?
[22:05:08] <Tom_itx> just trying to get the mill to run without errors
[22:05:22] <Willburrrr2003> ahhhh ok
[22:05:45] <Willburrrr2003> what errors have you been experiencing?
[22:06:11] <tjb1> postaL: ?
[22:06:17] <skunkworks> I had a hd go bad and cause all kinds of odd issues
[22:06:48] <Willburrrr2003> that sucks
[22:07:37] <skunkworks> long pauses - var file corruption.. and finally the mesa watchdog bit and sent the machine into estop. the smart data for the drive showed a high number of pending sector relocations.
[22:08:22] <skunkworks> it was a re-purposed hd... I wasn't suprised..
[22:08:35] <pfred1> skunkworks there is smart for Linux
[22:09:01] <skunkworks> yep. (well you can see it in the disk utillity...)
[22:09:10] <skunkworks> that is what I use...
[22:09:11] <Willburrrr2003> that reminds me, i should back up my config files after my recent upgrades
[22:09:13] <pfred1> I use gsmartcontrol
[22:10:09] <pfred1> I had one HDD go flakey on me and I ran fdisk on it and told it whatever as far as repairing errors go and that disk lasted for years after that
[22:10:17] <pfred1> well fsck
[22:10:35] <pfred1> but before i fixed it I couldn't even mount it
[22:11:16] <pfred1> as far as I know that HDD is still good in whatever system I have it in
[22:12:10] <pfred1> although I'm not sure I'd totally trust it myself for anything important
[22:12:39] <pfred1> overall Linux does seem to go pretty easy on HDDs
[22:13:20] <pfred1> I have a system with a 14 year old HDD in it and when it cracked 10 years i started getting a little worried about the HDD dying in it
[22:13:31] <pfred1> but it just kept working and working
[22:14:20] <pfred1> again as far as i know if I hit the power button on that PC today it'd still work
[22:14:43] <pfred1> but for the first 14 years that PC ran 24/7/365
[22:15:05] <ssi> you've had better luck with hard drives than me then
[22:15:25] <pfred1> ssi you run Linux all of the time?
[22:15:28] <ssi> yep
[22:15:31] <pfred1> wow
[22:15:41] <ssi> I had an ipcop router here that had an eight year uptime
[22:15:44] <ssi> finally got reset
[22:15:45] <pfred1> well when I ran Windows I did go through some HDDs back then
[22:15:49] <ssi> and it wouldn't come back up
[22:15:55] <ssi> hdd wouldn't make it through fsck
[22:16:16] <ssi> linux is definitely easier on drives than windows is
[22:16:29] <ssi> cause windows swaps constantly and thrashes the disk heh
[22:16:33] <pfred1> yup
[22:16:48] <pfred1> Linux manages hardware way better
[22:16:48] <RyanS> This is hilarious I bought this item a couple of years ago http://www.ebay.com.au/itm/Special-Magnetic-Base-0-40-0-Dial-test-indicator-Measure-Bool-/160789014681?pt=LH_DefaultDomain_15&hash=item256fc5a899
[22:17:24] <pfred1> RyanS what is so funny about it?
[22:17:34] <RyanS> it is literally identical to this http://www.uktools.com/mitutoyo-513-415e-dial-test-indicator-101mm-p-64736.html?osCsid=e992df032d9ea1a4e0fa71919d6faa35
[22:18:23] <pfred1> so they fell off of a boat
[22:18:32] <RyanS> even says made in Japan on the front, it's like the genuine Rolex of dial test indicators
[22:18:35] <Tom_itx> the errors i've been getting are all in this log
[22:18:37] <pfred1> their loss is your gain!
[22:19:15] <pfred1> until I dropped my dial calipers they worked great
[22:19:28] <pfred1> turned out they weren't too shockproof
[22:19:41] <pfred1> but they were some cheap Chinese off brand
[22:19:47] <RyanS> so it could be some grey market thing or it's literally a knockoff
[22:19:51] <ssi> even good ones don't survive mistreatment :/
[22:19:57] <ssi> I have a couple pairs of nice calipers that both need repair
[22:20:02] <ssi> or maybe are beyond repair, I dunno
[22:20:06] <pfred1> yeah it was a high drop onto a concrete floor
[22:20:06] <ssi> so I mostly just get cheap ones
[22:20:14] <ssi> i'm hard on stuff like calipers
[22:20:31] <ssi> sucks, cause I really like nice ones
[22:20:32] <pfred1> mine lasted quite some time 10 years?
[22:20:43] <ssi> but I'm hesitant to spend more money on new ones
[22:20:57] <RyanS> You think a phoney DTI is necessarily less accurate?
[22:21:10] <ssi> nah
[22:21:24] <ssi> it'll be less smooth, more stiction, which could impact fine repeatability
[22:21:29] <pfred1> well my work is never too close tolerance even cheap measuring equipment is better than I am likely going to do
[22:21:54] <pfred1> I have a micrometer set but they just sit on a shelf it simply doesn't come down to it
[22:22:09] <ssi> I'd rather have two cheap .001" DTIs and two cheap .0001" DTIs than just one nice .001"
[22:22:42] <jdh> I have 5 or 6 HF digi ones and one 0.0005 mituyo. Other than one flakey HF that won't old 0, they all read the same. Half a thou is better than I can do :(
[22:22:46] <pfred1> I don't know what brand my mikes are but they're Polish with standards etc
[22:22:59] <pfred1> supposedly better than chinese stuff
[22:23:10] <ssi> https://fbcdn-sphotos-g-a.akamaihd.net/hphotos-ak-frc1/303191_725209428042_2131740446_n.jpg
[22:23:19] <ssi> those are older but decent .001
[22:23:27] <ssi> I put chinese .0001" on it after it was dialed in there
[22:25:07] <RyanS> the thing with DTI I only ever see them being used to zero things so you sort of take a nominal value rather than literally taking a 'measurement'?
[22:25:15] <ssi> that's right
[22:25:21] <ssi> because they don't have a reference
[22:25:52] <ssi> they're for measuring runout, backlash, etc
[22:25:53] <pfred1> well they get used on surface plates with stands
[22:26:06] <ssi> yeah that too
[22:26:18] <ssi> although I mostly use a height gage for that
[22:26:20] <RyanS> So it's probably not as supercritical as something like a micrometer
[22:26:24] <jdh> it was 'it' with my right hand one key to the right
[22:26:28] <jdh> <urk>
[22:27:02] <pfred1> RyanS I worked in a tool and die shop and on one piece of equipment we built the final test we'd mount a carbide ball in the unit and indicate it
[22:27:20] <pfred1> it the needle moved more than a half a tick it was no good
[22:27:33] <pfred1> if even
[22:28:01] <pfred1> once i made one and the needle didn't even move I thought I broke the indicator
[22:28:18] <pfred1> turned out i just got lucky and the thing had no play in it at all
[22:28:39] <pfred1> I had 3 other people look at it
[22:28:47] <ssi> there's a vido of a guy making parts on an HNC and measuring them with I think a 50uinch passometer
[22:28:50] <ssi> it's pretty impressive
[22:30:23] <RyanS> So what's the big daddy of metrology equipment that if you can't trust its accuracy you can't trust any other piece of equipment (the master reference). Something you can't even look at the wrong way for fear of upsetting its accuracy?
[22:30:43] <pfred1> instead of the diamond we had carbide balls soldered to shafts we'd put in these things http://thumbs2.ebaystatic.com/d/l225/m/mLlg-pSA33bGrxOnEplJHfQ.jpg
[22:30:58] <pfred1> then move it through it's range of motion
[22:31:23] <RyanS> Or hurting its feelings if you misstreat it
[22:31:31] * pfred1 knows how to make one of those
[22:31:32] <ssi> RyanS: neat thing is if you have enough time, you can make all your references :)
[22:32:02] <pfred1> that one there new you're talking about $5,000
[22:32:03] <RyanS> But how could you make your own reference if you don't have another reference that you know is accurate
[22:32:12] <ssi> RyanS: by making more than one
[22:32:20] <ssi> RyanS: you can make a flat surface by scraping two surfaces together
[22:32:28] <pfred1> 3 surfaces
[22:32:28] <ssi> you can make reference squares by scraping THREE together
[22:32:33] <ssi> you can do 2 surfaces
[22:32:38] <pfred1> 2 one could be convex one concave
[22:32:38] <ssi> you jsut have to rotate one 90 degrees
[22:32:44] <ssi> yeah that's true
[22:32:46] <ssi> 3 surfaces then
[22:32:52] <pfred1> the rule of 3
[22:33:05] <pfred1> 3 surfaces they all have to be flat to each other
[22:33:07] <ssi> I dunno how you'd get an accurate inch
[22:33:09] <RyanS> but how do you know if those two are accurate if you can't reference this to something else?
[22:33:29] <ssi> RyanS: like pfred is saying, if there are three that all spot together, they're definitely accurate
[22:33:43] <pfred1> there is no combination of convex and concave that'll work out touching on all 3
[22:34:09] <ssi> well yes there is, if you don't check in 90 degree orientations
[22:34:11] <ssi> a potato chip shape
[22:34:14] <ssi> like a pringle
[22:34:22] <ssi> it'd be possible to spot all three together in a specific orientation
[22:34:29] <pfred1> all I'm saying is 2 never works
[22:34:48] <pfred1> or you can't be sure
[22:34:50] <ssi> right
[22:35:09] <pfred1> I see 3 work that is good enough for me!
[22:35:49] <pfred1> I know with the metric system now the reference is a wavelength of light
[22:35:54] <ssi> yeah
[22:36:03] <ssi> but that's hard to meausre with caveman tools
[22:36:16] <pfred1> it used to be a platinum bar but if you picked it up the heat from your hand would distort it
[22:36:43] <ssi> a decent screw-type micrometer is a remarkably accurate and durable tool to have around
[22:36:53] <ssi> but that'll distort by microinches from hand heat too
[22:37:03] <pfred1> what blows my mind is how accuratly they could measure things 100 years ago
[22:37:06] <ssi> and I dunno how you cut a 40tpi screw without another 40tpi screw
[22:37:08] <ssi> yeah
[22:37:45] <pfred1> I mean 100 years ago somehow they knew about the platinum bar problem
[22:37:55] <pfred1> they didn't have a solution but they were aware of it
[22:38:02] <RyanS> some youtube video did "Chinese spotting" and blued a banana and squished it on a way. i dont get it
[22:38:50] <pfred1> RyanS if you ever saw the gib out of my milling machine you'd get it
[22:39:26] <pfred1> thing looks like it got surfaced by being tossed in the gravel driveway of the plant
[22:39:36] <pfred1> and letting truck run over it repeatedly
[22:41:06] <RyanS> they just dont use an accurate reference so might as well be a banana ?]
[22:41:24] <pfred1> I think what the Chinese do is they have trade schools where they make the stuff they ship to us
[22:41:45] <pfred1> then once the students get good they get moved somewhere else where they keep the goods they make
[22:43:33] <pfred1> I used to buy a lot of chinese stuff not so much anymore
[22:43:58] <pfred1> I've gotten burned too many times sometimes literally
[22:44:20] <pfred1> once i was running this jig saw it was a bosch knock off the thing blew up in my hand
[22:44:26] <ssi> lol
[22:44:29] <jdh> I did some cuts this evening with a brand new chinese 6mm 2flute end mill. The cut liked like someone had used a screwdriver on it
[22:44:39] <pfred1> I did a forensic of it afterwards it was quite devlish how it happened
[22:44:49] <ssi> me and my friend once had the brilliant idea that we could make cheap brass tumblers by buying $7 oscillating sanders from harbor freigth and bolting them to tumbler bowls
[22:45:01] <ssi> it worked ok, but the things caught on fire after running continuously for an hour
[22:45:16] <pfred1> there was this ceramic piece in there kind of rough that the trigger wiped against
[22:45:30] <pfred1> well as the trigger wiped the copper built up on that roug hceramic
[22:45:40] <pfred1> once the copper built up enough kaboom!
[22:46:20] <pfred1> it was like a freaking time bomb
[22:46:35] <pfred1> every time you pulled the trigger yo ugot a little closer to it blowing up
[22:46:40] <ssi> awesome
[22:47:03] <pfred1> man when that thing blew up in my hand I threw it
[22:47:14] <pfred1> it was like an electrical explosion going on with it
[22:47:29] <pfred1> it was nothing you wanted to be holding let me tell you
[22:47:39] <RyanS> Imagined in 100 years however if meteorologists go "you know that master reference in the museum of 1 m", "well the super accurate instrument we measured it with and then sore the results in light waves - it was wrong"
[22:48:03] <RyanS> Every measurement we have taken since was wrong
[22:48:17] <ssi> well the beautiful thing is it can't be wrong
[22:48:21] <pfred1> the whole metric system was originally based on a flawed measurement of the Earth
[22:48:24] <ssi> because we DEFINED THE UNIT in terms of the wavelength
[22:48:25] <jdh> it's arbitrary anyway
[22:48:40] <pfred1> it was supposed to be 10,000 kilometers from the equator to the pole or something
[22:48:47] <pfred1> but it isn't even close
[22:49:03] <pfred1> they totally blew calculating the diameter of the planet
[22:49:06] <ssi> lol
[22:49:19] <ssi> well it'd be a quarter circumference
[22:49:42] <pfred1> well all I'm saying is they messed it up from the word go
[22:49:43] <ssi> and how the hell do you measure that accurately in 1800s tech anyway?
[22:49:51] <pfred1> apparently they didn't
[22:49:53] <ssi> :D
[22:50:04] <pfred1> but you can do it with trig
[22:50:06] <toastyde2th> didn't some egyptian get the circumfrence (hence the diameter) within 2%?
[22:50:10] <toastyde2th> like, waaay back in the day?
[22:50:11] <ssi> yeah I guess so
[22:50:11] <pfred1> some greek looked down a well and got close
[22:50:30] <ssi> shoot a star with an astrolabe, walk north, shoot it again
[22:50:33] <ssi> trig it out
[22:50:38] <pfred1> toastyde2th it was a greek
[22:50:39] <ssi> something like that
[22:50:50] <pfred1> in like 2,500 BC
[22:50:51] <ssi> I dunno, I never studied celnav
[22:50:59] <RyanS> I can tolerate decimal Imperial but I looked at a model steam engine plan, it was all in fractions and my brain melted
[22:51:10] <ssi> hahaha
[22:51:17] <pfred1> fractions are good for construction
[22:51:19] <ssi> I can deal with fractions
[22:51:28] <ssi> most of my airplane work is in fractions
[22:51:31] <ssi> to the 64th
[22:51:40] <ssi> it's easy
[22:51:44] <ssi> a 32nd is about 0.032
[22:51:47] <ssi> 0.03125
[22:51:52] <pfred1> 64s are tough for me to make out on a scale 16ths aren't bad though
[22:52:05] <ssi> 64ths aren't bad on a decent scale
[22:52:23] <pfred1> nah I'd rather use calipers
[22:52:26] <ssi> there's a couple of those key numbers you remember and everything's not too bad
[22:52:31] <ssi> 1/32 ~ 0.032
[22:52:35] <ssi> 1mm ~ 0.040
[22:52:46] <ssi> how big is 3mm? 3x40 = 120, so about an eight
[22:52:47] <ssi> not bad
[22:52:56] <toastyde2th> there's a secret to decimals
[22:53:13] <toastyde2th> what's good about them is they're sort of a quick engineering check
[22:53:17] <RyanS> I don't get the UK it appears to be a hodgepodge of metric and imperial.....
[22:53:20] <pfred1> nah you need the magic chart http://www.delorie.com/wood/DrillGuide.html
[22:53:29] <toastyde2th> an engineering print with fractions on it will likely have few errors
[22:53:40] <toastyde2th> if any
[22:53:43] <pfred1> I copied that off a steel supplier handout card
[22:54:08] <toastyde2th> a print with uniform tolerances is likely to be more erronious
[22:54:43] <RyanS> Or is model engineering for old people who can't bring themselves to work in metric regardless of whether it's the UK or Australia?
[22:55:03] <toastyde2th> generally it doesn't matter to a machinist what unit you're working in
[22:55:10] <toastyde2th> it's an engineering concern
[22:55:25] <pfred1> toastyde2th like if you're trying to land a probe on Mars
[22:55:29] <toastyde2th> yep
[22:55:46] <pfred1> and you're using miles but the computer is programmed in kilometers
[22:55:50] <RyanS> i read that machinists don't care
[22:56:11] <pfred1> that is what happened to one of our probes
[22:56:11] <toastyde2th> yep, it's just a number, and nothing is ever dead nuts so you have to deal with error anyway
[22:56:24] <toastyde2th> that .500" endmill is really .501" or something
[22:56:37] <pfred1> well it is going to deflect
[22:56:40] <toastyde2th> so it doesn't matter whether it's mm or inches
[22:56:51] <pfred1> I used to run endmills until they snapped
[22:56:52] <toastyde2th> because it's all just numbers you have to correct and bump around, the unit is meaningless
[22:56:58] <RyanS> Is anyone here from the UK? They use imperial for roads but metric for other things?
[22:57:00] <pfred1> they'd change size before they snapped
[22:57:35] <pfred1> RyanS I live in the USA and we're technically metric but we use Imperial for almost everything
[22:57:58] <RyanS> I thought the US was officially imperial?
[22:58:02] <pfred1> nope
[22:58:07] <pfred1> we are officially metric
[22:58:13] <ssi> well the military is
[22:58:14] <jdh> we are?
[22:58:16] <pfred1> being the primary member of NATO
[22:58:16] <ssi> I wouldn't say the country is
[22:58:36] <pfred1> yes we are officially metric
[22:58:42] <pfred1> but you'd never tell it by being here
[22:58:48] <jdh> officially how so?
[22:58:58] <pfred1> the government uses all metric
[22:59:04] <toastyde2th> yep
[22:59:06] <pfred1> like our bullets
[22:59:25] <RyanS> It's funny when you deal with so-called metric pipe it's all just Imperial written down as metric
[22:59:31] <pfred1> all have metric designations
[22:59:35] <ssi> RyanS: yeah but there's a lot of that the other way 'round too
[22:59:42] <ssi> go to home depot and try to buy 3/4" plywood
[22:59:44] <pfred1> RyanS yeah or a 13mm drive
[22:59:48] <ssi> you get 25/32
[22:59:58] <pfred1> ssi that is because of surfacing
[23:00:10] <ssi> it gets bigger when you surface it?
[23:00:15] <pfred1> ssi if you buy hardwood plywood 3/4 is 3/4
[23:00:17] <ssi> I'm pretty sure they're selling us 20mm plywood
[23:00:35] <pfred1> nah it is a funny thing with plywood grading
[23:00:44] <RyanS> I think we have 20 mm tube fittings and it's not even close
[23:00:51] <pfred1> softwood plywood is sold like draming timber
[23:00:55] <pfred1> framing timber
[23:01:02] <pfred1> like a 2x4 isn't really 2x4
[23:01:18] <toastyde2th> "let's measure it before we put it through the planer"
[23:01:23] <pfred1> they use the presurfacing measurement
[23:01:26] <pfred1> right
[23:01:48] <ssi> same way they measure quarter pounders? ;)
[23:01:58] <pfred1> yeah precooked weight
[23:02:11] <pfred1> same principal
[23:02:35] <pfred1> they even make router bits that can dado for cockeyed plywood
[23:03:58] <RyanS> Why does everyone on youtube seem to go on about 0.0001" "I have a thou runout,that's pretty good"but is that good ?
[23:04:06] <ssi> that's a tenth
[23:04:11] <ssi> and a tenth runout is quite good
[23:04:25] <RyanS> What's a 10th
[23:04:33] <ssi> a ten-thousandth of an inch
[23:04:35] <ssi> 0.0001
[23:04:37] <pfred1> RyanS when I worked in a machine shop we never expected greater than 0.003" accuracy out of a lathe or a mill operation
[23:04:41] <ssi> or a tenth of a thousandth
[23:04:53] <pfred1> RyanS when we needed stuff closer than that we surface ground it
[23:05:28] <ssi> yea I spin all my ballscrew bearing journals in the grinder
[23:05:34] <ssi> journal is the wrong word
[23:05:39] <ssi> but you know what I mean :)
[23:05:45] <pfred1> but a lot of non professioals seem to get really hug up on machine cutting accuracy
[23:06:02] <pfred1> hung up even
[23:06:16] <toastyde2th> jesus christ "runout"
[23:06:19] <pfred1> I guess because they're not making a living out of it I don't know
[23:06:21] <toastyde2th> i love pissing matchies about that
[23:06:27] <toastyde2th> *matches
[23:06:35] <ssi> well it matters for some things
[23:06:39] <ssi> grinder spindles for instance :)
[23:06:45] <pfred1> man we ran machines that were beat to death
[23:06:56] <toastyde2th> yeah, but people don't seem to realize you can't measure runout with a dti
[23:07:04] <pfred1> but if you kept an eye on them you could manage good work
[23:07:11] <Tecan> yay z axis done now to hook it to duino http://i.imgur.com/Yd1Dzf5.jpg
[23:07:13] <toastyde2th> you can measure runout of a surface with a dti, but you can't measure the runout of the bearing itself with a dti
[23:07:27] <ssi> pfred1: I only have one machine that I really feel like is so worn out it's bordering on unusable
[23:07:41] <ssi> and the ways are just shot, and the carriage moves around a lot, unexpectedly
[23:07:59] <ssi> it's a lathe, and if you turn a diameter over 6", youll get a ridiculously uneven surface
[23:08:14] <toastyde2th> haha i love that on bridgeports - "And suddenly the table moved half an inch"
[23:08:37] <ssi> toastyde2th: that sounds more like cutter grabbing the work and pulling it into the backlash
[23:08:45] <toastyde2th> ssi, that's exactly what it is
[23:08:55] <pfred1> yeah climb milling
[23:09:10] <RyanS> what so can't measure the spindle runout or chuck runout with a dti??
[23:09:17] <toastyde2th> ryan, nope.
[23:09:19] <pfred1> when I pocketed clamps we'd climb mill them and it worked
[23:09:25] <ssi> on my lathe I can just feed a cutter toward the chuck on the powerfeed and it'll cut fifteen different diameters with steps :(
[23:09:51] <toastyde2th> ryan: the idea is that you can measure a part's runout no problem - just stick a dti on it, fine
[23:10:24] <toastyde2th> but the spindle's center of rotation is an imaginary line, and that line moves around. it doesn't have anything to do with parts you can see or put a simple measurement tool on
[23:10:42] <toastyde2th> people like to mesure the runout of a spindle taper, and that does matter in some cases (small endmills)
[23:10:43] <RyanS> don't you use a dti to clock in a 4jaw?
[23:10:51] <toastyde2th> yeah, but that's a part.
[23:10:58] <toastyde2th> that's not telling you anything about the spindle.
[23:12:02] <pfred1> toastyde2th I know what you mean I think
[23:12:05] <toastyde2th> old shop had two air bearing lathes from waaaay back in the day. If you measured any visible part of the spindle, you'd get crazy runout. Cut a test bar and look at it, and it was under a millionth of an inch. and that wasn't bearing accuracy, that was just the concentricity of the part after being cut
[23:13:20] <pfred1> if you need stuff so accurate you have to question the design to begin with often
[23:13:38] <pfred1> the Germans had great guns until they took them to Russia
[23:13:44] <pfred1> then they weren't so great
[23:13:49] <toastyde2th> it was neat because you could put a sharpie clocking mark on the part and set the measurement oscilloscope to trigger on the massive spike
[23:14:06] <RyanS> I guess there's nothing I can do to improve run out of lathe spindle, bu how do I tell if mythree draw a scroll chuck is shit?
[23:14:07] <toastyde2th> because a sharpie line is like 30-40 millionths of an inch high
[23:14:27] <RyanS> jaw
[23:14:31] <pfred1> I thought you could recut scroll chucks
[23:14:43] <toastyde2th> cut yourself a few diameters of pins
[23:14:47] <toastyde2th> 1", 2", etc
[23:14:52] <toastyde2th> close the 3 jaw on the pins
[23:14:54] <toastyde2th> measure the runout.
[23:15:24] <ssi> yea 3 jaws are mostly about how repeatable the concentricity is
[23:15:26] <toastyde2th> a good 3 jaw will usually be less than .005" runout, and a new chuck will do around .001" (in my experience)
[23:15:36] <toastyde2th> BUT.
[23:15:38] <ssi> you want to be able to cut something, remove it, put it back in, close it, and have it be pretty true
[23:15:43] <ssi> mine is the opposite of that
[23:15:46] <ssi> so I use mostly 4jaw :)
[23:15:52] <toastyde2th> you can use ANY beat-ass three jaw if you make yourself a set of soft jaws
[23:16:02] <toastyde2th> and then your beat up 3 jaw will be more accurate than a 4 jaw.
[23:16:19] <ssi> on my hardinge I use a lot of machinable step collets
[23:16:20] <toastyde2th> downside is that you have to cut the jaws every time you make a part
[23:16:24] <ssi> and they're VERY accurate
[23:16:32] <toastyde2th> ya, collets are baller
[23:16:40] <ssi> I have a part that I run that's 2" raw stock, and I finish it to 1.985"
[23:16:42] <toastyde2th> i always use collets if i can
[23:16:56] <RyanS> that's too much hassle for the home
[23:16:57] <ssi> and I have a step collet that's got a 1/4" deep 2" bore, and 1/2" deep 1.985" inside that
[23:16:59] <toastyde2th> most of our lathes had 16C collets rather than 6C
[23:17:05] <toastyde2th> and it was brilliant
[23:17:08] <pfred1> I need to get a set of 5C collets
[23:17:24] <ssi> so I can pop raw stock in, finish most of the part, flip it over in the deeper section, and finish the other end
[23:17:28] <ssi> it's pretty awesome
[23:17:32] <toastyde2th> i think 16C holds like 2.5" max
[23:17:32] <toastyde2th> thru spindle
[23:17:33] <ssi> especially with the turret and the air closer :D
[23:17:37] <pfred1> I picked up a collet holder but I have only one collet for it
[23:17:38] <ssi> 16C is 1.625"
[23:17:44] <ssi> 5C is 1.0625"
[23:17:46] <toastyde2th> hm
[23:17:59] <ssi> my HNC is the 5C one, but the CHNC is 16C
[23:18:03] <ssi> I wish I had the bigger spindle
[23:18:06] <ssi> but the collets cost more
[23:18:14] <toastyde2th> is there another size after that
[23:18:19] <toastyde2th> I swear I had bigger collets
[23:18:20] <ssi> not that I know of
[23:18:22] <pfred1> ssi lots say the ywish they had a bigger spindle
[23:18:32] <ssi> yeah everyone wants more
[23:18:33] <Xfriend> how do fix a nena 23?
[23:18:34] <ssi> heheh
[23:18:46] <pfred1> Xfriend a stepper motor?
[23:18:47] <Xfriend> pfred1: Hi...!!
[23:18:47] <ReadErro-> lol wat
[23:18:54] <Xfriend> yes
[23:19:02] <pfred1> you can't fix stepper motors
[23:19:06] <ReadErro-> find trash
[23:19:08] <ReadErro-> insert motor
[23:19:12] <pfred1> as soon as you take one apart it is shot
[23:19:31] <pfred1> they use special factory magic to assemble stepepr motors
[23:19:44] <ssi> when you open it all the torque leaks out
[23:19:59] <pfred1> even if you can get it back together i hear they put out like a tenth the power they're supposed to
[23:20:05] <ssi> exactly
[23:20:08] <ssi> cause the torque leaked out
[23:20:38] <RyanS> Someone uses a mt3 er32 collet chuck through the lathe spindle with a drawbar for milling. is that going to screw up bearings?
[23:20:39] <Xfriend> I made a hole on the shaft
[23:20:43] <pfred1> yeah a messed up stepper motor is a paper weight
[23:21:05] <toastyde2th> RyanS, also, soft jaws really aren't that big a deal for the home shop, people just play it up like it's really hard
[23:21:27] <toastyde2th> what a lot of people do is mill a long bar with whatever pattern the lathe takes and just chop it up
[23:21:52] <toastyde2th> most small chucks don't take very wide jaws so you get a lot of jaws (and a lot of life per jaw) out of a small bar
[23:21:56] <Xfriend> pfred1: can you send me or tell me the BOM of the stepper motor you have in intru.!
[23:21:58] <toastyde2th> (aluminum, obv)
[23:22:34] <toastyde2th> others don't even do that and just drill a hole through aluminum and bolt it without any other mounting features
[23:22:45] <toastyde2th> so you'll get these weird chunks of alu bolted to a chuck
[23:22:52] <RyanS> How are you supposed to attach them to the hard jaws?
[23:23:02] <toastyde2th> are your jaws one-part or two part
[23:24:33] <RyanS> umm t looks like http://www.machineryhouse.com.au/C200 but crapper
[23:24:34] <ssi> I want to get a nice new 3 jaw with set-tru and 2 piece jaws
[23:24:56] <toastyde2th> k those are single piece/one part jaws
[23:25:02] <ssi> i have a beat up six jaw bison with set tru but I only have three jaws for it
[23:25:07] <toastyde2th> what you do with those is make a thing that slips over the jaw
[23:25:28] <pfred1> one scummy thing we did at the machine shop we worked at is we'd buy lathe chucks from harbor freight and rebuild them and sell them like we made them
[23:25:31] <toastyde2th> cut a blind slot in a chunk of metal that's an interference fit and slip them over
[23:25:44] <ssi> lol
[23:25:52] <toastyde2th> tap them on with a hammer
[23:26:02] <pfred1> but only the smaller ones the bigger ones we'd actually make
[23:26:19] <toastyde2th> then, to cut them, clamp an aluminum pin in the uncut jaws, and machine your holding diameter
[23:26:27] <toastyde2th> open the jaw, throw your part in, and go
[23:27:13] <toastyde2th> no part distortion (you can use a ton of force with thin-wall parts), very accurate, and if you put a clocking mark on the jaws and part, you can usually get .0002-.0003" repeatability
[23:27:25] <pfred1> ssi that was a different department than I worked in that did that but I saw it going on
[23:27:48] <ssi> not good
[23:28:16] <pfred1> ssi well to be fair they would get all torn down the bearings would be replaced and they'd be machined if they were out and all tested and everything
[23:28:31] <ssi> well they were likely better than harbor freight chucks afterward i guess
[23:28:31] <pfred1> they did it with the little live spindles too
[23:28:53] <pfred1> but we made huge live spindles
[23:29:01] <pfred1> like for the paper industry
[23:29:07] <RyanS> hmm, ok. If I use a milling attachment on the lathe I wonder if it's any good just to clamp endmills in the 3jaw?
[23:29:18] <pfred1> things were rated to 300 tons or something stupid
[23:30:47] <pfred1> RyanS sometimes I juse end mills in my drill chuck on my mill depends on the work I'm doing if I bother to change out to the collets or not
[23:31:21] <pfred1> it is a pain to put my collet holder into the spindle
[23:31:25] <RyanS> http://www.youtube.com/watch?v=wESyli48YqA
[23:31:27] <Tecan> (wESyli48YqA) "Milling steel on a small metal lathe" by "xynudu ." is "Howto" - Length: 0:04:33
[23:31:39] <ssi> I don't like using endmills in a drill chuck
[23:31:41] <RyanS> He uses a drawbar
[23:31:52] <ssi> I'll do it occasionally for quick and dirty stuff, but it's not so good
[23:32:05] <pfred1> ssi I'm not wild about it either but if I'm just cutting a little piece of plastic or something
[23:32:20] <ssi> yeea
[23:32:22] <pfred1> good enough
[23:32:22] <toastyde2th> RyanS, it's dangerous to clamp a small endmill in a 3 jaw
[23:32:45] <RyanS> They can come out easily?
[23:32:51] <toastyde2th> hard jaws have shitty holding force on other hard things (endmills), and they have a lot of runout, which risks a break
[23:33:19] <pfred1> I will say this for those imported end mills they do shatter like glass
[23:33:20] <toastyde2th> if you use a .500" endmill, and have .005" of runout (which is pretty good for a used 3 jaw) that's 1% of the diameter
[23:34:01] <ssi> so do US made carbide ones :0
[23:34:10] <toastyde2th> haha heyoooo
[23:34:22] <pfred1> I suppose but better HSS end mills seem tougher to me
[23:34:30] <toastyde2th> so any time you do feed calculations, that .005 is going to ADD to one tooth, and subtract from the other side
[23:34:34] <ssi> yeah I prefer us made HSS endmills
[23:34:35] <ssi> they work great
[23:34:37] <toastyde2th> HSS is more shock resistant
[23:34:44] <ssi> biggest problem I ever have with them is gumming them up with 6061
[23:34:56] <ssi> and then I can knock it out of the teeth, but it's a hassle
[23:35:58] <RyanS> what about clamping a custom straight shank endmill holder and c clamp in thre jaw
[23:36:09] <toastyde2th> endmill holders also suck, i'd use a collet.
[23:36:21] <toastyde2th> endmill holders are GREAT for bigass endmills and roughing endmills
[23:36:32] <toastyde2th> anything under 1" i would suggest not using them
[23:36:51] <toastyde2th> 3/4" roughing works pretty good on a holder if you're worried about pullout
[23:37:09] <ssi> pullout is obnoxious
[23:37:13] <ssi> I've ruined a lot of parts with pullout :(
[23:37:23] <toastyde2th> i'd get something like a ER32 or ER16 holder for your lathe
[23:37:29] <pfred1> pulling out? doesn't sound manly to me!
[23:37:43] <ssi> pfred1: you prefer having lots of unexpected children?!
[23:37:53] <RyanS> ok, I can just make a drawbar for my mt3 er32 chuck
[23:38:00] <pfred1> I say we leave it in there and get the job done!
[23:38:09] <ssi> that gets expensive
[23:38:12] <toastyde2th> yeah, ER32 is the right tool for the job, it's made to hold endmills.
[23:38:22] <RyanS> You prefer plunge milling?
[23:38:25] <ssi> what's the max size on er32?
[23:38:26] <pfred1> it was a Carlon skit about Vietnam
[23:38:27] <ssi> 20mm?
[23:38:37] <toastyde2th> plunge milling is a good roughing strategy on smaller machines
[23:38:37] <RyanS> or boring
[23:38:37] <ssi> I think ER16 is 10mm
[23:38:52] <RyanS> Sorry I was making innuendo
[23:38:56] <toastyde2th> oh
[23:39:01] <toastyde2th> haha, sorry
[23:39:27] <toastyde2th> ssi, not sure what the max is
[23:39:37] <pfred1> I had to make a drawbar for my mill
[23:39:55] <pfred1> I cut the threads off of a bolt and welded them to a rod and welded a nut on top
[23:40:17] <toastyde2th> a lot of er holders have a coolant hole
[23:40:22] <RyanS> I can't remember where I saw this, he makes a drawbar and a sleeve that centres the bar in the middle of the pulley
[23:40:24] <toastyde2th> so in theory a drawbar should be easy to make
[23:40:28] <RyanS> With a washer
[23:40:33] <toastyde2th> put a bolt through the coolant hole and attach a bar to it
[23:40:40] <pfred1> but I was real careful when I welded everything together you see so it was all lined up right
[23:41:19] <RyanS> Doesn't welding fuck up threads ?
[23:41:25] <RyanS> Stick welding anyway
[23:41:56] <pfred1> I only welded it on one end
[23:42:01] <pfred1> and i TIGed it anyways
[23:42:19] <toastyde2th> yeah, i was gonna say nearly everything in machining is tig
[23:42:36] <pfred1> I had my doubts when I made it but it's held up for close to 20 years now so I guess its OK
[23:42:37] <Xfriend> any who have a cnc machine assembly with v-slot ?
[23:43:38] <RyanS> There are so many people sticking endmills in a chuck .. Someone is going to end up with shrapnel in their face
[23:43:44] <pfred1> the drawbar that mill came with didn't even fit I had to put a brass sleeve on the top of it to take up the slack
[23:44:37] <toastyde2th> RyanS, lots of people run their shizz really slow?
[23:44:47] <pfred1> I mean it isn't even close it is almost an inch too long
[23:44:52] <toastyde2th> i have noticed the hobby trend is to run the endmill too quickly, but feed it too slowly
[23:45:05] <toastyde2th> that means they dull the shit out of their tools in short order, but nothing ever snaps
[23:45:30] <pfred1> toastyde2th yeah if you don't have your surface speed right you're not machining
[23:46:47] <RyanS> How on earth do you feed a mill at consistent rate with out power feed.
[23:46:59] <toastyde2th> you don't, you just overcompensate by running the machine slower
[23:47:13] <toastyde2th> you can run a tool SLOWER than it's recommended speed without much of an ill effect
[23:47:25] <toastyde2th> that way not much happens when the feed varies
[23:47:45] <toastyde2th> most endmills are quite tolerant of overfeeding by a bit, but dull rapidly with underfeeding.
[23:48:05] <toastyde2th> it's also hard to snap an endmill from overfeed if you're cranking it by hand
[23:49:12] <toastyde2th> small endmills, yeah, snap city
[23:49:23] <toastyde2th> anything over 3/8ths, probably not.
[23:50:17] <toastyde2th> a 3/4" endmill will take .015 of feed in most materials without a huge problem - even steel. it just becomes a power issue.
[23:50:48] <RyanS> toasy are you a professional machinist. but cnc too?
[23:50:52] <toastyde2th> i was
[23:50:58] <toastyde2th> now i'm back in school
[23:51:21] <toastyde2th> i'd call myself a junior machinist, not an old hand
[23:51:39] <RyanS> ok
[23:51:56] <toastyde2th> both cnc and manual, mills and lathes. a little grinding, no edm.
[23:52:03] <CaptHindsight> <--- more of a micro-machinist
[23:53:33] <toastyde2th> one of our machines was a big manual horizontal mill, and I always had trouble getting my co-workers to run it properly
[23:54:12] <toastyde2th> it ran .5" and 1" wide, 10" and 12" diameter cutters
[23:54:22] <toastyde2th> 20 or 40 teeth per cutter
[23:55:00] <toastyde2th> and so the proper way to run it was some insane feed rate - you could send it nearly a half inch per rotation
[23:55:22] <RyanS> Not sure why I'm interested in this stuff.... I literally can't use machine tools due to disability but it just seems logical how to operate them, (i have a dude employed to help me out with building shit). I'm thinking of getting a small CNC Mill so I can 'drive' it, seems very expensive for what you get with a hobby sized machine
[23:56:14] <toastyde2th> ya, hobby sized machines are the same price as like, used cncs in working order from the 80s
[23:56:20] <toastyde2th> that are much, much larger
[23:56:31] <toastyde2th> but that's the rub - most people can't fit a bigass vmc in their garage
[23:56:44] <roycroft> not to mention not many have 3 phase power
[23:57:03] <roycroft> and decent 3 phase converters are expensive
[23:57:23] <toastyde2th> older machines aren't as touchy about power cleanliness
[23:58:14] <toastyde2th> if you were gonna run a newer machine, yeah, no dice, but something old shouldn't have a problem
[23:58:26] <roycroft> space is my bigger concern, personally
[23:58:30] <toastyde2th> we had an OKK from the 60s they retrofitted, and it survived brownouts
[23:58:36] <toastyde2th> "derp derp derp still cuttin'"
[23:59:49] <RyanS> I don't know how important 'feel' is in manual machining, but if I know the theory I can explain "turn that handwheel" but I don't know if I could hear or see if they are doing that wrong if I'm not turning the knob
[23:59:59] <toastyde2th> super important