00:00:13 machine on should send a signal that turns the power on 00:00:25 (imagine that) 00:00:27 does that 00:00:45 even after you come out of estop? 00:00:58 yes because its jsut a hardware estop 00:01:35 what's the difference between 'just a hardware estop' and 'combined hardware and software estop'? 00:01:38 lerman has joined #emc 00:01:58 Hey what's that paste image url? 00:01:59 that emc knows there was a hardware estop 00:02:16 or a softlimit 00:02:18 imagebin.org 00:03:11 ok what about that mesa output... it latches on the F2 from the gui, and if the hardware isnt estopped that machine powers. 00:04:09 could it also know that there is a voltage going through it? 00:04:20 not an output 00:04:57 so if i put an input in parralel with the output 00:05:17 if voltage goes to the input machine on, if voltage doesnt machine off 00:05:50 my head hurts 00:06:05 did you draw a diagram? 00:06:20 right now its in my head 00:06:23 :| 00:06:26 Guest458 has joined #emc 00:06:29 well draw stuff 00:06:29 and on the machine 00:06:47 put emc on the left and the estop chain on the right 00:07:15 m5i20 in between 00:08:03 http://www.ucsc.edu/news_events/press_releases/text.asp?pid=1534 00:09:07 http://imagebin.org/10514 00:09:34 Is this a bit like you are thinking only without the limits in the circuit. 00:10:20 note that the 24V supply is switched only by the physical power switch, not the estop chain 00:11:16 Good thought. 00:11:29 that's pretty necessary IMO :) 00:11:59 unless you want to require several people to hold buttons down to start the machine ;) 00:12:01 yeah thats how we have it 00:13:15 Looks like I missed a line between +24 and the mesa in1 00:13:30 yeah, thats the thing 00:13:30 Must have been to thin to show up. 00:13:50 ours stops recieving 24V when the estop chain is broken 00:14:21 Let me finish that missing line cause there's a relay contact in there. 00:14:40 we need to get that to disable the machine in the software, and stop looking so we can start the machine up again from F2 00:17:01 im thinking that if our input is linked to a machine disable, then it wont let us power back up cause its voltage is missing.... is there a way to turn off that input after its been tripped... and turn it back on when we press F2? 00:17:26 http://imagebin.org/10519 00:17:41 That should show the estop return signal to the mesa. 00:18:55 is that a capacitor? whats its story? 00:19:00 relay contact 00:19:12 CR1 is the coil, the lower line is the contacts 00:19:16 That is a dry contact on control relay one 00:19:26 skunkworks has quit 00:19:31 oh, is that what CR stands fro? 00:19:33 for 00:19:41 You got it. 00:19:55 cool. learn something new (or old) every day :) 00:19:56 what does "dry contact" mean? 00:20:01 no power flow 00:20:06 at least, none expected 00:20:18 So the trick is to energize the coil of control relay 1 using the mesa out5 00:20:35 and it signals the emc through mesa in1 that estop is up. 00:20:58 why dont you just hook mesa-in1 to the estop chain? 00:21:02 Now you have an impossibility 00:21:25 fenn, it's hooked through the relay 00:22:25 fenn you could by wiring down to mesa in1 from the chain. 00:22:49 But that doesn't tell you the relay has been pulled 00:22:53 well - where would you connect the two contacts? 00:23:07 i'd hook mesa-in1 to where CR1 is now 00:23:10 you need to put some current through the input opto 00:23:26 hmmm. that could work 00:23:40 and get rid of CR1 CR1-2 00:23:52 Um no! 00:24:05 what does CR1 do? 00:24:08 CR1 is likely to have other contacts that do other things 00:24:15 cause that's a short circuit through the mesa card. 00:24:21 like contact 3-phase power to the motor 00:24:36 true - that should be a low current chain, but not that low 00:25:02 (something like 50-100 mA, instead of the 10 mA you want through the mesa) 00:25:04 You could put all of the estop chain into mesa in1 00:25:12 and not have an out 00:25:29 yeah, connect mesa-in1 to the left end of the estop chain 00:25:42 but that's always getting +24V arg 00:25:53 you still want an out, so hitting F1 will stop the machine 00:25:57 But I prefer, and most USA standards require an estop chain without silicon. 00:26:31 right. so this relay is a current sensor basically 00:26:32 Each input has a + and a - terminal... how does that work? is the internal resistor between them? i was expecting to just have a voltage input to one terminal 00:26:59 In the case I drew, you can kill power to the system, CR1 without the mesa doing anything. 00:27:49 does CR cut off power in other parts of the system? 00:28:19 other contacts from CR1 would probably turn off mains contactors for the spindle and servo power supply 00:28:28 ok that makes sense 00:28:39 or they'd turn off relays that then turn off those large contactors 00:28:39 (ive never hooked an input up to the 7i37... so bear with me on that input terminal question) 00:29:02 Adam_, it's an optoisolator. an LED (possibly with a resistor) 00:29:12 ah 00:29:13 * fenn hates the disconnected coil/contact stuff 00:29:15 you need to run a little bit of current through it to turn it on 00:29:27 so does it sense current, or voltage between the terminals? 00:29:34 supper bbiab 00:29:52 it doesn't sense anything. the LED turns on when some current is run through it 00:30:47 but the input has to tell emc that the stop has been pressed... what exactly is it measuring to get that signal? 00:30:54 current i take it? 00:31:15 RTFM :) 00:32:05 ahh 00:33:21 (i did, but its a lot of EE jargon... ill have to do some research to understand, for instance what an optoisolator is) 00:33:36 The opto-isolated inputs have 4.4K Ohm series resistors and reverse input 00:33:36 protection diodes across the opto-isolator LEDS. Input current an the maximum 24V input 00:33:40 is approximately 5mA. The isolated inputs will work with input voltages from 4 to 24V. 00:34:06 "an" should be "at" 00:34:17 yeah 00:34:20 V=IR 00:34:54 right, so i cant just supply the + terminal a voltage and assume the resistor is grounded... i have to complete the circuit with the -ve terminal 00:34:56 so if you give it 24V, I = 24/4400 == .005 00:36:46 nm, i got it i think 00:37:06 this way you have separate grounds for each opto 00:37:13 ive programmed a plc before, but its inputs didnt work the same 00:37:14 i see 00:38:18 an opto is just an led and a light sensor 00:38:46 and usually an amplifier to boost the output from the light sensor 00:38:58 gotcha 00:44:53 ribbito has joined #emc 00:46:28 so how do i do this in the config? 00:46:31 keroppi keropp meet ribbito 00:49:56 fenn: tomp testing a laptop wit DSL ( damn small linux ) 00:50:05 i guessed from your ip 00:50:12 clever 00:52:07 enuf 4 me... byebye (tomp) 00:52:19 ribbito has left #emc 00:55:43 Adam_: by "this" you mean ray's diagram? 00:57:56 i mean, im wiring it to my input1 on the 7i37.... what do i have to do in the config to use this signal? 00:58:28 dougztr has joined #emc 01:00:02 hook it to iocontrol.0.emc-enable-in i suppose 01:00:25 hi all. hey, i'm trying to create some sort of bootable USB version of linux that includes emc that will run on a 2Gb usb stick. i'm thinking about putting a stripped down version of debian on it and taking it from there, any ideas? 01:00:56 Adam_: " Should be driven FALSE when an external estop condition exists" 01:01:17 ok thanks, ill fumble my way through that :) 01:02:42 dougztr: use the emc livecd as an image. you might have to do some grub-install stuff on the usb key to get it to boot 01:03:16 then you can mount /home as unionfs or jffs from somewhere else on the usb key 01:04:37 jtr has quit 01:05:28 fenn: the livecd says it needs 2+ Gb though, I only have 2 total 01:05:56 how are you going to fit 2GB on a CD? that's silly 01:06:28 it's compressed 01:06:53 so, keep it compressed 01:07:14 skunkworks did thi - he has a bootable USB stick that's like the CD 01:07:34 I don't remember if he ever put up a wiki page, but there will be links to the web pages he used in the IRC logs 01:07:58 s/thi/this/ 01:08:04 that's whats neat about 'overlay' filesystems like unionfs, you dont have to uncompress everything else 01:08:08 fenn: i really don't need all that fluff from the livecd, yeah its just a pet project. 01:08:46 fenn: i'll have to research unionfs, i first experienced it just the other day toying with DSL 01:08:53 dougztr: the reason i suggested the livecd is that it already boots from a realtime kernel 01:08:53 http://pendrivelinux.com/2007/01/25/usb-x-ubuntu-610 01:09:06 fenn: i know how to copy a kernel, :-) 01:09:23 i have this so far (emc-enable-in is TRUE, and isnt driving FALSE when i cease input to 01) 01:09:24 -newsig EstopSense bit 01:09:25 linksp EstopSense <= m5i20.0.in-01 01:09:25 linksp EstopSense => iocontrol.0.emc-enable-in 01:09:25 - 01:10:28 dougztr: you could just install debian on a usb-key and download the emc .deb's then 01:11:27 SWPadnos, fenn: great! I'm working on preparing a R&D project at work to deploy emc2 on some big stuff 01:11:33 i never had much luck booting computers from usb keys 01:12:12 fenn: my new system doesn't have a floppy, and its panic time when i wipe my bootsector out 01:12:34 use netboot (pxe) 01:12:48 even old stuff that wont boot from usb seems to have pxe 01:13:02 fenn, if i had half a brain i'd get that set up on my toy server 01:13:30 its not that hard really 01:13:31 lerman has quit 01:14:15 anyways, guys, suppose i wanted to convert a old manual table mill to cnc, and use emc2. what HP steppers do you think i'd need? 01:14:31 steppers are usually rated in oz*in 01:14:57 also there's the frame size (nema 23 34 etc) 01:15:39 fenn: duly noted. i've seen some kits for bridgeports. i presume something like that would probably do the job 01:16:06 is it a BP sized mill? 01:16:20 thats the smallest one we have 01:16:56 ok. I have a BP 9x42, and the servos I have are 27 in-lb 01:17:11 I think newer ones are ~31 in-lb, due to nicer magnets and stuff 01:17:30 ty for the tip. 01:17:57 they're 2500 RPM, 9.7A continuous, 39A peak (meaning ~108 in-lb peak), max 150V 01:18:19 at full speed/full torque, they'd be 6KW, wihch is a lot 01:18:47 for steppers, you'd need higher initial torque, since the torque falls off the faster the motor goes 01:19:03 around 500 oz*in is good for bridgeport steppers isnt it? 01:19:17 that's undersized for a pro machine 01:19:22 ==450 watts or so 01:19:25 for a hobby or home shop, it's probably OK 01:19:49 my servos are 27 in-lb, which is 432 oz-in 01:20:00 but they're that torque continuously, and at 2500 RPM 01:20:16 what's peak torque? 01:20:22 ~108 in-lb 01:20:38 1728 oz-in 01:21:11 but again, they can do that when they're spinning at 1000 RPM or higher, which a stepper can hardly even reach let alone deliver any torque 01:21:33 i don't think i'd need 1000 rpm 01:21:33 right, so more like 2-3 times that value 01:21:48 unless your name is mariss :P 01:21:52 1000 RPM at the motor, through a 2:1 drive, with a 5 TPI screw, means 100 IPM rapids 01:22:15 this motor is intended to be used at 2:1 01:22:18 or higher 01:22:58 how would the stepper compare to the servo at 10 ipm cutting steel? 01:23:27 using the same drive rations (2:1, 5TPI), that means 100 RPM, so the stepper will probably be at full torque 01:23:47 a very latrge stepper might not be though 01:23:49 large 01:23:51 um.. wtf is this doing on the ubuntu website "Restart your computer, remove the CD and boot back into Windows" 01:24:03 heh 01:24:59 hey, i have 24V applied to input 1, but when i take it away EstopSense doesnt revert back to false... any idea why? 01:25:18 gotta get back to work,wiring up the usc tonite 01:25:44 Adam_, you'll need to check the manual. I think estop was turned into a read/write pin at some point - there's some handshaking that has to go on (I think) 01:25:58 yeah there is an EstopWrite 01:26:00 but i dont use it 01:26:03 im just sensing 01:26:27 what I'm saying is that it may need something more than just being hooked to an input pin 01:26:33 but I don't know what that something is :) 01:27:19 ok 01:27:48 no it's not an i/o pin 01:27:50 im new to these configs... its still a little mysterious to me... but i can see that EstopSense is a 'bit' signal 01:28:01 thus it has an ON/OFF state 01:28:12 and its linked to input 1 01:28:20 and to emc-enable-in 01:28:52 and emc-enable-in is an IN pin 01:28:53 if i throw not at the end of the input1 link, then emc thinks the estop is on and wont let me do anything 01:29:07 but i cant get the voltage at input 1 to switch this for me 01:30:37 linksp EstopSense m5i20.0.in-1-not right? 01:31:24 * fenn doesnt have an m5i20 to play aorund with 01:31:28 yeah. this is a simple way that i checked that emc-enable-in was going to false and disabling things for me 01:31:50 or is m5i20.0.in-1-not a parameter? 01:32:25 i dont think so 01:32:26 :S 01:32:40 i know im missing something simple 01:32:47 but im not fluent in this yet 01:32:48 ok google tells me in-01-not is a pin 01:33:06 newsig EstopSense bit 01:33:05 linksp EstopSense <= m5i20.0.in-01 01:33:05 linksp EstopSense => iocontrol.0.emc-enable-in 01:33:18 for HAL digital inputs, generally a regular and an inverted form is available: in-NN and in-NN-not 01:33:25 right 01:33:43 i want the non inverted, i just inverted it to test that the estop was in fact tripping 01:34:04 my problem is that the 24V at input 1 can be removed and EstopSwitch doesnt see it 01:34:09 it doesnt change to false 01:34:16 you might need a pull-down resistor 01:34:20 ahh 01:34:38 check the voltage on the actual pin 01:34:44 with a multimeter 01:34:47 i am 01:34:55 its definitely going from 24 to 0 01:35:04 i thought maybe i was missing some code 01:35:24 would those 3 lines above be enough? is that properly linking the input? 01:35:51 dougztr has left #emc 01:36:02 you have to add the m5i20 functions to a thread of course 01:36:21 besides that i think it should work 01:36:24 ok 01:36:45 addf m5i20.0.digital-in-read servo-thread 1 01:36:51 this wouldl be the line you want 01:36:55 ? 01:37:00 ya 01:37:04 ok, its there 01:38:03 did you attach a ground to the opto input? 01:38:07 if you short the two input contacts together, does the input change state? 01:38:21 (don't do this with the 24V signal still connected) 01:38:36 u mean, attach pin1 to pin 2? 01:39:07 yes 01:39:36 if you do that, and the input doesn't change, then the problem is somewhere between the terminals and EMC 01:39:39 what would that do.... doesnt it need power to activate the input? 01:39:54 if it does turn off, then the problem is in the circuit it's connected to 01:40:11 jtr has joined #emc 01:40:33 you said if you power it then un-power it, the input bit changes frmo 0 to 1, but never goes back to 0 01:40:35 (I think) 01:41:21 not exactly... the bit doesnt change either way 01:41:28 it doesnt seem to be seeing the input 01:41:43 oh. I thought you saw one transition and not the other 01:41:44 and shorting the terminals doesnt do anything.... (i dont know why it would) 01:41:47 neg 01:41:54 are you looking at the bit itself, or a signal connected to it? 01:42:17 both 01:42:29 err - so you have two halmeters open? 01:42:30 the signal is definitely switching (via voltmeter) 01:42:35 yeah 01:42:38 ok 01:42:43 Adam_: and there's current going through the opto right? 01:42:46 disconnect the bit from the signal 01:42:49 yeah there is 01:43:09 hold up i dont know if i checked the actual pin in the halmeter, ill do that 01:43:11 and you're probing the two input terminals, not one terminal plus a known ground? 01:43:23 (with the physical voltmeter) 01:43:58 ok, the pin switches to false when i connect the 24V 01:44:04 but estopsense doesnt 01:44:40 do 'halcmd show all *estop*' 01:44:49 no i have to probe both... the powersupply doesnt work as a node relative to the ground... it seems to be two out of phase ac signals that average to dc 01:44:58 ok 01:46:10 oh damn its case sensitive 01:46:56 show sig EstopSense 01:47:16 when i did that it stated 03 and 07 as the input and output pins... is that my problem? 01:47:28 probably 01:47:42 you want something like 01:47:55 bit TRUE EstopSense 01:48:09 <= m5i20.0.in-01 01:48:21 => iocontrol.0.emc-enable-in 01:50:41 hahha it works!! 01:50:48 i didnt realize it wanted pin 3 01:51:04 i was trying to use input1.. i switched it all to 3 and it toggles 01:51:09 thanks very much 01:58:52 Guest458 has quit 02:08:05 03fenn 07TRUNK * 10emc2/docs/man/man1/halui.1: added teleop/joint pins 02:12:57 skinnypuppy1334 has joined #emc 02:16:15 tomp2 has joined #emc 02:16:31 Adam_ has quit 02:20:34 roltek has quit 02:32:27 fenn: thanks for your halui improvements and for improving the docs with it 02:41:16 Ziegler has quit 02:44:31 i'm thinking about adding a man page for emcsh and friends 02:45:01 man pages are great - you just type 'man whatever' and it pops up with what you need 02:46:56 JymmmEMC has joined #emc 02:50:21 JymmmEMC has quit 02:50:40 JymmmEMC has joined #emc 03:07:23 hehe i found a bug that causes the robot visualization to go berserk 03:08:42 rayh has quit 03:12:57 03fenn 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: limit teleop jogging to XYZABC because UVW doesn't make sense 03:23:32 03fenn 07TRUNK * 10emc2/configs/scara/scara.ini: missing halfile causes emc not to load 03:58:01 does the wiki support tables? 03:59:17 the wiki - wiki says it can if certain plug-ins are installed, just wondering if they are. 03:59:58 I'm pretty sure that tables work 04:00:09 using the | | notation in the instructions 04:00:51 ||a||b||c|| 04:05:53 ok - I'll give it a shot once I have all the info condensed 04:06:23 skinnypuppy has quit 04:15:04 grarr i know there's some stupid c library function that gives me the length of an int 04:15:34 but google is being coy 04:16:19 aha! sizeof 04:21:40 heh 04:21:45 not just for ints :) 04:27:24 sizeof is not a function nor is it part of a library :P 04:28:16 it's a compiler function 04:28:51 or at least it really looks like one, because of the parentheses :) 04:33:02 technically it's an operator 04:33:02 i'm going to flaunt my apathy and call it a function though 04:33:42 jepler has quit 04:33:42 ds2 has quit 04:34:08 jepler has joined #emc 04:34:08 ds2 has joined #emc 04:34:43 Twingy has quit 04:34:43 Vq^ has quit 04:34:52 Twingy has joined #emc 04:34:52 Vq^ has joined #emc 04:44:19 Off_Namuh has quit 04:44:33 03fenn 07TRUNK * 10emc2/src/emc/task/taskintf.cc: new code doesnt break when you have axes out of order, for example XYZC which has a mask of 39 and 4 axes. this change tightens the error checking a bit too since axes cant be 0 04:50:31 Off_Namuh56 has joined #emc 05:12:18 03fenn 07TRUNK * 10emc2/configs/scara/scara.ini: it wasn't the R axis anyway; now works in teleop mode (again?) 05:36:19 <_sh3> _sh3 has joined #emc 05:40:33 x3rox has joined #emc 05:51:46 <_sh3> _sh3 has quit 05:55:44 <_sh3> _sh3 has joined #emc 06:05:21 hugomatic has left #emc 06:17:18 tfmacz has quit 06:25:17 jmkasunich__ has quit 06:31:04 jmkasunich__ has joined #emc 07:29:01 <_sh3> _sh3 has quit 08:11:56 <_sh3> _sh3 has joined #emc 08:46:49 toast has quit 09:12:12 skinnypuppy1334 has quit 09:48:19 robboplus has quit 10:20:25 dimas has quit 10:25:58 dimas has joined #emc 10:40:34 x3rox has quit 11:36:34 lerman has joined #emc 12:00:40 tomp2 has left #emc 12:09:10 skunkworks has joined #emc 12:18:49 Off_Namuh56 has quit 12:28:06 jlmjvm: http://www.cnczone.com/forums/showthread.php?p=345396#post345396 12:43:19 highoctanejg has joined #emc 12:43:42 Has anyone here ever installed emc2 on Feisty Fawn? 12:45:12 Can anyone hear me? Am I logged in OK? 12:47:03 highoctanejg: yes we can hear you 12:47:24 highoctanejg: you will have to compile it yourself, the official packages are for Ubuntu 6.06 LTS only 12:47:39 http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Why_aren_t_there_packages_for_Ubuntu_7_04_Feisty_Fawn http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Simulator_on_Ubuntu_Feisty_Fawn_7_04 12:48:02 if you want to control hardware and you are not a pro at patching and compiling kernels, I recommend you use Ubuntu LTS and the official packages 12:48:28 if you want to run the simulator only, the compile is much easier, so if you have some experience building from source code you should give it a try 12:48:39 if you don't have any experience building things from source, then you should use LTS 12:49:00 What is LTS? 12:49:07 Ubuntu 6.06 LTS, also called Dapper Drake. 12:49:24 It's the version that is supported for security bugfixes for 5 years -- LTS means "Long Term Support" 12:49:51 Oh. Yeah. I started with that. But it did not have proper support for my video drivers, intel 845g. So i had to update to Feisty Fawn. 12:50:26 I realize I need to compile it myself. I have been attempting to compile it, but am getting hung up at a certain point in the process. I am wondering if anyone here has ever done that. 12:52:02 I have built the simulator version on feisty, but not the real version suitable for controlling hardware 12:52:06 lerman has quit 12:53:13 was it pretty straight forward? 12:53:17 unfortunately I really need to drive in to the office .. I'm already late... 12:53:25 OK. Thanks for your help. 12:53:49 highoctanejg: I think there were some important problems keeping version 2.1.7 from working on feisty; the wiki instructions say to use CVS TRUNK. 12:54:08 besides whatever fixes I made in CVS TRUNK I don't remember any gotchas 12:55:08 I'll be around later, and a fair number of the other people here also have experience compiling emc, so if you show the error you're getting using http://pastebin.ca/ somebody may be able to help 12:56:44 Off_Namuh56 has joined #emc 12:59:09 highoctanejg: do you want to drive a machine with emc2 on feisty? 14:31:18 jmkasunich_ has joined #emc 14:33:17 jmkasunich__ has quit 14:35:13 alex_joni: yes, I want to drive a machine with emc2 on feisty 14:37:36 highoctanejg: is your video card 'on the motherboard'? 14:38:35 yes, it is. 14:39:42 Normally - on-board vidoe that uses shared memory do not work with realtime. You would be better off sticking a old video card into the computer which would allow you to run dapper 14:40:17 older video card 14:41:36 Oh, really. So, even if I got emc2 running on Feisty, I would still have video card issues because it would not work with real-time linux? 14:41:54 you would have issues because of the video chipset, not because of Feisty 14:44:15 I guess what I meant was... I have spent two days trying to get this video driver to display at 1024x768... finally got that right by updating to Feisty, and now emc2 doesn't work. So, I was going to try compiling it. Have I been wasting my time completely here? 14:44:29 it's possible :( 14:44:40 have you run any latency tests on that PC? 14:46:55 if you can boot from the LiveCD (even at the wrong resolution), you should run the latency tests as described here: 14:46:58 http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting 14:49:55 Guest534 has joined #emc 14:50:47 Guest534 has quit 14:52:00 SWPadnos, by the way. Does Feisty come with the real-time kernel installed, or do you have to compile it? If so, how difficult is that? 14:52:18 it does not come installed - you'll have to compile it 14:53:13 it isn't so hard to patch and compile, it's configuring the kernel "just so" that's time consuming 14:53:45 I don't think I've figured out a good formula for Feisty (and gutsy is coming in a few weeks anyway) 14:55:03 Even though I want to get 6.06 installed from CD so I can use emc2, I am interested in developing possibly. 14:55:23 welcome aboard! :) 14:56:29 if you want to get started with 6.06, then it may be best to install an old cheap video card - something from Matrox or a low end nvidia (using the open source driver - their proprietary one is fast but not compatible with RT) 14:58:03 O.K. - I'll see what I can dig up. By the way, what is emc2 like without realtime? 14:58:21 err - what do you mean? 14:58:43