#emc | Logs for 2010-07-12

[00:06:37] <mikegg> what does hit limit in home state 12 mean?
[00:17:55] <madsci44> lost position?
[00:19:23] <madsci44> i had something like that earlier today on my stepper gantry - it was missing steps because the base period was too high for the feedrate
[00:21:12] <mikegg> it was during a homing sequence...I didn't have HOME_USE_INDEX or HOME_IGNORE_LIMITS set properly
[00:21:13] <mikegg> ...
[01:17:20] <ries> apart from QCad, what other open source 2D CAD packages are popular?
[03:35:56] <Jymmm> FREE SLURPY DAY!!!
[03:43:33] <Jymmm> http://www.slurpee.com/
[03:44:53] <madsci44> hoo-hoo! I had a feeling this was a long weekend
[04:13:12] <madsci44> my internet is not cooperating tonight
[06:20:50] <ries_> ries_ is now known as ries
[08:01:06] <darren_> darren_ is now known as redwizard-kde
[14:14:14] <jepler> does anyone know where I can find this script referred to by steve b that allows jog during tool change?
[14:16:11] <jepler> ah, found it
[14:16:19] <elmo40> link?
[14:16:37] <jepler> http://mid.gmane.org/4BA0FC88.3000709@hbahr.org
[14:18:08] <elmo40> you want to jog something while doing a tool change?
[14:18:15] <elmo40> what is the benefit of that?
[14:23:29] <cradek> looks like it does the abort and then rfl for you
[14:23:58] <cradek> elmo40: imagine a machine with collets where you have to jog to a fixture to set the new tool length
[14:24:20] <jepler> elmo40: personally I don't need that feature, but I was just trying to make sure I knew what it did before I replied to a post in which Steve B. mentioned it.
[14:28:53] <elmo40> cradek: jog during a tool change... what axis can you jog? most tool carousels are stationary, meaning the spindle needs to travel toward it to change the tool. I know on my Mazak Horizontals, the Z and B are the only axis that move separate from the spindle. they move out of the way for clearance.
[14:29:32] <cradek> elmo40: this is most relevant for machines with just collets (without tool changers)
[14:30:00] <cradek> imagine using a retrofitted bridgeport with R8 collets
[14:31:45] <JT-Hardinge> or just about any router based machine
[14:35:09] <elmo40> I still don't get it. I am trying to visualize jogging and tool change at the same time. I guess with my machine the Z is the only axis on the spindle, therefore the XY can move to the tool probe thingy while the tool is being changed... would only make sense on that kind of machine. ones with multiple axis on the spindle I would think it would be useful.
[14:36:55] <JT-Hardinge> what if you don't have a tool probe...
[14:37:42] <elmo40> then why jog and tool change?
[14:38:23] <JT-Hardinge> to change your tool and set the z offset for that new tool
[14:38:29] <elmo40> unless you are machining a different section of the part... far away from the current tools cutting path... what is the need to jog
[14:38:56] <elmo40> I can see where it would work, I really do.
[15:13:57] <piasdom> hi
[15:14:31] <piasdom> installed hardy onto a new harddrive and now i get this error http://pastebin.com/2avWrxyW
[15:14:53] <piasdom> i upgraded to 2.4 after install
[15:15:20] <piasdom> it doesn't list a NML files, should it ?
[15:16:19] <JT-Hardinge> did you follow these instructions when you upgraded to 2.4? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UpdatingTo2.4
[15:17:52] <piasdom> yes
[15:18:31] <piasdom> went to source list and changed to 2.4 (both) and then update/upgrade
[15:18:42] <micges_work1> opengl driver error
[15:19:08] <piasdom> don't like my V card ?
[15:19:09] <micges_work1> check if you have installed any vga driver
[15:19:26] <piasdom> how ?
[15:19:33] <micges_work1> synaptic
[15:19:47] <piasdom> JT-Hardinge: micges_work1 ; thanks
[15:20:24] <micges_work1> sorry if you have fresh install and this error, then you HAVE to install some video driver, which one - depends on hw configuration
[15:20:44] <piasdom> micges_work1: thanks
[15:21:05] <micges_work1> piasdom: package with 'GLX' in name
[15:21:26] <micges_work1> (don't remember which one)
[15:22:15] <piasdom> k
[15:38:23] <pcw_home> JT-Hardinge so you got your turret encoder working again?
[15:55:12] <JT-Hardinge> pcw_home: yes, thanks it was a couple of diodes and rat shack had them :)
[15:55:39] <JT-Hardinge> I now know more about that circuit than ever before :)
[15:56:31] <JT-Hardinge> I still don't know how the coolant got in there but I have a sheet metal cover I wooped out to cover that area for now
[15:56:49] <pcw_home> Well, should be easy to fix next time...
[15:57:15] <JT-Hardinge> yea, I bought $3 worth of them, so I'm set for life I think
[15:58:01] <pcw_home> My lesson of the weekend is not to pound t-posts into the ground when the post is above my head
[15:58:34] <JT-Hardinge> ouch!
[15:58:58] <pcw_home> 4 stitches of ouch
[15:59:38] <JT-Hardinge> that will be a lesson you may not ever forget then :)
[16:00:44] <pcw_home> The scar should remind me...Mainly I feel stupid
[16:02:02] <JT-Hardinge> you should feel more experienced now... and smarter to boot
[16:03:42] <pcw_home> Ha! I probably needed a label on the post pounder with pictures of stupid things no-to-do
[16:04:21] <bosko> Hell everyone!
[16:04:26] <bosko> Hello
[16:04:40] <JT-Hardinge> somehow I don't think you need the the label now :)
[16:07:36] <skunkworks> bosko: hello!
[16:08:20] <bosko> What i9s good jitter time ?
[16:08:59] <skunkworks> IMHO usually <25000
[16:09:18] <skunkworks> depends on what you are wanting for step rates also.
[16:09:30] <bosko> i have around 10000 on Celeron 1800, 512 mb of ram, intel 815 plate
[16:09:41] <skunkworks> that is really quite good
[16:10:08] <bosko> TNX
[16:14:44] <skunkworks> that was a hit and run
[16:33:59] <skunkworks> so - 7 days going through yellow stone, grand teton, glacier national park was a whirlwind tour. :) It was pretty phenominal.
[16:36:08] <skunkworks> the drive on the 'going to the sun' road at glacier was, I think, the best.
[16:36:38] <skunkworks> biab
[16:47:39] <Jymmm> There's a glacier in Glacier NP ?
[16:48:48] <Jymmm> Ah... "Glacier National park was named for the glaciers that carved, sculpted, and formed this landscape millions of years ago. Despite the recession of current glaciers, the park's name will not change when the glaciers are gone. "
[16:50:11] <elmo40> :P humans... only thinking of the short term
[16:52:23] <Jymmm> Huh?
[17:21:07] <Fox_M|afk> Fox_M|afk is now known as Fox_Muldr
[17:35:50] <Jymmm> Heh, I like the name... http://sfbay.craigslist.org/sby/tls/1838813087.html
[17:37:49] <Jymmm> UG UG UG http://sfbay.craigslist.org/sby/tls/1838789991.html
[17:40:55] <Jymmm> ds2: http://sfbay.craigslist.org/sby/tls/1838749673.html and http://en.wikipedia.org/wiki/Meehanite
[18:44:22] <ej> hello!
[18:44:34] <skunkworks> Hello!
[18:44:37] <ej> any pyvcp gurus here?
[18:45:08] <skunkworks> ask your question
[18:45:46] <ej> I'm configuring my machine and thought that pyvcp would make nice spindle control (pluto-servo for XYZ and spindle)
[18:46:26] <ej> my question is: is it possible to have a scale element for spindle speeds (5k-20k, 500 increments) so that it would send MDI command e.g. "s12000"?
[18:46:56] <ej> the idea is to enable spindle speed control in the g-code files, but also to control the speed via the pyvcp
[18:47:35] <ej> At this point I'm guessing a system that would change the pyvcp element according to the S g-code command accordingly is too much :)
[18:55:28] <ej> ...so no ides here?
[18:56:26] <SWPadnos> I don't think you'll be able to have the S word output change a pyvcp control, since the HAL pins aren't bidirectional
[18:57:01] <SWPadnos> you also can't have pyvcp directly send an MDI command, though you can get that to happen with halui
[18:57:11] <SWPadnos> but you can't have halui send a variable value
[18:57:27] <ej> ok
[18:57:36] <ej> any ideas for that?
[18:57:58] <SWPadnos> well, you'd have to do some coding in halui and/or pyvcp
[18:58:26] <ej> any hints on that?
[18:58:32] <SWPadnos> not really
[18:59:07] <SWPadnos> (I meant modifying and recompiling the code, not making creative configurations, in case that wasn't clear)
[18:59:25] <ej> ok :)
[18:59:42] <ej> that I won't do
[18:59:47] <ej> not that I even could :D
[19:01:14] <ej> It seems I'll just do a pyvcp->PWM speed control
[19:01:40] <SWPadnos> yep, that seems the simplest thing to do
[19:02:02] <SWPadnos> you can add to that a checkbox or something that lets you switch a mux (though it's not recommended to do that)
[19:02:42] <ej> I'll go with the simple way
[19:03:16] <ej> thanks!
[19:04:42] <SWPadnos> oh well, no caution warning there then
[19:08:00] <alex_joni> hmm.. why not use spindle-speed-override slider?
[19:09:29] <SWPadnos> or the spindle speed buttons themselves ...
[19:09:35] <SWPadnos> or MDI commands
[20:25:52] <skunkworks> mozmck: you around?
[21:04:03] <andypugh> If a chip is rated at 600V max, I wonder how long it will survive at 680V?
[21:07:53] <Jymmm> what kind of chip?
[21:08:11] <Jymmm> and do you have a fire suppression system?
[21:08:36] <andypugh> It's a little 3-phase MOSFET driver.
[21:09:13] <andypugh> (Suplementary question is how you are meant to enforce a 3mm track spacing on a 0.05" pitch chip...
[21:10:47] <pcw_home> There are 1200V isolation gate drivers (unless you like the scent of vaporized epoxy/copper)
[21:12:44] <pcw_home> High side common mode range is a better term than isolation...
[21:13:18] <andypugh> True, but the 1200V ones are £8 each and I need 3, whereas this £3 6-output chip is on the table behind me.
[21:13:44] <andypugh> (When I bought it I thought I was going to have 440V)
[21:14:23] <andypugh> I will probably see how it works at 300V
[21:15:03] <andypugh> Some kind of opto-isolation between it and the Arduino might be called for.
[21:15:34] <pcw_home> unless you want to do the neutron dance
[21:15:56] <andypugh> ?
[21:17:06] <pcw_home> With a bridge rectifier both V+ and V- are hot relative to ground
[21:17:27] <andypugh> Ah, yes.
[21:18:02] <andypugh> My voltage-doubler has that problem, but then that does mean that neutral is like a 3-phase neutral, so that seems sensible.
[21:21:21] <geo01005_home> geo01005_home is now known as geo01005
[21:23:42] <pcw_home> Be careful and ground you low side well. a failed gate driver will likely take out your transistors (wear eye protection)
[21:23:44] <pcw_home> but more insidious, might overload an opto and cause loss of isolation
[21:24:33] <pcw_home> A charred opto is to longer a good insulator
[21:24:42] <pcw_home> (no longer)
[21:25:09] <andypugh> I have been deliberately charring insulator this evening
[21:25:54] <andypugh> I made a heating element for a melting pot by winding nichrome wire round a plastic rod, then widing the rod round a crucible
[21:26:23] <skunkworks> this is cool.
[21:26:25] <skunkworks> http://www.innovations-report.com/html/reports/physics_astronomy/large_binocular_telescope_achieves_major_156426.html
[21:26:27] <moop> yes ground is important, i blew a load of stk stepper drives last week cause the ground came off the current setting network
[21:28:11] <moop> but i fixed my problem with etch-servo locking my system, the problem was the servo periods
[21:28:39] <andypugh> I am beginning to be more and more sure that my plan is feasible, and less and less sure that I can do it.
[21:29:10] <pcw_home> Probably put a little more voltage into the current sense circuit than the stk was expecting...
[21:29:51] <pcw_home> terror at the 680V?
[21:30:15] <moop> when the ground came off the voltage went high at 5v the max should have been 2.5v
[21:30:57] <andypugh> Partly the voltage, but partly that as the voltage goes up the component choice goes down and the cost goes up.
[21:31:00] <moop> so the current doubled the max rated value, i think it fried the step sequence logic?
[21:32:20] <moop> it still works as a current limiter, but all the phases stay energised regandless of control pins
[21:32:28] <pcw_home> andtpugh: is it the IR gate drivers that are 8 pounds?
[21:33:52] <andypugh> I found a slightly cheaper one: http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=5412290
[21:36:02] <pcw_home> what about adum6132?
[21:36:48] <moop> any one know about pwmgen routing in .hal file?
[21:37:09] <andypugh> 6132?
[21:37:26] <andypugh> moop: I might know the answer, if I understood the question
[21:38:02] <moop> I am looking at the etch-servo config, and it seems there are three pin outputs a forward/reverse/ and a pwm signal
[21:38:26] <pcw_home> andypugh: http://www.analog.com/en/interface/digital-isolators/adum6132/products/product.html
[21:38:46] <moop> i want a forward and a reverse, which both output pwm
[21:39:51] <moop> so i use less pins
[21:40:03] <moop> i think i just solved my own query
[21:40:24] <moop> i should change the pwm output_type to 2??
[21:40:54] <SWPadnos> if that's "up/down", then yes, I think that's right
[21:41:16] <andypugh> pcw_home: It seems to have a 560V max?
[21:41:19] <moop> yes, sorry for wasting bandwidth
[21:42:06] <moop> i should RTFM
[21:42:13] <andypugh> moop: You could also assume forward if not reverse too.
[21:42:45] <andypugh> Or, even, use 50% PWM for stop, and only use one pin.
[21:43:32] <SWPadnos> hahaha. this is a great opening line for a 419 scam email: "Greetings in the name of God please let this not sound strange to you for my only surviving lawyer who would have done this died early this year."
[21:43:39] <andypugh> Of course, the consequences of an Etch-a-Sketch running away don't bear thinking about. You might never get that aluminium powder out of the carpet.
[21:43:55] <SWPadnos> surviving: obviously this word doesn't mean what you think it means
[21:45:06] <andypugh> I think there is an implied past tense there
[21:45:11] <moop> hehehe, im going to run my death ray using the etch-servo config
[21:45:54] <pcw_home> Well the 560V depends on class (you will have additional isolation from V- to logic)
[21:45:56] <pcw_home> transient isolation voltage is 6000V peak (10 seconds)
[21:46:13] <pcw_home> (try that on you 1200v chip...)
[21:46:18] <andypugh> As in: He was down to one surviving lawyer, then even that one died. I think that getting involved would be foolish on that basis alone.
[21:47:32] <moop> I wouldn't want to be his lawyer
[21:47:35] <SWPadnos> yeah, that would be a(nother) good reason to delete the email :)
[21:48:50] <moop> send it to 419eater.com
[21:52:20] <andypugh> pcw_home: This one says 630V max working isolation, 1181 test voltage and 6000V peak
[21:52:22] <andypugh> http://docs-europe.origin.electrocomponents.com/webdocs/0779/0900766b807796fc.pdf
[21:52:43] <andypugh> And is actually an opto-isolator too.
[21:54:29] <pcw_home> ok same UL isolation spec (3750V RMS) but no power (so needs bootstrap)
[21:54:50] <andypugh> Though I would need 6, and I was about to mention bootstrap
[21:55:39] <SWPadnos> moop, nah, I don't have time to do the baiting
[21:55:40] <pcw_home> and low side gate supply
[21:56:06] <andypugh> (The drawback with the ADuM6132 is that I cant find a supplier, incidentally)
[21:56:44] <pcw_home> Yes Analog wants to be the one and only supplier...
[21:58:55] <pcw_home> so you will need some HV bootstrap diodes
[21:58:57] <pcw_home> and a 15V 100 ma supply with good isolation
[21:59:56] <andypugh> I might just buy a new motor
[22:00:27] <pcw_home> probably safer
[22:00:38] <andypugh> Probably cheaper too.
[22:00:43] <pcw_home> or find a good deal on a VFD
[22:01:07] <andypugh> Voltage-increasing VFDs are quite expensive.
[22:01:34] <pcw_home> Well the low tech part is the voltage doubler...
[22:02:01] <andypugh> Though I have a sneaking suspicion that I could run a 3-phase - 3-phase one off of my 680V DC supply.
[22:03:38] <pcw_home> unless it has some missing phase fault detection
[22:04:38] <andypugh> Yes. That is the whole problem really, they are expensive because they are technically advanced drives. All I need is a phase converter.
[22:05:30] <pcw_home> though a 3phase - 3 phase converter they may just check dc-bus droput, but who knows...
[22:05:40] <pcw_home> (on a)
[22:11:53] <andypugh> Optical fibre transcievers are very cheap and would offer a lot of isolation :)-)
[22:17:28] <pcw_home> for higher voltages that make a lot of sense (also noise isolation)
[22:23:47] <pcw_home> I notice the 6132 has a 15 year lifetime with 800V peak and 50 year with 660V peak
[22:23:49] <pcw_home> Avago chip is probably similar. Note 6132 has power suppy but wimpy drive (200 mA)
[22:23:50] <pcw_home> so would need npn/pnp buffer. 6I32 has 2 drivers but low side is not isolated so you would
[22:23:52] <pcw_home> need 6 optos in addition to 3 6132s, or 6 Avago chips + bootstrap
[22:24:29] <andypugh> I was going to use an Arduino for this bit.
[22:24:59] <andypugh> Sorry, I think I mis-read.
[22:25:45] <andypugh> I assumed that the 6i32 up there was a Mesa product :-)
[22:27:02] <andypugh> I wonder what the lifetime of the 600V rated chip that I have already is? I mean, is it going to explode immediately due to breakdown, or simply die young due to stress.
[22:27:35] <andypugh> The latter is OK with me, as I would be re-engineering onto a proper PCB once the comcept is proven anyway.
[22:29:24] <pcw_home> What are the abs-max specs on the gate driver you have?
[22:29:33] <andypugh> 600V
[22:30:19] <andypugh> I would not be at all surprised to see the 680V drop to <600 under even slight load.
[22:31:00] <pcw_home> well power it up without your transistors and see what happens...
[22:31:34] <andypugh> And without the Arduino, too. :-)
[22:31:56] <pcw_home> I wonder if the specs show a typical breakdown voltage...
[22:31:57] <andypugh> I do seem to have 6 very nice PWM signals from the Arduino now.
[22:32:52] <andypugh> http://docs-europe.origin.electrocomponents.com/webdocs/0dca/0900766b80dca3c9.pdf
[22:41:54] <pcw_home> Nothing on what to expect on breakdown voltage but they do list 600V as a maximum working value which might give a little leeway
[22:42:59] <pcw_home> You said you had 1000V MOSFETS?
[22:43:03] <andypugh> 240V mains doubled should be perfectly safe...
[22:43:18] <andypugh> Yes, the MOSFETS are less of a concern
[22:43:41] <andypugh> And also pretty cheap
[22:45:08] <pcw_home> Beware DVDT turnon with MOSFETS much bigger problem with MOSFETS than IGBTs
[22:45:43] <andypugh> Time for bed now. I think tomorrow I will connect the driver to the supply and see if it shows signs of instant death.
[22:46:07] <pcw_home> 'nite
[22:46:24] <andypugh> If not, then I will put the optos on and see what comes out
[22:47:27] <pcw_home> remember for the MOSFETS: low turn on gate drive and high turn off drive
[22:56:03] <moop> has anyone integrated webcam into axis yet?
[22:56:12] <moop> I see this thread: http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/id,2198/catid,21/limit,6/limitstart,0/lang,english/
[22:59:33] <JT-Hardinge> Dang that bore is nice and smooth... looks like it was ground to size
[23:02:14] <moop> a wild boar?
[23:02:32] <JT-Hardinge> no a smooth boar
[23:02:53] <moop> where?
[23:02:58] <JT-Hardinge> it's so nice to have 8 stations again
[23:03:07] <JT-Hardinge> in my lathe
[23:04:48] <Fox_Muldr> Fox_Muldr is now known as Fox_M|afk
[23:09:31] <Valen> anybody used EMC to run a welder?
[23:09:39] <Valen> (thinking of some CnC welding)
[23:10:26] <JT-Hardinge> now to put a nice 0.010" radius on the edge and its ready to part off...
[23:13:53] <L84Supper> this should be interesting, attempting to use EMC to control a cartesian 3-axis inkjet with 5 printheads
[23:15:16] <JT-Hardinge> alex_joni: has/does I think
[23:26:26] <JT-Hardinge> speaking of supper I better get started cutting up the veggies for dinner