#emc | Logs for 2009-08-15

[08:22:49] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[11:20:33] <MediumJohnT> MediumJohnT is now known as jthornton
[11:56:16] <jthornton> WEEEE the base for the tower is now in concrete
[12:53:44] <Valen1> Valen1 is now known as Valen
[15:10:03] <skunkworks_> sound about right? http://cnczone.com/forums/showthread.php?p=652739#post652739
[15:15:09] <mozmck> skunkworks: yeah. the 'charge pump' signal is usually more like 12 Khz or so though
[15:16:10] <skunkworks_> mozmck: thanks
[15:16:54] <mozmck> you can get all kinds of noise on the parport when booting. I don't know if it polls the port or what, but if you have relays and stuff connected they'll go wild!
[15:47:00] <SWPadnos> when Windows boots, it may mess with the parallel port during its scan for new devices
[15:47:58] <SWPadnos> also, I think the only way to detect a port is to write to some registers and check the readback value, so Linux does the same thing if you load parport_pc (or whatever)
[15:49:08] <tom2> tom2 is now known as tom3p
[15:51:01] <good_kid> SWPadnos: What if I successfully run EMC2-RTAI in cegeda in, for example, MS Windows XP -- It seems it could drive the motors, but RTAI advantages will not be used, right?
[15:51:25] <SWPadnos> it will be terrible, but may appear to function
[15:52:31] <good_kid> SWPadnos: OK, I thought it could be a solution for nebooks
[15:52:49] <good_kid> *netbooks
[15:52:58] <SWPadnos> there's no point to doing that, unless you're developing a parallel port driver
[15:53:30] <SWPadnos> drivers are the only reason to use RTAI on a PC like that, since it can't run a real machine
[15:53:51] <SWPadnos> if you want to do testing, training, or other development, then sim is a better way to do that
[15:54:02] <good_kid> SWPadnos: I mean PCMCI-to-Par on netbooks (if it have pcmci)
[15:54:24] <SWPadnos> you will not be able to run any machine worth running when using RTAI inside a virtual machine
[15:56:43] <good_kid> SWPadnos: Can /etc/laptop-mode/laptop-mode.conf help if I thin tese are cache write which enlarges the latency from 15 ms to 100 at the first 10th or 15th minute?
[15:57:18] <SWPadnos> I saw what you posted regarding laptop-mode, and it looked like interesting data
[15:57:51] <SWPadnos> but I don't know that you can completely disable ACPI/APM, which is usually necessary in notebooks (for battery and heat management)
[15:58:55] <good_kid> SWPadnos: right, but I had to repair RAM module last time I used RTAI... One laptop with the Athlon XP 1600 GHz hadn't worked a minute in RTAI kernel
[15:59:30] <SWPadnos> RTAI is unlikely to cause damage, but its usage pattern amy make existing problems show up
[15:59:48] <SWPadnos> (a bit like running prime96 or similar as a burn-in test)
[15:59:52] <SWPadnos> err, prime95
[16:00:08] <good_kid> SWPadnos: my laptop works in RTAI perfectly well -- but I fear that turned off ACPI support heats up the CPU, motherboard and RAM
[16:00:20] <SWPadnos> yes, that is a problem
[16:00:40] <SWPadnos> you meed to leave ACPI on (and may not be able to disable it completely) for a laptop
[16:00:52] <SWPadnos> this is why they're usually not sutable as realtime machine controllers
[16:01:50] <tom3p> 110 fan plugged into ac outlet for my hp laptop w acpi biod defeated... overheats/shutdown anyway, cant reverse airflow thru it well
[16:02:19] <SWPadnos> tape it to the case in an exhaust configuration :)
[16:02:34] <SWPadnos> (just unplug the original fan first, you can blow the drivers if you spin them too fast)
[16:02:37] <good_kid> SWPadnos: And since I do not compile new RTAI kernel, I can not tune it up to use ACPI, what isn't a default behaviour.
[16:03:16] <SWPadnos> what I'm trying to explain is that if ACPI is enabled, there is no point in using RTAI, since you will not get acceptable latency all the time
[16:04:19] <SWPadnos> and you must get good latency all the time to make a reliable machine - if there's a 0.001% chance of something going wrong, you'll get about 1 error every minute and a half (using the standard 1 kHz servo thread)
[16:08:16] <good_kid> OK. Old PC with stsndard radiators will work stable either with acpi disabled or enabled.
[16:09:25] <good_kid> OK. Another question.
[16:11:58] <good_kid> Xemc vs AXIS interfaces. I know I can try it, but Xemc probably uses less CPU timings? Or is it exactly a consistency of the signals at the first order what makes the CPU to work?
[16:13:30] <tom3p> the time consuming stuff isnt in the gui, its in the common underlying control, so cpu burden isnt (very) gui dependant
[16:13:59] <good_kid> ok
[16:14:50] <good_kid> it needs something like 500 or more KHz for EMC2 routine
[16:48:51] <Valen> you could turn acpi off perhaps and underclock the cpu
[16:49:41] <Valen> oh on boot some motherboards use the parallell port to output "port 80" info
[16:50:15] <Valen> so expect lots of random weirdness as it starts until the OS takes over the port, then perhaps more weirdness until it gets going
[17:00:38] <tom3p> LawrenceG: yesterday you talked of using dspic33 fro encoder in & pwm out. i saw theres an rtos for it. would emc2 benefit by offloading tasks to rt objects?
[17:00:50] <tom3p> for
[17:19:49] <LawrenceG> tom3p, well... I have several projects running with the servo loop in the pic.... the biggest hassle with external devices is getting an easy reliable, real time high speed interface to the PC.
[17:20:12] <LawrenceG> the mesa stuff is the best I have seen so far for off loading the PC
[17:22:11] <LawrenceG> realtime network would be nice, but it has not really materialized.... not sure what is happening with the new pci-e serial bus.. it could be an easier way in/out of the PC, but the electronics is probably beyond the hobbyist
[17:22:55] <LawrenceG> USB still cant get it right in the PC world... I dont have any hopes for it in the real time end of things
[17:25:39] <LawrenceG> I really like the servo loop inside the PC... it makes tuning easier using halscope... the main issue is getting encoder information into the PC.... really only 2 choices.. hardware on the PP or the mesa card on PCI bus
[17:27:56] <tom3p> so synchronation is needed between the rt objects? async rt is not useful?
[17:29:10] <LawrenceG> a dedicated microcontroller has very good RT performance when doing 1 task like a servo loop... it is the realtime communications with the PC that sucks
[17:29:26] <jimbo655> Pid tuning M5i20?
[17:30:25] <LawrenceG> 5i20 is a very clean communication route in and out of the PC and is already supported (yay)
[17:33:12] <LawrenceG> its getting to the point that it could be possible to port emc to a cell phone and run a 5 ton cnc machine.... but its not high on my list of things to work on
[17:43:29] <Guest229> Hi everyone!
[17:43:49] <Guest229> I have a question about tkemc if someone has time - I am trying to get it to run 640x480
[17:44:08] <Guest229> but the docs that I found online are outdated
[17:44:59] <Guest229> The "Editing the GUI" page is what I am referring to
[17:45:17] <jimbo655> I am trying to work on the PID tuning. I am having trouble with Halscope triggering on PID*error. any advice?
[17:45:19] <Guest229> I would like to change the font size to 8 for everything
[17:48:41] <tom3p> Guest229: i dont think tkemc or its directory specify the default font, i think tk does that. IF so, then this may be of help http://www.tkdocs.com/tutorial/fonts.html
[17:49:19] <Guest229> cool man, thanks
[17:49:41] <Guest229> I am hoping there is a way to globally change the font, rather than resizing each component tho
[17:49:44] <Guest229> we'll see
[17:51:17] <tom3p> that page talks about the global ( default ) font used by all tcl/tk apps
[18:23:19] <LawrenceG> bummerl... ebay rip off of my work ! http://members.shaw.ca/swstuff/esrmeter.html http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=320378624786
[18:24:31] <LawrenceG> direct copy of my website description.... likely his meter is totally different inside... poor unsuspecting ebayers
[18:26:25] <archivist> likely ripped off the circuit as well, except for the display
[18:27:32] <tom3p> interesting "any ripple current through the capacitor causes the capacitor to heat up due to the resistive loses." i'm sitting here with 6pcs 300amp 105Vdc, i oughtta check my cap bank. tho i never felt 'em warm up
[18:28:43] <archivist> someone just pointed this cap on fleabay in ##electronics http://cgi.ebay.com/_W0QQitemZ350230250951QQcmdZViewItem
[18:31:13] <LawrenceG> now that is a cap.. ( 3 I suspect)
[18:31:18] <tom3p> oil filled caps ;) big time
[18:31:54] <micges> I have question: did any of you saw lathe with more axes than XZ ?
[18:32:05] <LawrenceG> power factor correction caps... they sit right across the bus bars in a motor control center
[18:32:08] <micges> like XZBC or sometning?
[18:32:43] <archivist> micges, there are plenty of the sliding head lathes with multiple axes
[18:33:18] <tom3p> http://www.bardonsoliver.com/machines/4axislathes.html
[18:34:02] <micges> archivist: thanks
[18:34:35] <archivist> micges, if I can get the hardware I will retro fit one
[18:35:01] <micges> me too :)
[18:35:43] <micges> but I'm asking becouse of modifcations of emc that will be made
[18:36:24] <archivist> this needs a home when we close http://www.archivist.info/cnc/works2008/P1010226.JPG
[18:36:42] <archivist> and with some live tooling
[18:37:15] <tom3p> 8 axis lathe? i never saw the workpiece spin! http://www.youtube.com/watch?v=vGEOY9aGJF8
[18:37:27] <micges> archivist: when we close?
[18:37:51] <archivist> we are shutting here
[18:38:00] <archivist> * archivist out of a job
[18:38:16] <micges> oh
[18:39:04] <tom3p> micges, what limitations do you see in emc2 for multiaxis lathes?
[18:39:22] <tom3p> archivist: sorry, hope you find work
[18:39:39] <tom3p> it could be worse, i have work and no pay!
[18:39:47] <archivist> http://www.youtube.com/watch?v=weJQS0WkqdY
[18:40:07] <micges> archivist: I hope you find work too
[18:40:42] <archivist> may work from home depends what toys I take from here
[18:41:19] <micges> tom3p: I'm reworking mapping of joints<-> axes
[18:41:46] <tom3p> archivist: cool vid, now mount the headstock on an a&B to get those bosses drilled at any angle :)
[18:42:22] <tom3p> micges: trivkins or more complex?
[18:42:28] <archivist> I keep looking at sliding head vids for inspiration
[18:43:42] <micges> tom3p: I'm actally writing that part of emc (branch joints_axes3 in git source) ;)
[18:44:23] <micges> don't have any lathe so I'm asking
[18:44:50] <tom3p> ah, i'm not a lathe user
[18:46:13] <archivist> I do use lathes, Im also milling round to save second op now
[18:46:49] <archivist> its a tooling v speed issue
[18:59:26] <archivist> and, spindle motor capable being rigid for milling against and fast for turning
[19:24:37] <micges> archivist: are you using emc ?
[19:24:50] <archivist> yes, 5 axis mill
[19:25:25] <micges> XYZAB ?
[19:25:36] <archivist> yes
[19:26:11] <archivist> set up trunnion style
[19:26:24] <micges> trunnion?
[19:26:57] <archivist> A mounted on B mounted on XY
[19:27:16] <micges> cool
[19:27:24] <micges> pics?
[19:27:44] <archivist> its a scruffy object
[19:29:58] <archivist> let me upload recent from the camera
[19:30:18] <micges> archivist: any lathes with emc?
[19:30:46] <micges> brb
[19:34:53] <Danimal> hi guys
[19:43:35] <alex_joni> hi
[19:48:43] <Danimal> anyone try pncconf?
[19:54:41] <archivist> micges, pic from 5 minutes ago http://www.collection.archivist.info/archive/djcpd/PD/2009/2009_08_15/102CANON/IMG_0277.JPG swarf made on it http://www.collection.archivist.info/archive/djcpd/PD/2009/2009_08_15/102CANON/IMG_0274.JPG
[19:55:07] <archivist> it is built from scrap and oddments
[19:58:12] <archivist> swarf is 2.4 mm od
[20:00:43] <Danimal> nice
[20:03:27] <archivist> not so nice it has some serious problems, backlash, rigidity of the rotary and restricted XY travel
[20:04:13] <Danimal> join the club in the backlash department
[20:04:27] <archivist> so one cannot use cam software on it, you have to program the gcode unidirectionally
[20:04:37] <Danimal> i gotta replace/rebuild my ball screw in the x axis
[20:04:48] <Danimal> ahh
[20:05:19] <Danimal> you must not have a ball screw, huh
[20:05:46] <archivist> no not yet
[20:05:57] <Danimal> tht'll help
[20:06:42] <archivist> best part of £350 each needed for X and Y
[20:06:59] <archivist> thats slides and ballscrew
[20:07:04] <Danimal> wow, that's a bit much
[20:07:19] <Danimal> oh, slides too.. not bad
[20:07:53] <Danimal> you dont have room for a full size machine like a bridgeport, huh?
[20:08:01] <archivist> not worth fixing just one problem, need to fix travels and backlash
[20:09:00] <Danimal> for 700 i'm sure you can get an old bridgeport cnc with a boss control that will have a fairly solid mechanical side
[20:09:37] <archivist> probably but the cost is way out of reach
[20:10:23] <Danimal> i dunno, around here they basically give them away
[20:11:02] <Danimal> usually they have some old outdated controller and noone wants to deal with them
[20:11:14] <Danimal> perfect cadidate for EMC2
[20:11:43] <archivist> the cost of retrofit, transport etc put them out of my reach
[20:12:52] <Danimal> i hear ya
[20:14:24] <Danimal> i need to change my stepper outputs to open drain
[20:14:51] <Danimal> i dont think my stepper drives like the 3.3v my mesa i/o board is putting out
[20:15:10] <Danimal> the manual says 4-5v
[20:16:27] <archivist> I use a buffer
[20:17:00] <Danimal> a buffer?
[20:17:18] <Danimal> like an amplifier?
[20:17:32] <archivist> a chip more capable of driving
[20:17:36] <archivist> yes
[20:18:06] <archivist> 7414 iirc off my parport
[20:19:31] <Danimal> so a transistor should work, huh?
[20:19:50] <archivist> yes
[20:20:47] <Danimal> hmm
[20:22:14] <Danimal> i really need to learn how to design and build circuits
[20:22:26] <Danimal> im new to transistors and all that stuff
[20:22:46] <Danimal> i understand how it all works, i just dont understand much as far as specs go
[20:24:05] <tom3p> a transistor solves one pin's problem, a chip solves 6 or 8, use a chip
[20:26:55] <Danimal> never dealt with a chip before
[20:27:06] <Danimal> but i suppose i can look into it
[20:27:58] <archivist> some are even designed as level translators 3.3 to 5v logic
[20:28:29] <tom3p> http://www.fairchildsemi.com/ds/74%2F74LCX244.pdf 3.3 to 5v xl8r
[20:28:38] <tom3p> translator
[20:29:10] <SWPadnos> what stepper drives are you using?
[20:30:58] <Danimal> leadshine m880a
[20:31:20] <SWPadnos> oh right
[20:31:25] <SWPadnos> those are optoisolated, aren't they?
[20:31:42] <Danimal> thanks tom3p and archivist
[20:31:46] <Danimal> i believe so
[20:31:48] <Danimal> yes
[20:32:27] <Danimal> just double checked and they are
[20:32:50] <SWPadnos> are there +/- terminals for both step and direction, or are they commoned?
[20:33:13] <Danimal> seperate for both
[20:33:18] <Danimal> not commoned
[20:33:27] <SWPadnos> ok, so you can do common + and have the FPGA pull low
[20:33:37] <SWPadnos> which most logic is prett ygood at
[20:34:04] <Danimal> yea, the manual for the drivers confirm that
[20:34:35] <Danimal> i was reading the emc2 manual today trying to figure out how to configure the hal for it
[20:34:44] <SWPadnos> no idea, but it should be possible :)
[20:35:06] <SWPadnos> you'll probably need to invert the output in HAL, and then set up whatever the Mesa card needs (is-opendrain or something)
[20:36:09] <Danimal> yea, that sounds about right
[20:36:52] <tom3p> so its not a 3.3v or 5v problem, the opto needs ~<20mA current and the prob is setting up mesa?
[20:37:47] <Danimal> the driver needs between 7 and 16 ma
[20:37:52] <Danimal> mA*
[20:38:16] <Danimal> well]
[20:38:48] <tom3p> then ignore all the stuff about trx and chips
[20:38:49] <Danimal> i guess i can do it the hardware way with a chip, or i can do it with an opendrain setup
[20:39:11] <Danimal> the chip wont work?
[20:39:54] <tom3p> not needed afik, if the mesa can drive 16mA per output, its just setup and wiring
[20:42:05] <Danimal> ugh i'm confused lol
[20:43:01] <SWPadnos> if you were dealing with inputs to the FPGA, or outputs to 5V logic, you might need a level shifter
[20:43:35] <Danimal> ah ok
[20:43:47] <tom3p> ? mesa cant drive an opto?
[20:43:48] <SWPadnos> since the input to the stepper drive is actually a current (to turn on the LED in the optoisolator module), and you can configure it so that the inputs need current "sunk) (ie, they get pulled to ground), you don't need any external hardware
[20:45:01] <Danimal> ok
[20:45:14] <Danimal> makes sense
[20:45:41] <Danimal> now just gotta figure out how to configure it for that
[20:46:22] <archivist> does the mesa card have a series R to limit current
[20:46:34] <SWPadnos> no, I don't think so
[20:46:42] <SWPadnos> but the drive should have a current limiting resistor
[20:46:52] <micges> archivist: very cool
[20:46:53] <archivist> some dont
[20:47:31] <Danimal> i think there is one in the drive according to the manual
[20:48:13] <micges> archivist: I'll try to build XYZAB myself
[20:49:23] <archivist> micges, ,ost of my work is round with a regular feature eg gear or escape wheel etc
[20:51:13] <Danimal> hey the max input frequency of my drives is 300khz, how would i know if i'm exceeding that?
[20:51:17] <micges> archivist: now we 're building XYV machine with two heads
[20:51:43] <archivist> Danimal, you wont be
[20:52:43] <micges> two heads welder
[20:52:44] <SWPadnos> the Mesa card can, but not if you set up the step length and space correctly
[20:53:33] <archivist> at that speed your steppers would stall
[20:54:35] <SWPadnos> unless you set the drive to bazillion-microstep mode :)
[20:54:37] <Danimal> hmm
[20:55:03] <Danimal> 2 bazillion. is that bad?
[20:55:08] <SWPadnos> heh
[20:57:01] <Danimal> i dont see any parameters for the stepgen to change it to open drain
[20:57:20] <Danimal> in the hal part of the manual
[20:57:21] <SWPadnos> it's not the stepgen
[20:57:29] <SWPadnos> it's a GPIO setting
[20:57:48] <tom3p> http://en.wikipedia.org/wiki/Open_drain
[20:58:21] <Danimal> oh i didnt know the gpio settings would effect the stepper outputs
[20:59:04] <SWPadnos> yes, the GPIO settings affect how the pin interfaces to the real world, but the output value is taken from the stegen (when the stepgen is enabled)
[20:59:12] <SWPadnos> stepgen
[20:59:35] <SWPadnos> you can still read the GPIO -in pin too, which will show you (sloely) the stepgen output
[20:59:39] <SWPadnos> gah
[20:59:42] <SWPadnos> slowly
[21:01:52] <Danimal> I/O pins on the board which are not used by a module instance are exported to HAL as "full" GPIO pins. Full GPIO pins can be configured at run-time to be inputs, outputs, or open drains, and have a HAL interface that exposes this flexibility. IO pins that are owned by an active module instance are constrained by the requirements of the owning module, and have a restricted HAL interface.
[21:02:00] <Danimal> from the manual
[21:03:37] <Danimal> (bit r/w) is_opendrain: This parameter only has an effect if the "is_output" parameter is true. If this parameter is false, the GPIO behaves as a normal output pin: the IO pin on the connector is driven to the value specified by the "out" HAL pin (possibly inverted), and the value of the "in" and "in_not" HAL pins is undefined. If this parameter is true, the GPIO behaves as an open-drain pin. Writing 0 to the "out" HAL pin drives the IO pin l
[21:04:13] <Danimal> hmm so it looks like you're right
[21:04:55] <SWPadnos> note the last sentence of the is_opendrain paragraph
[21:04:56] <Danimal> if im understanding it correctly
[21:05:04] <Danimal> yea
[21:05:06] <SWPadnos> "Only full GPIO pins and IO pins used as outputs by active module instances have this parameter."
[21:05:23] <Danimal> and stepgen is an active module
[21:05:29] <SWPadnos> with outputs
[21:05:45] <SWPadnos> if you don't activate it, then the pin will be a full GPIO, and will still have the parameter
[21:06:08] <SWPadnos> it's only pins that are shared with input finctions (like encoders) that don't have the setting when the function is active
[21:06:16] <SWPadnos> (since the encoder function needs inputs)
[21:06:29] <Danimal> ok makes sense
[21:11:36] <Danimal> so if i change the opendrain parameter for 1 gpio, will it set the rest like that as well?
[21:11:51] <SWPadnos> nope
[21:11:54] <SWPadnos> each is separate
[21:12:04] <SWPadnos> just copy/paste the line in the HAL file, and change the numbers
[21:13:55] <Danimal> ok i think i get it...
[21:15:02] <Danimal> cool, thanks for the help, i deffinately have a better understanding of it all now
[21:15:09] <SWPadnos> great
[21:16:45] <Danimal> cant wait till i can rapid with this machine
[21:16:57] <Danimal> it will save me tons of time
[21:19:07] <Danimal> i finally fixed my power drawbar this week
[21:19:38] <Danimal> damn rectifier was bad on one of the logic boards
[21:21:21] <Danimal> well i'm off to the store to get some connectors... thanks again for the help
[21:21:27] <Danimal> adios
[21:21:27] <SWPadnos> see ya
[21:21:31] <SWPadnos> work time here
[23:00:49] <SWPadnos_> SWPadnos_ is now known as SWPadnos