00:00:12 * robin_sz wakes up again 00:00:53 what? oh people where talking to me 00:12:46 picnet has joined #emc 00:12:52 shh. 01:47:00 thalx has joined #emc 01:55:33 picnet has quit 02:13:19 picnet has joined #emc 02:21:52 robin_sz has quit 08:18:58 alex_joni has joined #emc 08:22:58 jf353 has joined #emc 08:27:46 Guten Morgen jf353 ;) 08:33:59 robin_sz has joined #emc 08:34:20 Morning robin 08:35:24 Guten Morgen 08:35:42 any luck with rtai? 08:36:06 I am just looking for plx9052 chip schematics 08:36:16 for a D/A converter 08:39:25 nice ;) 08:39:28 PCI - board? 08:39:30 or ISA? 08:40:10 ISA is easy, Motherboards are still available, but the future? 08:41:25 maybe better: design a PCI board 08:42:30 or a plx9054: 24$ (plx9052 16$) 08:42:53 I'll be later on 08:42:53 bye 08:42:56 alex_joni has quit 10:22:18 picnet has quit 10:23:33 picnet has joined #emc 10:30:50 robin_sz has quit 11:19:54 * anonimasu yawns 11:28:27 picnet has quit 12:24:08 picnet has joined #emc 12:30:47 alex_joni has joined #emc 12:42:19 * alex_joni is frozen ... 12:50:51 picnet has quit 14:02:25 slomo has joined #emc 14:08:20 jf353 has quit 14:18:32 and not a peep was heard 14:18:40 indeed 14:21:08 jf353 has joined #emc 14:24:41 picnet has joined #emc 14:47:59 * alex_joni has left 14:48:00 alex_joni has left #emc 14:56:31 picnet has quit 14:57:23 jf353 has quit 14:59:18 anonimasu has quit 15:18:36 rayh has joined #emc 15:26:26 slomo has left #emc 16:11:51 Can someone answer a question regarding the generic.run script and 4.08? 16:12:36 I'll try 16:13:23 has it changed in 4.08 compared to CVS of a few weeks ago? 16:13:40 no idea 16:22:32 Bert is trying to start up 4.08 at his place in Eindhoven and it kills Linux. 16:23:05 He gets past the report that minimillio has started. 16:23:23 And then locks up. 16:24:25 Here the next line that I get is, "simDioInit (generic.ini)" 16:25:17 picnet has joined #emc 16:27:05 maybe his PERIOD is set too fast 16:27:36 I thought so also but freqmod completes and his 2.4 athalon should handle the default value. 16:27:53 is he using EMCMOT=freqmod? 16:27:59 oh 16:28:06 yeah it should be fine on such a fast machine 16:28:09 Right. 16:28:16 so the mouse pointer stops moving and everything? 16:28:45 or it reboots? 16:29:59 He says that it's dead. I'm guessing that he has to reboot. 16:30:24 I just ask because sometimes "dead" means different things to different people 16:30:36 maybe he should try running sim.run to make sure it's a realtime issue 16:30:41 Right. 16:30:51 Oh. That's a good plan. 16:30:55 Thanks. 16:30:56 if it is already running freqmod successfully, it may not be realtime at all 16:31:07 it might be a plain old hardware problem 16:31:20 Okay. 16:32:04 I see a fairly complete set of error messages in minimillio so I'd think it would report an ini file problem. 16:32:31 I think few ini file problems should crash the machine 16:33:08 you might also ask for sure what ini he's using and what changes he's made to it, and also what happens exactly when it hangs (mouse pointer stops? hard drive light? etc.) 16:33:50 It is generic.ini and it would appear he had two piled into the same file. He fixed that. 16:34:13 maybe he should get generic.ini from cvs to make sure it's not messed up 16:34:29 http://cvs.sourceforge.net/viewcvs.py/*checkout*/emc/emc/generic.ini?rev=1.18 16:35:31 he could also put "set -x" at the top of the generic.run 16:35:42 then he will be able to see exactly which command runs last 16:35:50 (not at the exact top - under the #!/bin/bash) 16:36:02 Ah. That will help. 16:40:06 grrrr 16:40:29 I could help better if there were bdi_4_08 tags in cvs 17:04:02 Thanks Chris. I sent a note to Bert. We'll see what happens. 17:04:24 you're welcome 17:21:29 picnet has quit 17:52:44 robin_sz has joined #emc 18:25:52 picnet has joined #emc 19:01:10 anyone home? 19:01:53 maintenance day for me...cleaning ballscrews, linear rails, etc 19:02:37 have to go to town...can't find my 90 degree zerk fitting and must buy another 19:14:52 I'm here... 19:15:11 how far is town? 19:15:26 what is a zerk? aren't they the aliens in the game Starcraft? 19:15:33 haha 19:15:59 it's that nipple thing you pump grease into, like on your car's steering 19:18:55 picnet has quit 19:21:14 rayh has quit 19:25:28 robin_sz has quit 20:11:59 alex_joni has joined #emc 20:12:10 hello 20:12:14 anyone around? 20:13:07 nobody 20:13:10 sorry 20:13:34 not me either 20:15:00 hmm.. just the man I need 20:15:11 * alex_joni assumes jepler is perfect 20:15:20 yes, it's true, I'm the perfect man. 20:15:24 taken into consideration that nobody is perfect 20:15:30 :P 20:15:35 anyways... 20:15:39 I think I won't bother to ask you guys 20:15:46 you probably don't know ;) 20:16:04 * alex_joni will have to figure it out all by himself :( 20:16:21 geek trap! 20:16:51 lol 20:17:06 it's such a simple thing to do... yet it defies all logic and common sense ;) 20:17:15 eh? 20:17:21 * alex_joni tries to send one lousy message 20:17:40 you would assume that's easy 20:17:48 sure, probably not hard at all 20:18:01 only one slight problem ;) 20:18:25 you're talking about sending nml messages I assume 20:18:37 I don't know enough about RCS/CMS/NML to make it work properly 20:18:44 jepler: unfortunately ... yes ;) 20:18:50 thing is the following 20:18:57 I started on IO for emc2 20:19:09 1st basic task (add estop to the iocontroller) 20:19:31 one way it's trivial (I added the HAL pins, and set the values) 20:19:59 the taskcontroller now sends estop, and the HAL pin (iocontroller.0.estop) gets set properly 20:20:18 the other way is what I don't seem to be able to manage 20:20:40 what I figured out so far: 20:20:46 what is the "other" way? 20:20:54 emc1's way is completely different from emc2 20:21:05 I know emc1's way if I know anything 20:21:13 getting a HAL pin to change should propagate to the task-controller 20:21:21 and this one should get into estop 20:21:32 * alex_joni talks about estop-in 20:21:48 this should be a pin connected to external HW which signals an external ESTOP 20:22:23 are you following so far? 20:22:40 the way axis learns about estop is by polling an RCS_STAT_CHANNEL frequently, but I don't have/use a machine with a hardware estop to know whether this actually works 20:23:14 jepler: this should be the right way to do it 20:23:23 the task controller does the same job 20:23:49 and when it gets an ESTOP from the iocontroller, it outputs an estop through a RCS_STAT_CHANNEL to the gui aswell 20:23:55 so axis gets into estop 20:24:29 ok... now for the communication between task and io 20:24:46 2 channels are opened (command and stat) 20:24:46 that is something I don't know anything about -- I've only done user interface 20:25:07 task-controller sends a command (through the command channel) 20:25:31 and io-controller executed it and ack's it (by sending RCS_DONE along with the current status) 20:26:08 now if I only change the hal-pin into the io-status, the task-controller will be aware of this only on his next command 20:26:28 picnet has joined #emc 20:26:34 because iocontroller doesn't send any status-messages on his own 20:26:43 at least until now 20:27:08 which thing is constantly updating the coordinates shown by the front-end? Couldn't it be changed to poll this value from the IO controller? 20:27:17 that's motion 20:27:29 other controller 20:27:59 I tried, and succeeded only partially to push a status_message, even if no command has arrived before 20:28:58 but it seems that the taskcontroller is coded up to only take return values (I didn't fully figure out the serial_number stuff in the messages) 20:31:14 now.. I can't take over code from emc1, because there it was completely different 20:32:02 there is a iomaincontroller, and a bunch of subordinate controllers (aux, cool, lube, spin, tool) 20:32:42 which all are classes derived from NML_MODULE (some strange class containing machine_states & NML messaging) 20:33:49 I hope you figure it out 20:33:51 so, my guess is that I'm doing something wrong with the values I add to the message, so that I break some NML-parsing inside taskcontroller 20:34:09 other duties are calling me, and I don't think I can help much anyway 20:34:19 thanks for listening ;) 20:34:25 anytime 20:34:31 I kinda cleared the whole picture somehow 20:34:55 I've been reading code from a lot of places ;) 20:41:47 jepler: still around? 20:43:48 cradek: you're gone too? 20:44:03 no 20:44:16 I'm just hiding because I probably can't help 20:44:20 unless you ask more gdb questions 20:44:42 you know any of emcmodule.cc ? 20:44:52 axis/extensions/ ? 20:44:59 that I know about 20:45:18 I see thatyou send NML commands 20:45:27 and every command uses a new serial nr. 20:45:33 right 20:45:36 that much I know 20:45:40 feel free to interrupt me if I'm wrong 20:45:48 no, that's right 20:45:52 you have a function called next_serial() 20:46:13 which returns a new serial nr. to be used for the NML message to be sent 20:46:19 yes 20:46:29 ok.. now... the QUESTION ;) 20:46:55 do you know any constraints to this serial? 20:47:06 what happens if 2 messages have the same serial? 20:47:21 don't know 20:47:25 I assume they should be always increasing 20:47:41 ok.. now 20:47:50 there are 2 channels (command & status) 20:48:19 those have different serial_nrs I think 20:49:11 yes I agree 20:49:32 so you're just trying to send a message to set the machine state to ESTOP? 20:49:40 yes 20:49:45 EMC_TASK_SET_STATE state_msg; 20:49:48 state_msg.state = EMC_TASK_STATE_ESTOP; 20:50:00 state_msg.serial_number = ++the_last_serial_number 20:50:05 emcCommandBuffer->write(state_msg); 20:50:19 on command? 20:50:26 didn't think of that 20:50:28 yes the command buffer 20:50:33 those are bidirectional? 20:50:38 * alex_joni checks the .nml 20:51:15 toolCmd = RW 20:51:22 what do you know ;) 20:51:26 you are probably right... 20:51:29 * alex_joni tries that 20:59:50 not fully the right thing... but it's worth a deeper investigation ;) 21:00:05 that's how it works in emc1 21:02:01 from the GUI... 21:02:29 thing is... the task controller enters ESTOP_STATE, which is ok 21:02:40 so far it works like it should 21:03:15 but then it starts complaining about not receiving RCS_DONE for some IO-commands it sent out (SET_LUBE & such) 21:03:39 I think the serial_number is not fully ok 21:09:23 hmmm.. it seems to work now 21:09:30 I did this: 21:09:58 msg.serial_number = emcioCommand->serial_number+1 21:10:02 not 21:10:07 msg.serial_number = ++emcioCommand->serial_number 21:10:20 wanna check it out if it works for you? 21:10:26 * alex_joni commits 21:14:06 03alex_joni * 10emc2/src/emc/iotask/ioControl.cc: modified the NML passing of the ESTOP coming from HAL. this works now, but it may not be the way to do it 21:14:14 I don't have a machine here, and I don't have a machine with estop anywhere! 21:14:38 I know.. me neither 21:14:45 but now estop is connected to HAL 21:14:51 so you can do whatever with it 21:14:59 I connected estop to a halmeter 21:15:28 and the other estop to a signal (halcmd newsig estop bit; halcmd linksp estop iocontrol.0.estop-in) 21:15:45 and the using (halcmd sets estop TRUE/FALSE) I can modify it to test it 21:17:08 rayh has joined #emc 21:17:27 hello rayh 21:18:17 alex_joni: it's great to have a start for digital IO 21:18:33 I hope it's ok... 21:18:46 if people say it works, I'll add other stuff aswell 21:18:50 my machine won't run without some kind of digital output now (I rebuilt the stepper drivers recently) 21:19:05 what kind of output? 21:19:26 well a bit on the port enables the choppers 21:19:47 I am using estop out now with emc1 (no estop-in or real estop button) 21:20:59 I see.. well with HAL you can use whatever to enable that pin 21:21:27 even estop or smthg else (a default value) 21:23:34 or just turn it on, I imagine 21:24:00 or hook a square wave to it or some other funny thing 21:24:17 hello all 21:24:21 hi les 21:24:48 finished inspection and most of lube 21:24:53 howdy folks 21:24:57 machine in good shape 21:25:07 hi fezr 21:25:15 gezr 21:25:17 :) 21:25:18 haha 21:25:18 les: any luck trying paul_c's potential fix for homing? 21:25:32 cradek: you had a point there 21:25:35 not yet ...but I think dave tried it 21:25:42 with that sqare wave 21:25:56 but you can PWM your motor power like that 21:25:58 les: I know your working tree doesn't have that in it 21:26:17 (my customer came sat and asked me to triple production) 21:26:19 hmm 21:26:24 les: but who knows what will fix it (I sure don't) 21:26:24 hi les 21:26:25 hope to try it soon 21:26:31 hi alex 21:26:33 howdy gezr 21:27:17 I've got to try to get paul_c to tag the cvs when he makes CDs 21:27:20 picked up a spring compression tool and a valve holder spinner thing today 21:27:30 cradek: it kinda is in CVS 21:27:31 I will be doing 24-36 hrs of machine running a week 21:27:35 but.... 21:27:40 not in emc1-CVS 21:27:45 alex_joni: I mean there needs to be a BDI_4_08 tag 21:27:52 it is in emc2 CVS 21:28:11 because emc1 from BDI_4_08 uses libnml 21:28:14 but ... surely it's emc1 on the BDI 21:28:23 and a lot of different stuff 21:28:29 so it doesn't fit in emc1 CVS 21:28:33 * cradek shivers 21:29:12 well whatever, I can't install every one of these darn CDs, and I should be able to use cvs to help people or make a fix for someone. 21:29:31 must go for a bit...customer is pulling up to pick up product 21:29:42 brb 21:29:43 if I have no idea what code is on the CD, I can't do that, and we can't roll back and fix something for a user later 21:29:54 darnit, the comression tool I got is either going back or in need of heavy modification 21:29:55 cradek: I _think_ it is there in CVS 21:30:00 I'm not really sure 21:30:25 alex_joni: rayh asked me about generic.run and I looked. No tag. 21:31:02 ok, forget what i was saying ;) 21:31:10 you could open the .deb :D 21:31:12 let me look somewhere else (some source) 21:31:20 Hi guys. 21:31:24 hey ray 21:32:11 you certainly have some experience with RCS... 21:33:13 cradek, rayh: about that set -x 21:33:26 Yes? 21:33:28 # Touch the file .debug_scripts in your home directory if you like 21:33:30 # to see all the commands echoed. 21:33:37 if [ -f ${HOME}/.debug_scripts ] ; then 21:33:38 set -x; 21:33:44 copied from generic.run 21:33:45 ah 21:33:52 Right 21:34:17 just instruct the user to `touch ~/.debug_scripts` 21:34:55 It looks like bert's emc is failing just after freqmod and 21:35:14 The last messages are 21:35:52 sudo /sbin/mpdprobe freqmod PERIOD=24000 IO_BASE_ADDRESS=0x378 SHMEM_KEY=100 21:35:52 done 21:35:52 sleep 1 21:36:30 did he try running sim? 21:36:48 So freqmod is going in - generic.run says done - -x replies sleep1 which is between freqmod install and minimillio. 21:38:19 after the sleep, he never sees the next if [ $emcio = simio ] ... ? 21:38:35 There are no sim modules available or any sim files. 21:38:42 sleep 1# Run emcio in backgroundecho -n "starting EMC IO PROGRAM -- $emcio..." 21:38:52 I tried it here using the existing sim files but nothing. 21:38:59 he should see a "starting EMC IO PROGRAM...." 21:39:07 rayh: that's nice 21:39:50 rayh: can he modprobe the freqmod by itself? 21:39:55 to see if that works? 21:39:59 The generic.run in 4.08 is a bit different. 21:40:14 Let me try to see if freqmod goes in. 21:40:15 I had a machine where modprobe freqmod completely hung the machine 21:40:23 I needed to reboot 21:40:39 I finally tracked the problem to be rtai-specific :( 21:40:46 and compiled/installed another kernel/rtai 21:40:57 think with BDI this shouldn't be 21:41:05 agreed 21:41:33 rayh: has he disabled all the APM in the BIOS? anything strange about the machine (laptop?) 21:41:41 guess that's why it's called BDI 21:41:56 NOAPM, NOAPIC, NOLAPIC 21:42:13 Yes it will and it brings in adeos, rtai_hal, rtai_lxrt, rtai_shm with it. 21:42:31 rayh: with those same arguments? 21:42:52 alex_joni: that hang might have been period in freqmod. 21:43:14 I agree 21:43:22 here it runs with PERIOD=50000 21:43:49 that's in CVS right now, I think 21:43:51 I see that paul loads several values into freqmod when it is installed by *.run. 21:44:15 I'll post a note to burt about the apm stuff. 21:44:32 not really sure it has anything to do with this... 21:44:39 you could also have him double or triple PERIOD just to be sure 21:44:45 that stuff usually keeps a puter from booting 21:44:51 add a 0 ;) 21:45:03 no, remove a 0 21:45:04 * alex_joni wants one of these: http://www.fordvehicles.com/cars/mustang/launch/ 21:45:50 oops... yes ;) 21:46:03 alex_joni: I saw one at a mustang club show 21:46:16 alex_joni: it's pretty nice. 21:46:17 looks nice... 21:46:22 the V8 anyways 21:47:01 I have an 01 21:47:05 cradek: how do you controll your mill? 21:47:16 spindle on/off ? 21:47:29 alex_joni: yes I have a spindle control 21:47:38 I meant from emc1 21:47:49 I don't understand the question 21:48:00 you just use spindle on/off 21:48:11 or you use spindle increase/decrease aswell 21:48:12 I use the spindle forward bit 21:48:18 no, just forward 21:48:31 so no analog control or such... 21:48:34 I might do a speed control sometime. 21:48:48 my spindle is speed-controlled with feedback but you set the speed with a knob 21:49:03 ok ... one more question 21:49:12 what keeps you from running emc2? 21:49:36 what is still missing in there to suit your needs? 21:49:40 well, that digital IO is the big thing 21:49:52 I want to use the segmentqueue planner 21:49:54 explain... 21:50:07 I want spindle control and estop output 21:50:15 (sounds like that's almost done) 21:50:20 estop works 21:50:22 ;) 21:50:25 I don't know the status of segmentqueue 21:50:32 segmentqueue.. is out of my reach ;) 21:50:39 but I'm not going back after seeing arc blending 21:50:49 lol... 21:50:52 it works *great* 21:50:53 fix it for emc2 too :D 21:51:55 the question I ask myself is: what advantage would I have if I switched to emc2? 21:51:58 I just had a brief look over at emc2/src/motion 21:52:04 you'd have HAL 21:52:05 :D 21:52:14 from a user standpoint, I don't care about that 21:52:33 I have fixed all the bugs in emc1 that bothered me 21:52:45 I have no real reason to switch 21:53:15 ok.. then from a developers point of view :P 21:53:34 a lot of things to be fixed in emc2 ;) 21:53:43 and you can always use them both :D 21:53:50 that's right, and I have 21:54:11 oh.. I almost forgot 21:54:32 you have a wonderful ./configure in emc2 ;) 21:54:35 lol 21:55:03 I know it's in emc1 now too... 21:55:12 don't know, haven't used it 21:55:30 * alex_joni wonders if anybody used ./configure on emc1/rcslib 21:55:44 Yep. Me. 21:55:44 (since I already know the secrets, emc1 is easy to compile) 21:55:51