Back
[02:08:56] <DJ9DJ> moin
[02:14:13] <uw> GROIN
[02:15:58] <DJ9DJ> good morning
[02:16:12] <uw> good morning to you DJ
[02:21:41] <kengu> good.. morning
[02:55:27] <gkwhc> Hey everyone! I was wondering, what 2D/3D software do yall use to model things to export them to gcode or other cnc format?
[02:58:00] <kengu> i use openscad, inkscape and .. umm.. i have seen autocad, blender and solidworks used
[02:58:51] <gkwhc> cool
[02:59:24] <gkwhc> sounds fairly simple, i might just be overthinking the whole process
[02:59:42] <kengu> for inkscape there are plugins to generate gcode
[03:00:03] <gkwhc> but depth data is constant right?
[03:00:07] <kengu> and 3d gcode.. well I usually print 3d objects so there are those slicer programs to generate that gcode
[03:00:27] <kengu> inkscape is more of 2d
[03:00:33] <gkwhc> right
[03:01:50] <kengu> there are milling, carving, plasma etc options in inkscape plugins but as I have not had need to figure out more 3d in that I don't know how proper 3d-milling would then work
[03:02:08] <kengu> i did some engraving but well.. constant depth.
[03:02:25] <gkwhc> i see
[03:02:37] <gkwhc> designing with inkscape sounds pretty straightforward
[03:02:46] <kengu> 2d atleast
[03:03:35] <gkwhc> yeah
[03:04:38] <kengu> gcodetools is one of the plugins
[03:05:35] <gkwhc> cool i'll look into that
[03:05:50] <gkwhc> have you used 3d softwares?
[03:06:01] <kengu> and then there is umm.. skeinforge to generate what ever gcode you need
[03:06:17] <kengu> http://replicat.org/generators something there
[03:06:27] <kengu> gkwhc: 3d what?
[03:07:19] <gkwhc> cnc milling 3d models?
[03:07:41] <kengu> um.. nope.
[03:08:13] <kengu> but skeinforge can do thta
[03:10:48] <gkwhc> sweet
[03:16:15] <gkwhc> thanks for the tips kengu!
[04:40:31] <kengu> guten abend eller evening (german, swedish and english. all in one) (:
[04:40:39] <kengu> and should be afternoon
[07:41:32] <jthornton> mornimg
[07:44:57] <skunkworks> morning!
[07:45:05] <skunkworks> did you guys get your mesa stuff sorted out?
[07:45:34] <jthornton> yes, it was a combination of errors LOL
[07:46:05] <skunkworks> heh
[07:46:07] <skunkworks> been there
[08:51:13] <aksr> hi archivist
[08:51:15] <aksr> :)
[08:56:07] <aksr> i have a few questions ..
[08:57:39] <aksr> tools, with color like silver, used for parting, anybody knows which material is this?
[09:03:50] <jdh> steel?
[09:06:32] <aksr> jdh: maybe, but it's not a regular steel, i'm sure of it...
[09:07:11] <jp__> tool steel?
[09:07:18] <aksr> ..because it parts other steel
[09:07:39] <aksr> again, not sure...
[09:08:47] <aksr> jp__: is there some specific name for it, other then _steel_?
[09:08:56] <jp__> http://en.wikipedia.org/wiki/High-speed_steel
[09:09:02] <jp__> start there
[09:09:09] <aksr> jp__: i already read that
[09:09:11] <aksr> :D
[09:09:18] <aksr> a few min. ago
[09:09:25] <aksr> :P
[09:09:38] <aksr> still, i'm not sure it's the same thing
[09:09:42] <aksr> it's very bright
[09:09:50] <aksr> silver-like
[09:10:05] <aksr> used (often) for parting
[09:13:55] <jp__> unless it's something exotic i would expect it's just hss.
[09:15:21] <jp__> cobalt tools usually have a dark tinge to them
[09:17:35] <cpresser> silver could also indicate be Carbide-Tools
[09:17:38] <aksr> jp__: is there some slang-word for it?
[09:18:28] <jp__> is the tool heavy for it's size?
[09:19:04] <aksr> i think so
[09:23:59] <jp__> weigh it and calculate it's volume. might be able to figure out if it's carbide based if the tool is large enough
[09:25:48] <aksr> i shall ;)
[09:26:49] <jp__> if it is carbide it should also be brittle
[09:27:12] <jp__> easy to chip compared to hss
[09:28:01] <jp__> but mot carbide i've seen is greyish matte like
[09:30:19] <aksr> jp__: this must be hss, it's not greyish
[09:30:33] <jp__> but it could be polished finish as well
[09:30:37] <aksr> anyhow, jp__ is there some other name for it?
[09:30:53] <jp__> name for what?
[09:30:59] <aksr> if we suppose it's hss
[09:31:19] <aksr> name for hss ...
[09:31:38] <jp__> yes look in that wikki link
[09:31:58] <jp__> T1
[09:32:06] <jdh> is this just for random knowledge?
[09:32:06] <jp__> M2
[09:32:09] * aksr is trying to remember how my neighbohr called it
[09:33:36] <aksr> s/my/his :P
[09:33:59] <jp__> like old names like kewattin?
[09:35:30] <aksr> some general name for it ...
[09:35:52] <aksr> i don't seem to recall it
[09:39:30] <aksr> jdh: well, yes and no :)
[09:39:41] <aksr> thanks for your help jp__ jdh cpresser
[09:39:42] <aksr> :)
[09:39:44] <jp__> http://www.varcoprecision.com/tools.htm
[09:40:07] <jp__> big list of trade names ^
[10:33:16] <Nick001> <aksr> put a magnet to it - mostly carbide is non-magnetic
[10:35:10] <aksr> Nick001: nice idea!
[10:37:41] <Nick001> I always have to fixture carbide to make a form tool - I can't use the magnetic chuck to hold it to flat grind it,etc
[10:38:17] <aksr> ;)
[10:44:24] <archivist> parting tools often HSS
[10:44:43] <aksr> hi archivist
[10:46:19] <archivist> carbide parting tools usually have replaceable tips
[10:48:52] <archivist> and I am biased, I like the Sandvik parting tips and blade
[10:50:53] <archivist> I use a modified one
http://www.greenwood-tools.co.uk/shopscr23.html
[12:10:36] <IchGuckLive> hi all around the globe B)
[12:30:01] <IchGuckLive> welcome the new members here ler_hydra abd jasen_
[12:32:43] <archivist> new?
[12:34:12] <ler_hydra> a matter of perspective I suppose... ;)
[12:34:23] <IchGuckLive> B)
[12:34:32] <archivist> I think you pre date me
[12:35:26] <ler_hydra> hmm, I seem to recall you being around when I first "joined" the IRC channel
[12:35:57] <ler_hydra> 7-ish years ago or so
[12:36:06] <IchGuckLive> O.O
[12:36:52] <archivist> I thought I joined 2007 ish :)
[12:37:35] <ler_hydra> hmm, well I remember being active on IRC when I had justed started university, and that was in 2006, so sometime around then =)
[12:37:39] <archivist> ish + or minus a bit
[12:37:48] <ler_hydra> seems to match up fairly well
[12:37:58] <awallin> anyone know how tapered pipe threads are made? someone at the lab wants to make a 600-bar pressure vessel with those threads..
[12:38:12] <IchGuckLive> then im the yungguy as old as i am
[12:38:12] <ler_hydra> IchGuckLive, it sure sounds like a lot, though fwiw I haven't exactly been very active
[12:38:20] <awallin> internal threads in the pressure vessel I guess
[12:39:32] <kwallace2> Other than using a pipe tap?
[12:39:40] <archivist> awallin, there are not too many standards, BSP and NPT
[12:39:57] <IchGuckLive> awallin:
http://www.fsg-furtschegger.de/Gewinderollen/files/page5_1.jpg
[12:40:52] <awallin> say you have a 30mm plate of stainless steel, how do you make an internal tapered NPT thread in that??
[12:41:06] <archivist> awallin, that is one thing I was going to add to linuxcnc but need more round tuits
[12:41:31] <JT-Shop> I like this sig line "The two most common elements in the world are hydrogen and stupidity."
[12:41:54] <archivist> you can use g33 and do some maths to fix its pitch
[12:41:57] <awallin> so the tapered threads are made on a lathe always?
[12:42:37] <archivist> I have a taper turning attachment on my lathe for that sort of work
[12:43:39] <kwallace2> You could also thread mill a tapered thread.
[12:44:16] <archivist> with a single point tooling
[12:45:54] <IchGuckLive> on a mill there is the simple G2/3 with the R and P Z word to do a tappert since 2.4.3
[12:46:02] <kwallace2> Or single row tool?
[12:46:54] <archivist> thread tips top as well
[12:46:57] <awallin> hm, so do you need a 4-axis mill to make the pilot hole before 1-point threading?
[12:48:04] <archivist> not sure how one would do a tapered helix without 4 axis
[12:50:42] <IchGuckLive> why 4 as the helix is always along Z
[12:50:53] <IchGuckLive> and so your movement is only XY Z
[12:51:22] <archivist> it is a tapered helix
[12:51:37] <IchGuckLive> http://www.supernec.com/images/taperedhelix1.png
[12:52:29] <archivist> the path for a pipe thread is a one liner with linuxcnc and a 4th axis
[12:54:41] <archivist> 5 axis it you want to clear the tool from the side of the groove properly
[12:55:05] <archivist> if milling
[12:55:48] <archivist> tilt to the helix angle but travel along the taper vector
[12:57:58] <IchGuckLive> http://www.youtube.com/watch?v=1GgxyfX0yuI
[12:58:00] <Tecan> (1GgxyfX0yuI) "konisches Gewinde fräsen" by "Steffen Maier" is "Tech" - Length: 0:01:32
[12:58:18] <archivist> top middle shows the tilt before I got the 5th axis an allen key
http://www.archivist.info/cnc/stage6/
[12:59:32] <archivist> at 600 psi he needs the correct form thread
[13:03:53] <IchGuckLive> im of by
[13:07:39] <ReadError> off by what
[13:07:42] <ReadError> kjk
[13:07:44] <ReadError> jk
[14:21:29] <false> PCW: you in?
[14:22:02] <PCW> Yeah
[14:22:49] <false> I connected the old NC and took the scope to the vcmd pin:
http://tinypic.com/r/2hi7148/5
[14:23:35] <false> This waveform shifts with command voltage, no variation in period or amplitude
[14:23:47] <false> the channel is at 100us
[14:24:06] <false> Something to take into account, or just normal?
[14:25:27] <PCW> Odd 2500 Hz modualtion
[14:29:04] <false> It's exactly the same on all vmcd pins
[14:29:15] <false> could this be some kind of noise rejection?
[14:29:59] <false> The drive would only sample the correct voltage, but that would require a clock sync between the drive and nc right?
[14:30:58] <PCW> so this 2500 Hz sine wave rides on top of the velocity command?
[14:31:42] <false> You could put it like that, if the output is 5V the sine is centered arount that
[14:31:52] <false> but still same amplitude
[14:32:05] <PCW> at its 400 mV P-P?
[14:32:11] <false> Yes
[14:33:29] <false> Fanuc might be going fanuc style on this one, meaning different :P
[14:33:56] <PCW> No idea really (its above the drive bandwidth)
[14:34:56] <false> So it should not have any effect, and by that I mean that the output of the 7i33 should be fine without this 'extra' signal?
[14:36:58] <PCW> I would guess so
[14:38:54] <false> Ok I'm gonna keep searching, sorry to keep bothering you with this, but I'm just very very curious as to why this isn't working like it should.
[15:17:13] <andypugh> I wonder if it is an anti-stiction dither?
[15:18:32] <andypugh> PCW: I am thinking of a parameter that defines the end bit of up to 4 fields in the SSI data.
[15:19:31] <andypugh> So, for my 1 fault bit and 24 data bits I would set a param to 0x1901 and end up with two active outputs.
[15:20:43] <andypugh> Though I don't then know what to do with the encoder scale. Simplest is probably to scale every field.
[15:21:53] <andypugh> Unless I do it as a load-time parameter. That might make more sense.
[15:22:03] <PCW> Yeah i was thinking of just begin/end for each field and a separate bit length
[15:23:36] <PCW> so fields are count, count-dup, opt1,opt2
[15:24:22] <andypugh> Madparam parsing breaks at commas.
[15:24:35] <andypugh> It's annoying that way
[15:25:27] <PCW> Ahh I thought these were just HAL parameters
[15:26:35] <andypugh> What did you mean by fields?
[15:26:56] <PCW> thats 10 hal parameters, a bit excessive
[15:27:42] <andypugh> And a lot of HAL pins if you output raw and scaled/wrapped and boolean for every field.
[15:28:41] <PCW> There is really only one field that needs scaling (the count field)
[15:28:56] <andypugh> Yes, but which one is that?
[15:29:54] <PCW> the one named count_field
[15:30:45] <PCW> 2 u32s would do for all bit fields
[15:30:49] <andypugh> "b1c24i12" might be a workable format. Creates a boolean pin which follows bit 1, a count field ending at pin 24 and a simulated index on bit 12.
[15:31:50] <andypugh> If there are more than 32 bits then the module fills the data2 registers, I assume?
[15:32:02] <PCW> Yes
[15:32:20] <andypugh> And so we really need to think about conctenating those both into a 64-bit reg before doing anything else.
[15:33:08] <PCW> normally this would only be done if you are doing double reads for verification
[15:33:39] <andypugh> The main limit is that you can't create pins at run-time, so you either have to pass in a format at load time, or cover all eventualities and drive the pins based on parameter values.
[15:34:14] <andypugh> That Fanuc code you had last night was more than 32 bits. (how many was it?)
[15:34:37] <PCW> Fanuc has 76 + a start bit
[15:34:44] <PCW> (its not SSI)
[15:34:57] <andypugh> For added excitment I am expecting this module (abs-encoder) to transparently cover the other formats too
[15:35:31] <PCW> You will like hearing that Fanuc has a CRC :-)
[15:36:08] <andypugh> I have just remebered that we have existing code for splitting out 96-bit smart-serial data.
[15:36:42] <andypugh> And I would anticipate the CRC being handled externally.
[15:36:54] <PCW> Yeah this is quite similar to sserials unpacking
[15:37:54] <andypugh> So, rather than querying the card, it has to pick up the same data from the modparam, and then it is all very similar.
[15:38:30] <PCW> Fanuc has CRC, count,turns, index, absolute commutation track, battery good, etc
[15:39:02] <false> just out of curiosity, you guys are talking about decoding the famous fanuc serial encoders?
[15:39:18] <PCW> I like your list idea so number,letter repeated
[15:39:24] <false> infamous that should read :P
[15:40:58] <andypugh> So, codes. b-boolean, c-counter, g-gray-code counter, i-index bit, anything else?
[15:41:07] <PCW> 53b1u24c24a3r1e
[15:41:09] <PCW> = 1 bit unused 24 count 24 alternatecount 3 crc 1 error
[15:41:42] <andypugh> index is a bit of a nuisance
[15:42:19] <andypugh> 16.8c for an index?
[15:42:37] <PCW> yeah index and gray are not part of the bitfileld string
[15:43:45] <andypugh> I think gray is probably a synonym for counter. They surely don't have gray-coded bit fields and CRCs?
[15:44:25] <andypugh> What am I talking about. I know they do. The device on my floor has that
[15:44:49] <andypugh> The LSB is an error bit, but only after gray-decoding.
[15:45:42] <PCW> Would not have expected that one
[15:46:01] <andypugh> Actually, I am not sure that it is. How annoying.
[15:46:04] <andypugh> Hmm, how to tell?
[15:47:06] <PCW> cause an error and see if its independent of the count?
[15:48:41] <PCW> unfortunately my one SSI encoder has no extra bits and its all gray (count+turns)
[15:48:51] <andypugh> If I turn off gray-decoding I only see even numbers with a good reading and odd numbers with the reflector covered.
[15:49:23] <andypugh> So no, the error bit is not included in the gray-code scheme
[15:49:49] <PCW> Thats what I would have expected
[15:51:05] <andypugh> So, 16.8g as a code means 16 bits per turn, 8 bits of muti-turn and an index on bit 16.
[15:51:17] <PCW> false: andy is working on generic absolute encoder code (above the lowest level, they are all pretty similar)
[15:52:28] <andypugh> I guess I could combine the format string handler in the LCD component with the parameter extractor in the sserial code :-)
[15:52:31] <PCW> Yeah I might do 8.16 since i thin of numbers that way
[15:52:46] <false> PCW: ah is see, this for addition to hm2?
[15:52:55] <PCW> yes
[15:53:33] <false> that will make a lot of people very happy I think, I'll let you to get to it. Thanks again!
[15:53:39] <PCW> HM@ has low level support for SSI, BISS and FanucSerial
[15:53:50] <PCW> HM2
[15:55:12] <PCW> I also have a modern Yaskawa motor/encoder to decipher
[15:56:15] <PCW> just 2 pins!
[16:37:36] <JT-Shop> see you guys Sunday afternoon
[16:39:41] <andypugh> Have fun, whatever you are doing
[16:40:26] <skunkworks> http://groups.yahoo.com/neo/groups/mach1mach2cnc/conversations/messages/141607
[16:59:46] <Parre> I've got a Teensy++ 2.0 and a few pololu stepper drivers. any idea if theres a readymade firmware that i could use in a laser cutter with that hardware?
[17:01:38] <CaptHindsight> Parre: not using Linuxcnc
[17:02:02] <Parre> okey :/
[17:02:24] <CaptHindsight> maybe you're thinking of Sprinter for glue gun printers
[17:03:58] <Parre> CaptHindsight: No idea. I just though it would be nice to be able to use what i have :)
[17:04:40] <CaptHindsight> Teensy has too teensy a micro to be supported by Linuxcnc
[17:05:24] <andypugh> Parre: LinuxCNC doesn't use any USB devices. What LinuxCNC is is a CNC controller that runs on a Linux box. The point is that all the motion planning happens in real time under Linux.
[17:06:04] <Parre> andypugh: and how are the stepper hooked up? Confused :S
[17:06:15] <andypugh> Anything with a USB interface has to run the motion controller on the far side of the USB connnection, because USB is too indeterminate
[17:06:58] <andypugh> And at that point the machine controller is in the USB device, not the PC/Beagle/Pi
[17:07:51] <andypugh> Parre: You can connect the stepper drives to any of several interfaces. PCI is very good. PCIe works too, as does Ethernet (in development).
[17:07:53] <CaptHindsight> Parre: using a Parallel port or a PCI/e fpga to output STEP and DIR for a stepper driver
[17:08:14] <Parre> Now i get it. Thanks.
[17:08:14] <andypugh> Or, as he said, Ye Olde Parallelle Porte
[17:09:37] <andypugh> We also have drivers for a couple of ISA cards
[17:11:05] <CaptHindsight> http://en.wikipedia.org/wiki/Industry_Standard_Architecture back when PC's only had 16 bits
[17:11:12] <andypugh> In theory the IDE hard drive port could have been used
[17:12:21] <andypugh> CaptHindsight: I wrote one of the ISA drivers only a year or two ago. Actually the card (PCL720) looks potentially quite useful and the PCI version ought to work with exactly the same driver.
[17:19:49] <CaptHindsight> my kingdom for a double row angular contact ABEC 9/ISO P2 50mm ID bearing block
[17:21:45] <andypugh> I have always made my own blocks. I buy the bearings though.
[17:22:31] <CaptHindsight> it's hard to even find the bearings in stock vs 60 days out
[17:23:52] <CaptHindsight> hardly anyone stocks any super precision and preloaded angular contact types
[17:25:21] <andypugh> What's the application?
[17:26:59] <andypugh> I am a cheapskate, and so the preload comes from the housing on the stuff I make. And I have never bothered with anything but normal precision. For a spindle I can see the point. I am not sure there is much to be gained in most other applications. (And this includes ballscrew mounts)
[17:33:36] <CaptHindsight> http://search.newport.com/?q=*&x2=sku&q2=RV120HAT similar to this
[17:37:00] <CaptHindsight> http://bearingsdirect.com/7910CTDULP4-Duplex-Precision-15-Degree-Angular-Ball-Bearing-50x72x12-ABEC7.html
[17:40:52] <andypugh> Pretty!
[17:41:20] <uw> $559 give me a fkin break
[17:41:33] <andypugh> But I would genuinely be asking what $500 bearings give you that $50 bearings done.
[17:41:44] <uw> ^seriously
[17:42:33] <uw> id find a cheaper way do what needs to be done
[17:42:36] <andypugh> Don't get me wrong, the super-precision beariings have their role.
[17:42:44] <CaptHindsight> less radial clearance and runout
[17:43:17] <andypugh> Radial clearance on any preloaded angular contact bearing is zero, isn't it?
[17:43:30] <andypugh> Runout is a fair point.
[17:44:11] <CaptHindsight> you can say the same for that rotary stage at $9k
[17:44:34] <andypugh> But runout on a rotary table (if it was in the race rather than the balls) would be something you didn't even notice. You will lose it when you centre the work.
[17:46:42] <andypugh> I think you would need an incredibly stiff machine to notice any defficiency in rotary stage bearings too.
[17:47:10] <CaptHindsight> yeah it's really stiff and non contact
[17:47:26] <andypugh> The stage itself is quite expensive, but those don't have a cheaper alternative really.
[17:47:54] <andypugh> (Or, more specifically, the cheap alternatives are definitely worse)
[17:48:34] <andypugh> Non-contact makes some difference, but how does the runout in a standard bearing compare to the spark gap?
[17:48:36] <CaptHindsight> plus no high volume applications yet to attract Chinese makers :)
[17:48:40] <uw> *yawns* could be done cheaper
[17:49:31] <CaptHindsight> bearings direct is on the high side, with some shopping you can find them for <$100
[17:49:39] <andypugh> uw: Well, yes, but a $50 rotary table probably won't do the job of that rotary stage. I think it is possible that the $50 bearings will.
[17:49:47] <CaptHindsight> i think the $500 is for a pair
[17:50:54] <andypugh> I would build it with cheap bearings then upgrade if they prove to be a weakness (especially if time is an issue)
[17:51:32] <uw> CaptHindsight, $100 is more reasonable...
[17:51:38] <uw> andypugh, i would do the same
[17:51:51] <uw> cheap until it breaks or proves not up to the job otherwise
[17:52:50] <CaptHindsight> http://en.wikipedia.org/wiki/ABEC_scale
[17:53:43] <CaptHindsight> price nearly doubles for every class
[17:55:08] <uw> I believe it, however i really dont know where an "ABEC 9" bearing would provide you with any better this or that
[17:55:20] <uw> accept stuff that spins at 200k rpm and up
[17:55:42] <uw> maybe it's because i've never worked with anything that spins at 200krpm?
[17:55:59] <andypugh> Quite. ABEC 1 is more precise than any single thing I have ever tried to make.
[17:57:42] <CaptHindsight> so for low speed you'd just use a brass bushing?
[17:58:07] <PCW> I think 30K and up is getting into the realm of air bearings
[17:58:07] <andypugh> And as I mentioned earlier, in the context of a rotary table you don't know where the bores are anyway, you find the centre of rotation and work to that. You could have a bearing with 1" of eccentricity and your work would still clock up true.
[17:58:30] <uw> well i've done alot of work on my atlas 6" and that has bushing for it's head?
[17:58:38] <uw> depends on the precision you need
[17:59:34] <andypugh> uw: Bad example, all the _really_ super-precision lathes had bushings.
[18:00:29] <andypugh> DSG did some with both a ball bearing (for high loads) and a white-metal bush (for precision) at the spindle end.
[18:00:54] <uw> yea my super precision is like ~ .0005" i'm proud of myself (lucky) when i can make something down to that LOL
[18:01:37] <uw> i was reading something on babbit bearings being pretty good but theres something else i dont have much experince with
[18:01:49] <uw> i think they are peuter alloy
[18:01:59] <uw> putier?
[18:02:02] <uw> idk how to spell it
[18:02:08] <andypugh> Whilst early versions of the 17T had plain spindle bearing, later models were generally fitted with a set of the highest quality Timken taper roller bearings - though the option of a plain bearing headstock continued to be available. Some specials were also built, notably for Rolls Royce, where the headstock was a special, larger casting with a spindle that ran in taper-rollers but also a hand scraped, plain white-metal
[18:02:08] <andypugh> bearing in front of the front taper-roller bearings. The bearing could take the fitter Donald Booth several days to scrape in (using red lead not "blue" and maintenance engineers report that these lathes never require any headstock work. The bearing was to ensure that, when a test piece was put on the Tallyrond (a machine to test surface finish and absolute roundness) no "bearing marks" were visible. The spindle and its
[18:02:09] <andypugh> three bearings (the front plain was 4.625 inches in diameter and five inches long) were of massive proportions - with the mandrel bored through to clear a 3.5-inch diameter bar.
[18:02:40] <andypugh> Pewter
[18:02:52] <uw> ah thats it
[18:02:57] <andypugh> But the bearings are mainly tin.
[18:05:01] <andypugh> Here is a white metal bearing I poured and then machined for the 1916 fire-engine:
https://picasaweb.google.com/lh/photo/q2WCf9DdVhy2IA-vBNcqZtMTjNZETYmyPJy0liipFm0?feat=directlink
[18:05:44] <uw> very cool do you make lathe head bearings?
[18:05:57] <CaptHindsight> http://inceptionforums.com/images/simons/5axistrunnion7.jpg any idea what HAAS uses for trunnion bearings?
[18:05:58] <uw> my lathe actually needs new head bearings
[18:06:49] <andypugh> And, by a curious coincidence, here is some Pewter work I did (a drinking cup lid)
[18:06:49] <andypugh> https://picasaweb.google.com/108164504656404380542/PotLid?authuser=0&feat=directlink
[18:07:29] <CaptHindsight> metal spinning on the lathe?
[18:07:38] <andypugh> Yes
[18:07:54] <uw> oh by your plugs i'm guessing you're not in the US
[18:08:11] <andypugh> http://youtu.be/IOBxmGZR3j0
[18:08:13] <Tecan> (IOBxmGZR3j0) "Metal spinning with improvised equipment." by "andy pugh" is "Tech" - Length: 0:06:53
[18:08:39] <uw> did you make the pewter yourself andypugh?
[18:08:42] <andypugh> No, I have proper manly plugs for proper manly electricity :-)
[18:09:03] <andypugh> uw: No, I bought it. I feel like such a fraud.
[18:09:35] <uw> LOL I'll take that. as it is bullshit how 220v isn't the normal wall outlet here in the US :(
[18:09:36] <CaptHindsight> thats right UK and Hong Kong use those giant plugs
[18:09:55] <uw> is UK 50hz?
[18:10:14] <andypugh> The plugs have a fuse in them. There is a reasonably good argument in favour of this.
[18:10:21] <andypugh> Yes, UK is 50Hz.
[18:10:50] <uw> ah so we at least got the 60hz i guess
[18:11:02] <CaptHindsight> better a fuse on the plug than the conductor as a fuseable link
[18:11:11] <uw> still id rather have 220v std outlets
[18:12:26] <andypugh> CaptHindsight: Yes, exactly. In principle if you are plugged into a 40A main then every appliance flex ought to be 40A rates.
[18:14:46] <CaptHindsight> http://www.amazon.com/Nachi-Angular-Contact-Bearing-Shielded/dp/B0045DUY9K $93.37 for a double row
[18:17:34] <CaptHindsight> and half that on ebay
[18:19:11] <andypugh> At half that I would give it a go and see how it works out. If only to get the equipment comissioned while waiting for the 'spensive ones.
[18:21:37] <andypugh> It might even be the case that $50 bearing on a 6-month replacement schedule are techically superior to $500 ones.
[18:25:43] <CaptHindsight> the bearing is mainly for radial loads, the axial moments are almost nil
[18:26:26] <andypugh> Have you looked at 4-point or crossed-roller bearings?
[18:27:03] <CaptHindsight> yeah, I might compare both, the crossed roller are even harder to find in stock
[18:28:38] <CaptHindsight> the bearing distributors are becoming like the auto parts stores, selling inferior substitutes
[18:29:39] <CaptHindsight> SKF, Timken, NTN have long lead times so they have no name or brands you've never heard of with no specs
[18:31:31] <andypugh> I think Nachi are K
[18:31:34] <andypugh> OK
[18:35:53] <CaptHindsight> have to head back to the shop before it gets even warmer
[19:24:22] <andypugh> Anybody got a big lathe and needing a tool changer?
http://www.ebay.co.uk/itm/Dorian-Smartdex-8-station-tool-post-tool-changer-Manual-or-CNC-lathe-99p-nr-/181211320420
[20:57:22] <uw> wow thats a good price
[20:58:09] <uw> id prolly try to turn that into a 4th axis for my mill
[20:58:38] <skunkworks> wow
[20:58:40] <skunkworks> http://www.practicalmachinist.com/vb/dmg-mori-seiki-gildemeister-maho-cnc/fix-your-maho-thread-270525/index3.html#post2111490
[20:58:56] <skunkworks> i guess I hear linuxcnc sucks....
[20:59:17] <uw> got a tldr for that thread post?
[21:07:47] <skunkworks> tldr?
[21:08:02] <Tom_itx> it's code
[21:08:10] <Tom_itx> you need a secret decoder ring to decipher it
[21:08:32] <uw> Top Less Dancer Rental
[21:08:49] <skunkworks> heh - sorry - had to look it up. No - you just need to read it.
[21:08:50] <skunkworks> :)
[21:17:36] <uw> mehh
[21:21:15] <Tom_itx> seems he's raggin on everybody
[21:26:48] <uw> I read the 4th paragraph because that had "LinuxCNC" in it. yea, it sounds like he was too dumb and/or lazy to figure out how to use it and it was more time-efficient to pay for some company's software.
[21:26:50] <uw> good for him
[21:30:26] <Tom_itx> what contouring doesn't lcnc do that others can?
[21:30:57] <Tom_itx> he said the mesa stuff had more io
[21:32:04] <Tom_itx> i think the focus of his bash was the camsoft and Galil card
[21:32:41] <Tom_itx> if he's looking for state of the art cad cam he's gonna pay state of the art prices as well
[21:32:54] <uw> beats me. I dont know what he was talking about but the overall theme was "I've done alot of internet reading and haven't really tried anything for myself and from that, i dont think I'm going to use linuxcnc"
[21:32:59] <Tom_itx> i bet he's not committed to that
[21:33:33] <Tom_itx> some ppl are doers and some are followers
[21:33:45] <uw> then you have the complainer
[21:33:59] <uw> and the moochers, who expect everything to be done for them
[21:34:41] <kwallace1> I think the problem is that 3D contour paths tend to be very short linear moves. LinuxCNC normally sees linear moves one at a time, or, accelerate from stop, travel, decelerate to stop. This goes very slowly, unless you tell LinuxCNC to merge the lines, which sacrifices accuracy.
[21:36:20] <kwallace1> For some people this is unacceptable due to very slow feed rate, or lack of accuracy.
[21:36:28] <uw> so i guess he wants a machine that can do 100parts/hr?
[21:36:53] <uw> welp i guess he can pay then
[21:37:37] <Tom_itx> how much deviation is experienced in reality with contouring on lcnc?
[21:37:58] <Tom_itx> you can expect some even in good quality cad cam generating cutter paths
[21:38:20] <Tom_itx> you can adjust it down but the files get alot bigger
[21:38:37] <kwallace1> I haven't study this, but it has been an issue for a long time, and sometimes heated.
[21:39:24] <Tom_itx> i can imagine
[21:40:18] <kwallace1> From my understanding, one can adjust the CAM to give longer lines, or create arcs. Or you can tell LinuxCNC to round the corners by an adjustable amount.
[21:41:58] <Tom_itx> generally, if i tell my cad cam to follow closer the lines are shorter and the file is considerably bigger
[21:42:13] <Tom_itx> we did alot of surface cutting
[21:42:18] <Tom_itx> aircraft stuff etc
[21:43:16] <kwallace1> A big issue has to do with how far in the motion que LinuxCNC can look ahead, some CNC controls have no problem with this.
[21:44:31] <Tom_itx> so it can predetermine the accel decel for the moves it would need a decent look ahead buffer
[21:45:06] <Tom_itx> i've seen machines chew thru a good size file very quickly but it was all sort contour line moves
[21:48:28] <Tom_itx> i'm sure it's tough to arrive at a balance
[21:51:16] <Tom_itx> i wonder if the guy really knows if he needs that accurate of contouring or not or if he's just belly aching
[21:54:57] <kwallace1> I have no idea, and I have only been on the sidelines watching others duke it out.
[21:55:59] <Tom_itx> me either but i can see the concerns. i just wonder if the accuracy is demanded or if it's just that they think they need it
[23:02:48] <someone972> Is there a resource for compiling LinuxCNC for Fedora?
