Back
[00:05:59] <Gamma-X> hey everyone whats up
[00:08:45] <eric_U> everybody put me on their idiot filter, what's up with you?
[00:12:05] <JymmmEMC> eric_U: you already are and have been for a very very very long time. ;)
[00:12:41] <eric_U> I see there is an ambiguity with the tense in that sentence
[00:13:00] <SWPadnos> ok everybody, just relax
[00:13:11] <JymmmEMC> SWPadnos: Got Beer?
[00:13:18] <SWPadnos> har de har har
[00:13:22] <eric_U> SW is buying!
[00:13:35] <JymmmEMC> eric_U: Yep, that's how I read it as
[00:13:41] <SWPadnos> I rarely buy beer, and it's usually Guinness when I do
[00:13:47] <eric_U> wait, I'm in your idiot filter
[00:13:56] <eric_U> Guinness is good
[00:14:02] <SWPadnos> but "usually" is waaaaaay too strong a word, considering that I buy beer about once every 5-10 years
[00:14:09] <JymmmEMC> Guinness is Fine by me, how about eric_U?
[00:14:47] <JymmmEMC> SWPadnos: 2008 sounds like a good year to me
[00:14:57] <SWPadnos> so - Costco had a really good deal on 2GB Cruzer USB drives
[00:15:06] <JymmmEMC> <$10?
[00:15:14] <SWPadnos> (actually, their regular price is excellent)
[00:15:20] <SWPadnos> slightly over - $13 each
[00:15:24] <SWPadnos> 3-pack for $40
[00:15:35] <SWPadnos> normally $50, but $10 coupon
[00:15:39] <eric_U> medical companies have started sending my wife 512mb flash drives instead of cd's
[00:15:59] <JymmmEMC> eric_U: Nice, gotta tell em you need more data
[00:16:01] <SWPadnos> yeah, I've gotten several of those at trade shows with slide shows on them or whatnot
[00:16:19] <eric_U> I wish they were bigger, that's for sure
[00:16:27] <eric_U> my 1gb filled up in 6 months
[00:16:51] <SWPadnos> heh
[00:17:32] <SWPadnos> I bought a 256M, won a 1G, got a 128M as trade show swag, then got another one or two 256M at other shows, then finally bought a larger one
[00:17:52] <eric_U> I got a 2gb from work
[00:17:59] <SWPadnos> now I've got enough that my wife can use them and I don't have to worry about her losing them :)
[00:18:25] <SWPadnos> it's kind of funny to be holding 28GB of flash in my hand
[00:18:26] <Gamma-X> anyone ever mess with the linux touchware drivers?
[00:18:43] <eric_U> should we?
[00:18:43] <SWPadnos> nope
[00:19:30] <Gamma-X> eric_U no not neccasary
[00:20:32] <SWPadnos> funny - they provide binary drivers for "some of the most common distributions"
[00:20:43] <SWPadnos> which is 6 red hat versions and 5 suse versions
[00:20:49] <SWPadnos> and nothing else
[00:20:58] <SWPadnos> (luckily there is also source)
[00:28:05] <Gamma-X> what kernel does emc run? on the latest release
[00:28:17] <eric_U> uname -a
[00:29:08] <Gamma-X> not on it currently
[00:29:40] <eric_U> I'm running 2.6.16
[00:29:49] <eric_U> 2.6.15
[00:30:02] <eric_U> that probably means you'll have to build from source
[00:31:44] <Gamma-X> ok
[00:31:52] <Gamma-X> i think the driver is actualy allready in the kernel
[00:32:16] <Gamma-X> i read sumthing stating that the driver is in kernels 2.6 and on
[01:50:27] <jmkasunich> as much as I dislike harbor fright stuff, I gotta admit their 4x6 H/V bandsaw is nice
[01:50:41] <jmkasunich> just made two cuts through 2-1/4 x 6/14 cast iron
[01:50:59] <jmkasunich> the motor is very hot and smells of varnish, but it will live
[01:51:06] <jmkasunich> 6-1/4 that is
[01:52:07] <SWPadnos> solid or tube?
[01:52:11] <jmkasunich> solid
[01:52:14] <SWPadnos> cool
[01:52:25] <jmkasunich> tube would be easy
[01:52:28] <SWPadnos> I've used the Grizzly version, and it's slow but it works
[01:52:43] <SWPadnos> much faster on the Big Johnson :)
[01:53:06] <skunkworks> we have broke the shaft that runs the bottom wheel twice now.. (not a HF but similar design)
[01:53:31] <jmkasunich> wow, how'd you manage that?
[01:53:38] <skunkworks> I think it has probably paid for itself may times over.
[01:53:44] <skunkworks> fatuge
[01:54:12] <skunkworks> we have probably had it for 20 or so years.
[01:54:17] <jmkasunich> ah
[01:54:26] <jmkasunich> and probably used it a lot more than I use mine
[01:54:29] <skunkworks> heh
[01:54:36] <jmkasunich> I suspect the motor will be the first thing to go on this one
[01:54:55] <skunkworks> yes - we have replaced that atleast once..
[01:55:04] <skunkworks> and yes - it gets hot :)
[01:55:28] <jmkasunich> the smell of unhappy motor
[01:55:36] <skunkworks> boil spit
[01:56:09] <skunkworks> I think we broke the lower bearing once also.
[01:56:33] <skunkworks> when the blade stops tracking - something is wrong ;)
[01:57:11] <SWPadnos> I found it tedious to cut angle steel
[01:57:32] <SWPadnos> small stuff - like 1.5x1.5x1/8 or something
[01:57:50] <jmkasunich> not as tedious as a hacksaw I bet
[01:58:00] <SWPadnos> not quite :)
[01:58:21] <SWPadnos> I was making V cuts for a bracket - bend then weld
[01:58:24] <skunkworks> when the blade stops tracking - something is wrong ;)
[01:58:30] <skunkworks> oops
[01:58:40] <SWPadnos> so two cuts, trying to get them to meet in the middle
[01:58:51] <SWPadnos> oh wait, it was square tube, not angle
[01:58:53] <jmkasunich> using it in vertical mode, cut along the line?
[01:58:55] <SWPadnos> 1x1x1/8
[01:59:07] <SWPadnos> yes, I think so
[01:59:25] <SWPadnos> horizontal to cut to length, vertical to do the angles "not quite through"
[01:59:31] <skunkworks> I ment to say.. I was yelled at more than a few times for taking the teeth off the blades by cutting too thin of material for the blade.
[01:59:47] <skunkworks> (when I was younger)
[01:59:51] <SWPadnos> heh
[01:59:51] <skunkworks> of cource
[01:59:53] <jmkasunich> vertical is tedious by its nature - you gotta do the pushing instead of gravity
[01:59:55] <SWPadnos> (20 minutes younger ;) )
[01:59:59] <SWPadnos> yeah
[02:00:11] <SWPadnos> it would have been just about impossible on the Johnson
[02:00:16] <SWPadnos> horizontal only
[02:00:36] <skunkworks> you put one foot on the table - and lean into the part.. just make sure nothing meat related is beind the blade..
[02:00:47] <skunkworks> behind
[02:00:57] <skunkworks> or sit on it..
[02:00:58] <SWPadnos> with the Johnson?
[02:01:09] <SWPadnos> or the little horizontals
[02:01:24] <skunkworks> little cutoff bandsaws
[02:01:27] <SWPadnos> yep
[02:01:59] <SWPadnos> it's bad to put body parts near the Johnson blade
[02:02:20] <jmkasunich> s/Johnson/any
[02:02:23] <SWPadnos> heh
[02:02:25] <jmkasunich> some worse than others of course
[02:02:28] <SWPadnos> yep
[02:02:41] <SWPadnos> it only has a 1HP motor or thereabouts, but it's a real HP
[02:02:48] <skunkworks> don't fall into this one..
http://www.electronicsam.com/images/KandT/DSC02968.JPG
[02:02:50] <SWPadnos> through a 10:1 reducer
[02:03:17] <jmkasunich> skunkworks: yeah, wood saws are much scarier than metal ones
[02:03:23] <skunkworks> yes.
[02:03:26] <SWPadnos> ah yes - logging saws
[02:03:29] <SWPadnos> yeah - big teeth
[02:03:30] <jmkasunich> something about 10x faster blade speed and really big teeth
[02:03:55] <skunkworks> thinking around 55mph blade speed iirc
[02:04:20] <skunkworks> around 1 tooth per inch
[02:04:24] <jmkasunich> zooming in on that pic it seems like about 1 tooth per inch
[02:04:27] <jmkasunich> ;-)
[02:04:36] <skunkworks> heh
[02:04:44] <jmkasunich> 1 finger per tooth
[02:05:04] <skunkworks> it is actually a metal cutting blade - but they work great and are cheap
[02:06:19] <JymmmEMC> skunkworks: what wood is that?
[02:07:01] <jmkasunich> your tax dollars at work: "We just received an RFQ (Request For Quotation) from a jobber here in the U.S. who supplies products to the military, along with the original U.S. Army RFQ that was sent to them. It is for one item. One, very well-known, standard item that has been in manufacture by one company and sold by the thousands to the general public for many years. We all know what it is, we all know how it works, we a
[02:07:01] <jmkasunich> ll know that it has a model number, and it is not terribly expensive
[02:07:01] <jmkasunich> The Army RFQ is 38 pages long-- that's 38 full, 8.5" x 11" pages."
[02:07:35] <JymmmEMC> a hammer?
[02:07:40] <JymmmEMC> or toilet seat?
[02:07:42] <jmkasunich> the posting doesn't say
[02:08:30] <jmkasunich> it does say that in the 38 pages, they give only a vague description of the item, a small photo, and an overall length (which doesn't match the photo)
[02:08:39] <JymmmEMC> bag of lays potato chips!
[02:09:24] <skunkworks> JymmmEMC: maybe oak.. I don't remember.
[02:09:52] <JymmmEMC> skunkworks: ah, ok. just cruious... seemd like it had a somewhat tight grain
[02:10:08] <eric_U> I recently saw something about a woman that went to jail for overcharging for washers
[02:10:08] <skunkworks> I am color blind.. could be cherry also.
[02:10:13] <JymmmEMC> =)
[02:10:28] <eric_U> $700 or so
[02:10:44] <JymmmEMC> skunkworks: you forgot walnut - then we got all the colors covered =)
[02:10:50] <jmkasunich> skunkworks: my non-colorblind eyes say it could be cherry
[02:10:58] <jmkasunich> definitely not walnut
[02:11:23] <jmkasunich> but the crosscut is darker than the rip - could be just the lighting, or could be that the crosscut has aged a bit
[02:11:25] <JymmmEMC> jmkasunich: it was a joke in respect to his "color blind" statement
[02:12:16] <eric_U> the bark should give that log away
[02:12:27] <JymmmEMC> skunkworks: like on the screen and such, are you able to distinguish red or blue?
[02:13:04] <JymmmEMC> skunkworks: MY friend says he has issues with the color green.
[02:13:44] <JymmmEMC> eric_U: thin bark. nfc myself.
[02:13:58] <jmkasunich> eric_U: the bark you can barely see?
[02:14:14] <jmkasunich> I'm thinking cherry
[02:14:14] <eric_U> that's the part that is a give-away
[02:14:25] <eric_U> I just don't know what tree has bark like that
[02:14:30] <jmkasunich> so what is it, o master of the wood?
[02:15:01] <skunkworks> the last few things cut where oak and cherry.. maybe cotton wood - but I would guess cherry
[02:15:51] <JymmmEMC> cherry bark
http://www.wildmanstevebrill.com/JPEG'S/Plant%20Web%20Images/BlackCherryTrunk.jpg
[02:16:06] <skunkworks> we de-bark the wood befor we saw it.
[02:16:09] <skunkworks> log
[02:16:18] <eric_U> that would explain things
[02:16:19] <skunkworks> less blade sharpening that way
[02:16:31] <eric_U> how do you do that?
[02:16:34] <skunkworks> plus good exersize ;)
[02:16:41] <skunkworks> double bit ax
[02:17:36] <eric_U> I only ever got to saw up one tree, walnut
[02:17:59] <eric_U> it was a lot of work, but I got more quartersawn wood from that than I'll probably ever get any other way
[02:19:06] <eric_U> I helped a professor who built his own saw
[02:20:20] <eric_U> the saw got at least one mention on ratemyprofessor.com
[02:23:46] <fenn> "students caught cheating will have to answer to the SAW"
[02:28:35] <eric_U> He only got "1"s because there was no "0" option
[02:28:43] <eric_U> students rate him highly
[02:29:52] <eric_U> "He's an ****. Very disrespectful to students during class. He loves his bandsaw, you'll see what I mean."
[02:39:30] <jmkasunich> alright, which one of you no-goods took my t-nut
[02:39:54] <skunkworks> I wouldn't touch that with a 10 foot cattle prod...
[02:40:18] <jmkasunich> good
[02:40:34] <jmkasunich> whoever took it put it in my shop coat pocket, the bastard
[02:41:55] <eric_U> I hid it just where you left it last, just like I do to my kids
[02:50:07] <jmkasunich> now I'm on to your nefarious plan!
[03:43:32] <cradek> I bet my bridgeport had never made futon parts before tonight
[03:45:26] <JymmmEMC> lol
[04:20:41] <jmkasunich> cradek: probably not, but you never know
[04:21:23] <cradek> it was very successful, but HSS on wood doesn't last very long
[04:21:34] <jmkasunich> ?
[04:21:46] <jmkasunich> I'd expect it to last just about forever
[04:21:46] <cradek> seemed dull when I was done - or maybe it started that way
[04:21:58] <jmkasunich> by woodcutting standards it is dull
[04:22:06] <cradek> if it did, router bits wouldn't be all carbide?
[04:22:14] <jmkasunich> hmm, true
[04:22:14] <cradek> (I don't know anything about cutting wood)
[04:22:39] <gezar> cradek how many rpm were you runing at?
[04:22:44] <jmkasunich> I've read that some woods have microscopic particles of silica or some other abraive substance in them
[04:22:48] <cradek> 3k - about top speed
[04:22:57] <gezar> thats why your hss bit got dull
[04:22:57] <cradek> 3/4"
[04:23:30] <cradek> it's all it's got...
[04:23:59] <gezar> jmk is right, all wood has grit in it, and because you werent able to shear the wood like a router would, your bit stayed in the cut longer, you cut, the would would swell into the cut area, and while doing that it would rub against the trailing edge of the cutter
[04:24:27] <cradek> I guess that's about 600sfm
[04:25:05] <gezar> but at the ultra fine level if you think about it, your cutting edge didnt have time to get away from material as it expanded, wood expands as its cut
[04:25:15] <cradek> what's a good surface speed for cutting wood?
[04:25:27] <gezar> warp drive
[04:25:54] <gezar> 20k or better, think of a 7" circular saw running at what 3600 rpm?
[04:26:09] <cradek> 20ksfm?
[04:26:15] <gezar> 20k rpm
[04:26:25] <gezar> planer tables, let me look
[04:26:27] <cradek> sure but not for 3/4" I bet
[04:26:35] <cradek> router bits are small
[04:26:49] <cradek> 3/4 looks pretty big at 20krpm I bet :-)
[04:27:17] <jmkasunich> cradek: 10" table saw blades do 3000+ rpm
[04:27:25] <cradek> wow ok
[04:27:32] <gezar> 3" dia raised panel cutter, run in a router at 9k rpm, 137fps
[04:28:04] <cradek> ok, warp speed is right
[04:28:09] <cradek> no wonder
[04:28:15] <gezar> wait, thats for a 4inch cutter
[04:28:32] <cradek> but, got my part made with what I had available, which is worth a lot
[04:28:32] <gezar> 1/2 inch in a router, 22krpm
[04:29:33] <gezar> its to get the cutting edge away from the material
[04:29:53] <gezar> almuinum is sorta the same way, with the exception of built up edge
[04:30:12] <gezar> and calculus sucks, at least right now it does, I couldnt answer 3 homework questions
[04:30:16] <cradek> I used an aluminum cutting endmill
[04:30:35] <cradek> well I know more about calculus than cutting wood, so we're even
[04:30:48] <gezar> I only know wood, cause ive worked with it a lot
[04:30:59] <gezar> but I havent done calculus in 17 years
[04:33:25] <gezar> i wont ask for help in here if I can avoid it
[04:33:37] <gezar> maybe ill find an irc math channel
[04:33:51] <cradek> going back to school?
[04:34:19] <gezar> yeah I am, 18 hours to start things off, I spent 3 on math homework today alone
[04:34:38] <cradek> I think 18 is a very heavy class load
[04:34:54] <gezar> I have soo much catching up to do, but im picking on quickly, we will see what my homework grade from today is like tommrow
[04:35:26] <gezar> if I can get up to speed in math quickly enough, ill be fine, if not, I dont know what ill do
[04:35:46] <cradek> I bet it'll come back to you if you've taken it before
[04:36:25] <gezar> but im doing the work, and maintaining, its just funky right now, limits and the likes, im having to relearn so many things, that havent been necessary for years
[04:36:49] <gezar> basic trig and algebra are about all you need for machining
[04:37:04] <gezar> calculus is needed for the motion of an axis
[04:37:48] <gezar> comp science 1 shouldnt be an issue at all, till a few weeks down the road, but im disapointed in the .net only stuff
[04:37:49] <cradek> many machinists don't know trig - that's how rotary tables are sold
[04:38:32] <gezar> heck, dividint head calculations arnt necessarly easy
[04:38:46] <gezar> well, if your wanting to come up with your own set
[04:39:19] <gezar> and with that, im going to hit the sac, ill look at those equations again in the morning and see if something doesnt come to me
[04:39:29] <cradek> me too, goodnight
[04:39:36] <gezar> night
[05:13:20] <tomp> regards the vivaek machines... no luck translating/finding but there seems to be a lot of similar machines
http://www.ad169.com/en/cpzs.asp?page=2
[05:13:55] <tomp> afaik the name is something like da sheng ming mumble mumble
[05:27:10] <tomp> this is the closest
http://www.ad169.com/en/cpzs.asp?FatherClassID=195&ProdClassID=207
[05:32:37] <tomp> closer yet
http://jinshengxing.manufacturer.globalsources.com/si/6008825605234/pdtl/Engraving-machine/1003848121/Machine.htm
[05:34:24] <tomp> spot on
http://jinshengxing.manufacturer.globalsources.com/si/6008825605234/LargeImage/CNC-Engraving-Machine/product_id-1003848126/action-GetProduct.htmhttp://jinshengxing.manufacturer.globalsources.com/si/6008825605234/pdtl/Engraving-machine/1003848121/Machine.htm
[05:36:06] <tomp> key contact Ms Fancy :-) all the chinese girls have western names, but they choose rather odd ones :-)
[05:41:35] <tomp> enuf enuf
http://www.hgcnc.com/ too many names for same things same company same copy
[07:37:12] <micges> hi
[10:04:01] <alex_joni> hi micges
[10:04:07] <alex_joni> got any luck with the translations?
[10:04:15] <alex_joni> if not, you can email them to me, I'll commit them
[10:27:58] <micges> it works but when I sudo make It generate a much larger pot file :)
[10:28:09] <micges> and I went to sleep :D
[10:28:19] <micges> today is the day :)
[10:28:44] <micges> ok I will send to you when ready
[10:35:27] <archivist> alex_joni, how do the G code html pages get built?
[10:40:40] <lerneaen_hydra> alex_joni; where are the untranslated .po files located in the cvs?
[10:42:57] <alex_joni> archivist: which ones?
[10:43:11] <alex_joni> lerneaen_hydra: some are in CVS , some get generated by the build system
[10:43:20] <alex_joni> let me look into it and refresh my memory
[10:43:20] <lerneaen_hydra> oh ok
[10:43:37] <alex_joni> lerneaen_hydra: emc2/src/po
[10:43:41] <alex_joni> there's a README there
[10:43:58] <lerneaen_hydra> ah, right
[10:44:16] <archivist> alex_joni, well the reasom I was looking is that the live cd poits to web hence it needs local html and I was going to look into what was needed
[10:44:38] <lerneaen_hydra> hmm, that didn't explain in any way I could understand where the original files are
[10:44:49] <alex_joni> you mean the g-code quick doc?
[10:44:54] <archivist> yes
[10:45:01] <alex_joni> that's a html file
[10:45:05] <alex_joni> in docs/ somewhere
[10:45:27] <archivist> yes but that links to the web site not local docs
[10:45:28] <alex_joni> archivist:
http://cvs.linuxcnc.org/cvs/emc2/docs/html/gcode.html
[10:45:57] <archivist> so the linked docs need to be generated for the live cd
[10:46:14] <alex_joni> hte Live CD contains the emc2 docs
[10:46:23] <alex_joni> under /usr/share/.. whatever
[10:46:23] <archivist> not everyone has a network to the workshop
[10:46:36] <alex_joni> but that is not a hardcoded path
[10:46:46] <alex_joni> it is set up based on what you selec on ./configure
[10:47:09] <alex_joni> when building the deb's it gets set to a certain value/path
[10:47:32] <alex_joni> so I'm not sure we can "cleanly" change the quick ref to point to a certain location locally
[10:47:45] <alex_joni> archivist: and you get the problems of installed vs. run-in-place
[10:47:47] <alex_joni> etc..
[10:48:08] <alex_joni> I think the best way would be to have a different file for installed systems, and put that on the menu
[10:48:30] <alex_joni> lerneaen_hydra: when you run make the .pot files get generated
[10:48:45] <alex_joni> lerneaen_hydra: axis.pot, rs274ngc.pot and tcl.pot
[10:48:46] <lerneaen_hydra> alex_joni; aha, ok, I see
[10:49:08] <lerneaen_hydra> some automagic spooky action? what are they made from?
[10:49:19] <alex_joni> texts in the source
[10:49:25] <lerneaen_hydra> neat
[10:49:28] <alex_joni> the ones with msgfmt
[10:49:55] <lerneaen_hydra> ok so the compiler reads through the source and grabs text strings with some tag
[10:50:03] <archivist> I was going to write something to do it, hence the need to know how the online (eg .lyx->html) web pages get generated so I can do a local set
[11:45:05] <alex_joni> archivist: the g-code quick ref is a manually written html file
[11:45:09] <alex_joni> it doesn't get generated
[11:45:21] <alex_joni> only the full g-code docs get generated from lyx
[11:45:29] <archivist> I realise that
[11:45:44] <alex_joni> using a lyx2html written by jepler
[11:45:51] <alex_joni> it converts to xml then to html iirc
[11:46:57] <archivist> ok will have a look later
[11:47:24] <archivist> it being work time now :((
[12:41:17] <alex_joni> hi rayh
[12:49:07] <rayh> Hi alex_joni
[12:49:34] <rayh> Did you get a holiday in the mountains this year?
[12:49:45] <alex_joni> yeah
[12:50:10] <alex_joni> a bit shorter than last year though
[12:50:48] <rayh> I was looking at that photo you sent just the other day. Awesome!
[12:51:11] <alex_joni> * alex_joni wonders which one
[12:51:51] <rayh> Long ago. Even before PC left.
[12:52:10] <alex_joni> winter time?
[12:53:20] <rayh> Yes.
[12:53:29] <rayh> I like winter time.
[13:22:21] <jepler> archivist: the gcode documentation is installed when you install the standard ubuntu dapper package, but it is in pdf format so the "quick reference" document can't link inside it.
[13:23:07] <archivist> jepler, Im considering doing a html version for the live cd
[13:23:31] <archivist> pdf is just too slow on the cd
[13:55:11] <jepler> it's easy to build the documentation -- make sure you have the right stuff installed, then pass --enable-build-docs=html to configure
[13:55:57] <jepler> it's an historical accident that the docs in the standard package ended up being pdf instead of html, though a lot of people prefer pdf for reasons I can't personally fathom
[13:56:47] <archivist> Im also contemplating smaller pages to reduce loading time, rather than a pointer to some location on a large page
[13:57:32] <jepler> familiarize yourself with the system that produces the documentation before you make all your grand plans..
[13:57:33] <alex_joni> one of the reasons was because lyx2html was pretty crappy back then
[13:58:01] <archivist> heh I know ive got reading to do!
[13:59:06] <archivist> jepler, yes I hate pdf as well
[14:00:19] <jepler> do you find that gcode_main.html is large enough to make loading times long? It doesn't seem that bad but I have an installed system with lots of RAM..
[14:00:56] <archivist> well loading from linuxcnc.org is slow for me on this box
[14:02:15] <archivist> a ram starved debian dual PII
[14:02:43] <archivist> and pdfs crash it now and again
[14:03:04] <rayh> Hi guys.
[14:04:07] <jepler> hi rayh
[14:04:08] <rayh> I wonder if we can get html pages fine enough grained to work as a help system?
[14:04:53] <rayh> Morning jepler
[14:05:23] <jepler> do you mean, broken up into many many more files? that's not easy with the documentation system we have, it doesn't break up single .lyx input files into multiple .html output files
[14:05:37] <jepler> however, it's easy to link to any section of any file through HTML anchors
[14:06:16] <archivist> could trigger pages from headings
[14:08:00] <jepler> yes, and it is technically possible to produce multiple output files by processing one .xml file with one .xsl file; somebody just has to write it
[14:08:17] <jepler> unfortunately, l2h.xsl turned into a bit of a trainwreck, it's hard to maintain or modify :(
[14:10:22] <jepler> bbl
[14:10:24] <jepler> real work calls
[14:11:38] <rayh> see you guys.
[14:23:30] <skunkworks152> logger_emc: bookmark
[14:23:30] <skunkworks152> Just this once .. here's the log:
http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-01-15.txt
[14:38:19] <tomp> sometimes finding a location is done well by eye ( enhanced by a borescope ) you set a reference with it ,then you cnc using that reference.
http://www.miketreth.mistral.co.uk/centrecam.htm
[14:43:54] <archivist> or a centering dial guage
[14:45:05] <cradek> the camera is neat. I have not found a better method than wiggler + loupe
[14:49:42] <archivist> I wonder how accurate he gets his centering
[14:51:06] <cradek> looks like maybe .1mm is possible (if the lines are actually on center)
[14:51:42] <cradek> I can do maybe 2-4x that well with a wiggler (.001-002 inch)
[14:53:48] <archivist> heh I need a few tenths for small gears (3mm)
[14:54:11] <archivist> usually trial and error after basic setting
[14:54:35] <archivist> as cutter profile affects it as well
[14:55:21] <archivist> "adjust till looks ok" method as standard in clock work
[14:57:52] <tomp> setting up with a mill scope and milling wave guides with .0013" wide detail
http://www.mmsonline.com/articles/040701.html using mill scopes is an old idea , is 'very' accurate
[14:58:27] <tomp> displaying it on a crt is just a spiff
[14:59:27] <archivist> heh that one is in a different league, proper toy
[14:59:51] <tomp> heh, the mill was made by a diamond lens turning co
[15:20:30] <tomp> steampunk edm/ecm using emc2 (not done yet, but a guy that does get things done)
http://steampunkworkshop.com/cnc-ecm.shtml
[15:21:58] <alex_joni> bbl
[15:30:25] <JymmmEMC> Heh... Got Vista?
http://schneier.com/essay-198.html http://technet2.microsoft.com/WindowsVista/en/library/417467e7-7845-46d4-85f1-dd471fbc0de91033.mspx?mfr=true
[16:55:58] <skunkworks152> so - I said to myself, self - I bet I have some non williamette pentium 4 processors in motherboards that will accept williamette ones. (I was right)
[16:57:40] <SWPadnos> and those users will never notice the speed difference :)
[16:58:28] <skunkworks152> nope.. I actually swapped a 2ghz northwood celeron for a 1.8ghz pentium 4 williamette. I don't think they will notice.
[16:58:51] <SWPadnos> yeah - whatever those names mean ;)
[16:59:08] <skunkworks152> you would have to ask intel ;)
[16:59:18] <SWPadnos> I just figured if some motherboard didn't support something, the something must be too new :)
[17:04:58] <skunkworks152> The new 478 socket pentium 4 motherboards I bought don't support the pentium 4 processors I had :)
[17:05:30] <skunkworks152> williamette was the first version of the pentium 4
[17:06:01] <SWPadnos> oh - see how logic sometimes gives you (me) the opposite answer?
[17:06:16] <skunkworks152> yes
[17:06:22] <SWPadnos> now if only I could remember why I'm not using this Tyan motherboard
[17:12:52] <skunkworks152> I don't think I have ever used a tyan motherboard.
[18:35:14] <skunkworks152> odd - dapper installed from livecd.. Login and then you get a blank screen with a mouse.
[18:35:27] <skunkworks152> odd'
[18:35:44] <archivist> my install crashes on exit
[18:35:58] <skunkworks152> mouse moves but you can't switch to any of the terminals
[18:37:02] <skunkworks152> archivist: crashes how?
[18:37:04] <SWPadnos> even with ctrl-alt-<terminal #>
[18:37:30] <skunkworks152> oops
[18:37:33] <skunkworks152> that works
[18:37:34] <skunkworks152> heh
[18:37:48] <skunkworks152> for some reason I thought it was just alt.
[18:37:52] <archivist> skunkworks, during the "tone" on shutdown it stops and the tone/musac repeats
[18:38:13] <skunkworks152> if you exit emc first - does it crash?
[18:38:34] <SWPadnos> you need to use ctrl-alt when in X, alt-arrows or alt-num works from a text console
[18:39:14] <skunkworks152> yes - I can get to the terminal
[18:39:25] <skunkworks152> how do you stop x from terminal?
[18:39:33] <SWPadnos> sudo /etc/init.d/gdm stop
[18:39:34] <skunkworks152> stopx doesn't work ;)
[20:03:31] <maddash> is it feasible to build a true rms voltmeter using AVR-based processing and sampling? like some atmega[large number here]?
[20:03:55] <SWPadnos> for low accuracy and low speeds, yes
[20:04:29] <fenn> the built-in adc goes to something like 15kHz
[20:04:50] <fenn> though i'd check the datasheet to be sure it can slew fast enough for whataver you're doing
[20:05:20] <maddash> hm, I waaas thinking along the lines of 24-36 samples per 60Hz AC cycle
[20:05:52] <SWPadnos> the conversion time is roughly 65 us on most of the AVRs, if you want the full "10-bit" resolution
[20:05:54] <maddash> you see, the rabbit I have uses an ADS7870, which claims to have a max rate of 52kHz
[20:06:00] <fenn> you can go faster with an external adc
[20:06:23] <maddash> the bottleneck seems to be the ucontroller, which is why I'm asking
[20:06:58] <fenn> well a moving average would seem easy enough to implmement quickly
[20:07:11] <fenn> add add add bitshift
[20:07:19] <maddash> fyi, the rabbit can't go beyond 3 channels on a 24*60=1440 Hz rate
[20:07:24] <maddash> er, bitshift?
[20:07:38] <maddash> oh, /2
[20:07:45] <SWPadnos> yes
[20:07:55] <SWPadnos> do a power of 2 number of samples per cycle
[20:08:02] <maddash> still, you need to multiply by the time interval between samples
[20:08:07] <SWPadnos> but you don't want averaging, you want RMS, which is different
[20:08:14] <SWPadnos> averaging should give you zero
[20:08:17] <SWPadnos> roughly
[20:08:26] <maddash> not a moving avg...
[20:08:28] <fenn> well, that's why you 'rectify' the signal first
[20:08:48] <fenn> i've never seen a good explanation how to do 'true rms' on an arbitrary waveform
[20:08:50] <maddash> SWPadnos: I think the moving avg is only zero at certain points
[20:09:09] <SWPadnos> no, for a sinewave averaged over a full cycle, it's zero for any one-cycle window
[20:09:16] <maddash> fenn: take your samples, and apply the discrete definition of RMS on the values
[20:09:39] <fenn> what's "discrete definition of RMS"?
[20:09:56] <maddash> SWPadnos: true, true. I was referring to avg over < full cycle
[20:10:09] <SWPadnos> you need a full cycle to get RMS
[20:10:27] <maddash> fenn:
http://en.wikipedia.org/wiki/Root_mean_square
[20:10:31] <SWPadnos> next problem: the ADC on the AVR is not bipolar
[20:10:36] <maddash> the definition for x_rms
[20:11:35] <maddash> well, that's not a problem, because I'll be using some simple intermediate circuitry to scale down the +/-120*sqrt(2)VAC to a range from 0 to 20V
[20:12:01] <SWPadnos> and then to 0-5V, like the AVR needs
[20:12:09] <SWPadnos> but make it 0.5-4.5V, so it really fits
[20:12:29] <fenn> the adc response is slightly nonlinear at the edges
[20:13:02] <SWPadnos> and it's a shitty ADC anyway - the 10 bits is really 7-8 bits, if you're lucky
[20:13:08] <fenn> wikipedia is terrible at math explanations
[20:13:31] <maddash> well, then I'll be using an ext. ADC. possibly ADS7870 again
[20:13:36] <SWPadnos> that's true for most microcontroller and DSP ADCs I've seen actually, not just the AVR
[20:13:57] <maddash> SWPadnos: you mean the non-linearity?
[20:14:03] <fenn> i'm happy with it for most stuff
[20:14:07] <SWPadnos> no, the non-resolution
[20:14:31] <fenn> as long as you arent trying to build a video digitizer out of it :P
[20:14:32] <SWPadnos> they say 10 bits, but if you look carefully, you'll see how that goes down to 7-8 bits in actual use
[20:14:56] <maddash> is 12-bits (on the ADS7870; 1 bit for sign) high enough?
[20:15:02] <fenn> SWPadnos: obligatory mumble about oversampling
[20:15:13] <SWPadnos> heh
[20:15:35] <SWPadnos> that's just enough, if you compare e.g. to a 4000-count radio shack multimeter
[20:16:38] <fenn> maddash: depends on your application now doesn't it?
[20:16:57] <maddash> er, measuring 120 VAC? heh.
[20:17:08] <maddash> well, <=120VAC
[20:17:09] <fenn> why are you measuring 120VAC with an AVR?
[20:17:59] <SWPadnos> because a multimeter (with a serial port) is too hard
[20:18:46] <maddash> well, I plan on doing other things as well
[20:19:14] <maddash> I guess the main point here is to figure out whether I can get a microcontroller to play nice with an ADC @ >=1.44kHz
[20:19:20] <SWPadnos> do you need true RMS for arbitrary waveforms, or is 60-Hz a reasonable assumption
[20:19:43] <SWPadnos> sure, you can do 15 KHz with an AVR, twice that with one of the ones with dual ADC units
[20:19:52] <SWPadnos> but it's not so great
[20:20:00] <SWPadnos> probably good enough though
[20:21:22] <rayh> "AVR is not bipolar" but some of us are!
[20:21:28] <SWPadnos> I resent that
[20:21:31] <SWPadnos> I love that
[20:21:48] <maddash> 15kHz using, "main() { while (1) { sample(); }}", you mean?
[20:22:07] <maddash> heh, the sampling's gotta be uniform,
[20:22:07] <fenn> you can do other stuff while you sample, the adc isn't part of the CPU
[20:22:27] <archivist> use an interrupt
[20:22:45] <SWPadnos> if you use C for the sampling, you're - how should I put this - an idiot :)
[20:22:54] <archivist> hehe
[20:23:10] <fenn> i resemblove that statement!
[20:23:11] <SWPadnos> you need to set the ADC in free-run mode, and read the value (int a buffer maybe) in the ADC complete interrupt
[20:23:28] <SWPadnos> that aslo is the only way to get the higher sample rates
[20:23:29] <SWPadnos> also
[20:23:39] <fenn> er, how is C bad for that?
[20:24:00] <archivist> assembler fer speeeeeeed
[20:24:12] <fenn> you can do interrupts in C, and surely reading an int is something the compiler knows how to do
[20:24:13] <SWPadnos> it may not be all bad, if it's reasonably easy to do interrupt functions
[20:24:40] <archivist> depend on the C you use
[20:24:46] <SWPadnos> but what I've seen in C compilers isn't all that efficient - for example saving/restoring all registers, using external pointers for everything...
[20:25:21] <SWPadnos> the AVR compilers I've used (keil, IAR, a little with gcc) aren't too savvy about the various memory spaces in the AVR
[20:25:36] <fenn> well, that's all there is really
[20:25:54] <SWPadnos> I think tasking may have one too, but still
[20:26:08] <maddash> I'm fine with interrupts. In fact, on the rabbit testbed, I hooked up a timer to knock off the interrupts for me. except the rabbit has tied down parts of the ADC so that the only way to get a sample from it is through a serial port.
[20:26:22] <maddash> so things get slow as hell
[20:26:28] <SWPadnos> fora simple function that reads a register pair and sticks it in memory somewhere, assembly is much more efficient
[20:26:33] <SWPadnos> wow, that's really stupid
[20:26:33] <maddash> SWPadnos: what's this "freee mode"?
[20:26:44] <fenn> samples continuously
[20:26:45] <SWPadnos> free run mode - check the datasheet
[20:26:48] <SWPadnos> right
[20:27:30] <maddash> hm, but a) will the ADC interrupts be in sync with each sampling cycle? and b) is this ADC-specific?
[20:27:46] <SWPadnos> the ADC is what generates the interrupt
[20:27:48] <fenn> otherwise you'd have to tell the ADC to sample each time, and it wouldn't start sampling until the CPU got around to it
[20:28:00] <archivist> someone has not read the data sheet
[20:28:05] <maddash> fenn: which exactly what is happening right now
[20:28:06] <fenn> so this way the cpu can read the (latched) buffer at its leisure
[20:28:20] <SWPadnos> oh - do you want to sample at some part of the AC cycle?
[20:28:43] <maddash> SWPadnos: yes...twenty-four times each 1/60 sec
[20:28:53] <SWPadnos> ok, so you're assuming a 60-Hz waveform
[20:28:54] <fenn> the adc interrupts at the end of the sampling cycle
[20:28:58] <archivist> timer kicks adc then
[20:29:00] <maddash> hm, this makes sense
[20:29:14] <SWPadnos> yes, you can tie the scaled A/D to the analog comparator, which can start a sampling cycle
[20:29:19] <archivist> zero crossing starts timer
[20:29:20] <SWPadnos> (even if it's done in software)
[20:29:49] <archivist> or run async/near enough
[20:29:54] <SWPadnos> indeed - I think one or two people haven't read the datasheet ;)
[20:30:17] <maddash> * maddash checks ads7870 (
http://measure.igpp.ucla.edu/hemisphere/Manuals/ads7870.pdf) for free mode
[20:30:39] <maddash> maybe i didn't read carefully enough
[20:30:43] <archivist> * archivist hasnt but I do hardware assembler in PIC chips
[20:32:23] <maddash> is there a synonym to "free mode" besides "running continuously"?
[20:32:50] <SWPadnos> free-run is pretty common, continuous is another word to look for
[20:33:08] <maddash> whoa, cool 'continuous mode'
[20:33:18] <maddash> !
[20:34:16] <maddash> crap, continuous mode only works on one channel at a time
[20:34:26] <fenn> that's because there's only one ADC
[20:34:51] <maddash> er, so is this a standard constraint among ADC's?
[20:34:51] <fenn> when you switch channels it takes a while for the mux/amp to settle
[20:35:17] <fenn> it's because there is only one ADC and it has a mux on front..
[20:36:30] <maddash> god, just lookking at the spec makes me angry
[20:37:07] <archivist> there are allways trade offs in low cost chips
[20:38:09] <SWPadnos> or to rephrase that, there are alwas tradeoffs
[20:38:13] <SWPadnos> always
[20:40:33] <maddash> fuck. just fuck. fucking rabbit.
[20:41:09] <fenn> that's what rabbits do
[20:42:30] <maddash> http://www.rabbit.com/documentation/docs/manuals/RCM3400/5soft.htm#977377
[20:42:55] <maddash> "CONVERT - Input - Pulled down"
[20:43:06] <SWPadnos> "This function will address the A/D converter in Register Mode only, and will return an error if you try the Direct Mode"
[20:43:25] <SWPadnos> at least they tell you it'll be useless
[20:44:06] <archivist> just as bad with pic chips, try this try that.... wtf....
[20:44:14] <maddash> no, no, that just applies to anaInConfig(); the actual gateway call to read from the ADC is anaInDriver(), which ALWAYS uses dIRECT MODE
[20:44:17] <SWPadnos> and even better, I bet those functions are serializing, so the CPU gets to busy-wait while the data is transferred (could be bit-banged anyway)
[20:44:30] <maddash> exactly
[20:44:42] <BigJohnT> what do you guys think about the cubloc?
[20:44:53] <SWPadnos> well, there's a reason I'm designing an ARM-based rabbit replacement
[20:46:18] <maddash> there are at least three, "int i=10; while (i>0) { i++; checkReady(); }" style loops in anaInDriver(), which is the "low-level" ADC function
[20:46:24] <maddash> i--*
[20:46:40] <archivist> * archivist wonders if ARM has improved since I last worked with them,
[20:46:59] <SWPadnos> how long ago was it?
[20:47:24] <archivist> back when they chips were in Acorn computers
[20:47:29] <SWPadnos> heh
[20:47:30] <archivist> the
[20:47:38] <SWPadnos> yes, I'd say there's been some improvement
[20:47:42] <archivist> only had short branches
[20:48:10] <SWPadnos> they still have short ranches, but there's a "high half setup" instruction, so a 2-word instruction can do a long jump anywhere
[20:48:13] <SWPadnos> branches
[20:48:15] <maddash> do I have any other choices besides getting a fast ucontroller and wiring in an ADC?
[20:48:19] <archivist> and Acorns C screwed up on the switch statement because of it
[20:48:40] <SWPadnos> and "short" means within 64k, I think
[20:48:50] <SWPadnos> but maybe I'm thinking of PowerPC
[20:48:51] <archivist> heh thats long!!
[20:49:10] <SWPadnos> well, the $10 ARM microcontrollers are 32-bit these days
[20:49:37] <fenn> maddash: you could get a slow ucontroller...
[20:49:49] <archivist> I wrote a printer driver for Riscos
[20:50:14] <fenn> maddash: or with sufficient parport hackery you could use the ADC directly from a PC
[20:50:26] <fenn> i never managed to get that to work thoguh
[20:51:25] <skunkworks152> in the robot builders banaza book - they show you how to directly interface to the isa slot.. :)
[20:51:36] <jepler> hm, on an ebay auction (!?) I find that there is something called 'rabbit - A development system for the Rabbit2000/3000 microcontrollers on Linux. A cross C-compiler, assembler, and linker, base on SDCC. For all Linux/BSD/UNIX type systems. Available persuant to the GNU General Public License.'
[20:52:06] <archivist> someone selling gcc?
[20:52:26] <jepler> you can do that and follow the GPL at the same time (but this item refers to sdcc, which is not gcc)
[20:52:45] <maddash> like hell
[20:52:55] <maddash> rabbits don't work with linux at all
[20:53:01] <maddash> got a link?
[20:53:34] <jepler> maddash: this is the ebay auction that refers to the software .. all I know about it is the line I pasted above. no relation to the seller, don't endorse the product.
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=270202703199
[20:54:08] <maddash> I'll keep that in mind when I feel the urge to slander the seller.
[20:54:43] <maddash> 'bakefile'. real nice.
[20:56:17] <jepler> there's nothing wrong with selling CDs of GPL software. but when it's something that would be of huge interest to a fair number of nerdy types, it's surprising that I can't find any other information about this supposed compiler on the internet
[20:57:19] <archivist> http://sdcc.sourceforge.net/
[20:58:07] <archivist> heh AVR and gbz80 ports are no longer maintained.
[21:00:43] <maddash> how can a ramp generator aid in analog voltage measurement?
[21:01:04] <SWPadnos> uh - where did that come from?
[21:01:14] <maddash> "On-board analog features include a ramp generator that, coupled with inexpensive comparators and the pulse capture capabilities of the Rabbit 3000 microprocessor, allows measurement of analog voltages for applications such as temperature measurement using the on-board thermistor."
[21:01:16] <maddash> http://www.rabbit.com/products/PowerCore/
[21:01:23] <SWPadnos> heh
[21:01:34] <SWPadnos> that's a cheap, inaccurate, poor-man's ADC
[21:01:41] <jepler> maddash: one (poor!) way to measure an analog voltage is with one analog comparator and a variable voltage source
[21:01:44] <SWPadnos> you use an analog comparator and time how long it takes to trip
[21:01:52] <jepler> when the comparator trips, you know approximately what the voltage was
[21:02:37] <maddash> [sigh]
[21:02:39] <SWPadnos> that's good for systems where you need at most a few tens of updates per second, and where the real resolution can be down in the 3-5 bit range
[21:03:08] <SWPadnos> like using several switches on a single analog input instead of a key matrix
[21:04:21] <maddash> hm, so i'll definitely be doing some soldering with that board
[21:04:36] <maddash> (attaching an external adc)
[21:05:07] <maddash> hey wait a sec -- can't I use a heat sensor to calculate RMS?
[21:05:31] <SWPadnos> how does the heat information get into the microcontroller?
[21:05:36] <SWPadnos> (hint - you use an A/D)
[21:06:20] <archivist> or a d/a and a comparator (maintain thermister temp)
[21:06:44] <archivist> used in HP microwave heads
[21:07:15] <SWPadnos> sure - that's for a control system, not a sensor
[21:07:15] <archivist> so a thermal balance
[21:07:26] <archivist> yes a sensor
[21:07:58] <SWPadnos> that only works when the feedback is well-calibrated
[21:08:09] <archivist> well yes
[21:08:28] <SWPadnos> and also the D/A
[21:09:25] <archivist> * archivist trying to remember the whole topology , but it dates back a long time to the analog meters
[21:10:15] <SWPadnos> I think I see how it would work, but then you're depending on the accuracy of the feedback and the D/A, and it would be better for slow-response systems
[21:10:33] <SWPadnos> (maybe a 1-10 Hz loop)
[21:10:39] <archivist> also ised in the Fluke rmsmeter 931
[21:10:49] <archivist> no d/a in them
[21:11:21] <SWPadnos> there has to be a variable output reference
[21:11:27] <SWPadnos> or an A/D
[21:11:53] <archivist> fluke did have a d/a as it drives the constant current iirc
[21:12:12] <archivist> loooooong time since I read the manual
[21:13:49] <Gamma-X> my shipment should be here from mesa.
[21:13:59] <archivist> effectivly a bridge ac driving input thermister and d/a driving the other
[21:14:06] <Gamma-X> and i also found a computer that has a p3 730mhz processor in it with 256 ram
[21:14:24] <maddash> 'found'?
[21:14:41] <Gamma-X> yeah.. lol
[21:14:46] <archivist> we luv "found"
[21:14:47] <Gamma-X> at work believe it or not
[21:15:10] <Gamma-X> it was bein trashed. an old dell. it runs xp fine.
[21:15:47] <maddash> * maddash has never found a working PC
[21:16:13] <archivist> * archivist "found" a rotary table in the basement today. need to convince the boss to give it to me
[21:16:51] <Gamma-X> lol
[21:17:21] <archivist> he claims he paid £60 30+ years ago
[21:17:38] <Gamma-X> once i have the machine up and running, and the input in the mesa cards, how would I configure emc to use the linear scales?
[21:17:43] <maddash> ugh, the avr library for ads7870 also uses a serial protocol
[21:17:47] <maddash> http://www.mil.ufl.edu/~chrisarnold/components/microcontrollerBoard/AVR/avrlib/docs/html/ads7870_8c-source.html
[21:18:15] <fenn> maddash: how else are you gonna get data out of it? parallel 12-bit data lines?
[21:18:46] <SWPadnos> yeah - in case you didn't notice, it's a serial-interface A/D
[21:18:50] <maddash> sure, why not? it'd be faster, and simpler too
[21:18:58] <fenn> serial is fast enough for most sample rates, and it's automatic in most micro's
[21:19:00] <SWPadnos> then get a parallel-interface A/D
[21:19:15] <maddash> these angels exist?
[21:19:18] <SWPadnos> of course
[21:19:35] <fenn> "serial" doesnt necessarily mean "slow" - sata for example
[21:19:48] <SWPadnos> search digikey, analog devices, burr-brown, linear technologies, microchip ...
[21:20:25] <SWPadnos> indeed - the 6-channel, 16-bit ADC I have is serial, but I still get 175k samples per second from each channel (over 1MSa/sec total)
[21:20:31] <fenn> i even foudn parallel adc's at radio shack
[21:20:42] <SWPadnos> and it's bipolar, +/-10V input range
[21:20:49] <Gamma-X> any ideas on the linear scales?
[21:21:07] <SWPadnos> simultaneous sampling on any of 3 pairs, or all 6 channels if you want
[21:21:08] <fenn> Gamma-X: you hire a programmer to add the necessary code to emc
[21:21:17] <Gamma-X> ! lol
[21:21:33] <SWPadnos> Gamma-X, I'm not sure how much you expect to have changed since the last time you asked the question ... :)
[21:21:51] <maddash> SWPadnos: have you actually tested out the 175khz?
[21:21:56] <SWPadnos> especially considering that none of the programmers have linear scales at the moment
[21:22:15] <SWPadnos> maddash, yes, it can go faster, but that's the limit of the FPGA with the long cable between
[21:22:25] <maddash> ah, fpga*
[21:22:43] <SWPadnos> well, you can talk to it with an SPI port on a microcontroller too
[21:22:52] <Gamma-X> SWPadnos i honestly dont remember askin lol ive been sick for a while. dont remember things itss wierd.
[21:22:58] <maddash> but it'd be a significantly slower
[21:23:10] <SWPadnos> at ~20 Mz, it would still give you over 1Msa/sec at 16 bits
[21:23:12] <SWPadnos> MHz
[21:23:26] <SWPadnos> you use the actual hardware SPI port on an AVR or something, and it'll be plenty fast
[21:23:31] <fenn> * fenn thinks Hz should be shortened to z
[21:23:42] <SWPadnos> no, that's impedance ;)
[21:23:49] <fenn> fff impedance is Z
[21:23:56] <maddash> er, impedance is Ohms
[21:24:02] <SWPadnos> no, resistance is in Ohms
[21:24:08] <maddash> ditto impedance
[21:24:15] <fenn> anyway, "impedance" isnt a unit
[21:24:18] <SWPadnos> impedance includes inductive or capacitive reactance
[21:24:22] <fenn> it's a unit type
[21:24:29] <maddash> yes, yes, but it amounts to the same unit
[21:24:46] <SWPadnos> well, that depends on whether you're doing AC analysis or DC analysis
[21:25:04] <maddash> so, you'd say, "6 ohms of impedance," and not, "6 Z'z of impedance"
[21:25:20] <maddash> the latter would make you sound sleepy
[21:25:21] <Gamma-X> http://cgi.ebay.com/Galil-Motion-Control-DMC-1040-ISA-4-Axis-Servo-Encoder_W0QQitemZ300190497783QQihZ020QQcategoryZ71395QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[21:25:26] <SWPadnos> heh
[21:25:27] <Gamma-X> good deal?
[21:25:29] <SWPadnos> ok, ;oint taken
[21:25:31] <SWPadnos> point
[21:25:32] <fenn> resistance reactance inductance
[21:25:43] <maddash> SWPadnos: :P
[21:26:06] <SWPadnos> Z is the symbol for impedance, how's that? :P
[21:26:15] <fenn> Gamma-X: but what's it good for?
[21:26:24] <SWPadnos> you can use it with Mach, maybe
[21:26:28] <Gamma-X> ewww
[21:26:32] <maddash> but can it run linux?
[21:26:33] <SWPadnos> if you have an ISA slot
[21:26:33] <Gamma-X> nasty. never mind then lol
[21:26:58] <fenn> well you dont need a fast computer
[21:27:17] <Gamma-X> i wonder if i have to just buy encoders for my servos.
[21:27:39] <SWPadnos> you can certainly try without them, but I think everyone has pointed out that tuning may be difficult
[21:27:44] <fenn> you could do a cost/benefit analysis compared to hiring a programmer :P
[21:28:01] <fenn> but encoders are a sure bet
[21:28:18] <SWPadnos> if the machine has no backlash, then the linear encoders should work nicely
[21:29:11] <Gamma-X> SWPadnos dont know if it does...
[21:29:16] <Gamma-X> i guess ill find out wont I?
[21:29:25] <SWPadnos> probably
[21:30:25] <Gamma-X> can anyone recomend a site for encoders?
[21:30:32] <fenn> i think you'd still want to use information from the tacho's for velocity
[21:30:53] <fenn> since linear scales will have a lot of quantization error
[21:30:57] <SWPadnos> usdigital.com is popular, also renco.com
[21:31:10] <Gamma-X> ok thanks
[21:31:19] <SWPadnos> if you want expensive industrial encoders, beiied.com
[21:31:27] <Gamma-X> nope inexpensive lol
[21:31:55] <SWPadnos> expensive/industrial = sealed, good bearings, high quality connectors, metal housings, etc.
[21:32:04] <SWPadnos> but it also means 10x the price
[21:32:39] <fenn> (YAGNI)
[21:32:58] <Gamma-X> im assumin rotary hollow bore encoders?
[21:33:06] <SWPadnos> depends
[21:33:21] <SWPadnos> you'll probably need to remove the tachs and put encoders in their place
[21:33:35] <SWPadnos> that means you'll have to match whatver the tach is, which may be a shaft with a coupler
[21:34:50] <cradek> surely the amps need the tach signals as-is
[21:34:57] <SWPadnos> probably true
[21:35:34] <SWPadnos> I'm betting the motor enclosure doesn't have extra room for an encoder though, which could be a problem
[21:35:47] <skunkworks152> biostar motherboard 2ghz pentium 4 - rock solid jitter at 10079
[21:35:59] <skunkworks152> rock solid?
[21:36:01] <skunkworks152> ;)
[21:36:02] <SWPadnos> single-core though
[21:36:04] <SWPadnos> ?
[21:36:05] <skunkworks152> ye
[21:36:06] <skunkworks152> yes
[21:36:18] <SWPadnos> hmmm. then a do-nothing app may not help :)
[21:36:22] <skunkworks152> heh
[21:36:23] <cradek> the encoder could go on the screw, on the far end, or using a toothed pulley and belt
[21:36:27] <SWPadnos> but you could try it
[21:36:44] <skunkworks152> where is it?
[21:36:47] <skunkworks152> and how?
[21:37:06] <SWPadnos> skunkworks152, in another terminal, try while true ; do echo "nothing" > /dev/null ; done
[21:37:20] <Gamma-X> well my servos had an option of bec755 and a sumtak lht model
[21:37:32] <SWPadnos> whatever that means
[21:43:24] <skunkworks152> SWPadnos: will closing the terminal stop the do nothing process?
[21:43:55] <SWPadnos> should, or you can hit ctrl-c in that terminal
[21:44:02] <SWPadnos> or top then kill
[21:48:43] <skunkworks152> just goofing around.. It may help just a bit - 10079 vs 8924 (went back and forth a few times now)
[21:49:58] <skunkworks152> it goes to above 10k quickly - while with the do-nothing running it starts at 4K and works up to almost 9k
[21:50:54] <SWPadnos> interesting
[21:51:19] <SWPadnos> is this latency-test or the RT kernel latency test?
[21:55:16] <skunkworks152> latency-test
[21:55:48] <SWPadnos> ok
[21:55:55] <skunkworks152> jitter
[21:56:53] <skunkworks152> oops - 250000
[21:57:02] <skunkworks152> something happens after a few minutes
[21:57:02] <SWPadnos> ouch
[21:57:19] <skunkworks152> (this isn't probably going to run anything.. just playing)
[21:57:32] <skunkworks152> it has a on-board video - s3
[21:59:19] <skunkworks152> heh - max shared memory.. 256mb
[21:59:37] <skunkworks152> that could be it.
[22:11:36] <maddash> 1440Hz = .6944 ms or less per interrupt
[22:13:08] <maddash> 20000 consecutive samples takes 1687 ms = .08435 ms per sampling
[22:16:14] <maddash> which means that 1440Hz is way too fast for sampling 8 channels
[22:23:31] <alex_joni> good night all
[22:23:46] <alex_joni> maddash: gotta say your level of tinkerness surely has improved in the last year :P
[22:30:21] <fenn> 215 xp required to advance to next tinker level
[22:49:56] <Gamma-X> my mesa cards came in but theres no termination blocks on them
[22:54:00] <Gamma-X> where do i get those terminal blocks?
[22:54:54] <BigJohnT> doesn't the 5i20 use ribbon cables?
[22:56:00] <Gamma-X> yes but the 7133t and 7137 need termnial blocks
[22:56:16] <BigJohnT> ok didn't know
[22:56:29] <BigJohnT> call mesa?
[22:57:26] <BigJohnT> don't see any terminal blocks on those cards looking at the pictures...
[22:58:10] <BigJohnT> the 7i33T is the one with terminal blocks
[22:58:20] <BigJohnT> not the 7i33
[23:02:13] <Gamma-X> turns out u need to spend an extra 30 dollars per card to have termnial blocks from mesa.... nice right lol
[23:04:08] <skunkworks> umm - I am sure SWPadnos told you that..
[23:05:14] <Gamma-X> im clueless as usual
[23:05:45] <Gamma-X> hahaha nice. the lady from mesa sent me the item number from digikey where they order them from
[23:05:50] <Gamma-X> 7 bucks each on digikey
[23:07:04] <fenn> still no warning about that from mesa?
[23:07:18] <Gamma-X> ?
[23:07:23] <Gamma-X> come again
[23:08:03] <fenn> Two TB24 pluggable terminal blocks are required for use with the 7I33T (not included).
[23:08:18] <Gamma-X> yeah i saw that.
[23:08:29] <Gamma-X> this is gunna sound stupid...
[23:08:48] <fenn> it's ok, we already have a low opinion of you :)
[23:08:49] <Gamma-X> i thought they were reffering to mounting that thing on brass standoffs! lol
[23:09:08] <Gamma-X> ehhh... what u gunna do.
[23:09:42] <fenn> i'm gonna sit here on my couch until parts arrive for my car
[23:10:17] <skunkworks> I had the local tech school overhaul the engine for our subaru.. Now I need to find the time to put it back in..
[23:10:22] <fenn> i tell ya, internet mail order is so nice for doing car stuff
[23:10:34] <skunkworks> Yes
[23:10:35] <fenn> you can actually shop around and get what you want
[23:10:46] <Gamma-X> seems the lady sent me the wrong item number... she sent me the link to the connector soldered into the board already
[23:11:03] <fenn> and unlike most weird industrial-ish stuff i do, there's hundreds of businesses that actually want my business
[23:11:29] <fenn> Gamma-X: you have to solder them on
[23:11:46] <skunkworks> really?
[23:11:59] <fenn> it's a two-piece connector, part of it solders in, the other part plugs into that part
[23:12:03] <skunkworks> ah
[23:12:04] <skunkworks> ok
[23:12:24] <fenn> i think it's really weird that it doesnt come with it soldered on already
[23:12:34] <Gamma-X> fenn they already are... im just looking for the terminal blocks that connect to that thing.
[23:12:39] <fenn> because without those connectors it's basically useless
[23:12:40] <Gamma-X> it does...
[23:13:07] <fenn> oh well, it's still useless then, but slightly less inconvenient
[23:13:29] <skunkworks> heh
[23:13:46] <maddash> alex_joni: "tinkerness"? how?
[23:14:29] <maddash> oh my god. I've been here for over a year already.
[23:14:39] <fenn> scary isn't it
[23:14:49] <fenn> at least you have a running CNC..
[23:14:55] <maddash> you don't?
[23:15:02] <fenn> no, not currently
[23:15:07] <maddash> don't you have some ecstasy?
[23:15:10] <maddash> six-axis thing
[23:15:15] <fenn> hextatic
[23:15:19] <fenn> it's not built yet
[23:15:29] <fenn> mainly due to my laziness
[23:15:43] <fenn> or ... something (CIA mind control waves)
[23:15:59] <maddash> life's too short
[23:16:06] <fenn> too short for what?
[23:16:25] <maddash> doing all the things I want the way I want
[23:16:38] <fenn> yeah
[23:17:13] <maddash> the problem with me is that I like to "tinker" and simmer with my creations
[23:20:29] <fenn> i think what he meant by 'tinkerness' is your competence in electronics/software
[23:32:41] <SWPadnos> Gamma-X, there are multiple types of pluggable terminal block
[23:32:53] <SWPadnos> one of them is DK part number 281-1762-ND
[23:33:07] <SWPadnos> I think that's the vertical orientation
[23:33:43] <SWPadnos> there's also a horizontal otrientation header that will work - you may want one of each because the headers are very close together
[23:34:00] <SWPadnos> (so it's hard to route wires from the "back" connector around the "front" connector)
[23:34:24] <SWPadnos> if you run the wires on top of the front connector, leave extra room for when you want to pull that strip out
[23:36:01] <Gamma-X> SWPadnos im just lookin for the things that plug into them.
[23:36:09] <Gamma-X> im on the phone with digikey to see about pricing.
[23:36:54] <fenn> if it says "call" it's going to be expensive
[23:36:58] <Gamma-X> http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=281-1062-ND
[23:37:00] <Gamma-X> that one
[23:37:21] <Gamma-X> there trying to get the price lower apparently...
[23:39:42] <SWPadnos> yes, the part number I gave you is the pluggable strip that plugs into the header
[23:40:00] <SWPadnos> oops - no it isn't :)
[23:40:12] <SWPadnos> it's the right angle header (I guess I should have looked inside the box :) )
[23:40:34] <SWPadnos> the vertical plug strip is 281-1026-ND
[23:41:00] <SWPadnos> (that stands upright in the header, with the screw terminals on top and the wires exiting to the right)
[23:42:23] <eric_U> call means digikey doesn't have it
[23:43:01] <SWPadnos> or its EOL and they don't want to list it any more
[23:43:28] <eric_U> I interpret "call" as "you aint gonna get it from us, bonfuto"
[23:43:40] <SWPadnos> I think the right angle terminal strips are 281-1062-ND
[23:44:05] <SWPadnos> yep
[23:44:10] <SWPadnos> http://rocky.digikey.com/WebLib/Weidmuller/Web%20Photo/1620440000.jpg
[23:44:26] <eric_U> what are the two missing pins?
[23:44:35] <SWPadnos> ?
[23:44:48] <SWPadnos> on the 7i37-T?
[23:44:52] <eric_U> I have 50 pins, that thing has 24
[23:45:03] <eric_U> 2*24 = 48
[23:45:14] <SWPadnos> they're either no connects or power/ground on the 7i37-non-T
[23:45:25] <eric_U> probably grounds
[23:45:28] <Gamma-X> damn
[23:45:35] <Gamma-X> 53 dollars on terminal blocks...
[23:45:49] <eric_U> didn't we tell Gamma to order them from Mesa?
[23:46:00] <Gamma-X> there 15 from mesa
[23:46:05] <Gamma-X> so i actualy got them cheaper...
[23:46:16] <Roguish> try to find any other 'inexpensive' breakout boards!
[23:46:21] <Gamma-X> like like a dollar or 2 but im gettin raped in shippin, I am AN ASSQWHOLW
[23:46:22] <eric_U> are you buying spares?
[23:46:30] <SWPadnos> you're lucky - it's my fault they're cheaper at DK, and that Mesa sells them at all :)
[23:46:41] <SWPadnos> they were $22 each when I bought them
[23:46:42] <Gamma-X> ?
[23:46:53] <SWPadnos> I mentioned to PeteW that $22 was a lot, and they should sell them
[23:47:06] <Gamma-X> there still expensive at dk, i talked the girl into giving me them for 14.23 what arrow is sellin them for
[23:47:08] <SWPadnos> he called DigiKey and said "Arrow has them for $9", so they lowered their price
[23:47:09] <eric_U> $22 is outrageous
[23:47:14] <SWPadnos> strange
[23:47:35] <SWPadnos> yes, and I even got relatively high quantity - like a 10-pack
[23:47:55] <eric_U> too bad automation direct doesn't carrry them
[23:47:56] <Roguish> i just picked up a 5i20 and a 7i37t at Mesa yesterday. spent hours this morning looking for 'inexpensive' 50 pin breakout boards.
[23:48:06] <eric_U> Roguish: good luck
[23:48:07] <Gamma-X> the girl said is there anything else i can do for u today? I replied. " uhhh noooo ur in minnesota so that wouldnt work"...
[23:48:19] <tomp2> spent the day debugging stuff over the phone... then got this email ( beware just stupid joke about troubleshooting )
http://pastebin.ca/856709
[23:48:55] <eric_U> Gamma-X: I used to live in Wisconsin, and I have an improper comment that I'm not going to make
[23:49:12] <Gamma-X> go ahead! lol
[23:50:26] <Roguish> Industrologic board is $50. DAQstuff board kit is $28. Check www.winfordeng.com $45.
[23:50:46] <eric_U> how much you offering for 2?
[23:51:32] <SWPadnos> eric_U, the 7i37 manual says that pins 49 and 50 are unused
[23:51:50] <eric_U> i haz been robbed!
[23:52:03] <SWPadnos> Roguish, and then you have to add the cable, and the extra cabinet space
[23:52:03] <Gamma-X> ?
[23:52:20] <Gamma-X> SWPadnos why wouldnt he just have a termination block?
[23:52:21] <SWPadnos> so even with expensive terminal strips, the 7i37-T is still a better deal
[23:52:29] <SWPadnos> who?
[23:52:39] <eric_U> what where and why?
[23:52:43] <Gamma-X> Roguish
[23:52:50] <Roguish> Mesa's T version board with 2x$15 for plug in connectors are not too bad. the T version is +$10, so for $30 ya get the works on one board.
[23:53:01] <Roguish> $40.
[23:53:01] <eric_U> T version much better from space perspective
[23:53:06] <SWPadnos> because those breakouts need to connect to the Mesa card, and that's done with a cable
[23:53:20] <Gamma-X> wich is more money.
[23:53:21] <eric_U> it's worth a lot
[23:53:23] <SWPadnos> the breakout takes much more space than the extra area added to the mesa boards
[23:53:32] <Roguish> yeah. no extra board and cable.
[23:53:34] <tomp2> ander's m5i20 DIY breakouts
http://www.anderswallin.net/2006/08/optoisolator-cards-for-mesa-5i20-servocard/
[23:54:02] <eric_U> no servo board though, can we smack anders until he does one of those?
[23:54:03] <Gamma-X> http://www.cablestogo.com/product.asp?cat%5fid=1212&sku=02744&p=e
[23:54:43] <Gamma-X> got 2 of them also..
[23:55:02] <SWPadnos> I think you need the version with long ejectors if you want to use a connector with strain relief
[23:55:12] <eric_U> I have a lifetime supply of 50 pin breakout boards, but I don't remember if that's why I didn't get the T version
[23:55:26] <eric_U> If they had them, I wish I had
[23:55:27] <Roguish> Anders are ok, but the parts are still expensive. plus ya gotta make it!!
[23:56:09] <tomp2> oh you wanted cheap and easy
[23:56:22] <Gamma-X> Roguish if u bought the t version why are u lookin for a breakout board?
[23:56:26] <eric_U> "aint no free"
[23:56:28] <Gamma-X> for the 5i20?
[23:56:42] <SWPadnos> to connect the 7i37 to the 5i20
[23:56:53] <Roguish> yeah, i need a couple of setups.
[23:57:02] <Gamma-X> ahh ok
[23:57:34] <Roguish> yesterday Mesa was out of stock on the 7i37t's. should be in next week.
[23:58:28] <Roguish> hey they had a nice looking 3 phase motor drive on the table under development. 3kw.
[23:58:44] <eric_U> Mesa did?
[23:58:47] <Roguish> ya
[23:58:58] <eric_U> 3Kw is hard
[23:59:23] <eric_U> 1Kw is starting to look fairly reasonable
