#emc | Logs for 2009-10-06

[00:00:24] <i-pink> if i put the phototrnsistor in dark id 300K and in light is 20K
[00:02:28] <andypugh> what happened to the LED from the mouse?
[00:02:49] <andypugh> You probably need a test circuit, something like this
[00:02:51] <andypugh> http://imagebin.ca/view/QIn5XyvJ.html
[00:03:28] <andypugh> Though the mouse sensor is probably meant to work with a slot, not white/black card.
[00:03:34] <i-pink> no
[00:03:55] <i-pink> the led from the mouse? is IR
[00:04:17] <i-pink> and i need to see black and white
[00:04:40] <andypugh> I used those reflective sensors I linked, and they are Infra red, and work with black-white paper
[00:05:16] <andypugh> The encoder will be black and white in infra red too
[00:06:03] <andypugh> But it probably has a filter, and yellow LED light won't get through.
[00:06:52] <i-pink> maeby to put alominum foile on the "white" area?
[00:07:16] <andypugh> See if you can use the mouse LED. You have a digital camera, you can use that to check that the LED is lit, as camera detectors can see the Infra red. Use the camera viewfinder as an infra-red viewer.
[00:07:50] <andypugh> You shouldn't need to, though your card was less white than one might hope for
[00:08:08] <i-pink> i know it
[00:08:50] <andypugh> Can you laser-print the wikipedia gray-code disc and cut off the outer edge?
[00:09:06] <i-pink> is very very very weak led
[00:09:29] <andypugh> Yes, you really want a focussed one for reflective sensing.
[00:10:19] <andypugh> But is is worth trying with the one the phototransistor is optimised for
[00:10:26] <i-pink> what do you mean focussed?
[00:10:49] <i-pink> if the IR light is very weak is not work...
[00:11:34] <andypugh> The ones I used have a little lens in front of the photoransitor and the led. There is a spot 3.8mm in front of the sensor where they converge. Anything there makes the transistor conduct, but further away (or closer in) doesn't
[00:12:34] <andypugh> It might work, at the right distance. As long as you get _some_ change in conductance you can do the rest with a comparator (see that circuit diagram I posted)
[00:12:53] <i-pink> i replace the led now to very high power white led
[00:13:22] <andypugh> The 20k pot lets you adjust the sensitivity
[00:13:44] <andypugh> I don't think while LEDs give out any infrared, but I might be wrong
[00:14:34] <andypugh> But anyway, good luck. I need to sleep.
[00:14:49] <andypugh> Let us know how you get on tomorrow.
[00:15:17] <i-pink> it look good
[00:16:20] <AchiestDragon> i-pink: do you have a time limit on how long its going to take you to sort out a driver and optical encoder the z axis motor ?
[00:17:49] <i-pink> time limit ???
[00:22:09] <AchiestDragon> no disrespect but given what electronics questions you are asking i suspect not ,, it can take many months to learn enough to be able to undersand enough to put together a driver and encoder needed to drive that motor , and in the end would cost the same if not more than just buying a stepper for that and using drivers like you have on the other axis's
[00:23:52] <i-pink> but in my place to bay stepper is expencive
[00:24:02] <i-pink> and is my hobby
[00:24:05] <AchiestDragon> so i would reccomend eather the stepper option or some prebuilt servo drivers
[00:25:07] <AchiestDragon> yes i do understand that ,, but it can often cost as much in parts to build as it would do to buy it prebuilt unless you also have access to cheap parts
[00:25:41] <i-pink> yesssssssssssssssss
[00:25:46] <i-pink> it worrrrrrrrrkkkkkkkkkkkkk
[00:26:27] <AchiestDragon> theres a few places in china that that you can get steppers from cheaply depending on your countries import tax ,, but that tax would affect parts also
[00:28:29] <AchiestDragon> arround £40GBP would buy a stepper that i think may be sutable for your needs
[00:28:37] <AchiestDragon> it could cost you that for just the fets to drive that motor
[00:28:44] <i-pink> but my hobby, take junk and build something
[00:29:35] <skunkworks> :) that is great!
[00:30:26] <AchiestDragon> my hobby also but it helps if you have sutable junk
[00:31:02] <AchiestDragon> http://cgi.ebay.co.uk/CNC-Mill-Router-NEMA-23-Stepper-Motor-High-Torque-28kg_W0QQitemZ390099529133QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item5ad3be6dad&_trksid=p3286.c0.m14
[00:31:30] <i-pink> US $58.90
[00:31:58] <AchiestDragon> £37.06 gbp
[00:32:20] <i-pink> this is my motor
[00:32:21] <i-pink> http://www.youtube.com/watch?v=MA_kd_3oM3k
[00:32:27] <i-pink> 0$
[00:34:13] <AchiestDragon> cool
[00:35:05] <i-pink> you see the DC motor?
[00:35:45] <AchiestDragon> the one with the gearing on it
[00:36:01] <i-pink> no
[00:36:17] <skunkworks> http://www.youtube.com/user/samcoinc#play/uploads/5/eUmFKOVepYY
[00:36:32] <i-pink> in youtube all is bipolar stepper
[00:37:10] <i-pink> cat?
[00:38:41] <i-pink> Servo h-bridge running with emc2 - coool
[00:38:44] <jymm> http://www.youtube.com/watch?v=6drMZqmyXQc&NR=1
[00:38:52] <AchiestDragon> browser here does not like that picture share site you uploaded it to , wants to install tracking cookies
[00:39:27] <i-pink> heheee
[00:39:49] <i-pink> easy mill...
[00:39:55] <i-pink> smalllllllll mill
[00:40:12] <i-pink> i want A4 table
[00:41:47] <AchiestDragon> http://www.whipy.demon.co.uk/Cnc1.jpg < thats my other home made cnc
[00:42:04] <AchiestDragon> 24" by 12" table travel
[00:42:27] <i-pink> this is your CNC!!!!!!!!
[00:42:28] <i-pink> OMGGGG
[00:43:14] <AchiestDragon> built that about 3 years back
[00:43:48] <i-pink> i can marry with you??
[00:44:40] <AchiestDragon> built this on it http://fc09.deviantart.com/fs21/f/2007/290/f/e/tinybug_in_the_garden_oct2007_by_AchiestDragon.jpg
[00:45:50] <i-pink> i what servo tree!!!!
[00:46:22] <AchiestDragon> but too much flex on the legs and the drives were a bit slow so i scraped that and just finished making my new cnc from it http://www.whipy.demon.co.uk/dscf2342.jpg
[00:47:38] <AchiestDragon> neather machine would be up to doing what i need long term so got another in the pipeline just planning it
[00:49:15] <i-pink> and how you make the liniar scale?
[00:50:18] <AchiestDragon> a certain amount of parts i may be able to get from scrap but some parts have to be bought
[00:51:11] <i-pink> ok
[00:51:33] <i-pink> i think to bay the liniar axis
[00:53:07] <AchiestDragon> i used stainless bar and skateboard berings on the old machine ,, the new one i skimped and used 19" rack mount sliders since i had them , but the machine i am planning i will need proper rails and bearings
[00:54:01] <i-pink> is 70$ for one rails
[00:54:59] <AchiestDragon> £600 gpb for all the rails bearings , leadscrews and ballnuts needed in total
[00:55:49] <AchiestDragon> probablay going to take me a year or so to save for those
[00:56:21] <i-pink> i nees 1 or 2 raill for axis?
[00:56:57] <AchiestDragon> 2 rails on x 2 on y and 4 on z
[00:57:15] <AchiestDragon> 2 leadscrews on y and z and one on x
[00:58:17] <i-pink> 4 on Z???
[00:59:42] <AchiestDragon> a 2 foot by 1 foot by 1 foot travel
[00:59:44] <AchiestDragon> then want to fit 2 rotating tables to make it 5 axis
[00:59:50] <AchiestDragon> yea 4 on z ,, the gantry with the x axis on it is raised and lowered to give the z
[01:01:02] <AchiestDragon> reduces tool jitter when the z is at full extent
[01:01:12] <i-pink> i what 3 axis
[01:01:34] <AchiestDragon> need to be able to machine brass and steel on it
[01:02:12] <AchiestDragon> but only at the rates i can get from a 500w 27000 rpm spindle
[01:06:05] <i-pink> 500w 27000 rpm is for the???
[01:06:36] <AchiestDragon> spindle motor ( the cutting tool motor)
[01:06:42] <i-pink> haaa
[01:07:14] <i-pink> i think to use dremel
[01:08:44] <AchiestDragon> the one i just built is ment for a dremel
[01:09:48] <i-pink> this is my plan
[01:09:49] <i-pink> http://hwzone.co.il/community/index.php?action=dlattach;topic=424284.0;attach=121950;image
[01:13:43] <AchiestDragon> lots of changes since i did this but http://www.whipy.demon.co.uk/main-assembly.jpg thats the basic plan i used for the one i just built and the one i am planning
[01:15:00] <i-pink> mmm is solid?
[01:15:15] <AchiestDragon> but just found out that the place that sends all the nice aluminum channel section to the local metal scrap yard here is going bust ,, so not going to get a cheap supply of it by next year
[01:15:30] <AchiestDragon> yea would be ,, just drawn wire frame
[01:16:12] <i-pink> i love google skcach
[01:16:25] <AchiestDragon> previous drawing i did http://www.whipy.demon.co.uk/cnc6.pdf
[01:16:45] <AchiestDragon> but incompleate
[01:17:31] <i-pink> wowwww
[01:17:56] <i-pink> how you work with solid on ubuntu??
[01:19:02] <AchiestDragon> i have 2 main reqirements ,, one is to be able to do 2foot by 1 foot by about 1" thick pannels on 3 axis ,, and need it to work on 8" cube in 5 axis
[01:19:41] <AchiestDragon> cheated , used winxp and autocad
[01:19:56] <i-pink> xp?? OMGGG
[01:20:11] <i-pink> no in my home..
[01:22:16] <AchiestDragon> feel the same way ,, but i have quite a few packages that i cant run on linux ,, mainly altum designer , autocad , and some others that unforunatly theres not a sutable linux replacment for
[01:24:49] <i-pink> but XP is from microsoft
[01:27:11] <AchiestDragon> will try to refrain from slagging microsocks off today , im not a big fan eather
[01:28:03] <SWPadnos> I don't know if they've screwed it up recently, but SolidWorks around 2004-2005 worked very well under VMWare
[01:28:57] <i-pink> how i can get SolidWorks around 2004-2005??
[01:29:00] <AchiestDragon> only have xp installed on one machine here , the other 6 run linux
[01:29:07] <SWPadnos> i-pink, you can't
[01:29:47] <i-pink> :(
[01:29:58] <SWPadnos> and it's a $5000 program anyway
[01:30:04] <SWPadnos> maybe 6000 now
[01:30:21] <i-pink> google is free
[01:30:34] <SWPadnos> solidworks is not
[01:31:34] <ds3> You can get solidworks for free
[01:31:43] <ds3> they have a unemployed user program
[01:31:43] <i-pink> how???
[01:32:06] <ds3> but AFAIK, it is crippled like the EDU versions
[01:32:14] <i-pink> but i need SolidWorks around 2004-2005
[01:32:28] <ds3> why do you need solidworks?
[01:33:24] <i-pink> for modeling
[01:34:19] <ds3> but why that specific version?
[01:34:47] <i-pink> run it on VMWare
[01:34:53] <ds3> oh
[01:34:56] <AchiestDragon> try http://brlcad.org/
[01:35:11] <ds3> I bit the bullet and have a xp cad machine :/
[01:35:14] <SWPadnos> I don't know if it's been screwed up, the last time I trried it was on the 2004 or 2005 version
[01:35:34] <ds3> does alibre run in vmware?
[01:36:03] <i-pink> mmm
[01:36:05] <AchiestDragon> if they got archer sorted for linux the brlcad front end was comming on nicely last time a looked a couple of years back
[01:36:39] <i-pink> google sketchup 7 work good
[01:36:50] <i-pink> but the file is SKP
[01:36:52] <ds3> sketchup is horrible
[01:37:19] <AchiestDragon> if blender had a absolute cordinate sytem that would be usefull also ,,, although theres a couple of tricks you can manage with blender
[01:39:54] <AchiestDragon> create a flat mesh using a cad program like autocad , export as dfx , import it into blender and edditing a mesh in blender gives you a bit better control ,, export it again back to the cad program , and found it helps get arround doing some surfaces that way
[01:40:44] <AchiestDragon> although the blender interface is a total pita to use unless you know all the shortcut keystrokes
[01:41:42] <AchiestDragon> both blender and brlcad are free linux programs
[01:42:02] <AchiestDragon> worth having a play with
[01:45:02] <AchiestDragon> need to hunt for a dfx to g code converter
[01:46:42] <i-pink> but blander is overkill for cnc
[01:46:51] <AchiestDragon> as i usualy create my paths in cad and export as dfx , then inport that to nc
[01:48:15] <AchiestDragon> maybe , but when you get the hang of blender it does things that would take an age in say autocad
[01:48:44] <AchiestDragon> especialy say trying to create a 3d mesh of a face
[01:49:15] <AchiestDragon> without scanning it
[01:51:08] <AchiestDragon> the down side i find with blender is if you only use it once every 3 or 4 months you find you forget half of the commands , so spend longer serching the manual and tutorials than actualy editing images
[01:51:19] <i-pink> OMGGGG 3:50 AM
[01:51:39] <ds3> does blender project 3D objects onto 2D drawings?
[01:52:54] <AchiestDragon> yea advantages of beeing young free and single , although is more old avalable and unwanted now , is no one to tell you to go to bed
[01:56:14] <SWPadnos> err - blender might wrap 2D drawings on 3D objects
[01:57:57] <AchiestDragon> blender has a nice way of letting you edit 3d meshes that i have not seen in cad programs
[02:07:14] <ds3> SWPadnos: I was refering to being able to generate 2D DXF's of an object automatically
[02:07:26] <SWPadnos> ah
[02:07:30] <ds3> even Rhino3D can do the standard 3 view projections
[02:07:59] <ds3> SWPadnos: how's business?
[02:08:03] <SWPadnos> ah - automatically drafting 3D parts - no, I bet it doesn't do that
[02:08:05] <SWPadnos> busy
[02:08:22] <SWPadnos> I've been on the road most of the last 4 weeks, and I'm heading out again tomorrow
[02:08:30] <ds3> lucky you
[02:09:06] <Valen> making lotsa moneys SWPadnos?
[02:09:07] <SWPadnos> heh, yeah
[02:09:13] <SWPadnos> err, yeah to that too :)
[02:09:19] <Valen> noice
[02:09:28] <Valen> now give me some for being so awesome lol
[02:09:47] <SWPadnos> I just added up the expenses I'm billing for the last 3 months, and that's $25k
[02:09:51] <ds3> is EMC (the project) officially related to pico systems in someway?
[02:09:54] <SWPadnos> let alone actually charging for my time
[02:09:57] <SWPadnos> no
[02:10:07] <ds3> unofficially?
[02:10:18] <SWPadnos> Jon Elson was the second or third non-NIST user of EMC, so he's been around a long time
[02:10:30] <SWPadnos> I think he's also the first person to have designed hardware specifically to be used with EMC
[02:10:42] <ds3> looks like they are trying to get EMC into the Beagle board
[02:10:42] <SWPadnos> and of course he was on the board for 5 or so years
[02:10:46] <SWPadnos> yes, he is
[02:10:50] <SWPadnos> you could help ;)
[02:11:16] <ds3> I'd if I wasn't so busy staying out of the Chapter 7/13/etc courts :(
[02:11:20] <SWPadnos> ugh
[02:13:23] <Valen> I wonder what the latency would be like
[02:13:40] <ds3> it should be able to do as well as any PC if done right
[02:14:01] <L84Supper> why are they keeping the ARM port so secret?
[02:14:08] <SWPadnos> jepler had done some experimentation, and there were some many-millisecond pauses (like 400 or more)
[02:14:29] <SWPadnos> but I think that may have been the -RT kernel, not RTAI
[02:14:43] <ds3> -rt exists for a while
[02:15:08] <ds3> the "ARM port" doesn't makes sense... for this stuff, you really need to be more chip specific
[02:16:20] <Valen> where is this discussion?
[02:16:31] <L84Supper> Beagle is TI OMAP, the other popular is the Freescale imx515, what ubuntu-arm supports
[02:16:50] <ds3> L84Supper: and the I/O for the iMX and the OMAP are quite different
[02:17:07] <L84Supper> Ubuntu-ARM supports ARM11 and up
[02:17:36] <L84Supper> yeah, the ARM instruction sets vary
[02:17:51] <L84Supper> and IO is not consistent from SOC to SOC
[02:18:17] <Valen> so where are people talking about this?
[02:18:48] <L84Supper> not sure, I thought i was the only one planning and EMC port to ARM
[02:19:00] <ds3> might be interseting to have EMC (headless) running on ARM926's
[02:19:05] <L84Supper> sp/and/an
[02:19:34] <Valen> any paticular reason for trying to run it on arm?
[02:19:51] <L84Supper> ARM11 soc's are under $10 with 3D/2D accel
[02:20:23] <L84Supper> small board, low cost, acceptable performance
[02:20:29] <ds3> ARM's are cheap
[02:20:33] <ds3> low power
[02:20:36] <Valen> dual core atoms arent that expensive
[02:20:52] <L84Supper> no coreboot for ATOM yet
[02:20:54] <Valen> though i spose if your aiming at hexapods or something rather than mills
[02:21:04] <ds3> more consistant....just because a board has a Pentium it doesn'tmean it will behave the same as another Pentium board
[02:21:28] <ds3> Atoms and x86 has too much crap to keep legacy bios happy... SMM for example
[02:21:38] <ds3> L84Supper: why bother with core boot?
[02:21:49] <Valen> they seem to be posting pretty good latencys though
[02:22:23] <L84Supper> crappy Award bios
[02:22:28] <Valen> I spose what i'm asking is what you are getting out of an arm board over an atom board
[02:22:37] <ds3> on ARM, you have total control
[02:22:48] <L84Supper> smaller size and lower cost
[02:22:50] <ds3> U-boot is a common equiv
[02:22:58] <Valen> yeah, but whats lacking from the atom boards?
[02:23:07] <Valen> they seem to do the job with much less work
[02:23:08] <ds3> sanity
[02:23:17] <ds3> atom boards require a lot more work
[02:23:21] <L84Supper> there is $65 ATOM board but it's mini-itx
[02:23:33] <ds3> there is? got URL?
[02:23:44] <Valen> more work?, compile SMP kernel and emc and its done?
[02:23:45] <L84Supper> newegg
[02:23:50] <L84Supper> 1 sec
[02:24:57] <L84Supper> http://www.newegg.com/Product/Product.aspx?Item=N82E16813121342
[02:25:41] <ds3> hmmm that is too big
[02:25:51] <ds3> if it was only PicoITX sized...
[02:26:01] <Valen> too big for what?
[02:26:10] <L84Supper> yeah, no pico for Atom yet that I've seen
[02:26:20] <Valen> they do exist i believe
[02:26:24] <ds3> there is a PicoITX board for the Atom
[02:26:39] <ds3> think ESC/Boston was giving them away to paid participants
[02:26:42] <L84Supper> sorry I meant for under $100
[02:26:52] <ds3> yes, that is a problem
[02:27:07] <L84Supper> http://www.orbitmicro.com/global/pico820vga-z510-p-10896.html?ref=base
[02:27:17] <L84Supper> $430
[02:27:41] <L84Supper> almost 2x a VIA pico-itx
[02:28:37] <ds3> the other problem with x86 boards is you got to use a monitor which thesedays is 15" or bigger
[02:28:37] <ds3> and touch screens are a pain
[02:29:28] <L84Supper> I guess it depends on what you want to see on screen
[02:29:29] <Valen> a) why do you need a monitor? b) i use an eee all the time which is a non touch screen 8"
[02:30:45] <L84Supper> we're using 7" LCD touchpanel for machine UI, matchbox keyboard is big enough for even fat fingered users
[02:31:13] <Valen> stick a new eeepc in ;->
[02:31:40] <Valen> I'm just trying to understand your need for physical smallness though
[02:31:46] <L84Supper> just don't bump into it
[02:32:23] <L84Supper> mini-itx is larger than 7-8" panel
[02:32:32] <Valen> so?
[02:32:58] <Valen> i have an extended atx motherboard running my mill
[02:33:10] <L84Supper> so the enclosure becomes larger and costs more to make
[02:33:31] <Valen> vs spending heaps and heaps of time getting it all to work on arm?
[02:33:36] <L84Supper> if you had to sell 100k mills, what would you do?
[02:33:47] <Valen> your making 100k of them?
[02:34:27] <L84Supper> not mills, but other things
[02:34:45] <Valen> If your making 100k of them I'd get somebody to make an embedded dual core atom board for me that fit into the space ;->
[02:35:04] <L84Supper> I could roll my own
[02:35:12] <Valen> dual core so nice low latency, and easy to develop on/compatible with stuff
[02:35:55] <L84Supper> ARM's not too hard, been doing it since the mid 90's
[02:36:48] <L84Supper> Linux development exploded in the last year
[02:37:06] <Valen> true
[02:37:40] <Valen> you just need to do more work that has already been done I spose if your trying to make it so embedded
[02:38:48] <Valen> but yeah 100k of them I can see why you would do it the way your doing it
[02:39:03] <L84Supper> yeah, I wish x86 was lower cost, would be simpler, but not by much anymore
[02:39:26] <Valen> atom cores are pretty cheap
[02:39:39] <Valen> its the chipset that costs the $ on them these days
[02:41:39] <L84Supper> it's close but Intel hasn't been too open firmware supportive, they did an opensource QNX fastboot bios for Atom but kept the actual register configs and ram init closed
[02:41:41] <ds3> a 7" panel, sealed touch screen,no fans
[02:41:53] <ds3> the only exposed thing would be a click wheel for jogging
[02:42:31] <ds3> x86 generally need a fan which is a source for sucking in chips
[02:42:41] <L84Supper> maybe in 6 months things will look different
[02:43:40] <L84Supper> there is enough surface are in a 8" lcd enclosure to passively cool an Atom mainboard
[02:44:09] <ds3> except an 8" LCD backlight can generate enough heat to prevent proper cooling
[02:44:19] <L84Supper> once you start counting every penny the x86 BOM is higher than ARM
[02:45:25] <L84Supper> if we needed to do lots of floating point and multimedia then it would be another story
[02:46:26] <L84Supper> I though the first C in CCFL meant cold!?
[02:47:40] <Valen> coolish
[02:47:42] <Valen> ;->
[02:47:54] <Valen> get one with an LED backlight if you want low power
[02:48:37] <L84Supper> yeah the led versions are almost at price parity with the ccfl small panels
[02:48:55] <Valen> probably longer life too
[02:49:01] <ds3> Cold is relative...compared to something that is radiating incandencent light....
[02:49:32] <ds3> LED versions can get pretty warm too
[02:49:54] <ds3> the one I am looking at sucks 10watt peak
[02:53:26] <L84Supper> maybe the PixelQi displays will be more readily available by next year
[02:54:14] <L84Supper> http://www.pixelqi.com/
[02:58:42] <L84Supper> Atom netbooks in Taiwan are down to $200 USD with 10.1"LCD, 1GB ddr and 160GB HD
[02:59:08] <ds3> and the battery life is.....
[02:59:48] <L84Supper> will be interesting to see the street price of the OMAP or iMX515 netbooks this xmas
[03:00:05] <ds3> not enough volume to really drive it down
[03:00:29] <L84Supper> ds3 : yeah that will be the big difference even if the prices are about the same
[03:03:28] <L84Supper> the iMX515 netbooks are supposed to launch at under $200
[03:04:40] <ds3> you can build one off OMAP laptops yourself so I donno how interseting it would be to buy something with specs that someone thinks you want
[03:05:38] <L84Supper> the OMAP designs are to bee in the same price range, I'm waiting to see what actually starts to ship
[03:06:04] <ds3> the availablility of the i.MX51 chips is an unknown
[03:06:51] <L84Supper> you can follow the dram and rumors at digitimes.com
[03:07:32] <L84Supper> drama even
[03:08:06] <Valen> I have a guy who wants a LCD screen based dash
[03:08:24] <Valen> that pixelqi mob seems close but they don't seem to have any actual products
[03:08:31] <Valen> got any ideas on suppliers?
[03:08:31] <cradek> Valen: he's just mistaken :-)
[03:08:49] <Valen> mistaken how?
[03:09:00] <cradek> can't see them in the daylight
[03:09:13] <Valen> thats why you get a sunlight readable screen ;->
[03:09:23] <L84Supper> the pixelqi screen is going into the new OLPC
[03:09:32] <Valen> http://www.pixelqi.com/products
[03:09:35] <cradek> ahh!
[03:10:05] <L84Supper> it uses the sunlight or it can use the backlight
[03:10:06] <cradek> oh can we get those in real computers now? that would be very cool.
[03:10:10] <L84Supper> it changes modes
[03:10:22] <Valen> there are sunlight readable screens around currently
[03:10:26] <Valen> they are just somewhat spendy
[03:10:31] <Valen> take a look at toughbooks
[03:10:57] <L84Supper> Valen : yeah, but the pixelqi does it by using the sunlight
[03:11:04] <Valen> yeah transflective
[03:11:07] <Valen> tis cool
[03:11:10] <L84Supper> digital paper mode and LCD
[03:11:20] <Valen> some phones have transflective screens too
[03:11:32] <Valen> PDA type phones
[03:12:21] <eric_unterhausen> someone on ebay is trying to sell moog tool holders as 30 taper -- sometimes ignorance is no excuse
[03:12:23] <L84Supper> there are some youtube videos of the pixelqi screens
[03:12:55] <L84Supper> they look better than the e-paper screens in the Kindle
[03:19:34] <Valen> I was wondering what sort of tool holding system we should be looking at for our mill
[03:19:45] <Valen> I'd like something compatible with auto tool changers
[03:20:00] <eric_unterhausen> cat 30
[03:20:01] <L84Supper> the ARM SOC's don't have PCI, so if you want a high bandwidth interface to an FPGA there is only the DDR/2
[03:20:22] <Valen> that would be a sucky way to talk to things
[03:20:23] <ds3> there are ARMs with PCI
[03:20:32] <ds3> and there are other fast interfaces
[03:21:02] <L84Supper> I forget which OMAP is in the beaglboard
[03:21:03] <Valen> they have GPIO so you could presumably just treat it like a parallel port
[03:21:08] <ds3> OMAP3530
[03:21:22] <L84Supper> no PCI as I recall
[03:21:34] <L84Supper> same for iMX515
[03:21:56] <ds3> GPMC is a fast interface on theARM
[03:22:02] <ds3> s/ARM/OMAP3/
[03:22:26] <ds3> if you had a board with an ARM and a fast link to a FPGA, what would you use it for?
[03:22:55] <L84Supper> depends on the app or you could just use SPI
[03:23:42] <Valen> load mesas hostmot2 firmware into it ;->
[03:23:43] <L84Supper> that is where all the bickering will take place for EMC on ARM
[03:24:05] <ds3> Valen: and how much would you pay for such a board?
[03:24:27] <Valen> I wouldn't, I'd use a dual core atom and stick a mesa card into the PCI slot ;->
[03:24:44] <ds3> and how much is a dual core atom + mesa card?
[03:25:01] <Valen> about the same price as a decent face cutter
[03:25:11] <ds3> which is?
[03:25:16] <L84Supper> the mesa parallel fpga card is $99
[03:25:28] <Valen> so ~$180 then
[03:25:36] <Valen> cheaper than the facing tool
[03:25:38] <SWPadnos> um, not quite
[03:25:50] <ds3> you can get dual core atoms for $90?
[03:26:07] <SWPadnos> an Atom motherboard is $80-90, but then you need memory, disk, and power supply
[03:26:10] <eric_unterhausen> you need another $150 worth of interface boards for the mesa
[03:26:15] <SWPadnos> and then you need actual I/O for the machine
[03:26:42] <Valen> memory is ~$25, disk is <$50 if you want a new one, case is $25 incl psu
[03:26:48] <ds3> I think you can get an OMAP + FPGA board for about $400-$500
[03:26:52] <SWPadnos> and if you want to hold a PCI card, you also need some riser card thing, or a PC enclosure (or parts of one) to hold the PCI card
[03:27:04] <Valen> that depends on the case
[03:27:20] <Valen> put the mini-itx in a standard atx and you use a normal PCI card
[03:27:40] <SWPadnos> sure, a mini-ITX case will have a riser or hold a PCI card (if you look around), but they're not $25
[03:27:57] <SWPadnos> and a full ATX case isn't nice and small, like the atom board almost is
[03:28:30] <SWPadnos> I think the point ds3 was trying to make is that the atom solution is in the same cost range as a more "embedded" kind of board
[03:28:41] <SWPadnos> *but I could be wrong, I only saw a few lines before I jumped in)
[03:29:44] <Valen> its similar in terms of cost, the atom being more expensive
[03:29:57] <Valen> difference I see is i can do the atom today
[03:30:05] <eric_unterhausen> same here
[03:30:10] <Valen> making an embedded system is going to take a whole bunch of development
[03:30:33] <ds3> that and x86 is obsolete ;)
[03:30:41] <Valen> and unless you are making 100k of them it seems like lots of hard work for not so much benifit
[03:30:47] <Valen> anahoo i gtg do some shopping
[03:30:49] <Valen> bbl
[03:31:11] <eric_unterhausen> trying to buy lathe tooling, too much of a time sink
[03:34:20] <L84Supper> I haven't seen any OMAP or iMX515 boards with FPGA's out of Taiwan yet, but the ARM11 boards with 400K fpga are ~$100
[03:35:56] <eric_unterhausen> L84Supper: sorry to ask, do you have a link to such a board?
[03:48:39] <L84Supper> http://www.embedian.com/index.php?main_page=product_info&cPath=73&products_id=230
[03:49:02] <L84Supper> fpga on daughtercard
[03:51:07] <L84Supper> the Samsung S3C6410 ARM11 and the s3c2240 ARM9 boards are real popular in China
[03:51:37] <eric_unterhausen> thx
[03:53:05] <L84Supper> http://www.mini-box.com/ARM-Systems has ARM9 boards in the US for pretty low
[03:55:33] <L84Supper> from what I hear the pico-itx sized ARM cortex8 boards should be shipping by the end of the year
[03:57:41] <L84Supper> there are a few board houses in Shenzhen that are making ARM boards like crazy for the asian market
[05:21:32] <Mire> computer P/s was a smashing succes. my machine runs quiet except for servo noise now.
[05:22:22] <Mire> thanks all who helped me sort that out.
[05:41:54] <Mire> by the way, i don't know what an
[05:43:26] <Mire> ARM board is, but I put an $85 atom board in my Hardinge and it spins like a top. using a $10 4Gb usb stick for a hd
[05:45:45] <Mire> $40 stuffed it with 2 gigs ram. all the rest was on hand. I'd recommend it to anyone. Best computer I've built is also cheapest!
[05:48:19] <eric_unterhausen> Mire: no problems with latency?
[05:48:28] <Mire> could have gotten by w $60 motherboard(w processor, fan, video, sound, usb, network,...)but I got stuck on bios settings and returned it before I sorted out my mistake. good ol fry's.
[05:48:41] <Mire> none so far, but mine is servosystem...
[05:48:53] <eric_unterhausen> mine is too
[05:49:00] <eric_unterhausen> are you using mesa?
[05:49:31] <Mire> latency test was around 8 or 9 uS, 16 or so when I scribble w mouse and run other stuff.
[05:49:39] <Mire> Yes Mesa 5i20
[05:49:52] <eric_unterhausen> tell me how you did it :)
[05:50:10] <Mire> If there is trouble I've not noticed it. it would stop emc, right?
[05:50:25] <eric_unterhausen> real time error or some such
[05:51:23] <Mire> if they dump silently to dmesg then i might not have noticed...
[05:51:25] <eric_unterhausen> I need to change mobos, the one I have been using has too much latency
[05:51:35] <eric_unterhausen> pretty sure it throws up a dialog
[05:52:39] <Mire> I've had zero problems, except for obscure(to me)bios setting for USB drives.
[05:53:04] <eric_unterhausen> I'm a little surprised that the usb drive doesn't screw things up
[05:54:08] <eric_unterhausen> I need to install a newer version of emc to see if it fixes anything with my comp
[05:56:14] <Mire> well, it did, at first. It wouldn't recognize it as bootable. also, there was a funny bug which wrote bootsector to wrong drive. I don't actually understand why it eventually worked exactly as planned. but I've been very pleased.
[05:57:31] <eric_unterhausen> that's good
[05:58:05] <Mire> It is slow to load at times but has enough ram to be quite fast when loaded. I wonder if there are swap settings which I should change.
[06:02:49] <Mire> My design is a crude copy of cradek's HNC; same Mesa boards and Pico resolver converters. I followed his configs fairly close, wiring wise, but now that I'm getting the system dialed there are changes I mean to make.
[06:04:58] <eric_unterhausen> is that on the wiki?
[06:05:45] <Mire> dunno. i asked him special because I have identical machine.
[06:07:13] <Mire> having working set of configs changed my task to reverse engineering. that saved this forum many hours of tiresome questions, no mistake.
[06:09:27] <eric_unterhausen> what amps do you have?
[06:09:52] <Mire> original GE amps, came w machine.
[06:14:17] <Mire> I'm thinking of making a PyVCP with manual override controls; handles work with servoamps disabled. I could make entries for carriage/crossfeed and axis locks. Good for onesies, and for touching off tools.
[06:16:25] <Mire> then i get to make handles...
[06:20:41] <Mire> say, does anyone know if parport driver can be used alongsided Mesa 5i20? seems I'm out of output lines...
[06:24:28] <eric_unterhausen> do you have 2 7i37 boards?
[06:54:36] <MattyMatt_> hi all, is there a channel for hardware builders?
[06:56:25] <MattyMatt_> " and general machining" I guess it's OK here too :)
[07:01:41] <Mire> eric: yes 2 7i37s, 1 7i33.
[07:04:21] <Mire> MattyMatt, here is usually not bad for any of those topics. emc is the focus, but general machining and design get discussed all the time.
[07:04:47] <MattyMatt_> has anyone made a moving table with filing cabinet shelf slides? They seem stiff enough to me, if not full extended so the tool is always working over the overlapped part
[07:07:37] <Mire> i played with shelf slides and allthread years back. then i got a job in a machine shop.
[07:08:15] <Mire> I would drop the 80 bux or so for a little xy table now that i understand the need for stiffness.
[07:08:45] <MattyMatt_> I only want to run a dremel through wood :)
[07:10:17] <MattyMatt_> I've seen a nice x-y vice, but I can't machine the adaptors to the steppers
[07:11:10] <Mire> Mite be okay. finish can be sanded. buzzing can be a prob because of high rpm of dremmel, but it's nothing a success oriented attitude can't beat.
[07:13:21] <MattyMatt_> the slides are quite springy in one axis, so I'm thinking one flat under the table too, but I'll try it without first
[07:14:44] <MattyMatt_> I only got one meaty motor from 2 dead printers so far.
[07:15:05] <MattyMatt_> they don't make printers like they used to
[07:15:59] <Mire> xy vice is nice(indispensible?)for drillpress, but you want plain t-slot table, if it has the travel. how much travel do you need?
[07:16:49] <MattyMatt_> I'm aiming at 375mm 15"
[07:17:28] <MattyMatt_> yeah Tslot table would be nice. the 20mm stuff isn't too ruinous
[07:18:56] <Mire> that's pretty good size.
[07:19:20] <MattyMatt_> 575mm shelf slides, with 200mm working overlap
[07:19:45] <MattyMatt_> 540 mm wide, IIRC
[07:20:10] <MattyMatt_> it'll be twin screw eventually
[07:20:29] <MattyMatt_> or triple. I'll drill 3 sets of holes
[07:20:36] <Mire> do you have machining experience? do you know how to measure the stiffness w indicator?
[07:21:26] <MattyMatt_> nope. I can clamp one end and waggle it aginst an upright ruler
[07:22:01] <MattyMatt_> but I don't want to do that to the bare slides, they're only pressed sheet
[07:22:44] <MattyMatt_> if I get 0.1 mm accuracy out of this thing I'll be happy
[07:23:35] <MattyMatt_> 4 thous. got enough for woodwork :)
[07:23:49] <Mire> of course. stiffnes when built is what matters. remember that perpendicularity is also important.
[07:24:39] <MattyMatt_> yeah I'm bolting the table to heavy bench
[07:25:25] <MattyMatt_> well, a desk I never finished, shortened and reinforced
[07:27:24] <Mire> You will need a dial indicator before you are done, for perpendicularity and stiffness and for tramming(carefully squaring)vices or workpieces. dreaming up ways to use simple tools is all the fun of machine work.
[07:27:39] <MattyMatt_> I'll get the table straight and true first, do all my controller building, and then start the gantry
[07:28:52] <Mire> You will have major backlash trouble... plan on all cuts moving the same 2 directions.
[07:29:40] <MattyMatt_> stiffness of the Z motion is my big concern
[07:30:00] <Mire> oh, and from actual shelf slide experience, make sure your table doesn't fall away in Z. Ours was shorter when pushed all the way to one side.
[07:30:16] <MattyMatt_> I've got 8 inches of travel drawn up but I think that's trouble
[07:30:58] <Mire> drilling holes? no wait, you just figure the gantry won't be as stiff as the table?
[07:33:20] <MattyMatt_> dunno yet really :) this is all a bit experimental. the gantry upright will be box section at the motor end at least
[07:34:06] <MattyMatt_> all in 19mm plywood.
[07:35:44] <alex_joni> Mire: no problems using parport + mesa
[07:35:49] <Mire> I'd love to hear how it goes. I work in metal, so plywood isn't stiff enough for me anymore. what will you make with it?
[07:36:09] <MattyMatt_> the mk2 :)
[07:36:17] <Mire> thanks alexjoni.
[07:36:43] <Mire> I like how you think, mattymatt.
[07:38:34] <Mire> where are you?
[07:39:15] <MattyMatt_> England
[07:39:33] <MattyMatt_> near Liverpool
[07:39:53] <Mire> long swim from california
[07:40:28] <MattyMatt_> boat parts is something I might try
[07:41:13] <MattyMatt_> not many people brav ethe Mersey in a wooden boat tho
[07:42:59] <MattyMatt_> little models of boats, that's more like it. I could sell them alright
[07:43:41] <MattyMatt_> to buy some refractory, to melt some metal
[07:45:18] <Mire> I'd aim at a router table; bigger xy, less z. cuts patterns out of flat panels. It would be a better mk1... and would cut out a righteous mk2. It's my next project after my lathe.
[07:46:23] <Mire> you melt metal? foundry? metallurgy?
[07:46:32] <MattyMatt_> yeah 8inch isn't much use without u&v or whatever they're called
[07:47:16] <MattyMatt_> melting metal looks easy :)
[07:47:46] <MattyMatt_> with a lionellabs(tm) waste-oil burner
[07:48:03] <Mire> aluminum is easy. lost foam casting.
[07:48:19] <Mire> I'd like to learn about brasses and steels
[07:48:38] <MattyMatt_> same thing, hotter flame
[07:48:50] <MattyMatt_> more expensive crucibles
[07:49:38] <Mire> we hacked together aluminum foundry from items onhand. lost foam is a neat trick.
[07:50:16] <MattyMatt_> ah but then I do need 8inch of z travel to carve the foam
[07:51:16] <MattyMatt_> if I want the headstock on my lathe to be any size \o/
[07:51:45] <Mire> and an overhead crane to move the crucible...
[07:52:00] <MattyMatt_> good point
[07:52:15] <MattyMatt_> engine hoist would do :)
[07:52:47] <Mire> im enjoying this but i have to go. hope to see you tomorrow. goodnite, or for you, good afternoon.
[07:53:13] <MattyMatt_> okdoke. morning I tink (me looks at sky) yeah
[07:53:27] <Mire> or morning, i didn't check. small world after all.
[09:09:58] <piasdom> g'mornin all
[10:00:15] <archivist> robh most likely be at the show on the Friday
[10:02:35] <robh> hi archivist might be able to do the fri
[10:36:12] <Unifilarpole> Lol
[10:49:57] <AchiestDragon_> AchiestDragon_ is now known as AchiestDragon
[11:22:16] <i-pink> hii
[12:20:34] <i-pink> hii all
[12:42:13] <micges_work> hi
[13:17:16] <AchiestDragon> morning
[13:33:41] <tris-> tris- is now known as tris
[14:11:51] <Guest362> Hi All
[14:12:41] <Guest362> Anyone know how to get frequency out of the parallel port instead of PWM
[14:12:57] <Guest362> for the spindle
[14:13:54] <jymm> You want a sinewave from the paraport?!
[14:14:28] <Guest362> square wave
[14:14:57] <archivist> I understand even if jymm doesnt
[14:15:14] <Guest362> need frequncy for my frequency to voltage converter -> VFD
[14:15:51] <archivist> VFD stands for voltage to frequncy
[14:16:15] <archivist> you probably need 0-10v
[14:16:24] <archivist> therefore pwm is ok
[14:16:30] <SWPadnos> VFD == Variable Frequency Drive
[14:16:47] <archivist> and that
[14:17:09] <Guest362> yep it only accepts 0-10v
[14:17:39] <archivist> pwm is what you want then, just make a filter to convert pwm to voltage
[14:17:49] <Guest362> my frequency to voltage converter can accept a wide range
[14:18:20] <Guest362> tried it but its not as linear as the F2V
[14:20:19] <Guest362> http://cnc4pc.com/Files/EMC2.txt
[14:20:45] <Guest362> this is suppose to be able to do it from 0-25Khz
[14:20:57] <Guest362> but not sure how to set it up
[14:21:16] <Guest362> still new to emc2
[15:12:41] <christel> [Global Notice] Hi all, as you might remember we informed you a few days ago that we would be restarting some of our ircds due to a memleak issue -- I'm about to do round two of this and it will be slightly noisy as they restart. Apologies for the inconvenience and thank you for using freenode.
[15:50:36] <AchiestDragon> http://www.whipy.demon.co.uk/dscf2343.jpg well done the frist test on the machine ,,, not too bad concidering its not alligned as yet
[15:51:54] <AchiestDragon> done using a 3mm ball nose cutter so the error shows up quite well
[16:34:05] <AchiestDragon> 3/4 of a degree out on the xy allignment and just under 1/8th of a degree out on z
[16:37:33] <archivist> AchiestDragon, you coming to midland model engineer?
[16:38:41] <AchiestDragon> would of liked to but insufficiant funds atm
[16:39:10] <archivist> I have no job or funds but im going
[16:39:41] <AchiestDragon> not got funds to get there
[16:40:27] <AchiestDragon> would be about £70 in petrol with my gas gussler
[16:41:15] <archivist> its about 2 gals of deseasal for me
[16:42:45] <AchiestDragon> 4L jeep here ,,, get about 200 miles on a full tank , takes about £64 normaly to fill it
[16:43:20] <archivist> heh not a practical toy for long distance
[16:45:55] <AchiestDragon> true ,, but localy its usfull to have the 4wd ,, and only do 2 to 3 200 mile long trips a year and about 3000 miles a year so it does not get a lot of use
[16:47:11] <archivist> I do about 10k per year, will be cutting down uless I pick up some jobs
[16:48:11] <AchiestDragon> will be at local radio ham club tonight theres a few there that are interested in moddle helecopters and airoplains maybe some of them may be interested ,
[16:48:54] <AchiestDragon> will have to find the flyer i have for it to show them
[16:48:59] <EbiDK> EbiDK is now known as EbiDK|AWAY
[16:54:18] <AchiestDragon> anyway time to go out ,,, bk late tonight
[17:49:57] <piasdom> anyone uses FrontPanel Express ? ... i need to convert a *.fp file to *.dxf
[18:01:38] <BridgeportII> cradek: Hi, this is KimK, still in Minnesota. Have been doing other things, now back on Bridgeport II with John. Got time for another quick Q on CL? EMC2 is objecting to this: classicladder.0.s32out-2 does not exist. Shouldn't s32out-2 be %QW2 ? Got any advice for us, or things to check?
[18:17:03] <cradek> BridgeportII: comment out the line causing trouble. then you can start emc and use one of the hal tools to see the available pins and troubleshoot.
[18:17:28] <cradek> darn, if you're still in MN that means you won't be having dinner with us tonight on Omaha.
[18:18:47] <BridgeportII> John says, "Are you buying?", LOL
[18:19:53] <cradek> uh-oh
[18:21:35] <BridgeportII> We have been able to start by commenting out the bad line, but we'll go back and have another look at the HAL display. We'll let you know of more results in a bit. Thanks!
[18:22:07] <Didier> Hello everybody
[18:22:51] <eric_unterhausen> Hello Didier, in general it's best to ask a question vs. saying hi
[18:23:01] <cradek> also, hi
[18:23:22] <Didier> Somebody can say me if it possible to drive stepper motor with encoder ?
[18:23:47] <eric_unterhausen> it has been done
[18:24:19] <Didier> there is link to see, or to have explanation
[18:24:20] <cradek> can you say in more detail what problem you are trying to solve where you think using steppers and encoders together is a good solution?
[18:25:24] <Didier> it's just to known feasibility
[18:26:16] <Didier> if closed loop is possible with new stepper motor equiped with encoder
[18:26:19] <cradek> you can use encoders to detect when a stepper motor stalls due to overload
[18:26:45] <Didier> it's just detection of error or really driving it
[18:26:46] <cradek> in that case, EMC will abort with a following error instead of blindly continuing
[18:27:11] <cradek> stepper motors do not need encoders to drive them. I don't know what "really driving" means so I can not answer that question
[18:27:44] <jymm> I think by "really driving" he means closed loop
[18:28:06] <Didier> then it's a open loop (but stepper) with error detection, it's that's ?
[18:28:12] <cradek> jymm: that's still a fuzzy question and why I asked for a more specific one.
[18:28:43] <cradek> Didier: steppers are inherently open loop. you give them a step, they move, or if they are overloaded, they don't move. it is true that you can detect which thing happened, using an encoder
[18:29:28] <eric_unterhausen> you could get better microstepping
[18:29:28] <cradek> if they do not move, there is NOTHING you can do to make them move, so the best you can do is stop and error
[18:29:41] <Didier> i have my answer
[18:29:47] <cradek> if this is what you mean by closed loop, I guess the answer is yes
[18:30:25] <Didier> no closed loop for me is like i have : cc motor+encoder
[18:30:26] <cradek> I think it is a very useful addition to a stepper system to detect these kinds of stalls and stop. it may help you avoid ruined work.
[18:30:48] <cradek> what is cc motor?
[18:31:06] <Didier> dc motor, sorry
[18:31:15] <Didier> i'm french
[18:31:29] <cradek> oh ok, dc servo + encoder is by all definitions closed loop
[18:31:47] <Didier> yes
[18:31:53] <cradek> the PC constantly adjusts the servo's command to follow the programmed path
[18:32:11] <Didier> it is possible with stepper ?
[18:32:23] <cradek> steppers don't do this. they either step, or don't step, when you tell them to step. if they do not step, you can't tell it to "push harder"
[18:32:43] <Didier> right
[18:32:58] <Didier> then it's like a closed loop
[18:33:16] <cradek> even in a servo system, if your motor cannot follow the commanded path, it is STILL true that the only option you have is to stop
[18:33:36] <jymm> Uh boy... http://news.yahoo.com/s/ap/us_odd_child_in_box;_ylt=AnW7_HsLcEMdgMmZrpnykH4jr7sF;_ylu=X3oDMTJwMXJ0MmM2BGFzc2V0A2FwLzIwMDkxMDA2L3VzX29kZF9jaGlsZF9pbl9ib3gEcG9zAzEEc2VjA3luX3BhZ2luYXRlX3N1bW1hcnlfbGlzdARzbGsDYWxhd29tYW5sZXRz
[18:33:51] <Didier> it's certainly less expensive than dc motor
[18:34:06] <cradek> yes
[18:34:08] <andypugh> Considering that steppers have a very strong torque reduction with speed, I have wondered if you could let a slipping stepper catch up by slowing all axes
[18:34:34] <cradek> andypugh: the problem is that steppers don't get "a little" behind when overloaded - they just stop
[18:34:54] <cradek> if one motor abruptly stops, you are already screwed, because the others are still moving, and you have deviated from the programmed path
[18:35:26] <cradek> servos DO get a little behind - if they are short on power, they do keep turning but fall behind
[18:35:52] <Didier> the difference i see is dc motor increase torque when there is a problem
[18:35:56] <andypugh> Yes, I don't think that I have ever seen one miss one step, once they lose phase lock that does seem to be it
[18:36:03] <cradek> andypugh: exactly
[18:36:17] <celeron55> if you can accelerate fast with them and can accept a bit of error, maybe it might be possible to continue
[18:36:47] <andypugh> So we need EMC to see where it lost position, reverse the step sequence, pick it back up again and carry on. How hard can it be?
[18:36:49] <archivist> once a stepper gogs its lost it
[18:36:54] <cradek> celeron55: better to stop and fix your badly-tuned system than to keep turning out bad work
[18:37:04] <celeron55> cradek: of course, yes
[18:37:36] <cradek> andypugh: (can't tell if you are making a joke or not)
[18:37:40] <Didier> what do you prefer, dc motor or stepper ?
[18:37:52] <cradek> there are tradeoffs
[18:37:53] <andypugh> That's the problem with text. It was a joke.
[18:37:53] <archivist> steppers must never be driven close to their limits so they gog/lose steps
[18:38:01] <cradek> andypugh: :-)
[18:38:34] <cradek> Didier: servo systems are much more expensive and difficult to set up. They also perform much better.
[18:39:08] <Didier> yes, but i have a little problem when it works slowly
[18:39:09] <cradek> if one was better in all ways, the other would have died off by now
[18:39:19] <andypugh> Servos are much better, but much more expensive. Though there was a link posted here a few days ago where the price difference was not huge. Servo + encoder + driver for $250 I think it was
[18:39:25] <celeron55> i can show you a cheap servo system!
[18:39:26] <celeron55> http://celer.oni.biz/~celeron55/random/2009-10/totallyrandom/01102009.jpg
[18:41:03] <andypugh> Isn't i-pink trying to build a cheap (or free) servo system?
[18:41:10] <Didier> it's was a pleasure to talk with you since a long time
[18:41:40] <Didier> thank you, and talk you later, bye
[18:42:22] <christel> [Global Notice] Hi all, we appear to be having some routing issues with our server in Luxembourgh, I'm going to jupe it for now to prevent the constant bouncing. Affected users around 2300! Apologies for the inconvenience.
[18:42:46] <cradek> seemingly-simple question, very complex answer
[18:43:05] <cradek> neither of "yes" or "no" is correct
[18:43:19] <andypugh> I think I missed the original question?
[18:43:34] <cradek> Didier> Somebody can say me if it possible to drive stepper motor with encoder ?
[18:45:07] <archivist> re read andypugh 10 lines up, nearly right answer anyway
[18:45:16] <archivist> cradek, ^^
[18:47:08] <cradek> I think fewer hobbyists would succeed getting their first machines to work if they didn't just use steppers. servos can be a really big pain. so many things to know, complicated wiring, noise problems, fragile encoders, ...
[18:47:48] <cradek> compare with "parport to driver board, power supply to driver board, 4 wires to motor"
[18:48:04] <i-pink> andypugh, yes..
[18:48:16] <i-pink> i make the encoder now..
[18:53:07] <andypugh> Aha! searching the archives I found those cheap servos again. I am almost tempted.
[18:53:09] <andypugh> http://www.dmm-tech.com/Pricing.html
[18:56:22] <andypugh> I wonder how long it would take to machine a register for tooling on my rotary axis, using the axis itself as a spindle, and holding a lathe tool against it? I should get guaranteed concentricity that way.
[18:57:17] <archivist> I do round on my mill by rotating A
[18:57:46] <andypugh> Ah, don't bother answering, I have now doubled the number of brain cells allocated to the problem to two, and I can use the mill to do the machining work while the axis twirls.
[18:58:31] <archivist> should I post more brain cells to you
[18:58:48] <andypugh> I would only pickle them in red wine, it would be a waste
[18:59:58] <archivist> today I was drilling small holes .27mm and .3
[19:00:09] <andypugh> How deep?
[19:00:31] <archivist> about 1mm
[19:00:50] <archivist> bleed holes in plumbing fittings
[19:00:58] <andypugh> I spent quite some time a lot of years ago drilling 0.3mm holes all the way through 20mm long bits of 3mm brass rod
[19:01:19] <archivist> pecking for a while!
[19:01:34] <andypugh> Well, sometimes all the way through. Sometimes in the end and out the side....
[19:02:00] <archivist> getting a good start is fun
[19:02:48] <andypugh> Tiny centre drill?
[19:03:08] <archivist> power station carbon percentage in the ash measuring machine
[19:03:36] <andypugh> How does that work?
[19:03:49] <archivist> dunno, we just drill the holes
[19:06:53] <andypugh> Subsidiary question is why they care, but it is probably so that they can claim not to have thrown it into the air.
[19:07:42] <archivist> to adjust the burn aparantly
[19:08:08] <archivist> looking for less than 8%
[19:09:02] <andypugh> Makes sense. You are making the power station equivalent of a Lambda Sensor
[19:11:55] <archivist> just mad the samples in brass, they wand about 60 in stainless
[19:14:30] <andypugh> Carbide PCB drills might work well.
[19:15:52] <archivist> may but mine dont go down to that size, will have to buy some
[19:16:20] <andypugh> http://cgi.ebay.co.uk/PCB-Carbide-Turbo-Shank-Micro-Drills-0-30-NEW-UNUSED_W0QQitemZ370266490341QQcmdZViewItemQQptZUK_Home_Garden_PowerTools_SM?hash=item56359a45e5&_trksid=p3286.c0.m14#ht_500wt_956
[19:16:56] <cradek> ouch, 0.3mm is small (about #80 for us 'mericans)
[19:17:00] <cradek> do they even make pcb drills that small?
[19:18:11] <archivist> I did .24mm yesterday
[19:18:16] <andypugh> There were some at that link
[19:20:28] <archivist> that seller has some interesting toys
[19:21:17] <archivist> I use carbides for re pivoting clock pinions
[19:22:01] <archivist> one does break the odd drill, but seems easy enough to break out of the hole
[19:23:22] <andypugh> Sorry, I was poring over eBay stocklists
[19:23:29] <AchiestDragon> holes sizes down to 0.1mm on some pcbs
[19:24:05] <AchiestDragon> alhtough not shure if there drilled or cut by other means
[19:26:19] <AchiestDragon> never had need to go that small yet though
[19:27:12] <andypugh> I am curious how they drill the holes in fuel injectors. Considering the fuel is at 2000bar (more than the yield strength of mild steel) and they run red hot it must be something a bit special.
[19:27:52] <AchiestDragon> maybe tungsten ,, they do cost a far amount
[19:27:52] <archivist> edm probably these days
[19:29:46] <andypugh> EDM would seem the only way. Possibly laser?
[19:30:16] <andypugh> They have to be very exactly the right size, as that is all that controls the amount of fuel.
[19:30:35] <AchiestDragon> maybe ,, drilling would take too long for production quanaties
[19:31:10] <BridgeportII> cradek: OK, I have a little more info now. I was trying to send an s32 value from CL to the scale input (float) on the pico A/D board. So I tried to add a s32 to float converter, but now get "conv_s32_float.0 not found, addf failed". Can I send float values out of CL? That would be great if so.
[19:31:22] <AchiestDragon> i a pcb 100mm square (3") could have over 2000 holes in it
[19:32:40] <andypugh> I seem to recall overhearing that you can't add floating point functions to the base thread. But don't quote me.
[19:34:05] <BridgeportII> andypugh: We may not have a base thread, we have an external step gen board. Is there a default base thread if none specified?
[19:34:56] <cradek> BridgeportII: what are you actually trying to do?
[19:35:24] <cradek> oh is this scaling spindle speed command for the gear change?
[19:36:29] <cradek> (wild-ass guess)
[19:37:34] <cradek> CL doesn't have good support for floats, so you might want an s32 out to select between some float values. You can use mux2 (mux4, mux8) for that.
[19:38:18] <BridgeportII> Yes, scaling. I have written in CL (using two-party stop-start rungs) a gear change selector that can be activated either by PyVCP front panel buttons or user-M-codes. Need CL scaling outputs to Pico-systems A/D option board (scale, float, >1) and to PyVCP %RPM meter (scale, float, <1). Would like to fix that and use one value but haven't figured out 1/x in CL.
[19:39:05] <cradek> er, not s32out, you'd use a regular bit out of CL into mux2
[19:39:33] <cradek> http://www.linuxcnc.org/docs/html/man/man9/mux2.9.html
[19:39:58] <BridgeportII> I wrote mine using rising-edge contacts that initially assign (one-time) a scale value to an s32 word.
[19:40:24] <BridgeportII> That part is actually working, according to hal tester (?)
[19:40:53] <cradek> well you can use conv_s32_float or however it's spelled, but you can only have integer-valued floats in that case
[19:41:04] <cradek> might be fine for spindle speed
[19:41:54] <cradek> "conv_s32_float.0 not found, addf failed" is a consequence of spelling conv-s32-float.0 wrong (if the man page is to be trusted)
[19:42:31] <BridgeportII> Yeah, that's OK for Jon's board, our numbers are all integer and from 5 (we got lucky) to 350. The Py meter is 1/x though, so .000x whatever.
[19:42:42] <BridgeportII> What, hyphens and not underscores?
[19:43:08] <cradek> yes apparently
[19:43:18] <cradek> picky picky
[19:43:56] <BridgeportII> LOL! OK, we will try this. What time is dinner in Omaha?
[19:45:00] <cradek> 6ish I think. Airplanes are involved so it's hard to say exactly. Are you guys seriously going to come?
[19:46:11] <BridgeportII> Doesn't look like it, we'd have to drive too fast for too long. Wouldn't work. But you guys have a good time there. Where?
[19:46:29] <cradek> McFoster's
[19:46:49] <cradek> maybe he said 6:30 - sorry I can't find the conversation
[19:47:06] <cradek> yeah, long drive just for dinner, if you are in minneapolis
[19:48:09] <BridgeportII> No problem, we still couldn't make it. Maybe another time. We'll go try your advice and be back in a bit. Or if you are on your way to the dinner soon, we'll report our results and you can scroll back later.
[19:51:58] <andypugh> If you are using the MDI, is there any reason to use M6, or is G43 enough?
[19:53:30] <BridgeportII> andypugh: Well, we don't have a tool changer, if that's what you mean. Do you have some advice for us?
[19:53:46] <andypugh> No
[19:54:01] <andypugh> I was asking from a position of ignorance
[19:54:22] <archivist> you need to select a tool for g43 to mean something
[19:54:48] <archivist> as in g43 on its own
[19:55:05] <BridgeportII> No problem. We do have a PyVCP add-on to the AXIS front panel, with 10 buttons for "gear change" to take better advantage of the BP2 vari-speed drive.
[19:55:38] <andypugh> I don't have a toolchanger, but I do have tool offsets set up. I was wondering if it is M6 or G43 H<whatever> that tells Axis to touch-off into the right line of the tool table
[19:55:54] <cradek> T1 M6 G43 is the thing I learned to type
[19:56:12] <BridgeportII> You set the vari-speed to one of 5 agreed positions and set the proper scale factor for the VFD and Py meter.
[19:56:12] <cradek> once you M6, G41/G42/G43 automatically use the right tool information
[19:56:38] <andypugh> Aha, OK
[19:56:59] <andypugh> So I need the M6, but not the H-word
[19:57:32] <archivist> read the latest doc there was a clarification last week
[19:58:09] <andypugh> Which doc? pdf, html, wiki?
[19:58:18] <archivist> html on the site
[19:58:27] <archivist> all should match
[19:59:36] <andypugh> OK, I don't mind searching but I like to think I am looking in the right place
[19:59:42] <cradek> http://www.linuxcnc.org/docs/devel/html/gcode.html
[19:59:47] <cradek> for gcode questions I always start here ^^
[19:59:58] <cradek> in this case, click on G43 for details
[20:00:24] <BridgeportII> Now we are getting "can't find module 'conv-s32-float' in /usr/realtime/2.6.24-16/rtai/modules/emc2". Are you sure about the hyphens?
[20:01:13] <cradek> BridgeportII: back up a sec - look at http://www.linuxcnc.org/docs/html/man/man9/conv_s32_float.9.html again
[20:01:20] <BridgeportII> OK
[20:01:57] <cradek> notice the loadrt line wants _ and the addf/net lines want -
[20:02:12] <cradek> I think a lot of hal components are this way
[20:02:18] <BridgeportII> Ah, so an assortment. OK, we'll try again.
[20:02:34] <cradek> that's a very kind word for it
[20:03:14] <andypugh> I am still slightly unsure whether touch-off into the tool table uses T or H to choose the table entry.
[20:03:36] <andypugh> (Of course it is the work of a moment to find out experimentally)
[20:03:51] <cradek> andypugh: loaded tool (the one in the spindle), not the prepped tool
[20:03:54] <BridgeportII> Reminds me of a Linus Torvalds quote (from program comments) someone was using for a sig, it went something like this: "If you have had lunch recently, please skip over the next section".
[20:04:25] <cradek> andypugh: so it's NEITHER the T or H tool. it's the one you last M6ed.
[20:04:58] <andypugh> OK, but M6 not G43 tells it which table entry to modify?
[20:05:25] <cradek> yes the _loaded_ tool
[20:05:47] <andypugh> I might be failing to make myself clear.
[20:05:48] <cradek> imagine three lines like this: T1 M6 / G43 H2 / T3
[20:05:56] <cradek> tool table touch off will set the offset for T1
[20:06:10] <cradek> even though you are currently using the offset from tool 2 and tool 3 is currently prepped
[20:06:15] <andypugh> OK, that was similar to the example I was going to type as an example.
[20:06:52] <andypugh> So I do need the M6 then.
[20:06:56] <cradek> it's a little swirly in this situation isn't it
[20:07:47] <andypugh> Which reminds me, did you find a way to access the tool offsets in G-code? You can set them, but it seems a write-only process
[20:08:40] <cradek> no they are not available
[20:09:34] <andypugh> :-(
[20:10:18] <andypugh> Now, where did I put my copy of Kernigan and Richie?
[20:11:28] <archivist> andypugh, lookup recursion in the index in the 2nd ed
[20:12:56] <andypugh> I am almost sure I can guess what it says.
[20:14:58] <BridgeportII> OK, things are looking up here, we have advanced a couple of lines. Now error is "pin 'ppmc.0.DAC8.00.scale' does not exist". This is the scale input for Jon's A/D add-on board, and we could previously say "setp ppmc.0.DAC8.00.scale 350.0" or whatever value, and it would work. Now I'm trying to let CL/HAL set the scale.
[20:15:32] <kanzure_> kanzure_ is now known as kanzure
[20:15:53] <BridgeportII> The HAL line is "net spinscalef conv-s32-float.0.out => ppmc.0.DAC8.00.scale"
[20:17:05] <cradek> I do not have ppmc stuff to test, but the same thing applies: comment out the offending line. start emc. use the hal tools to check exactly how the pin is spelled and if it exists. you seem to have a frequent problem with misspellings. this is the easiest way to troubleshoot and fix them.
[20:17:56] <andypugh> I get confused about pins, parameters and the rest, but I don't think that you can link to parameters. ie if you can setp it, you can't net it.
[20:17:58] <BridgeportII> OK, thnaks I wlli od thta!
[20:18:16] <cradek> haha
[20:18:35] <cradek> andypugh: setp will set either a pin or parameter, so that is not a sufficient test
[20:19:01] <cradek> andypugh: but halmeter, halscope, halcmd show, halshow will all tell you which are which and show how they are spelled
[20:19:12] <andypugh> OK, i was going by my embarassment of telling someone they could net the PWM frequency, then finding I was wrong.
[20:19:24] <cradek> andypugh: ah
[20:20:01] <BridgeportII> Ah, so I may not be able to "net" the parameter? OK, I will watch for that.
[20:20:34] <cradek> hm, maybe you are right that it's a param - not sure.
[20:22:52] <BridgeportII> If this turns out to be a parameter, and I am not allowed to net it, then the mux components would not fix it either, correct?
[20:23:01] <andypugh> Would a workaround be to create a DAC8 for each scale you need, then choose them with a mux? (I am assuming here that there is a mux)
[20:23:26] <SWPLinux> logger_emc: bookmark
[20:23:26] <SWPLinux> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-10-06.txt
[20:24:18] <andypugh> I am also assuming that you only have a few scale values to manage.
[20:24:24] <SWPLinux> BridgeportII: you'll need to calculate the scale elsewhere in HAL, and just set the PPMC scale to 1 (or some other convenient number)
[20:24:51] <SWPLinux> using muxes and whatnot to select the scale value, and a scale block to do the actual multiplication
[20:24:57] <cradek> yes there is a scale component with input pin "gain"
[20:25:11] <SWPLinux> see you later
[20:25:17] <BridgeportII> SWPLinux: OK, thanks, that sounds good. Have a good dinner
[20:26:31] <andypugh> I came home early expecting the delivery of my new stepper and driver. It wasn't here, and no card. Just now (at 2120) my neighbour has come round saying that they had tried to deliver it to my old address, not found me in, and left it with her.
[20:27:30] <BridgeportII> andypugh: congrats, let us know how it tests out.
[20:28:07] <andypugh> But here is an odd thing, the driver has a table of pulses per rev on the top. Is this assuming that every stepper in the world is 200ppr?
[20:28:40] <cradek> probably
[20:31:02] <andypugh> The first thing I am going to do is take the back off the motor to modify it. I wonder if it is worth putting an ER32 collet on the shaft, hard against the motor face in an attempt to stop the rotor touching the stator?
[20:31:40] <cradek> seems like it couldn't hurt, but I know nothing
[20:32:11] <andypugh> I am not sure it even matters, but I have had dire warnings that it is bad to remove the rotor from the motor.
[20:32:21] <cradek> I have also heard that
[20:32:22] <archivist> 99.999% of us not mod the motors that way
[20:32:46] <andypugh> Is that a "do" or a "don't"?
[20:32:59] <archivist> nont
[20:33:02] <archivist> dont
[20:33:11] <archivist> tyop day
[20:33:27] <andypugh> So you just have 8 bare wires dangling in the breeze?
[20:33:43] <archivist> no I cover outside
[20:34:09] <archivist> and also yes for Z because its out the way
[20:34:13] <andypugh> How do you strain-relieve the cable entry?
[20:34:45] <archivist> I tie wrap the cable cover to the body
[20:36:09] <andypugh> Hmm, no. I have my pride :-)
[20:37:00] <archivist> I still have my original motors with magnetism
[20:37:08] <archivist> :)
[20:37:28] <archivist> and no grit in them
[20:38:04] <andypugh> You are over-valuing function over form.
[21:52:39] <i-pink> hii all
[22:00:04] <andypugh> Hi there.
[22:02:55] <ZeroFlex> hallihallo
[22:17:13] <alex_joni> good night all
[22:41:10] <andypugh> Aye, goodnight from me too.