#emc | Logs for 2010-05-08

[00:02:21] <andypugh> DOS 3.3 I remember when that was new-fangled.
[00:02:56] <Jymmm> This is IBM, not MS DOS
[00:04:11] <andypugh> Much the same thing, isn't it?
[00:05:00] <andypugh> Same code until DOS 5 http://en.wikipedia.org/wiki/IBM_PC_DOS
[00:05:36] <skunkworks_> interesting
[00:11:50] <MattyMatt> I've got a shrink wrapped NT 4.0 here
[00:14:48] <skunkworks_> http://imagebin.ca/view/p_DGZsi.html
[00:15:44] <skunkworks_> other than the mouse pad seems a bit random
[00:15:51] <skunkworks_> the install was painless
[00:18:06] <MarkusBec> MarkusBec is now known as MarkusBec_away
[00:22:57] <skunkworks_> heh - windows uses 1.something gig of memory - lucid is using 239megs right now. ;)
[00:24:43] <skunkworks_> logger_emc: bookmark
[00:24:43] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-05-08.txt
[00:26:31] <skunkworks_> mouse pad randomness seems to be from my thumb actually touching the pad while clicking...
[00:36:15] <skunkworks_> other than.... I installed the 32 bit version.
[00:48:27] <MattyMatt> a laptop with 3 jogwheels would be nice
[00:49:50] <skunkworks_> So - do you think I can install 64bit version of lucid without screwing up the whole system?
[00:50:36] <MattyMatt> an you make an image of the drive before you do?
[00:50:55] <skunkworks_> right now it is dual booting between lucid and windows7 with no issues.
[00:51:13] <skunkworks_> well - I could... but what fun would that be? ;)
[00:52:00] <MattyMatt> I have very few issues with 64 bit lenny
[00:52:52] <MattyMatt> SWF playback was broken until Adobe released the 64 bit plugin alpha, which works OK
[00:53:54] <MattyMatt> I didn't expect to keep this OS when I installed it. most of my hdd is an unused XP partition
[00:54:05] <skunkworks_> maybe I will play with this install for a while
[00:54:43] <skunkworks_> it is working ;)
[00:55:18] <MattyMatt> they're the best kind :)
[00:56:31] <MattyMatt> I might try Chrome on this. Iceweasel is the weak point in this machine
[01:07:10] <MattyMatt> cool, mGcodeGenerator.py works if you comment out 'from turtle import right'. python dropped it and the script doesn't use it
[01:14:38] <MattyMatt> this is for blenders later than 2.43 but no later than 2.49 :) always gotta qualify to 2dp when talking about blender scripts
[01:30:43] <skunkworks_> wow - I like lucid
[01:30:56] <skunkworks_> it 'just works'
[01:31:10] <andypugh> What is it?
[01:31:19] <skunkworks_> lucid lynx
[01:31:28] <skunkworks_> 10.04lts
[01:32:04] <skunkworks_> seems to have found everything on my relatively new laptop
[01:33:42] <Jymmm> Seems to have choked on mine. Not drivers but things like the clock/calendar, external LCD, etc.
[01:34:07] <Jymmm> It's very OSX like
[01:34:24] <Jymmm> well, specifically Aqua.
[01:34:37] <skunkworks_> Jymmm: http://imagebin.ca/view/p_DGZsi.html
[01:34:40] <skunkworks_> :)
[01:34:45] <andypugh> I find it hard to disapprove of that
[01:35:49] <Jymmm> eh
[01:37:19] <andypugh> I am sat here working in Xcode. The compilation happens on the Ubuntu machine in the garage, I just have the repository linked
[01:46:11] <andypugh> Right, that's the Mesa card LED driver working again, but this time for an arbitrary number of LEDs (future-proofing)
[01:46:42] <andypugh> Time to sleep.
[01:46:55] <dareposte> andypugh: Hey the INI file you sent me worked great all day long, thanks again
[01:47:02] <dareposte> err, HAL file
[01:47:24] <andypugh> Glad to hear it. I do think there might be a neater way though.
[01:47:36] <andypugh> But what works, works.
[01:47:47] <dareposte> I took the time to read through it and it makes pretty good sense, though I wouldn't have figured it out alone for sure
[02:09:53] <cradek> hi jmk!
[02:10:24] <cradek> oh, I bet he's not real
[02:22:23] <skunkworks_> I know you may think he is not real anymore... but I am sure he still is.
[02:22:27] <skunkworks_> ;)
[02:30:40] <cradek> I want to talk to him...
[02:32:15] <skunkworks_> he seems to be very very busy.
[02:33:28] <cradek> yep
[02:33:42] <cradek> I think he even does non-computer things sometimes, which is awfully bizarre
[02:34:08] <skunkworks_> I wonder what that involves.. odd
[02:34:31] <cradek> I think sometimes you even exit all human-made structures (and then not even get in a car)
[02:35:35] <skunkworks_> now you're just making stuff up.
[02:36:02] <cradek> speaking of which, I hope it's not windy tomorrow so we can go bicycling
[02:43:37] <skunkworks_> we have a nice bike trail down the mississippi and 2 bikes... we have not made it down it yet. this summer though - we will.
[02:44:21] <cradek> sounds nice - we have a trail on the north side of town that goes through some parks and stuff. it's more interesting than heading out into the (flat, boring) country
[02:46:23] <skunkworks_> I bet. I would love to ride the bike to work.. but I am way out of shape for that.
[02:47:04] <renesis> i want to ride my bike on trails around work
[02:47:33] <renesis> but it still doesnt have crank or seat post spacer/clamp =(
[02:48:22] <renesis> (6061 frame 24" BMX with carbon fiber and machined steel forks)
[02:50:27] <cradek> my bike is off-the-shelf :-)
[02:50:35] <skunkworks_> same here
[02:50:45] <cradek> sometimes it's tempting to try welding up something bizarre for fun, but then I come to my senses
[02:51:08] <skunkworks_> or battery powered - but what is the point of that..
[02:54:14] <cradek> yeah, may as well do a car
[02:54:19] <cradek> (which I want to do someday)
[03:28:18] <MattyMatt> " I would love to ride the bike to work.. but I am way out of shape for that." only for the first 2 weeks :)
[03:30:08] <MattyMatt> my bikes are mostly out-of-the-canal, if you include Holland :)
[03:31:24] <MattyMatt> ik was een fietsmakker
[03:32:36] <MattyMatt> hills and bikes don't mix tho. I want an engine for hills kthx
[04:27:07] <elmo40> ok. making a lathe :)
[04:27:52] <elmo40> haven't ran a CNC lathe in over 9 years!
[04:30:10] <elmo40> I was thinking of making it like a Hardinge
[04:31:27] <elmo40> no tailstock. for smaller/shorter things.
[05:17:45] <elmo401> elmo401 is now known as elmo40
[05:48:36] <elmo40> http://www.handlewithlinux.com/linux-laser-cool Interesting...
[06:32:04] <Jymmm> http://www.youtube.com/watch?v=jaCW1AHfcYw&feature=player_embedded
[06:41:04] <ries_> ries_ is now known as ries
[08:40:54] <ua9hbw> ua9hbw is now known as Athlocatle
[09:21:10] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[09:58:12] <piasdom> g'mornin all
[11:11:51] <JT-Dev> YEA! a poker run today
[12:28:37] <ries> can I have two axis with HOME_SEQUENCE = 1 ??
[12:28:52] <ries> Ahh yes...
[12:34:33] <Dave_911> logger_emc:bookmark
[12:34:33] <Dave_911> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-05-08.txt
[12:57:39] <ries> ok guys, whish me luck... going to make my auto Z and home switches today :9
[12:57:52] <ries> I think I did all teh prep work....
[13:26:05] <piasdom> ries; Good luck with that
[15:08:42] <elmo40> wow, expensive accelerometers! http://www.inertialsensor.com/
[15:16:25] <telmnstr> Check out Analog Devices isensors
[15:25:00] <geo01005> Anybody have some experience with heat sinking TO-247 packages?
[15:26:53] <DaViruz> a little
[15:28:42] <geo01005> So I'm looking at using a few 100 Watt thick film resistors in a TO-247 package. Looking at the heatsinks available for these packages I'm having a hard time believing that I can really dissipate 100 watts from the package.
[15:29:21] <geo01005> How would I cool these chips to get close to 100 watts dissipation from them?
[15:29:41] <geo01005> An example resistor: http://www.vishay.com/docs/50051/lto100.pdf
[15:30:10] <DaViruz> usually you want to look at the heatsinks C/W rating, and the rated temperature of the device
[15:30:39] <DaViruz> if a heatsink has 0.8 C/W it means the temperature will rise 0.8 degrees celsius above ambient for every watt you put in to it
[15:31:40] <DaViruz> say you want to allow a maximum temperature rise of 60 degrees celsius, at 100W, you would need at most 0.6 C/W
[15:31:55] <geo01005> I think I understand that... if the max temp is 175 C and ambient is 40 C then I can have a 135 degree rise.
[15:32:17] <DaViruz> yeah
[15:32:47] <geo01005> ok, so does anybody know where I can find a heatsink made for a TO-247 that has that low of a resistance?
[15:33:06] <DaViruz> this is oversimplification though, you get some thermal resistqance in the interface between the device and the heatsink, and the heatsink C/W is usualy rated at this and that airflow
[15:33:49] <geo01005> Yeah I have looked into most of those parameters.
[15:33:57] <DaViruz> pretty much any heatsink with a drilled and tapped hole will do
[15:34:05] <DaViruz> i guess
[15:35:54] <geo01005> hmm, most of the heatsinks that I could find have resistances no less than about 3 C/W (forced convection at 100 Ft/min).
[15:36:35] <geo01005> I'm wondering at what point It is more effective to just mount it to a chunk of aluminum.
[15:38:24] <DaViruz> most cpu heatsinks has a C/W of well below 1, sometimes down towards 0.3
[15:38:49] <DaViruz> but that is with a fan mounted on top of it of course
[15:39:28] <DaViruz> but getting rid of 100W passively porbably requires quite a lot of surface area
[15:39:57] <geo01005> I don't mind having fan, infact I expect to.
[15:40:55] <geo01005> My current application has 1500 Watts of wire wound resistors, but I would like to switch to thick film resistors because the inductance from the wire wound resistors is giving problems.
[15:45:21] <geo01005> I just want to be sure that I can dissipate about 700 Watts in a reasonable space on my PCB.
[15:55:19] <ries_> ries_ is now known as ries
[16:16:51] <andypugh> geo01005: Can't you get non-inductively wound resistors?
[16:17:12] <geo01005> Yeah, but now "non-inductive" are they?
[16:17:28] <geo01005> How not now...
[16:17:34] <MattyMatt> even a long straight wire has induction
[16:18:04] <andypugh> But a long straight wire next to another running back the same way isn't.
[16:18:16] <geo01005> I talked to some of the manufactures and they said that their non-inductive products were only about half the inductance of the regular type in most applications.
[16:18:34] <andypugh> (actually, they might need to be co-axial to have _no_ inductance.
[16:18:45] <MattyMatt> ah OK, long loop of nichrome wire then
[16:18:55] <geo01005> Was I told the wrong thing?
[16:19:04] <MattyMatt> red hot = yes I'm dissipating 700W
[16:19:36] <MattyMatt> I think nichrome wire is much cheaper than calibrated resistors too
[16:19:55] <andypugh> No, I imagine that the manufacturer knows a lot more about resistors than I do
[16:20:51] <MattyMatt> 700W is lot to get rid of from a PCB
[16:21:09] <MattyMatt> look at the size of a cpu cooler for 120W
[16:22:22] <geo01005> well a blow dryer put out like 1500 watts.
[16:22:50] <andypugh> There is a fair bit of cooling air there :-)
[16:23:10] <MattyMatt> domestic ones melt very quickly. the ones they use in salons are heavy
[16:23:26] <andypugh> In fact, use a hairdryer as your resistor :-)
[16:23:48] <MattyMatt> toaster is less induction
[16:23:59] <MattyMatt> and you get toast
[16:25:14] <geo01005> I considered a heating element from a stove.
[16:25:15] <MattyMatt> yeah use a cheap hairdryer, and be careful not to impede the airflow :). noise will be the biggest drawback
[16:26:21] <geo01005> However stove heating elements are not in the right resistance range for my application (.5 ohms)
[16:26:42] <MattyMatt> you want .5 ohm?
[16:26:56] <geo01005> yeah.
[16:28:07] <geo01005> I currently have 8 4ohm resistors in parallel.
[16:28:30] <andypugh> You can put higher-value, lower power, resistors in parallel to get the resistance you want.
[16:28:37] <MattyMatt> nichrome wire is ~1 ohm per inch I think, so if you had a few parallel lengths of a few inches, you should get something that doesn't melt
[16:32:33] <MattyMatt> I saw a stewart engine on a heatsink a while back. the fan is powered by the heat :)
[16:33:08] <andypugh> I think I saw that.
[16:33:35] <MattyMatt> I don't think it was a miracle cure for anything, but a neat idea
[16:34:51] <andypugh> I don't think Stirling engines are self-starting, which seems a serious drawback for a CPU cooler.
[16:35:57] <MattyMatt> rather :)
[16:36:10] <MattyMatt> not a showstopper on a 24/7 server tho
[16:37:44] <L84Supper> are Sterling engines for sale anywhere?
[16:38:19] <MattyMatt> stewart/sterling I knew it was scottish
[16:38:48] <MattyMatt> L84Supper: yeah but only as toys, or as 3rd world water pumps
[16:38:53] <L84Supper> sp Stirling
[16:39:13] <L84Supper> yeah, I saw the water pumps by the guy in NH
[16:39:59] <L84Supper> I always wondered if I were to try and sell them if I'd be bought out by the oil co's or just disappear some day
[16:40:02] <andypugh> I bought my dad one for christmas. It runs a treat when sat on the raditor.
[16:41:19] <andypugh> http://gyroscope.com/StirlingEngines/
[16:42:44] <L84Supper> I just want a quiet generator that runs on natural gas and outputs ~3KW
[16:42:48] <andypugh> Unlikely, they never really found a niche other than working off of garden waste.
[16:46:01] <L84Supper> https://share.sandia.gov/news/resources/releases/2008/solargrid.html Sandia, Stirling Energy Systems set new world record for solar-to-grid conversion efficiency
[16:46:21] <L84Supper> 31.25 percent efficiency
[16:46:27] <L84Supper> not too shabby
[16:48:03] <MattyMatt> yeah that's pretty good, although the GaAs PV are nearly 30%
[16:48:22] <MattyMatt> or whatever the green ones are
[16:49:03] <andypugh> Mirrors and heat engines are probably a lot cheaper though
[16:49:34] <MattyMatt> only until thin film PV is cheap
[16:50:12] <L84Supper> I was surprised that I could not find a natural gas powered generator except for internal combustion types
[16:51:05] <MattyMatt> turbines don't become efficient until they're the size of a 747 one
[16:52:39] <andypugh> I think you might find a natural gas fuel cell. I googled and they exist, but I didn't find any prices.
[16:53:03] <L84Supper> I haven't looked in a year
[16:53:05] <MattyMatt> you can get static gas turbines of a few KW
[16:53:30] <L84Supper> <50db sound level operating?
[16:53:34] <MattyMatt> oh no
[16:53:47] <MattyMatt> they sound like a jet taking off :)
[16:54:16] <L84Supper> yeah, it's for in the home or shop for the non-deaf
[16:54:54] <andypugh> Fuel cell seems like the way to go then
[16:55:04] <L84Supper> they sell natural gas powered AC units here, not sure of the sound levels
[16:55:33] <MattyMatt> millpond and overshot wheel. got room for one?
[16:56:31] <MattyMatt> use that for storage with a windmill and PV, supplemented by free rain
[16:57:16] <L84Supper> 100k rpm flywheels delivered right to your door.... don't see anything like that in the phone book
[16:57:26] <andypugh> I got a letter from British Gas trying to sell me a PV system last week.
[16:58:28] <MattyMatt> it a french firm does the flywheels
[17:00:01] <andypugh> Interesting bit of physics, what's the optimum shape of flywheel for a given mass?
[17:00:13] <L84Supper> http://cgi.ebay.com/ebaymotors/DTR-Hydrogen-Fuel-Cell-_W0QQcmdZViewItemQQhashZitem3eff766fd8QQitemZ270573924312QQptZMotorsQ5fCarQ5fTruckQ5fPartsQ5fAccessories
[17:00:33] <L84Supper> fuel cell that looks remarkably like a water filter
[17:01:02] <andypugh> It's not a fuel cell at all
[17:01:03] <MattyMatt> andypugh, it's all about losses. if there's 100% eff, then no shape is prefered
[17:01:09] <L84Supper> ;)
[17:01:41] <andypugh> It's an electrolytic cell to squirt hydrogen and oxygen and added snake-oil into your car manifolds.
[17:02:21] <MattyMatt> the french ones are a steel clinder about 50cm high and 30cm diameter, on magnetic bearings, in a vacuum, under the pavement
[17:02:44] <andypugh> MattyMatt: As the diameter goes up the centrifugal stress increases faster than the moment of inertia, so it turns out that long and thin and very, very, fast is best.
[17:03:33] <MattyMatt> yeah you don't often hear about shafts exploding :)
[17:04:09] <MattyMatt> well you do, but it's usually misalignment
[17:05:24] <MattyMatt> I'm worried about my ability to detect a scam. those fuel cells do have a lot of testimonial evidence
[17:05:36] <andypugh> There are probably a few of these going spare at the moment.
[17:05:37] <andypugh> http://www.flybridsystems.com/F1System.html
[17:06:48] <andypugh> So, you use electricity to electrolyse water, then burn it in your engine, convert it back to eleectricity with your alternator and make a net gain?
[17:08:51] <MattyMatt> the only thing I could think of, was that the oxygen means less nitrogen is wastefully compressed/heated/convertedtoNO2
[17:13:37] <andypugh> It is helpful to consider the nitrogen in the air-charge as the working fluid of an internal combustion engine.
[17:14:15] <andypugh> And the way to not make NO2 is to not have too much oxygen in there. Which is what EGR is for.
[17:14:54] <andypugh> That device is going to add free oxygen as well as the hydrogen, so I can't see it making any less NOx
[17:15:36] <MattyMatt> if they displace N2 as the working fluid, they would afaics
[17:16:12] <L84Supper> 3kw and it's enormous http://www.acumentrics.com/products-power-generators.htm
[17:16:38] <andypugh> Consider how much gas an engine breathes in a minute, and how much water you can reasonably expect to electrolyse in the same time.
[17:16:43] <MattyMatt> there is exactly enough H to mop up any O, of course
[17:16:58] <L84Supper> Peak efficiencies between 30-50%; near 80% with waste heat recovery.
[17:17:09] <andypugh> Yes, spookily exactly enough :-)
[17:18:10] <andypugh> L84Supper: It is fairly big, but looks otherwise ideal.
[17:18:18] <MattyMatt> yeah it smells like snake oil to me, but I read a long story from a mechanic who really (appeared to) have the right skeptical attitude & experiments, and it worked for him
[17:19:05] <MattyMatt> he plainly wasn't deluded. if that was fake it's simple fraud
[17:19:46] <andypugh> A local where my folks lives made one and swears by it. There might be something in it. I just don't see what.
[17:20:05] <MattyMatt> I haven't seen someone do serious experiments that debunk it, either
[17:20:13] <andypugh> Consider that an engine breathes 1kg of air in a minute @ 1000 rpm
[17:21:25] <MattyMatt> but also consider how easily stuff ignites at 25% oxygen compared to 21%
[17:21:27] <andypugh> Right, time to electrocute myself.
[17:22:13] <MattyMatt> if there's anything in it, it's the improved O:N ratio, I'm sure
[17:22:16] <andypugh> The time has come to take the servo motor to 250VDC, which means making a power supply.
[17:22:27] <andypugh> Could be free radicals.
[17:22:58] <andypugh> (but EGR already adds free radicals)
[17:23:02] <MattyMatt> well NO is a free radical :) that's injected deliberatley
[17:24:00] <andypugh> http://dwolsten.tripod.com/articles/jan97.html
[17:24:48] <MattyMatt> H2 gas might be a powerful catalyst for hydrocarbons in those conditions
[17:26:00] <MattyMatt> water vapour is. they impoved the efficiency of coal burners by spraying in water
[17:27:07] <MattyMatt> anyway, on the evidence I've seen, I have to say I want to see some more evidence :)
[17:30:05] <MattyMatt> PSU = 4 big diodes and a very big cap?
[17:32:08] <MattyMatt> neat, I think I've got end-to-end objects to gcode in blender old API
[17:33:06] <MattyMatt> I can now lay out instances of parts, and output gcode for the whole layout
[17:33:38] <MattyMatt> new blender will have to wait
[17:43:09] <andypugh> Yes, mu PSU is exactly that (plus an NTC resistor)
[17:43:59] <L84Supper> what's the best way to get real time position info from a robot since you can't hang 3 linear encoders from it's arm to measure xyz
[17:44:21] <MattyMatt> a variac and one of those V shaped sparky things on top would complete the effect
[17:44:59] <L84Supper> we always called them "V shaped buzz things"
[17:46:10] <MattyMatt> L84, rotation sensors in the joints is the normal way, but you could supplement that with accelerometeres now
[17:46:34] <MattyMatt> or even visual feedback
[17:47:31] <L84Supper> have you ever come across a robot that gives you access to the encoders at the joints?
[17:48:31] <MattyMatt> nor one that doesn't :)
[17:48:37] <L84Supper> visual feedback isn't accurate is you're working with few micron precision across 1m or travel
[17:48:41] <L84Supper> or/of
[17:48:53] <MattyMatt> depends how good your vision is
[17:49:09] <L84Supper> it's not practical
[17:49:30] <MattyMatt> you could laser track the hand
[17:49:58] <MattyMatt> you know roughly where the reflector is, from the joint encoders
[17:50:33] <MattyMatt> 3 laser rangefinders would give total accuracy
[17:51:50] <L84Supper> you actually need more if your moving the arm around a 3d part since the arm and part would be in the way at times
[17:52:11] <L84Supper> and it's not accurate enough
[17:52:11] <MattyMatt> OK 8. one at each corner of the room
[17:53:04] <MattyMatt> 6 big coils in the room, and 3 in the sensor. that worked well at nasa but no good in a steel machine
[17:53:07] <L84Supper> the catch has been how do you do this without voiding the robot warranty by opening up the robot and accessing the encoders directly
[17:53:46] <L84Supper> how fast was the arm at NASA?
[17:54:35] <MattyMatt> it wasn't an arm. the sensor was attached to clothes for VR
[17:55:19] <MattyMatt> A steel arm would cause reflections and ruin the signals
[17:55:33] <MattyMatt> so that won't work
[17:56:09] <MattyMatt> I don't think. if this is an academic exercise it may be the 'correct' answer
[17:56:43] <L84Supper> most robots spit out position info every few milliseconds, if the arm is moving at 1m/sec that's too much time for +- 5 micron accuracy
[17:56:44] <MattyMatt> so it was fast enough for VR, that's all I know
[17:58:02] <MattyMatt> OK. cameras on the robot wrist, barcodes on the walls
[17:58:37] <MattyMatt> massive gpu power and you'll never get better than 66ms :)
[17:59:52] <MattyMatt> you can buy 1ms cameras, but not with my money
[17:59:56] <L84Supper> heh, well we are working on a supercomputer that will have 2K of the HD5970 cards :)
[18:00:16] <L84Supper> 4 pflops peak single precision just might do it
[18:01:42] <MattyMatt> I think there's a reason the makers of linear slides are still in business :)
[18:02:12] <MattyMatt> one day we'll buy hand tools for our asimo
[18:02:58] <andypugh> Molex KK heades are rated at 5A. Blimey!
[18:03:22] <MattyMatt> yes asimo, scraped and lapped, and no you can't have any more oil until lunchtime
[18:03:47] <L84Supper> I've had to piggy back a controller just for real time position info but it means taking the robot apart
[18:08:54] <MattyMatt> sounds like a job for MillYourOwnRobotArm Man
[18:09:45] <MattyMatt> I got a neat book which shows how to put strain (pain) sensors in the joints too
[18:10:31] <sealive> hi to all
[18:14:00] <andypugh> Hi
[18:14:08] <sealive> hi andy
[18:15:05] <sealive> ok i go to the tv
[18:15:16] <frallzor> *phew*
[18:15:23] <frallzor> only 16 more holes to drill manually
[18:16:11] <frallzor> and 48 nuts and bolts to put in the holes
[18:16:19] <frallzor> that will be a bitch
[18:20:59] <piasdom> how do i get "run from selected line" to work in axis? or maybe i'm not selecting the line correctly(just clicking on a line in the bottom window))
[18:27:16] <L84Supper> MattyMatt: was that for a "Holodeck" type project?
[18:28:07] <cradek> piasdom: right click
[18:35:17] <piasdom> cradek; i've right clicked and used the "machine" dropmenu.. it go to the last three or so moves,runs them and stops
[18:37:11] <piasdom> i don't even click in the osub area,just the main program
[18:38:05] <cradek> run from line might not be possible when subroutines are involved. it only really makes sense in a straight-through type program
[18:39:26] <piasdom> cradek; cool .... thanks... just didn't want to start the program again
[18:40:16] <piasdom> i can edit the file and remove to where i want to start
[18:40:30] <cradek> yeah that sounds foolproof
[18:54:16] <dmess> hi all
[19:18:44] <andypugh> With cat-like reactions he replies: Hi dmess.
[19:23:06] <frallzor> f*king hell, Im done drilling!
[19:24:13] <frallzor> took its sweet time but the result was surprisingly easy to reach
[19:25:29] <toastydeath> ?
[19:25:39] <toastydeath> what were you drilling
[19:25:45] <toastydeath> (the machinist wants to know)
[19:27:33] <andypugh> Hold-down holes?
[19:37:21] <frallzor> kind of
[19:37:38] <frallzor> holes trough the bottom board that attaches to the machines
[19:37:54] <frallzor> drill > MDF and steel profiles
[19:37:57] <frallzor> now
[19:38:08] <frallzor> I just have to figure out how to bolt them :P
[19:38:15] <frallzor> dont wanna crawl under the table
[19:40:50] <frallzor> at least this was the part I feared the most, that turned out not to be that much of a bitch =)
[19:42:20] <andypugh> Did you consider tapping the holes? Too late now, but would have saved crawling.
[19:42:34] <andypugh> (And a cordless drill makes a great tap wrench)
[19:43:42] <frallzor> pretty thin material so I didnt wanna tap M8
[19:43:50] <frallzor> but I thought about it =)
[19:44:17] <andypugh> Use a really blunt drill and you gain tapping depth :-)
[19:45:34] <frallzor> but I think i will be ok, I just need to tighten the nuts so the reach the profile, then they will selftighten when screwing them from above
[19:46:25] <frallzor> then the fun begins, leveling =)
[19:46:28] <frallzor> 1hr drive
[19:47:23] <andypugh> CNC program to hold a socket on the bolts from the top while you tighten from underneath.
[19:55:55] <andypugh> I found myself randomly reading an old Practical Mechanics. And I think I disagree with their opinion on how to use an adjustable spanner.
[19:55:56] <andypugh> http://books.google.co.uk/books?id=C9wDAAAAMBAJ&lpg=PA183&ots=tHfP8i5GZv&dq=thin%20sheet%20tapped%20holes%20forming&pg=PA187#v=onepage&q&f=false
[19:58:39] <MattyMatt> I tapped M8 in 2mm steel no problems. that's my leadscrew nuts :)
[19:59:01] <MattyMatt> it's nearly 2 whole turns of thread
[19:59:42] <andypugh> For a rather broad definition of "nearly"
[20:02:06] <frallzor> the hell with this, Im going in again to bolt! :P
[20:02:37] <MattyMatt> I milled a layout \o/ blender is now blenderCAMpro
[20:04:48] <MattyMatt> I forgot to turn on compensation on the perimeter, but I'll call that styling
[20:09:15] <Jymmm> http://sfbay.craigslist.org/sby/tls/1730882077.html
[20:13:52] <andypugh> Nice machine. Can you wrap it nicely and send it to me?
[20:15:49] <frallzor> all in all 12 nuts I cant tighten myself, pretty ok =)
[20:16:25] <MattyMatt> superglue
[20:17:34] <frallzor> still need a 2nd person then =)
[20:18:15] <frallzor> pretty solid feeling now that most of the board is fastened
[20:19:03] <MattyMatt> naah, put the bolt in, then go underneath and put the nut on, with superglue on the top face, then goto the bolt and put it up to press the nut against the frame
[20:19:31] <MattyMatt> hold it for 10 seconds, like in the ad
[20:19:45] <frallzor> cant get the nuts on without wholding the bolt
[20:19:53] <frallzor> tried that when I did the others
[20:20:06] <MattyMatt> chewing gum will hold the bolt in place while you get the nut started :)
[20:20:22] <frallzor> chewing gum on mdf doesnt sound too good =P
[20:21:56] <MattyMatt> so packing the hole with thermite to weld it is out? :)
[20:22:09] <frallzor> yup! =P
[20:22:53] <MattyMatt> superglue is the answer. it doesn't leave residue and breaks easily
[20:23:04] <MattyMatt> much residue
[20:23:24] <MattyMatt> a spot of woodglue on the bolt shaft?
[20:23:36] <MattyMatt> you'd have to wait for that to dry
[20:24:19] <frallzor> I will try calling a friend instead and ask if he want to crawl around in mdf dust and metall shaving
[20:24:20] <frallzor> s
[20:24:29] <MattyMatt> that's what I'd do. tack all the bolts in place with woodglue, and put the nuts on in the morning
[20:26:10] <frallzor> will be fun to see which answer I get on my question now =)
[20:27:59] <frallzor> http://www.lathes.co.uk/arboga/page5.html how would you value one of these?
[20:31:47] <MattyMatt> with all the toys?
[20:31:48] <frallzor> top version
[20:31:48] <frallzor> someone here is selling one and I think its waaay overpriced
[20:31:48] <frallzor> so im curious =)
[20:31:53] <MattyMatt> several thousand with all the toys, if it's still in precision condition
[20:35:33] <MattyMatt> if it's just a nice usable mill, then a few hundred? 4-500 E
[20:36:23] <frallzor> seller want 1000E
[20:36:34] <frallzor> I think thats overpriced =)
[20:37:07] <MattyMatt> unless it's immaculate, in which case it seems fair
[20:37:43] <andypugh> Looks very valuable to a model maker, all machines in one
[20:37:54] <MattyMatt> you can get a new sherline or tieg for that price
[20:38:37] <andypugh> Yes, but that is made of lumps of cast iron.
[20:39:21] <frallzor> you can get a nice optimum for the same price
[20:39:36] <fenn> * fenn misread that as 'opium'
[20:40:44] <MattyMatt> for less than that, I could make a furnace capable of melting 12lbs of iron :)
[20:41:27] <MattyMatt> there's a foundry in town already I've heard. I should go get some prices
[20:42:27] <andypugh> Cast iron parts are very cheap. If you make the patterns we are talking tens of pounds.
[20:42:56] <andypugh> That machine is probably worth more than it is worth, if that makes any sense
[20:43:23] <MattyMatt> yeah it looks collectable, not abusable
[20:43:35] <MattyMatt> it's too pretty :)
[20:51:29] <MattyMatt> I like modular tools in general. I wouldn't mind a set of sherlines to play with
[20:54:44] <MattyMatt> I've been thinking how to make a jig for hobbing a large gear with an M8 tap to use an M8 wormscrew
[20:55:43] <MattyMatt> a general purpose bed to bolt the motors too would do the trick, instead of a one-off jig
[20:56:48] <MattyMatt> I should just buy a rotary I think :) a plywood one is silly
[20:57:56] <MattyMatt> M8 probably means I was thinking reprap at the time
[20:58:13] <andypugh> If you want anything hobbing, you could post it to me.
[20:59:33] <MattyMatt> it would be the whole rotary faceplate, if it's plywood
[21:00:45] <MattyMatt> duh I could hob it in place
[21:01:41] <MattyMatt> build rotary. put tap where work goes, attach motor #2 to shaft (maybe by temporary gear)
[21:01:52] <MattyMatt> ^where worm goes
[21:06:42] <MattyMatt> there isn't really room under my gantry for the kind of rotaries you see on ebay
[21:07:02] <frallzor> I won!
[21:07:20] <frallzor> went another round with the nuts and bolts
[21:07:31] <MattyMatt> yehaw
[21:07:31] <frallzor> pretty comfy under the machine
[21:07:43] <frallzor> had some spoild medf stored there
[21:07:52] <frallzor> pretty nice to lie on and tighten
[21:08:02] <frallzor> *spoil mdf :P
[21:09:00] <frallzor> now its all done and ready for leveling
[21:10:20] <MattyMatt> plastic gear is almost certainly a better idea than plywood. I need some nice thick delrin chopping boards
[21:13:15] <MattyMatt> I'm gaining confidence this machine will mill steel one day
[21:13:34] <MattyMatt> so maybe I'll leave the 4th axis till then
[21:14:03] <andypugh> MattyMatt: I find your approach refreshing :-)
[21:14:39] <andypugh> I think I would have done it the same way several years ago, but now spending money is easier.
[21:15:03] <frallzor> My machine can mill steel I think
[21:15:37] <frallzor> it ate a part of a clamp by mistake =)
[21:16:03] <MattyMatt> yeah easily, I reckon, up to the point where the chatter lifts the carriage off the rack
[21:16:09] <andypugh> Some steel is easier to machine than some Alumium.
[21:16:37] <andypugh> I guess that some steel is easier to machine than some (very exotic) wood.
[21:17:55] <MattyMatt> I can always put a mounted grindwheel in my dremel and get through anything with patience
[21:19:23] <andypugh> Diamonds? Neutronium?
[21:20:03] <MattyMatt> neutronium wheels are the purple ones? :)
[21:21:17] <MattyMatt> I've seen the saws they use on diamonds, not much bigger than dremel cutoffs
[21:21:49] <MattyMatt> thin steel wheels, constantly redressed with diamond powder
[21:22:16] <andypugh> I am assuming Neutronium is hard, it might be soft, but you would need a very rigid machine.
[21:22:53] <MattyMatt> well yeah, the mk2 would be made of neutronium :)
[21:25:40] <andypugh> It would be a rubbish machine for ordinary machining though, you would never get the swarf off the surface. Or the work. You wouldn't need clamps though.
[21:26:14] <MattyMatt> the xrays coming off the cutting edge would kill you
[21:26:50] <MattyMatt> a neutronium chuck guard chould solve that
[21:27:31] <MattyMatt> steel will do fine for now :) and carbon fibre like that F1 flywheel
[21:27:45] <frallzor> * frallzor is feeling satisfied
[21:33:05] <MarkusBec> MarkusBec is now known as MarkusBec_away
[21:33:28] <andypugh> MattyMatt: I think trying to let go of the handles would kill you.
[21:34:37] <MattyMatt> no, because the handle would be on the back shaft of the motors, which would be enormous
[21:35:15] <MattyMatt> even with superconducting neutronium coils
[21:36:30] <andypugh> I used to know a neutronium chemist, but he tended to work with quite small quantities.
[21:36:42] <andypugh> No, wait, that was muonium.
[21:38:37] <MattyMatt> there were events in neutrino detectors that are thought to be neutronium. it smashes its way through the planet instead of glided through like neutrinos
[21:40:13] <andypugh> I thought it needed gravity pressure to hold it together?
[21:40:47] <MattyMatt> dunno, maybe if it's dense enough it provides its own
[21:41:43] <MattyMatt> so any arbitrary size lump of neutrons can exist
[21:42:11] <MattyMatt> I presume spherical
[21:43:00] <andypugh> Ah, yes.
[21:43:24] <andypugh> If this ship sinks I will google. (gaming)
[21:43:57] <MattyMatt> hoist me gizzards to the mizzen
[21:44:23] <andypugh> If ye insist.
[21:45:20] <MattyMatt> yarrr
[21:48:48] <MattyMatt> hmmm, tool compo. can I be bothered working out whether a perimeter is drawn clockwise or ccw?
[21:49:01] <MattyMatt> not tonight fosho
[21:49:24] <andypugh> It's a simple vector thingy isn't it?
[21:50:05] <MattyMatt> yeah, but which one? :)
[21:50:27] <MattyMatt> I'm sure google knows. I'll ask it sometime
[21:51:38] <MattyMatt> I think I'd have to walk the perimeter, adding up the angles. if it's +360 at the end, it was cw, -360 = ccw
[21:51:57] <MattyMatt> anything else = mobius
[21:55:39] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[21:58:33] <MattyMatt> this old mGcodeGenerator.py script is full of useful stuff now I'm back on the old version of blender
[22:00:44] <andypugh> I seem to recall a much neater maniupulation to work out if a polygon points p or down.
[22:02:00] <andypugh> It is easy if you know the polygon is convex.
[22:02:19] <MattyMatt> the angles at each corner will be handy to know anyway, once this gets more sophisticated
[22:02:32] <MattyMatt> no, poly may well not be convex
[22:03:31] <andypugh> Is it faster to find the convex hull and the normal of that?
[22:04:22] <MattyMatt> could be. would that work with internal perimeters tho?
[22:04:28] <cradek> the well-known algorithm is to march around the polygon summing all the cross products
[22:04:51] <cradek> you can't trust any particular one, but you can trust all of them taken together
[22:06:03] <MattyMatt> I'll do angles :) that's how I do front faces in GLtriangles even though it's slower, I understand it
[22:06:46] <cradek> triangles are easy, always being convex - you just do one cross product
[22:08:26] <MattyMatt> well I need to set the side manually anyway, because this setup doesn't know internal perimters
[22:09:47] <MattyMatt> when they're made automatically, I can use a convention, but atm they're all freeform
[22:11:21] <MattyMatt> I suppose the convention is ccw for external, cw for internal, so your always cutting forwards?
[22:12:13] <andypugh> For conventional milling, yes. But conventional milling is becoming unconventional.
[22:13:11] <MattyMatt> so a checkbox for that
[22:17:38] <MattyMatt> these blender curves have some other parameters to abuse. the taper feature is neat
[22:18:15] <MattyMatt> and each vertex has UV
[22:19:32] <MattyMatt> if they have normals, then job's done
[22:40:51] <MattyMatt> doh, I didn't leave space for the tool between the items on this test piece
[22:41:11] <MattyMatt> good job it's only piece of softwood
[22:42:06] <MattyMatt> I wonder if public prefer MDF or plywood? they're about the same price
[22:42:41] <JT-Hardinge> I like 12ply plywood
[22:43:15] <MattyMatt> me too, but it's what the paying public, who'd rather have plastic, want
[22:43:57] <MattyMatt> ah, they're getting ply, if they want the first batch :)
[22:45:24] <MattyMatt> the mahogany I've got isn't thick enough to do all the parts
[22:45:31] <MattyMatt> nor the oak
[22:46:25] <MattyMatt> I'll buy oak if the demand is there for wooden reprap sets :)
[22:46:35] <JT-Hardinge> what are you making?
[22:46:46] <MattyMatt> sets of mendel parts
[22:46:55] <MattyMatt> there's still a clamour
[22:47:03] <JT-Hardinge> got a photo?
[22:47:07] <MattyMatt> $500 on ebay type clamour
[22:47:24] <MattyMatt> ebay have a good pic of a printed set
[22:49:02] <MattyMatt> http://cgi.ebay.com/RepRap-Mendel-Printed-Parts_W0QQitemZ280484297586
[22:50:45] <MattyMatt> that kit came with 2 working extruders, which are half the value
[22:51:33] <MattyMatt> $99 is the normal price for an extruder
[22:52:04] <JT-Hardinge> so your making the parts from wood instead of hot glue?
[22:52:12] <MattyMatt> yep :)
[22:52:19] <MattyMatt> I might paint them green tho
[22:52:53] <JT-Hardinge> I like this line "The parts are pretty good but will need a little cleanup, like most RepRap parts."
[22:53:28] <JT-Hardinge> and no photo of the extruders
[22:55:44] <MattyMatt> like an airfix kit. most kids don't bother :)
[22:56:19] <JT-Hardinge> what is an airfix kit?
[22:56:51] <MattyMatt> plastic airplane model kit
[22:57:03] <JT-Hardinge> oh
[22:57:28] <ndnihil_> ndnihil_ is now known as ndnihil
[22:57:29] <MattyMatt> I'd get the same response from a british kid nowadays
[22:57:49] <MattyMatt> we all made crappy spitfires in my day
[22:59:45] <MattyMatt> I think it's odd that all RP machines seem to be treated as office equipment
[23:00:39] <MattyMatt> "instant plastic stuff maker" could be handy in any workshop
[23:02:03] <JT-Hardinge> * JT-Hardinge wonders if there is a way to tell in Axis if there is a coordinate system offset in effect other than the funny blue/green ball that is never in view
[23:02:41] <MattyMatt> look for the GCODE in the state sting
[23:03:18] <MattyMatt> properties tab iirc
[23:04:03] <JT-Hardinge> the EMC Status window?
[23:04:45] <toastydeath> i read that as "Starbucks window"
[23:04:50] <MattyMatt> MDI [F5]
[23:04:52] <toastydeath> and thought, "man, I'm going to have to install emc on something now"
[23:17:07] <MarkusBec> MarkusBec is now known as MarkusBec_away
[23:19:57] <cradek> JT-Hardinge: I saw you gave a wrong answer on the web forum, but I forget what it was about now. How's that for helpful? :-/
[23:20:11] <JT-Hardinge> today?
[23:20:23] <cradek> oh right - absolute arc centers - G90.1, I think it's on the gcode quickref
[23:20:38] <JT-Hardinge> ah ok thanks I forgot I think
[23:20:48] <JT-Hardinge> or some other reason LOL
[23:21:23] <cradek> er no, it's not on there
[23:21:30] <cradek> maybe it's new in 2.4, I don't remember
[23:21:37] <JT-Hardinge> I'll check
[23:21:44] <cradek> thanks, sorry I'm little help
[23:21:57] <JT-Hardinge> cradek: can you cut multiple start threads in EMC?
[23:22:11] <JT-Hardinge> you gave me a clue that is usually all I need
[23:22:19] <cradek> of course, just change the Z start point
[23:23:56] <JT-Hardinge> that's what I was afraid of... would not work for what I was thinking of for a fancy smancy thing with 2 grooves and some spirals between the two
[23:24:37] <cradek> just rotate your encoder? :-)
[23:25:00] <JT-Hardinge> lol
[23:26:38] <JT-Hardinge> while were chatting is there any way in HAL to know if any machine offsets have been applied?
[23:27:15] <JT-Hardinge> what I'm looking to do is to make a led light up if any offsets other than tool offsets are in effect
[23:27:40] <JT-Hardinge> the blue/green ball seems to be out of my sight most of the time in the backplot
[23:30:08] <cradek> halui has offset (pos-relative) and not offset (pos-commanded) available
[23:30:20] <cradek> you'd have to do some nasty subtraction etc but you could theoretically get what you want that way
[23:31:16] <JT-Hardinge> thanks I'll check that out
[23:31:34] <cradek> or you could add that directly to halui (it already has halui.tool.length_offset.[abcuvwxyz] pins)
[23:32:07] <cradek> huh motion has tool offset output pins too, no idea why
[23:33:56] <JT-Hardinge> cool, the halui.tool.length_offset will work great any number other that 0 for ZX means I have an offset loaded
[23:34:49] <cradek> yep
[23:36:13] <JT-Hardinge> I noticed the status bar does not say if the tool loaded has the offset in effect
[23:36:40] <cradek> yeah, it shows the lengths whether or not they are active
[23:39:25] <JT-Hardinge> I just loaded an offset to Z for G54 and halui.axis.2.pos-relative jumps between -5 and 6
[23:40:10] <JT-Hardinge> If I Zero Coordinate System it settles down
[23:42:05] <JT-Hardinge> * JT-Hardinge heads in to cook some chow
[23:42:06] <MattyMatt> duuuh. if I'm on a plane in Z, then finding the angle of each segment will reveal arcs
[23:42:46] <MattyMatt> if they are evenly spaced
[23:43:44] <MattyMatt> cool, I can do the gcode importer now too
[23:44:17] <MattyMatt> oh no, that was to improve the quality of the exporter
[23:57:14] <Dave911> Are there any other examples of driving EMC2 via NML similar to what Jepler did with the jdi.py program? I finally tried out the jdi.py program and that is pretty slick. I don't recall seeing much as far as documentation on the NML interface?
[23:58:44] <steve_stallings> steve_stallings is now known as steves_logging
[23:59:04] <andypugh> Have you looked in the Developers manual?
[23:59:28] <andypugh> (in contrast to Users or Integrators)
[23:59:55] <Dave911> I think I have read just about everything and I don't recall any docs on NML... do you think I am wrong .... (hoping so ....)