#linuxcnc | Logs for 2012-09-26

Back
[01:39:29] <uw> does anyone here know of adam farkas?
[01:39:44] <uw> Hungarian guy who does CNC
[02:05:32] <theos> nop
[02:11:16] <DJ9DJ> moin
[02:23:27] <Loetmichel> mornin'
[03:01:25] <theos> o/
[06:47:47] <r00t4rd3d> wow i feel like balls
[06:48:37] <Valen> round and hairy?
[06:48:55] <DJ9DJ> lol
[06:49:56] <r00t4rd3d> more like blue and swollen
[06:50:34] <Valen> the rope is too tight then duh
[06:50:36] <r00t4rd3d> i think i gotta call in sick
[06:50:50] <DJ9DJ> lol
[06:50:56] <r00t4rd3d> which i havent done in over 8 years
[09:33:01] <JT-Shop> rotflmao http://www.brownebagracing.com/pictures/hf_tool_sale.pdf
[09:42:17] <cradek> JT-Shop: noise & fume generator!
[09:55:32] <JT-Shop> yea, I like that one
[09:56:17] <JT-Shop> and "the exact same crap at the exact same prices as in last week's flyer, just arranged differently!
[10:16:11] <skunkworks> wow... http://groups.yahoo.com/group/mach1mach2cnc/message/136901
[10:16:24] * skunkworks hugs linuxcnc again
[10:19:09] <JT-Shop> all of that for $100 plus $$$ for motion hardware?
[10:19:52] <JT-Shop> how can you remove something that doesn't work anyway... threading
[10:21:16] <jdh> Also we will have Linux and Mac version of Pro.
[10:22:10] <skunkworks> yes - which is just a gui to run some sort of motion hardware.
[10:22:30] <archivist> did they take out all the buggy stuff so they can at least get something out the door :)
[10:22:40] <JT-Shop> http://www.cnczone.com/forums/product_announcements_manufacturer_news/162843-mach4-lite_software_ethernet_smoothboard_screw.html
[10:23:36] <JT-Shop> http://www.automationtechnologiesinc.com/products-page/cnc-cam-software/mach4-lite-software-with-ethernet-smoothboard-with-screw-terminals
[10:25:06] <skunkworks> I think pcw_home should have a mach buster sale...
[10:25:29] <syyl> is there any information if mach4 will support closed loop movement?
[10:26:42] <jdh> seems unlikely
[10:26:50] * JT-Shop needs to quit rotflmao and get back to work
[10:26:51] * syyl hugs linuxcnc
[10:27:02] <syyl> then, noone needs it
[10:29:38] <jdh> $299 for HW step controller, simple sofware for windows types, easily wired terminal strips. Could be a potential winner for a lot of people.
[10:57:12] <skunkworks> jdh, for 200 dollars you get 48 i/o 5 hardware step generators, encoder counter and analog output. (plus it is expandable)
[10:57:24] <skunkworks> (and configurable)
[10:58:48] <jdh> sure, and use emc with is superior. But, that's not the target.
[10:59:23] <jdh> for $0, you can use a p-port and emc. But it is far less friendly for dabblers.
[10:59:24] <skunkworks> it should be ;_
[10:59:32] <skunkworks> ;)
[10:59:44] <JT-Shop> seems the target is folks who can hardly turn on a computer
[10:59:54] <jdh> JT: there are a lot of them.
[11:00:00] <JT-Shop> yea
[11:00:09] <JT-Shop> and most will never run a part
[11:00:16] <jdh> I bet there are a lot of people running CNC machines who fit that category also.
[11:00:31] <pcw_home> We should have a Ethernet version of the 5I25/7I76 (7I76E) sometime this year
[11:00:33] <pcw_home> (God willing the beans hold out)
[11:00:41] <JT-Shop> yea, machinest and computers don't get along
[11:00:47] <JT-Shop> cool
[11:01:21] <skunkworks> pcw_home, awesome
[11:01:56] <jdh> pcw: plays nicely with regular ethernet? Or dedicated segment?
[11:02:18] <pcw_home> has to be dedicated for real time
[11:02:59] <pcw_home> it plays nice in general (though it has no security)
[11:03:17] <jdh> will it work with Atom's and whatever servo thread latency issues they have?
[11:03:56] <pcw_home> yeah it should (its a little faster than the 7I43)
[11:04:54] <jdh> it will be a single board combo of 5i25 + 7i76? Will it have terminal strips?
[11:05:42] <pcw_home> looks like a 7I76 but instead of DB25 for host connection it has RJ45 Ethernet
[11:06:12] <jdh> does it come bundled with Mach4-lite? :)
[11:06:20] <JT-Shop> lol
[11:06:46] <pcw_home> Well its probably supportable with SOftDMC firmware
[11:07:25] <pcw_home> (same for the 7I77E)
[11:09:21] <JT-Shop> pcw_home: the 5v power supply came in today and no more errors on the 7i77 when using the MPG
[11:10:59] <skunkworks> pcw_home, there is a mesa forum on the mach support site where someone was making a driver for it. There has been no activity for quite a while
[11:11:10] <pcw_home> Probably drooping 5V (by the time we get analog power and you power the external encoders it easy to draw more 5V current more than the PTC in the 5I25 will allow)
[11:12:08] <pcw_home> Yeah The guy that wrote that driver now works for Mach
[11:12:13] <skunkworks> heh
[11:12:16] <skunkworks> smurf?
[11:12:17] <JT-Shop> it's steady at 5.0v now and the power supply is adjustable...
[11:12:24] <pcw_home> Yes
[11:12:34] <awallin> pcw_home: I was reading a paper where they have PLLs for higher clock-rates on an FPGA. Up to 3GHz internal clock for some simple logic on the FPGA. Have you done any of that? Doable with the FPGAs on the mesa cards? This was for detecting the phase-difference between two input square waves
[11:12:41] <skunkworks> wow - I hope they can sustain themselfs...
[11:13:44] <pcw_home> I use the DLLs all the time ( even the 5I20 firmware multiplies the clock so the base PWM clock is 100 MHz)
[11:14:16] <pcw_home> The Spartan6 chips do have PLLs but I have not used them yet
[11:17:57] <pcw_home> The Spartan3 and 6 chips have programmable delays on their inputs that allow ~35 ps step delay setting up to a few ns so if you calibrate these you can get very fine phase measurements
[11:19:31] <jdh> how do they do that?
[11:20:01] <pcw_home> tapped buffers I think
[11:20:46] <pcw_home> (sort of like the DLL which is a big ring oscillator)
[11:22:55] <pcw_home> (imagine 255 ultrafast 7404s in series)
[11:23:14] <jdh> doesn't that add a lot of extra HW?
[11:23:22] <jdh> or at least internal space?
[11:23:52] <jdh> and a really fast clock?
[11:24:03] <pcw_home> so tiny now I dont think they are even as big an the bonding pads
[11:25:45] <pcw_home> the DLL _is_ the clock (the ring oscillator has a variable length) but its phase locked to the external clock
[11:26:53] <jdh> uhm. OK. Perhaps I'll stick with "it's magic"
[11:27:40] <pcw_home> It is , rather
[11:30:13] <Jymmm> Just a bunch of flip flops =)
[11:30:24] <pcw_home> The spartan6 is fast enough that the PWM clock for it is now 200 MHz
[11:30:26] <pcw_home> and Hostmot2 with all its register decoding and muxing runs at 100 MHz
[11:30:28] <pcw_home> so ClockLow is now 100 MHz.
[11:31:27] <pcw_home> Also little processors run at 100 MHz easy, Spartan6s are very nice
[11:32:59] <pcw_home> well flip flops and gates (and hardwired things like multipliers and memory)
[11:33:35] <Jymmm> pcw_home: You playing with DDS at all?
[11:33:53] <pcw_home> The stepgen is a DDS
[11:34:00] <Jymmm> ah
[11:34:31] <pcw_home> Our UARTs use a DDS as well
[11:35:53] <pcw_home> (compared to the normal programmable divider baud rate generator the DDS gets better setting resolution at high baud rates instead of low baud rates)
[11:36:02] <Jymmm> pcw_home: I'm going to start playing around with a AD9850 module, just too cheap not to.
[11:36:26] <Jymmm> $6 shipped
[11:39:23] <pcw_home> AD has been making DDS's for radios for a long time
[11:40:35] <pcw_home> We have a similar Hostmot2 module (DDS --> Waveform RAM -->PDM)
[11:41:41] <Jymmm> pcw_home: Well, to well motive someone, I got my General last saturday, and working on gettign my Extra this saturday. So just a lil treat to myself to build a lil QRP rig with one.
[11:41:54] <Jymmm> s/well/help/
[11:42:28] <Jymmm> before he went to antartica
[11:47:59] <pcw_home> I was playing with TTL DDS's in the 70s (74LS283 FTW)
[11:49:10] <Jymmm> pcw_home: There are a LOT of neat projects now using the AD9850, one example http://www.youtube.com/watch?v=mrDYEbRGEds&feature=player_embedded#!
[11:49:25] <pcw_home> the software Stepgen is also a DDS (or phase accumulator if you wish)
[11:53:09] <archivist> ... --- ... .. -. . . -.. .-- --- .-. -.- ..-. --- .-. -- -.-- -.-. -. -.-.
[11:55:28] <frallzor> ahoyhoy
[11:55:30] <jdh> archivist: you could make me a Z ballscrew mount
[11:56:08] <archivist> you could make yourself one
[11:56:49] <jdh> yeah, getting started on things is one of my many character flaws.
[11:57:23] <archivist> never do today what you can put off till tomorrow
[11:57:46] <archivist> I have been procrastinating too
[11:58:16] <jdh> weather has been too good outside, or too bad inside for me.
[11:59:13] <Jymmm> Make a cnc router kit like everyone else does and sell it on ebay!
[11:59:51] <jdh> I'd have to clean first. I have 0 space at the moment and most things are > 1 layer dee.
[12:00:22] <frallzor> hmmm cnc router kits...
[12:00:48] <archivist> to make a kit one needs capital
[12:01:07] <Jymmm> Just make/sell plans for one
[12:01:12] <frallzor> not much capital needed
[12:01:31] <frallzor> make a sturdy mdf-design with easy to get components =)
[12:02:10] <archivist> different people have different definitions of "not much capital"
[12:02:58] <Jymmm> Fine, make GhettoCNC plans
[12:03:01] <jdh> some guy on craigslist was looking for "wanted machinests mini milling mach." so I offered him my X2. Can't decide if he is legit though.
[12:03:02] <frallzor> pretty sure not much is the same for all when the word MDF is involved =)
[12:04:15] <L84Supper> does anyone make a mini mill with a real fast movement, say 1m/sec?
[12:04:30] <L84Supper> or flatbed vinyl cutter
[12:05:15] <frallzor> 1000mm/sec
[12:05:44] <frallzor> 60000mm/min never seen
[12:05:56] <IchGuckLive> Hi all B)
[12:06:14] <frallzor> or well ive seen kick-ass pick and place machines doing similar speeds
[12:06:40] <IchGuckLive> speed is good precision is better
[12:06:59] <IchGuckLive> speed and precise is the goal
[12:07:23] <pcw_home> Speed is easy, acceleration/precision is harder
[12:08:07] <frallzor> nah
[12:08:24] <frallzor> slow precision is easy, fast precision is just a hell lot more expensive =)
[12:08:27] <L84Supper> linear motors easily handle it
[12:08:30] <IchGuckLive> pcw_home: how long is the max cable lengthe from your 7I76 to the driver
[12:08:47] <IchGuckLive> step/dir Enable
[12:08:57] <pcw_home> Linear motor are fine for small forces
[12:09:01] <L84Supper> I'm moving 0.5kg loads
[12:09:10] <IchGuckLive> if been asked for a 7m+ router
[12:09:27] <L84Supper> non contact
[12:10:38] <pcw_home> we normally supply 10 or 6 foot cables 15 or even 25 should be OK but you will likely have to supply 5V power to the 7I76
[12:11:29] <IchGuckLive> ok so 8m cable shoudt work
[12:11:46] <pcw_home> Should (or wait for 7I76E)
[12:11:51] <IchGuckLive> i will tel you if i run into problems on this
[12:12:26] <IchGuckLive> when it is ready for shipment
[12:12:36] <IchGuckLive> next year or within this
[12:13:55] <pcw_home> Within this year (though the main work is HostMot2 driver work)
[12:14:34] <IchGuckLive> ok i guess this E version can handle more then 5V on the AXIS outputs
[12:14:52] <pcw_home> ?
[12:15:20] <IchGuckLive> the Gecko and leadshine also accept 12V to the Step/dir/enable pins
[12:15:39] <IchGuckLive> via resistor
[12:16:18] <IchGuckLive> the card only outputs the 5V
[12:16:26] <IchGuckLive> on the 5Axis connectors
[12:17:24] <pcw_home> No i dont think i will change this (differential is the way to go)
[12:17:40] <IchGuckLive> ok
[12:19:02] <pcw_home> I'm pretty sure Leadshine will work with differential
[12:19:20] <IchGuckLive> for a sutch big mashine i need to figure out the best way to place the cables
[12:19:56] <IchGuckLive> X 7m Y 3m and Z3-4m
[12:20:28] <IchGuckLive> maybe i ride the hole controle with the X axis
[12:21:05] <pcw_home> differential has better noise immunity even than 12V single ended (due to sloppy/unknown OPTO thresholds)
[12:21:22] <IchGuckLive> agree on that
[12:21:36] <IchGuckLive> it works perfect with the 5I20+7i76
[12:21:54] <IchGuckLive> leadshine amps
[12:22:26] <IchGuckLive> leadshine is pretty easy to get here in germany
[12:22:37] <IchGuckLive> only 1or2 geckos available
[12:23:03] <IchGuckLive> also your cards are pretty rear here around in austria i got them
[12:55:54] <IchGuckLive> i added a Howto CRC linuxcnc to youtube as lot of people requested this
[12:55:59] <IchGuckLive> http://www.youtube.com/watch?v=yM_nP9jEB24
[12:56:50] <IchGuckLive> this is common to most of us but gives my students lots of brainactivity
[12:59:44] <JT-Shop> what makes offsets stand out is to have the same path without an offset
[13:01:44] <JT-Shop> first time I saw heeks working
[13:01:49] <IchGuckLive> maybe someone how speeks better english then me shoudt redo this
[13:02:04] <JT-Shop> I understand you fine
[13:02:11] <IchGuckLive> Thanks
[13:02:47] <IchGuckLive> Live from the Oktoberfest in Munich -> http://www.muenchen-tv.de/livestream.html
[13:03:09] <JT-Shop> are you running heeks in 10.04?
[13:03:14] <IchGuckLive> yes
[13:03:46] <JT-Shop> is heeks still being developed?
[13:03:51] <IchGuckLive> yes
[13:04:09] <IchGuckLive> he works most know on heeksart
[13:04:27] <IchGuckLive> but the reprap team is up as 0.22 is out
[13:05:26] <IchGuckLive> best is still theposts and all the functions are in python and can be modified to your own need
[13:08:34] <IchGuckLive> to ALL the New Draftsight from 3DS is out and available for free !!!
[13:08:54] <IchGuckLive> best to get dwg to dxf
[13:10:07] <IchGuckLive> also installable on linux
[13:11:22] <jdh> is there a change log?
[13:11:40] <IchGuckLive> on draftsight
[13:12:36] <jdh> yes
[13:13:59] <L84Supper> Acceleration to 1g (9.8m/s) http://hiwin.com/html/ball%20screws/supers.html or forced cooling http://hiwin.com/html/ball%20screws/cool.html
[13:15:04] <jdh> draftsight starts about 10x faster than autocad2012 on my winbox at home. Also doesn't get as laggy
[13:16:33] <IchGuckLive> jdh: its lighter then your full Acad2012
[13:16:54] <IchGuckLive> L84Supper: i dont think you will reatch this
[13:18:10] <L84Supper> I ever had a need to with ballscrews. I generally use belt drives or linear motors for those speeds
[13:18:16] <L84Supper> ever/never
[13:19:12] <IchGuckLive> im off BY
[13:37:32] <jdh> geez.. plugged in my 'secure, encrypted' thumbdrive and windows blue-screened
[13:38:15] <FinboySlick> jdh: It's secure, it doesn't let you use windows ;)
[13:56:00] <frallzor> I hate myself right now
[13:56:35] <frallzor> lots of sanding left for a frame I cut out just now, why dont I just cnc the parts holding it down....
[13:58:37] <frallzor> but no, I had to do it by hand leaving lots left to sand =P
[14:12:06] <mrsun> frallzor, you deserve it ...
[14:12:54] <frallzor> http://imageshack.us/a/img62/3990/img4485wt.jpg shitload to sand down (according to me)
[14:14:24] <mrsun> maww
[14:14:27] <mrsun> for a girlfriend ?
[14:14:55] <frallzor> yup, she wanted a cute frame for a regular 10x15 photo, so I made onw
[14:15:00] <frallzor> *one
[14:15:23] <mrsun> hahaha you hear that guys ? .. frallzor is playing with girls!
[14:15:29] <mrsun> what a wuzz!
[14:15:46] <andypugh> jealous?
[14:15:49] <mrsun> frallzor, nah looks good =)
[14:15:51] <mrsun> andypugh, yeah :(
[14:16:12] <frallzor> I love Birch, its a kickass wood
[14:16:29] <Jymmm> mrsun: What, you want him to play with boys instead?
[14:16:38] <frallzor> lol
[14:17:01] <FinboySlick> frallzor: I make all my Minecraft houses with Birch.
[14:17:01] <frallzor> TMI for ve7it
[14:17:16] <Jymmm> frallzor: Why is the edge of the large heart so course?
[14:17:39] <frallzor> its the software. when outlining a STL
[14:17:50] <mrsun> Jymmm, yeah and leave the girls for me =)
[14:18:04] <frallzor> for some reason it doesnt like curves made from vectors
[14:18:17] <Jymmm> frallzor: No, I mean the rough edge; it's like the cutter "stuttered"
[14:18:24] <mrsun> hand sawed
[14:18:25] <frallzor> that is it
[14:18:29] <mrsun> ? :)
[14:18:34] <frallzor> it isnt "smooth" in the software either
[14:18:40] <Jymmm> frallzor: weird
[14:18:45] <mrsun> what software +
[14:18:47] <mrsun> ?
[14:18:56] <frallzor> it models it like that due to the fact it doesnt like to outline STLs
[14:19:05] <frallzor> if I made it from scratch it would have been fine
[14:19:11] <jdh> you left those giant tabs on purpose?
[14:19:12] <frallzor> but the stl came from the web =)
[14:19:18] <frallzor> yup
[14:19:21] <Jymmm> frallzor: Why didn't you just clean up the files before machine it?
[14:19:32] <frallzor> clean up stls?
[14:19:42] <Jymmm> or export to another formet
[14:19:42] <JT-Shop> dang found a clausing surface grinder for sale not far from me
[14:20:02] <jdh> with the mag?
[14:20:03] <frallzor> STL is pretty much the only usable format for this
[14:20:16] <JT-Shop> yea and a freq drive
[14:20:19] <mrsun> frallzor, what software are you using ?
[14:20:25] <frallzor> vectric aspire
[14:20:56] <frallzor> Jymmm its really a non-issues, just have to make things either myself in modelers or in-software =)
[14:20:59] <mrsun> frallzor, hmm, then you could just load an image of it and convert it to vectors
[14:21:11] <JT-Shop> dang it's sold
[14:21:25] <jdh> how did you cut under the hearts?
[14:21:30] <mrsun> is there any way to send serial data over parport ... using one output and one input pin ...
[14:21:46] <frallzor> jdh turned the piece
[14:21:53] <jdh> oh, that's cheating.
[14:22:22] <frallzor> http://imageshack.us/a/img856/7270/p1000120d.jpg
[14:22:32] <frallzor> using 2 plugs for guidance
[14:22:42] <frallzor> simple turn =)
[14:23:05] <Jymmm> frallzor: WOW, that is some rough cuts there.
[14:23:08] <jdh> that's a lot of wood ro remove
[14:23:54] <frallzor> best way to machine, 3D rastering =)
[14:24:06] <Jymmm> frallzor: Seriously, you might consider some other software if it's doing shit like that.
[14:24:37] <frallzor> its not the software =) its low-res stls that causes it
[14:24:59] <frallzor> if I make something from scratch in the software its fine
[14:25:07] <frallzor> I export high-res stl from solidworks = fine too
[14:26:34] <frallzor> the cuts on that roughing is as it should be with that type of machining, quicker to work with
[14:27:17] <mrsun> if its less then 100mm cut depth its not roughing!
[14:27:40] <frallzor> http://imageshack.us/a/img37/903/img4046lr.jpg this piece made the same way =)
[14:27:50] <frallzor> but it was 100% from scratch in solidworks
[14:40:03] <pcw_home> mrsun: you could make a bit banged UART comp that runs at the base or servo thread baud rate (set servo thread to 1200 Hz for 300 baud and 4x oversampling)
[14:41:02] <pcw_home> good exercise in understanding how a UART works...
[14:41:14] <frallzor> one understand why high-res stls for machining costs a fortune =)
[14:43:05] <mrsun> pcw_home, hmm
[14:43:22] <andypugh> mrsun: what data are you trying to send?
[14:43:48] <mrsun> andypugh, well i want two signals of the parport to be serial data for non realtime stuff like set spindle speed, start spindle, stop spindle, coolant etc
[14:43:52] <mrsun> =)
[14:44:23] <andypugh> Talking to what?
[14:44:44] <mrsun> andypugh, linuxcnc sending the data, then stuff like spindle rpm etc could go back on serial data also =)
[14:45:00] <andypugh> Sending the data to what?
[14:45:07] <mrsun> andypugh, huh ?
[14:45:25] <andypugh> What is acting on the serial data?
[14:45:35] <andypugh> What is interpreting the serial data?
[14:45:36] <mrsun> linuxcnc -> breakoutboard (spindle speed etc) breakoutboard -> linuxcnc (spindle rpm, jogbuttons etc)
[14:46:19] <andypugh> So, what will the UART be on the breakout?
[14:47:01] <andypugh> It seems to me that _sending_ the data is the wasy part.
[14:47:08] <andypugh> (easy)
[14:47:24] <mrsun> well i guess both sides need the same capabilities
[14:47:38] <mrsun> linuxcnc to send serial data to the bob, the bob to send serial data back to linuxcnc
[14:48:28] <andypugh> Microprocessor on the BOB?
[14:48:51] <andypugh> Seems like it would be simpler to use an Arduino on USB
[14:49:15] <mrsun> andypugh, yeah was thinking that... =)
[14:49:34] <mrsun> those arduino is so overpriced tho ...
[14:49:35] <mrsun> :P
[14:49:45] <mrsun> and i freakin HATE that they dont make them with screw terminals
[14:49:53] <mrsun> sure you can buy screw terminal boards for it but ...
[14:49:58] <andypugh> Have you seen: http://axis.unpy.net/01198594294
[14:51:28] <jdh> http://www.jameco.com/1/1/671-dfr0060-screw-shield-converts-header-pins-screw-terminals-arduino.html
[14:51:50] <mrsun> yes ...
[14:52:29] <mrsun> does the arduino hal module work good ?
[14:53:23] <andypugh> skunkworks: is using it
[14:53:48] <skunkworks> seems to. No issues.
[14:54:37] <skunkworks> http://www.youtube.com/watch?v=h-CdFd2Zakc
[14:55:14] <skunkworks> I had some problems initally related to old firmware in the uno.. (but that info was added to jeplers blog)
[14:57:30] <FinboySlick> skunkworks: Always awed by your big mill. What do you mostly use it for?
[14:57:43] <skunkworks> tinkering
[14:57:45] <skunkworks> ;)
[14:58:00] <skunkworks> odd jobs here and there..
[14:58:01] <FinboySlick> skunkworks: Best type of use. I wish I had this kind of tinkerspace.
[14:58:39] <Jymmm> This is for JT-Shop http://hooniverse.com/2012/09/24/truth-in-advertising-hazard-fraught-tools/
[14:59:01] <mrsun> http://www.electrokit.com/arduino-nano-v3.49162 <-- would for example that work with the arduino module ?
[15:00:02] <taiden> evening gents
[15:02:00] <jdh> mrsun: looks like any arduino supported hw would work (that one included)
[15:02:10] <taiden> any built in pocketing routines?
[15:02:20] <taiden> just circles
[15:02:55] <jdh> taiden: check http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Simple_LinuxCNC_G-Code_Generators
[15:03:24] <taiden> okay that should work
[15:03:36] <taiden> but out of curiosity there is nothing built in?
[15:04:11] <jdh> no, it's not the unix way.
[15:04:31] <taiden> :]
[15:04:42] <jdh> you can embed code generators in axis
[15:04:57] <taiden> i still need to get to that point of understanding
[15:05:02] <syyl> or you can create your own routines
[15:05:07] <syyl> that can be called
[15:05:25] <taiden> i'd like to create a GUI with radio buttons and whatnot
[15:05:39] <taiden> that called different pieces of code
[15:05:40] <taiden> eventually
[15:06:02] <taiden> so any dummy can operate the production line
[15:06:06] <taiden> including this dummy
[15:06:16] <andypugh> taiden: Modify this? http://www.bodgesoc.org/lathe/lathe.html
[15:06:37] <syyl> at the moment i am tcreating a slotting routine, inspired by a routine of a haidenhein control
[15:06:38] <syyl> https://dl.dropbox.com/u/24396704/o210.png
[15:06:40] <taiden> nice!
[15:06:57] <taiden> yeah that's a good one
[15:07:03] <syyl> slot from centre, with given angle, depth, with and length
[15:07:05] <taiden> i do slotting in a particularly terrible way
[15:07:14] <jdh> you should think about making it an ngcgui subroutine
[15:07:14] <taiden> it works fine though
[15:07:30] <andypugh> ngcgui is certainly worth a look too.
[15:07:46] <taiden> i honestly haven't even had the time to stab into it
[15:07:58] <syyl> mh, i feel ngcgui a bit arkward to use
[15:07:59] <andypugh> Though I think ngcgui is perhaps a little more flexible than taiden wanted?
[15:08:05] <taiden> especially since my shop is now separate, i can't just walk downstairs and hack away anytime
[15:08:18] <syyl> i prefer to call the routine with a single line in the mdi
[15:08:32] <taiden> i wish i could run an instance of linuxcnc on windows/os x just for hacking purposes
[15:08:40] <syyl> taiden
[15:08:48] <andypugh> syyl: That's pretty much what my lathe system is, except buttons link to the MDI commands
[15:08:50] <syyl> look closely on my screenshot :D
[15:09:02] <taiden> haha!
[15:09:03] <taiden> derp
[15:09:10] <syyl> its windows, running virtualbox with emc
[15:09:14] <andypugh> taiden: I run linuxCNC unde VMware on my Mac.
[15:09:15] <syyl> for simulation purposes
[15:09:26] <taiden> no kidding
[15:09:32] <taiden> ask and ye shall recieve
[15:09:48] <taiden> how much disk space does the entire deal take up?
[15:11:17] <syyl> i set it up with 3gb space
[15:11:20] <syyl> seems to work
[15:12:20] <andypugh> taiden: 2.4.7, 2.5 and 2.6 all at the same time: http://imagebin.org/229938
[15:15:28] <Jymmm> ok, 8W from the solar blanket! woohoo
[15:15:36] <syyl> hey!
[15:15:38] <syyl> thats enough for
[15:15:39] <syyl> hmm
[15:15:42] <jdh> Got the shrimp for $680
[15:15:45] <jdh> <urk>
[15:15:47] <syyl> well
[15:15:54] <taiden> alright
[15:15:55] <syyl> not very much ;)
[15:15:56] <taiden> i am inspired to do this
[15:16:09] <Jymmm> syyl: right now, 4 blowers =)
[15:16:22] <syyl> to cool the solar blanket?
[15:16:30] <Jymmm> exactly
[15:16:44] <syyl> ;)
[15:16:57] <Jymmm> 7W to go!
[15:17:23] <syyl> you could support the sun with a big flashlight
[15:17:23] <syyl> :D
[15:17:28] <andypugh> I think the D525 is 18W..
[15:17:34] <Tom_itx> Jymmm, are you trying to get apollo 13 back?
[15:17:51] <Jymmm> andypugh: @ 12V ?
[15:18:01] <syyl> that will fail without tom hanks.
[15:18:05] <jdh> prepping for the impending zombie apocalypse
[15:18:13] <syyl> oh
[15:18:14] <andypugh> Yes
[15:18:22] <syyl> to burn the zombies away with a 15W laser?
[15:18:37] <Jymmm> andypugh: sucks, the solar blanket is max 15W
[15:18:57] <andypugh> syyl: With capacitors you might get 100w pulses.
[15:19:04] <syyl> ;)
[15:19:16] <syyl> i would prefer a projectile weapon
[15:19:26] <syyl> "old school"
[15:20:02] <Jymmm> ALRIGHT (now have full sun)... 14.34V @ 13.1W
[15:20:23] <syyl> hmm
[15:20:33] <syyl> i think we lost apollo13
[15:20:39] <syyl> a few seconds ago
[15:21:39] <Jymmm> syyl: I'm going to use it to recharge battery pack for portable transceiver use
[15:22:20] <Jymmm> (that's actually it's original usage for guberment)
[15:22:30] <syyl> ah ok
[15:23:13] <Jymmm> syyl: guestimate around 5 hours to fully charge battery, but we'll see
[15:24:06] <Jymmm> I've never really put the blanket to load as I don't have a "real" charge controller
[15:26:17] <Jymmm> syyl: Solar_blanket --> watts_up_pwr_meter --> LiIon_bat --> load
[15:26:50] <Jymmm> syyl: http://www.hobbyking.com/hobbyking/store/__10080__Turnigy_130A_Watt_Meter_and_Power_Analyzer.html
[15:32:08] <skunkworks> http://www.youtube.com/watch?v=hPE3Qr-ECtQ
[15:35:38] <syyl> wow
[15:35:43] <syyl> an working hexapod
[15:36:33] <syyl> but such a machine gets just big, for pretty small ax travels
[15:46:22] <DJ9DJ> gn8
[15:51:14] <frallzor> ah 5 minutes of sanding and a flawless frame
[15:56:51] <skunkworks> I think they are cool... I am going to have to build one some day
[15:56:59] <skunkworks> (hexapod)
[16:26:29] <BHSPiMonkey> hey all
[16:27:20] <micges> hi
[16:28:10] <BHSPiMonkey> (skimming my scrollback) wow r00t4rd3d, that box is amazing
[16:34:11] <r00t4rd3d> you should see the ones CarveOne makes ....
[16:34:25] <r00t4rd3d> makes mine look like roadside trash
[16:34:57] <andypugh> Have you seen Dewey's stuff?
[16:35:13] <r00t4rd3d> i dont think so
[16:36:28] <andypugh> http://www.deweygarrett.com
[16:37:39] <tjb1> hey r00t4rd3d
[16:37:45] <andypugh> Then there is bill ooms: http://billooms.com
[16:37:46] <tjb1> and andypugh how are ya fellows
[16:38:00] <andypugh> (Both using LinuxCNC)
[16:38:58] <BHSPiMonkey> Freud would be proud of the OT2 section of deweygarrett.com
[16:39:16] <BHSPiMonkey> Well, perhaps not proud, but he'd feel some way or another about it
[16:40:30] <tjb1> I have a question for the electronic gurus here
[16:40:40] <BHSPiMonkey> Man, his ornamental turning setup is amazing though
[16:40:57] <tjb1> I have all my homing switches on 1 input and they are all NPN, I want a light for each axis so I can see that they homed or if they are on the limit
[16:41:21] <tjb1> I am using the NPN switch to ground the light, is there something I can put so when one switch activates, it doesnt ground all the lights?
[16:41:26] <andypugh> I don't think you can do that.
[16:42:04] <r00t4rd3d> think they would all have to be on separate inputs to do that
[16:42:11] <tjb1> Yeah I was afraid of that...
[16:42:27] <andypugh> You my be able to arrange diodes so that only the "upstream" lights are on.
[16:42:31] <tjb1> I could do it with relays but thats over complicated
[16:42:32] <r00t4rd3d> or use an arduino and run another wire to each
[16:42:53] <tjb1> Maybe I will just do one light right now and upgrade later
[16:43:05] <andypugh> Hmm. I think I just changed my mind.
[16:43:30] <andypugh> I think maybe you can do it with diodes.
[16:43:56] <andypugh> If all the switches are connected to the p-port pin with a diode, then they can't "see" each other.
[16:44:13] <tjb1> Then just hook the light ground before it hits the diode
[16:44:22] <tjb1> so the light and switch are isolated from each other
[16:44:34] <tjb1> Isolated from the other lights/switches
[16:44:44] <andypugh> In fact, you might be able to use LEDs as the diodes..
[16:45:31] <tjb1> How would I use a LED? Wouldnt it require power?
[16:45:45] <andypugh> (That requires that the Vf of the LED be less than the 0v detection level of the parport)
[16:47:06] <tjb1> Ill give it a try with some diodes
[16:47:26] <andypugh> 1k resistor (possibly smaller) to +5V. Led + to the end of the resistor. P-port to the end of the resistor. LED - to the ground switch
[16:50:11] <tjb1> Ill try diode, that doesnt work Ill try the led, I dont have any 5v available currently, only 12 and 48
[16:50:20] <tjb1> nor a chip to get either one to 5 :)
[16:52:07] <andypugh> USB? There is 5V in there
[16:53:29] <tjb1> I have no USB in my box
[16:53:45] <tjb1> I can get it there and I have 5v wallwart but if not needed, less work ;)
[17:07:25] <andypugh> The p-port inputs can't source anything like enough current to light an led.
[17:08:19] <andypugh> But 12V on the opposite side of an LED will stop current flow out of the pin, and 0V there will allow current flow, so that ought to be parport-safe...
[17:08:28] <andypugh> But it feels a bit scary.
[17:08:54] <Jymmm> tjb1: laptop?
[17:09:12] <tjb1> For what Jymmm
[17:09:28] <Jymmm> tjb1: tjb1: I have no USB in my box
[17:09:29] <tjb1> The prox and leds have there own 12v power supply
[17:09:47] <tjb1> Yes, I have no run any USB to my box and dont want to add more power supplies than needed
[17:10:04] <tjb1> or cables
[17:10:04] <Jymmm> tjb1: is "YOUR BOX" a laptop?
[17:10:14] <tjb1> No its the box with the power supply, g540...
[17:10:18] <andypugh> Odd to have a parallel port and no USB
[17:10:45] <tjb1> The electronics box on the machine where everything is wired together.
[17:11:23] <Jymmm> tjb1: what type of computer conects to YOUR BOX? a deaktop or laptop?
[17:11:31] <tjb1> A desktop
[17:11:40] <tjb1> I dont not want another cable running into it if not needed.
[17:12:15] <Jymmm> tjb1: The HDD power cable red wire is +5VDC
[17:12:43] <tjb1> Yes I know, I also have a spare computer power supply but were back to I dont want another cable :)
[17:12:54] <tjb1> All I want from computer is the parallel cable(s)
[17:12:59] <Jymmm> tjb1: then you dont get 5V
[17:13:11] <Jymmm> you need another cbale SOMEWHERE =)
[17:13:13] <tjb1> I have a wallwart that if I need 5v I will stuff iti n there
[17:13:49] <tjb1> I have the 48v power supply for the gecko and a 12v wallwart for the 3 homing sensors and the 3 led indicators
[17:14:12] <tjb1> 1 power wire in, pull the 120 for wallwart off the 48v input :)
[17:14:18] <tjb1> I like to keep it simple
[17:16:31] <tjb1> I have to paint tonight and maybe get it all assembled tomorrow
[17:16:49] <tjb1> Actually get the full x axis racks on and both homing switches and maybe draw on some paper!
[17:18:56] <Jymmm> tjb1: Why not just grab a LM7805?
[17:19:05] <tjb1> That cost money!
[17:19:11] <Jymmm> 35¢
[17:19:25] <tjb1> Where
[17:19:26] <Jymmm> YOU CHEAP BASTARD!
[17:19:36] <Jymmm> anywhere
[17:19:37] <tjb1> You forgot shipping!
[17:20:15] <Jymmm> Radio Shack $1.99
[17:20:22] <tjb1> Thats $2 not $.35
[17:20:34] <Jymmm> Well, it's RS, that's implied
[17:20:36] <tjb1> I already gotta cough up $300 something for the water table
[17:20:59] <tjb1> College + no job = poor
[17:21:24] <Jymmm> you can collect cans for that
[17:22:00] <tjb1> Haha screw that.
[17:22:53] <Jymmm> YOU CHEAP LAZY BASTARD!
[17:23:00] <tjb1> You ever collect cans?
[17:23:07] <tjb1> It takes a full pickup load to get like $20
[17:23:12] <Jymmm> No, copper =)
[17:23:28] <tjb1> Ive burned much insulation off dynamite wire to turn in
[17:23:37] <Jymmm> arm full of copper is like $115 =)
[17:23:43] <tjb1> I like taking batteries in, they fetch a good price for being worthless :)
[17:25:07] <tjb1> class over, peace.
[17:40:54] * JT-Shop forgets how to tell what gpio on the 5i20 is what pin on the 7i37ta :(
[17:42:55] <frallzor> hello there mr JT
[17:43:03] <JT-Shop> hi
[17:43:17] <frallzor> oh sorry, Mr JT sir
[17:43:21] <PCW> they go input 0 through input 15 and the output 0 through 7
[17:43:25] <frallzor> ;)
[17:43:28] <PCW> then
[17:44:39] <JT-Shop> so in 1 on TB2 is
[17:45:12] <PCW> so the inputs are GPIO 0..15 or 24..39 or 48..63
[17:46:04] <PCW> and the outputs are GPIO 16..23 or 40..47 or 64..71
[17:46:51] <JT-Shop> thanks
[17:47:47] <PCW> Which 5I20 connector is the 7I37TA connected to?
[17:49:42] <JT-Shop> it is in the P3 connector
[17:50:27] <JT-Shop> I think it is hm2_5i20.0.gpio.025.in
[17:50:32] <PCW> so input 0 = GPIO24
[17:50:33] <PCW> and output 0 = GPIO40
[17:50:42] <JT-Shop> if I have it cyphered right
[17:51:08] <JT-Shop> ok so input 1 is gpio25
[17:51:16] <PCW> Yes
[17:51:25] <JT-Shop> dang I need to make a chart for this :)
[17:51:40] <JT-Shop> PCW: thanks again
[17:52:13] <PCW> just need to subtract multiples of 24 in your head
[17:52:44] <JT-Shop> math is a struggle for me if you have not noticed LOL
[17:53:49] <alex_joni> 4 sixpacks?
[17:54:26] <JT-Shop> I'd be asleep by the end of 2 sixpacks for sure if not sooner
[17:54:48] <alex_joni> 4 sixpacks = 24
[17:55:00] <PCW> Maybe this HAL metadata stuff will allow nicer names at some point
[17:55:51] <JT-Shop> I don't mind the names so long as I can figure out what name goes to what terminal
[17:56:39] <PCW> with the right kind of macro foo they could be 7I37.0.in3
[17:57:02] <jdh> write something to parse the output and assign terminals?
[17:57:03] <alex_joni> you can alias them
[17:58:46] <PCW> Yeah (its more easily done with sserial devices that identify themselves and have seriall numbers)
[18:00:00] <jdh> revo has DSV not BOV
[18:01:27] <PCW> were looking in to an alternate naming scheme for these so instead of the current "path name"
[18:01:28] <PCW> we have an card name so a 7I70 would be 7I70.SN.signalname rather then 5i25.0.7i77.5.signalname
[18:02:00] <PCW> much less painful if you swap cables around
[18:03:31] <LatheBuilder> Anyone on? New to this
[18:04:35] <PCW> the assumption being that its much easier to swap host interfaces than all the field I/O
[18:04:37] <PCW> and upstream host interface changes should be transparent
[18:05:20] <PCW> LatheBuilder didn't have much patience...
[18:14:11] <andypugh> No. I always feel bad when that happens
[18:15:03] <andypugh> PCW not got far with that idea. Distracted by an LCD component at the moment
[18:28:31] <PCW> Yeah its just a good idea at this point
[18:29:04] <PCW> LCD component?
[18:41:12] <andypugh> PCW: Something to link HAL pins to the 7i73
[18:42:08] <andypugh> I thought it would be quick and easy and fun. I have just found out that kernel-mode sprintf (and friends) are very limited, so I might need to write my own..
[18:50:40] <PCW> Yeah no hal strings
[18:51:30] <andypugh> I knew that part, and had already worked round it. My idea is actually quite neat, but bolloxed up by the kernel code.
[18:53:21] <PCW> regular printf is quite powerful (and big and slow)
[18:53:45] <andypugh> That is why I wanted to use it.
[18:54:13] <PCW> yeah back to stone knives and bearskins
[18:55:03] <andypugh> The idea was that loadrt lcd string_fmts="X Position = % 4.2f" would create a float-HAL pin and insert the formatted value of that pin into a string and pass it to the LCD.
[18:55:56] <PCW> That looks really cool
[18:56:13] <andypugh> Then I had a format for multiple pages, and a "page number' pin to switch between them. It would work nicely if kernel mode printf could actually deal with floating point formats..
[18:57:05] <PCW> just going to say the depends on printf magic
[18:58:03] <PCW> so the servo thread can access floating point functions but has no easy way to format them into strings
[18:58:12] <andypugh> Indeed
[18:58:28] <PCW> format FP numbers I mean
[19:00:19] <andypugh> It's not a disaster, I can write a limited printf. It also means that I can explicitly convert \n into 0x18 0x0A 0x0D
[19:00:26] <PCW> I wonder if there's a way to add a better printf to the kernel functions (or are these RTAI functions?)
[19:01:20] <andypugh> It has to be a deliberate choice not to support it, presumably there is a good (security?) reason
[19:02:57] <PCW> Did you notice the the 7I73 LCD wants zeros (null chars) when theres nothing changed?
[19:04:02] <PCW> (since sserial is a synchronous interface and you always send something)
[19:04:51] <andypugh> Yes. Though I am not sure whether it is worth bothering. It's actually easier for me to send a full screen repeatedly.
[19:05:40] <micges> rtapi_print is wrapper to rt_printk from rtai
[19:05:53] <micges> which also don't have floats
[19:06:04] <PCW> You can overflow its little (1024 byte) buffer if you send many slow commands in a row
[19:06:55] <andypugh> Do any commands take more than 1.024 seconds?
[19:07:30] <PCW> I think clear screen takes maybe a ms or so
[19:08:24] <andypugh> OK, I can put a bunch of nulls after a clear
[19:08:31] <PCW> but you would normally send cursor addressing, erase to EOL and new line data for each line and those are fast
[19:09:15] <PCW> You dont need the nulls Theres a input side FIFO, just dont send 1024 clear commands in a row
[19:09:24] <andypugh> Anyway, time to sleep.
[19:09:47] <PCW> also for a non-flickering display you dont normally clear
[19:10:00] <PCW> 'nite
[19:30:12] <Thetawaves> djdelorie, have you ever done any work with the rx arch from renesas?
[19:30:36] <djdelorie> yes, I wrote the GNU assembler and simulator for it
[19:31:00] <djdelorie> and I've done lots of hardware projects with them, including my BLDC controller, which uses an RX62T processor
[19:32:49] <djdelorie> why?
[19:33:50] <Thetawaves> i'm curious about endianness
[19:34:13] <djdelorie> it has both. Use little endian :-)
[19:34:23] <Thetawaves> i *have* to use little endian
[19:34:27] <djdelorie> unless you're doing CAN, that peripheral is endian-weird
[19:35:00] <Thetawaves> i noticed that instructions only seem to be stored in big endian, while data can be stored as little endian
[19:35:09] <djdelorie> (you can still do little endian with CAN, but you have to use different code to read/write some of the registers)
[19:35:14] <Thetawaves> all of my little endian dependent logic will still work?
[19:35:31] <djdelorie> instructions are stored as a byte stream
[19:35:44] <djdelorie> my projects are always little endian
[19:36:05] <Thetawaves> cool
[19:36:18] <Thetawaves> is it a relatively fun processor to work on?
[19:36:26] <djdelorie> yes :-)
[19:36:45] <Thetawaves> ok great
[19:36:54] <djdelorie> lots of peripherals, single-cycle FPU, wide range of chip sizes, etc.
[19:36:59] <djdelorie> linux tools :-)
[19:37:09] <Thetawaves> linux tools is a huge advantage
[19:37:20] <djdelorie> (some of which you get from me, like the flash downloader)
[19:37:24] <Jymmm> What happens to feed point impedance at the base of a fixed-length HF mobile antenna as the frequency of operation is lowered?
[19:37:29] <djdelorie> all my RX development is linux-only.
[19:37:32] <Jymmm> sorry
[19:37:39] <Thetawaves> i didn't select the pic32 because the compiler isn't integrated into gcc mainline
[19:37:58] <djdelorie> RX is, I use gcc "head" as my regular compiler
[19:38:00] <Thetawaves> what kind of programmer do you use? do you have jtag?
[19:38:17] <djdelorie> I use the serial bootloader. The JTAG interface isn't (and likely won't) be documented.
[19:38:37] <djdelorie> the serial bootloader is documented in the many-thousands-of-pages-long hardware documentation.
[19:38:45] <Thetawaves> don't use jtag anyway, so not a deal breaker
[19:39:14] <djdelorie> I have a schematic for a standard FT232R-based interface, too, which gives you the linux tool support and a printf console
[19:39:15] <Thetawaves> serial bootloader... i'll take a look
[19:39:29] <Thetawaves> bootloader comes with the chips?
[19:40:10] <djdelorie> http://people.redhat.com/~dj/m32c/ - the r8c/etc programmer includes RX too. The bootloader is pre-flashed in a separate flash area in the chip, it's always there.
[19:40:47] <djdelorie> http://www.delorie.com/electronics/rx/ & http://www.delorie.com/electronics/rx-stick/ have most of the stuff I've done hardware-wise.
[19:41:07] <Thetawaves> how do you configure it to boot directly to your code?
[19:41:32] <djdelorie> there are two mode pins on the chip; their values at reset determine the chip's mode
[19:41:46] <djdelorie> it chooses between serial bootloader, USB bootloader, and your program.
[19:41:57] <Thetawaves> very cool
[19:42:30] <djdelorie> if you haven't gotten hardware yet, consider http://www.delorie.com/electronics/rx/rx62n-breakout.html - I still have plenty and they're linux-friendly.
[19:42:55] <djdelorie> (plenty of blanks, that is, I build the full ones as needed)
[19:43:31] <djdelorie> here's the RX62T bldc controller board: http://www.delorie.com/electronics/bldc/
[19:43:31] <Thetawaves> how much for the parts kit?
[19:43:54] <djdelorie> dunno at the moment, my digikey BOM script is broken.
[19:44:22] <djdelorie> the chip is around $16, plus $5 for the usb chip, add a few dollars for connector and discretes as needed, etc
[19:45:22] <Thetawaves> i certainly would like to get a reference board
[19:45:41] <djdelorie> Renesas has a bunch, but they're not all linux-friendly.
[19:47:23] <djdelorie> http://www.digikey.com/product-detail/en/YRDKRX62N is a good one
[19:47:54] <djdelorie> but it needs a resistor swap to do USB bootloader
[19:50:09] <Thetawaves> i'd rather get your breakout board
[19:50:20] <Thetawaves> i just don't want to spend the time trolling digikey to get the parts
[19:50:54] <djdelorie> there's a BOM for digikey right on that web page
[19:51:23] <djdelorie> unless they obsoleted some of the parts on me :-P
[19:52:33] <djdelorie> heh... the bom includes everything *except* the MCU
[19:53:19] <Thetawaves> why do you have two boms with different size smd?
[19:53:24] <djdelorie> also, I've been having bad luck with the RTC clock circuit on my boards, don't know why.
[19:53:33] <Thetawaves> :O
[19:53:49] <djdelorie> The PCB allows for either size part, so you choose whatever size you mostly have in stock. My stock is mostly 0603, for example.
[19:54:04] <djdelorie> I suspect it's my choice of crystals
[19:54:49] <djdelorie> the renesas board uses an SMT crystal, but that's hard to hand-solder so I had to go with a canned crystal instead.
[19:55:03] <Thetawaves> i downloaded your schematic for the breakout board, and man is that a weird clock circuit
[19:55:25] <djdelorie> which clock?
[19:55:39] <Thetawaves> rtc, 32.768khz osc
[19:55:45] <Thetawaves> x2
[19:56:06] <djdelorie> it's a standard crystal clock circuit...
[19:56:09] <Thetawaves> never seen a clock with R14
[19:56:22] <Thetawaves> what is the load of X2?
[19:57:25] <djdelorie> depends on the crystal. you can always put a 0R there if that's what your crystal wants
[19:58:17] <djdelorie> if I ever respin that board, both clocks get both a series and parallel resistor footprint, "just in case"
[19:58:49] <djdelorie> next flaky board, I'll try a 0R and see if it helps.
[19:58:55] <Thetawaves> but that will put them unnecessarily far from the chip
[19:59:06] <Thetawaves> causing excessive stray capacitances that will throw the freq off
[19:59:40] <djdelorie> they're not that far from the cpu anyway, and the stray pF should be accounted for in the load caps anyway
[20:00:06] <Thetawaves> i'm trying to find the equation
[20:00:13] <Thetawaves> damn, i've got tons of pdfs open
[20:00:22] <djdelorie> you need more monitors :-)
[20:01:09] <Thetawaves> but it's like ((C1 + C1s) * (C2 + C2s))/((C1 + C1s) + (C2 + C2s))
[20:01:20] <djdelorie> I've seen inconsistent information on how to go from the "load capacitance" in the crystal spec, to the actual capacitance on the pcb
[20:01:49] <djdelorie> AFAIK, the caps are just under twice the spec'd load
[20:01:56] <djdelorie> (in general)
[20:01:57] <Thetawaves> so in that schem, you've got a crystal with about 16pf load?
[20:02:06] <djdelorie> something like that
[20:02:29] <Thetawaves> i've seen a lot of missinformation too
[20:03:07] <Thetawaves> you CAN calculate it, and the capacitor values are not 22pf when the crystal calls for 22pf load
[20:03:18] <Thetawaves> moving on
[20:03:24] <djdelorie> the RX spec says 15pF per cap including strays Their eval board uses two 12pF caps.
[20:03:52] <djdelorie> but that assumes you get a crystal that works with those values, too.
[20:04:03] <Thetawaves> definitely
[20:05:09] <Thetawaves> so you've got the cpu split up into a bunch of functional blocks in that schem
[20:05:14] <djdelorie> yup
[20:05:25] <Thetawaves> ahh cool
[20:06:00] <djdelorie> well, one block for power/download, and the rest are just one block per row of pins on the chip
[21:31:47] <r00t-Shed> mhaberler, fix ur stuffs
[21:50:58] <Loetmichel> sooo... for the highlight at the end: a the last monitor i had to rework had a loosened treaded rivet... so i couldnt dismatle it :-(... all done... now i will go home, sleep about 2 hrs, and go back to the company... that was a LOOONG day... startet yesterday 0800, now ist 0438 cver here...
[21:52:56] <Loetmichel> *workday
[23:45:16] <uw> does anyone have a CNC shaper?
[23:47:51] <toastyde1th> i've seen cnc shapers
[23:47:56] <toastyde1th> shit's crazy
[23:49:26] <toastyde1th> waste of time in my opinion
[23:49:28] <toastyde1th> but crazy
[23:50:55] <archivist> I can see a use for one off gear shaping
[23:51:22] <toastyde1th> true
[23:52:04] <toastyde1th> i mean that's essentially what production gear shapers are
[23:52:54] <archivist> I am waiting for the right request then I shall write the gcode to shape a bevel gear
[23:52:55] <toastyde1th> now a vertical shaper with an intgral vertical rotary table and maybe a tracer
[23:53:07] <toastyde1th> and we'd be in business
[23:53:15] <toastyde1th> but a normal shaper, I don't know how much I'd be into that
[23:53:35] <toastyde1th> also with a good shaper you don't need cnc to shape gears
[23:53:39] <toastyde1th> or any special cutters