#linuxcnc | Logs for 2015-12-23

[00:29:33] <Flipp_> any solidworks gurus around to help with a (probably dumb) assembly mate question?
[00:30:44] <zeeshan|2> :P
[00:31:13] <zeeshan|2> whats up
[01:01:17] <Flipp_> sorry, wife summoned me :D
[01:01:23] <Flipp_> I'm trying to mate a worm to a gear. I've got the pitch diameter of both
[01:01:39] <Flipp_> but I can't seem to figure out how to mate the worm to the worm gear at the right distance
[01:02:01] <Flipp_> I have the pitch diameter sketched out as construction lines in both of them
[01:02:21] <Flipp_> which result in two "perpendicular" circles
[01:03:55] <Flipp_> I can eyeball it, but I'd like to make it so that the worm is mated perpendicular to the worm gear's axis (which I can do) and at a certain radius from the center of the worm gear's axis (which I can't figure out)
[01:19:33] <CaptHindsight> Flipp_: post a screen shot, that might help
[01:20:48] <Flipp_> gah, yeah, true enough :)
[01:20:48] <Flipp_> http://i.imgur.com/zTq7Ltb.png
[01:20:59] <CaptHindsight> are these two parts that you want to mate or are you realy trying to mate the worm part of a gear to the body of the gear?
[01:21:34] <CaptHindsight> ok so 2 gears
[01:22:14] <Flipp_> I want to set up some mates exactly like they are in the picture, so that the worm is always perpendicular to a line drawn from the center point of the worm gear
[01:22:32] <t12> put a plane down the axis of the worm gear
[01:22:38] <Flipp_> I've kludged something together that works, which is a distance mate from the two centers ath rGear + rWorm
[01:22:40] <t12> and a plane through the midpoint of the gear teeth
[01:22:47] <t12> and mate them coincident?
[01:23:13] <Flipp_> t12: I can do that, but then that mate won't allow the two gears to form a "gear" mechanical mate
[01:23:15] <t12> then i guess a distance constraint ?
[01:23:18] <Flipp_> so I wouldn't be able to rotate them
[01:23:20] <t12> ooh
[01:23:20] <t12> sorry
[01:24:04] <Flipp_> np, gives me an idea though. I wonder if I can do a Plane1/centerline-of-the-worm-gear-bore coincident mate
[01:25:12] <Flipp_> yup, that does it
[01:25:18] <Flipp_> sorry, still getting used to advanced mates
[01:25:20] <Flipp_> thanks for the idea :)
[01:25:25] <Flipp_> cheers!
[01:25:30] <t12> accidental advice success!
[01:25:50] <CaptHindsight> now to animate it :)
[01:26:00] <CaptHindsight> post the vid
[01:26:17] <Flipp_> will do! step 1: learn to animate... haha
[01:26:23] <Flipp_> but it's somethin' I gotta learn :D
[01:28:33] <CaptHindsight> what's the hole in the worm teeth for?
[01:29:08] <Flipp_> set screw for nema 8 motor
[01:37:08] <t12> that is a wee little motor
[01:54:35] <Flipp_> t12: definitely. has to be, if I'm to stick it on the end of a delta 3d printer :/
[01:54:45] <Flipp_> capthindsight: animation rendering...
[01:57:42] <Flipp_> capthindsight: disclaimer: http://i.imgur.com/PyniXs2.png
[01:58:00] <Flipp_> https://www.youtube.com/watch?v=-eQccUxICFw
[01:58:11] <Flipp_> youtube is still rendering it in higher res, but it works :D
[02:11:25] <archivist> I love gear mates inSW
[02:18:46] <Deejay> moin
[04:47:55] <enleth> Ordered 6i25+7i77
[05:13:43] <archivist> I cant make my mind up for what I want
[05:14:45] <archivist> I want to make a test jig for gearing, possibly only temporary though
[05:15:29] <XXCoder> heys
[05:16:54] <archivist> partly to write a web page to show how bad some dividing heads are :)
[05:40:19] <SpeedEvil> I suppose optical with a webcam is cheating.
[05:41:25] <XXCoder> funny http://www.dump.com/pigeonleave/
[05:43:16] <archivist> SpeedEvil, to do what?
[05:44:43] <archivist> I have tested dividing heads with optical gear (precision polygon and angle dekkor)
[05:49:27] <XXCoder> dividing head = indexing thing?
[05:51:03] <archivist> indexing things are fixed angles, but also need testing/verification
[05:52:06] <archivist> rotary table uses a dividing disk to get any division within its limits
[05:52:14] <XXCoder> cool
[05:52:37] <archivist> cnc version is with the resolution of the system
[05:54:18] <SpeedEvil> I mean more using a simple image and massive oversampling
[05:55:08] <SpeedEvil> Keller tracks stars using images on the CCD that are many pixels across and then averages the disc to get precise positions.
[05:55:13] <SpeedEvil> Kepler
[05:56:19] <SpeedEvil> You can track a twenty pixel across circle very very well.
[05:56:20] <archivist> too much like hard work, I just want to log two encoders, input and output
[05:56:38] <SpeedEvil> :)
[05:57:17] <XXCoder> lol ok
[05:57:49] <archivist> which looks like (log,value of variable 123 is: #123) in the gcode
[06:00:16] <archivist> if I also grab the stepper command I can debunk microstep accuracy at the same time
[06:21:28] <XXCoder> archivist: how do you test that?
[06:22:07] <archivist> log commanded angle and measured angle
[06:22:39] <XXCoder> cool
[06:24:30] <archivist> unfortunately many think commanded=actual with steppers and rotary tables
[06:29:44] <XXCoder> if tolence isnt very tight it dont matter I guess
[06:32:51] <archivist> it does matter when noobs fool themselves into thinking they are making a silk purse out of a sows ear
[06:33:53] <archivist> and the net perpetuates the incorrectness in sales blurb
[06:33:56] <XXCoder> I do plan to test my machine accurate
[06:34:00] <XXCoder> and precision
[06:34:17] <XXCoder> but seems I cant run it for a while yet as garage is still messy and icy cold now
[06:38:31] <archivist> I was going to link to a page claiming impossible accuracy, seems it has been fixed
[06:38:47] <XXCoder> nice
[06:38:57] <XXCoder> atomic precision? ;)
[06:39:55] <archivist> he used the worm division ratio and stepper count to give an accuracy, now only states the resolution
[06:40:36] <archivist> the late owner replied to me when I pointed it out, Im too old.....
[06:43:41] <archivist> actually its is still up! http://www.sherlineipd.com/8700inst.htm
[06:45:21] <XXCoder> LATE owner? whats he or she die from
[06:45:29] <archivist> old age
[06:45:55] <XXCoder> so thats "zombie" site?
[06:46:17] <XXCoder> I have seen a few, and I also saw a few site "die" and "zombie" for few years before disappear
[06:46:18] <archivist> no I think the sons run it or something like that
[06:46:31] <XXCoder> in one case it was because owner was hit by truck
[06:48:38] <archivist> I can show a research article in finding hobbing errors, and then point at that mechanical accuracy paragraph :)
[06:48:53] <XXCoder> lol
[06:49:20] <SpeedEvil> archivist: yeah - it would be lovely if more things supported sin/cosine errors
[06:49:22] <SpeedEvil> archivist: err
[06:49:36] <SpeedEvil> archivist: I mean that there is not zero THD in steppers, if that makes sense
[06:56:00] <SpeedEvil> I cant' find anything on this on the web.
[06:56:04] <SpeedEvil> Maybe I suck at searching
[06:57:05] <archivist> here is the article testing hobbing setups http://www.collection.archivist.info/archive/proceedings/PR22/
[06:57:50] <archivist> they are using spectrum analysis though
[06:59:18] <SpeedEvil> Can't easily read that alas ATM
[06:59:53] <SpeedEvil> Though I suppose converting it to a time, not angle problem does make it easier to log in some ways
[07:00:40] <archivist> he discusses the tooth rate v spectrum for diagnosis
[07:00:59] <SpeedEvil> yeah - ivnert the FFT, and bang
[07:01:05] <SpeedEvil> FT
[07:01:23] <archivist> but telescope makers need even better to point at stars
[07:01:45] <SpeedEvil> For proper telescopes :)
[07:01:57] <archivist> they discuss worm/wheel errors a bit more on the web
[07:02:42] <archivist> I first learnt when I got some gears I made rejected
[07:02:44] <SpeedEvil> yeah - unless the harmonic errors in the stepper output excite a resonance mode, or the errors themselves are under your seeing, you probably don't care much
[07:03:22] <archivist> was using an index plate made on a crappy dividing head
[08:08:27] <JT-Shop> so much for todays plan... rain and thunder storms all day
[10:07:54] <Magnifikus> toastydeath, i found a tool manufacturer that offers 8mm single flute hss mills for ~8€ (coated +3€) :)
[10:08:09] <Magnifikus> some little company
[10:09:22] <enleth> good price
[10:09:54] <enleth> what types of coating are available?
[10:10:30] <Magnifikus> TiN / TiCN / TiAlN
[10:10:34] <Magnifikus> for 3,1€
[10:10:45] <Magnifikus> special coatings 4,65€
[10:11:38] <Magnifikus> http://snag.gy/EmFF8.jpg sory its only german :)
[10:12:14] <enleth> Magnifikus: can you ask for zirconium nitride?
[10:12:51] <enleth> if they can do it, I might want to buy some
[10:13:56] <Magnifikus> its that the perfect stuff for aluminium? :)
[10:14:11] <enleth> apparently so
[10:14:19] <Magnifikus> will contact them in january will ask
[10:14:35] <Magnifikus> is that poison? ^^
[10:15:55] <enleth> should be completely biologically inert
[10:16:48] <enleth> zirconium itself is and the nitride is insoluble in anything except some nasty fluoric acids
[10:17:32] <enleth> if you get anywhere close to that, you've got bigger problems than some zirconium ions
[10:18:01] <Magnifikus> hehe
[10:18:07] <Magnifikus> nah just careful with cobalt etc
[10:18:20] <Magnifikus> cause we use it in our mechanical workshop in the lab
[10:30:34] <lair82> Good Morning guys, What is a reasonable value to put in for an axis's "DEADBAND" in the ini?
[10:31:19] <archivist> that depends on your machine methinks
[10:33:29] <lair82> It is regarding my Cincinnati VMC with the rotary that I just finished, the Y axis is always a solid 0.0015" off from what ever the commanded position is. In the INI I have 0.001" set for the deadband.
[10:34:35] <archivist> I would have thought as small as possible that avoids hunting
[10:35:17] <lair82> So I would not be out of line putting a value of 0.0001" ?
[10:36:41] <archivist> heh https://forum.linuxcnc.org/forum/10-advanced-configuration/22678-deadband
[10:37:19] <pcw_home> Is there a problem with 0 deadband?
[10:37:44] <lair82> No, that was kinda what I was wondering?
[10:38:46] <pcw_home> For encoders I guess 1-2 encoder count equivalent may stop some jitter when not moving
[10:39:36] <lair82> This is a resolver machine, with a scale of RESOLVER_SCALE =0.1999999
[10:40:05] <pcw_home> Ahh then I would try 0 first
[10:41:50] <lair82> that was my thread, LOL, I will go set it to "0", and see what happens,
[10:42:05] <lair82> Not sure why that was in there to begin with,
[10:42:38] <pcw_home> if its too noisy when not moving, then increase slowly ( say starting with .00005)
[10:45:27] <pcw_home> resolvers will have "white noise" of some magnitude ( roughly between 1/4000 to 1/40000 of a turn with the 7I49 )
[11:40:24] <frank__> Hey guys, sorry to bother, but does anyone knows exactly how the 2 motors per axis thing is going right now? i understand there have been a few mods this last few days, but trying to understand developers messages are really a headache for the mortal person. Thanks!
[11:42:36] <frank__> FYI if i dont appear instantly its my freaking internet connection which is working like #€@=
[11:43:37] <CaptHindsight> frank__: comments in the source or on the dev mail lisy?
[11:43:45] <CaptHindsight> lisy/list
[11:45:47] <frank__> i saw a few updates from the emc-developers/commit (dont remember) mailing list (saw from sourceforge) --- im new, sorry for misinformation
[11:46:53] <frank__> Dewey Garret talking on update for Joint_axes
[11:48:47] <os1r1s> pcw_home Does it matter if the quad encoder signal runs through a buffer chip (IE the one in the G540)?
[11:50:19] <CaptHindsight> frank__: Dewey is working in changes in an experimental branch
[11:50:50] <CaptHindsight> frank__: what is your actual question about using 2 motors per axis?
[11:52:42] <frank__> well, currently i havent got yet to the electronics part, but i am figuring out if linuxcnc would be a good choice for my router 3 axis with 2 motors on the gantry the question would be if its viable, as todays version
[11:53:29] <frank__> i guess i missed some info, the build will be with panasonic servos, and mesa cards.
[11:53:40] <archivist> frank__, I think some are using it
[12:04:04] <pcw_home> os1r1s: I suspect the G540 inputs may be too slow for anything but low res encoders
[12:04:45] <os1r1s> pcw_home low res being less than 500 ppr?
[12:06:45] <os1r1s> pcw_home I guess I'm confused then. On the G540 XML/config, it puts a quadrature encoder on 3 on the inputs that run through the G540. It seems like that will cause problems ...
[12:06:55] <pcw_home> hard to tell there, are no g540 timing specs on inputs
[12:07:24] <pcw_home> how would that cause problems?
[12:07:45] <os1r1s> It says 250 KHz
[12:07:51] <os1r1s> On one of the datasheets
[12:08:20] <pcw_home> I think that the step rate not the misc input bandwidth
[12:08:44] <os1r1s> Well, what I mean is if you can't use a quad encoder because of the G540 speed, it seems like people would try and it wouldn't work well
[12:10:00] <pcw_home> you can always use a quad encoder, it just may be low res (50 PPT or so if really slow)
[12:10:42] <pcw_home> Gecko has no specs so its pretty much impossible to tell without trying
[12:10:56] <os1r1s> I'm calling them to ask :)
[12:11:11] <os1r1s> May not get anywhere, but I will try
[12:12:06] <pcw_home> if you have a 500 line encoder, you can just run it until it stops counting to get an idea of the input capabilities
[12:12:34] <pcw_home> ( by ramping up the speed I mean )
[12:12:57] <os1r1s> It looks like a LTV846 encoder which is 10 KHz
[12:14:24] <pcw_home> possible...
[12:14:50] <os1r1s> Yep, Gecko confirmed
[12:15:05] <os1r1s> So I'm guessing that is too slow for a spindle encoder
[12:15:34] <pcw_home> low res spindle encoder would be fine
[12:15:48] <zeeshan|2> os1r1s: my factory mill came with a 10 tooth wheel
[12:15:54] <zeeshan|2> :P
[12:16:20] <archivist> mine came with 48 slots on the spindle
[12:16:28] <archivist> lathe
[12:16:41] <zeeshan|2> ive been running a 70 tooth wheel with no problems
[12:17:00] <zeeshan|2> but then again im limited to 3150 rpm
[12:17:04] <pcw_home> if its actually 10 KHz, thats maybe 40K counts per second
[12:17:16] <os1r1s> Well, 2800 RPM x 500 ppr would be 1.4M
[12:17:29] <os1r1s> So it would be way too high
[12:17:39] <zeeshan|2> why do you want such a fine encoder on your spindle??
[12:18:27] <maxcnc> hi from germany ;-) xmas is around the shop
[12:18:30] <DaViruz> i've been working on creating spindle synchronized motion using a microcontroller. one of my lathes has no power feed, and id like to add it, with threading capability
[12:18:31] <pcw_home> 2800RPM 500 line is 93 KHz
[12:18:32] <os1r1s> zeeshan|2 Thats how it came on my lathe for threading
[12:18:48] <zeeshan|2> that is super fine :)
[12:18:53] <os1r1s> Its an amt-102 though, so I think I can just lower the dip switches to 100
[12:18:53] <DaViruz> i figured it would be fairly trivial to create step signals synchronized and scaled to an encoder
[12:18:56] <zeeshan|2> my lathe had a 100 tooth encoder
[12:18:57] <DaViruz> i was sorely mistaken..
[12:19:04] <os1r1s> But that would still be too high ..
[12:19:33] <CaptHindsight> os1r1s: how high is not too high?
[12:19:45] <pcw_home> 100 lines would be fine to 6000 RPM assuming 40K counts/s max
[12:20:20] <os1r1s> pcw_home Then thats probably more than sufficient. for my use ...
[12:20:51] <maxcnc> pcw_home: i solved all issues now no more com error on 5i25 7i76
[12:20:58] <pcw_home> for threading even 25 or 50 a lines (100/200 counts/turn quadrature) are fine
[12:21:12] <archivist> I put a speed indicator on my display to warn me about counting rates
[12:21:32] <pcw_home> maxcnc: great!
[12:21:50] <skunkworks> maxcnc, what worked?
[12:22:08] <maxcnc> half inc double lim
[12:22:54] * archivist runs that through google translate
[12:23:08] <maxcnc> on plasma firer zhe counts got up to 275
[12:23:17] <CaptHindsight> http://www.harborfreight.com/digital-photo-sensor-tachometer-66632.html counts accurately as long as you have enough contrast and hold it in it's window
[12:23:22] <maxcnc> im now on 400 lim
[12:24:38] <maxcnc> i will now move up linuxcnc to 2.7 stabel
[12:24:41] <CaptHindsight> and the reflective tape doesn't fly off the part you are spinning :)
[12:25:36] <Jymmm> archivist: isigamu inc lim kabili
[12:25:46] <maxcnc> its the warmest xmas ever recordet in germany
[12:26:30] <zeeshan|2> global warming :)
[12:26:32] <zeeshan|2> or........
[12:26:34] <zeeshan|2> weather is shifting
[12:26:40] <zeeshan|2> i remember last year, the winter stayed around for a long time
[12:27:00] <maxcnc> skunkworks: we are talkinh on this http://linuxcnc.org/docs/html/man/man9/sserial.9.html
[12:27:12] <frank__> here Lat America is raining more than ever
[12:27:41] <maxcnc> you got spring there
[12:27:50] <maxcnc> its suposed to be that
[12:27:56] <CaptHindsight> zeeshan|2: I think it was right up until it got warmer :)
[12:28:14] <frank__> its supposed to be hot, not supposed to rain more than ever before
[12:29:20] <frank__> capthindsight, i think my conection disabled or something, i was asking about the gantry mode, u think i will be able to carry on with it?
[12:29:27] <frank__> the main problem is the squaring right?
[12:29:39] <maxcnc> in the UK they got filled up to the chest with water not beer
[12:29:48] <frank__> lol
[12:30:10] <maxcnc> frank__: on servos or steppers
[12:30:13] <CaptHindsight> frank__: several people use 2 motors on a gantry
[12:30:22] <frank__> servos panasonic diff encoders
[12:30:41] <maxcnc> here all mashines above 6ft got more then one motor on a axis
[12:30:46] <Jymmm> 37F/24F (3C/-4C) with rain and snow tomorrow
[12:31:16] <CaptHindsight> frank__: more people have problems with keeping the gantry square die to poor machine design
[12:31:20] <frank__> i am new in the linuxcnc world, i am finding it quite likeable, lots of people help eachother
[12:31:24] <CaptHindsight> die/due
[12:31:42] <frank__> i see
[12:31:57] <frank__> i am going with rack and pinion (not the best option) and hiwin 25mm linear guides
[12:32:03] <CaptHindsight> they think that their strong will is enough to hold it in place
[12:32:06] <maxcnc> CaptHindsight: ypu only need one button to reorder it on your request
[12:32:10] <frank__> for about 100kg of weight from the gantry
[12:32:15] <t12> capt: ever work with compact c-mount microscopes
[12:32:17] <ReadError> finally snagged some servos! http://www.ebay.com/itm/Applied-Motion-Products-Brushless-Servo-Motor-48V-5-7A-200W-Encoder-/331611273222
[12:32:19] <t12> like jammed in tight near some work
[12:32:29] <maxcnc> frank__: thats a good setup
[12:32:55] <ReadError> got those + http://www.ebay.com/itm/331611226220?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT for about 110$ shipped each (combo)
[12:33:24] <frank__> thanks, trying to do my best, first build ever
[12:33:42] <maxcnc> frank__: what precision or step performence ate you looking fore
[12:33:48] <archivist> frank__, to cut what?
[12:34:19] <frank__> not a lot actually, i plan on cutting wood
[12:34:22] <frank__> maybe some acrylics
[12:34:24] <frank__> its a router
[12:34:43] <CaptHindsight> t12: compact and c-mount? That's like Jumbo Shrimp! :)
[12:34:44] <maxcnc> oh so you can go for 0.1mm
[12:34:48] <frank__> 2.4mts x 1.2mts size and a 15cm of z clearance for max stiffness
[12:34:57] <t12> haha
[12:35:12] <frank__> awesome, i just hope the rack and pinions made here are good enough,
[12:35:12] <maxcnc> thats a good 7,5I gear on 30tooth pinion
[12:35:21] <t12> looking to get 10-15x right near some work
[12:35:26] <t12> c-mount could be changed
[12:35:38] <maxcnc> frank__: the size of a standard plasma
[12:35:47] <t12> by right near i mean LWD but trying to avoid big pendulous lens tubes and such
[12:36:17] <maxcnc> frank__: get a 1.5kw hyanjang on it and ypou are fine
[12:36:18] <archivist> t12 with the right design you dont need to be close
[12:37:02] <frank__> exactly what i was going to ask you, do u know an estimate for top speed? i am getting the 1kw motors
[12:37:08] <frank__> a little bit too much i think
[12:37:20] <t12> i found this vendors stuff
[12:37:21] <t12> http://www.infinity-usa.com/products/proximity/InfiniTube-FM.aspx
[12:37:31] <maxcnc> frank__: why 2motors its better to use a long pin on the y to control 2sides of the X
[12:38:04] <maxcnc> im running at 10m/min top speed on wood
[12:38:18] <maxcnc> with a 6mm cutter
[12:38:33] <frank__> okey i got confused on that one, how would that go
[12:38:49] <maxcnc> hit my nick
[12:39:31] <frank__> nothing happens lol
[12:39:33] <CaptHindsight> t12: have you scoured ebay?
[12:39:41] <frank__> i hitted another members name and i can see info
[12:39:59] <maxcnc> no problem
[12:40:02] <CaptHindsight> some people give them away since they either sit for years or they don't know what they are
[12:40:12] <t12> i have not
[12:40:17] <t12> more looking for vendor suggestions
[12:40:22] <t12> haven't really bought this stuff before
[12:40:54] <maxcnc> frank__: a 20mm fix pin runs throu the y axis on the end is a arm like plate
[12:40:59] <CaptHindsight> run the numbers, field of view, distance from work to lens etc
[12:41:39] <archivist> t12 I use a stereo zoom 7-30 magnification, all depends on usage
[12:41:45] <maxcnc> frank__: on both sides that hold the pinion
[12:41:49] <CaptHindsight> t12: all pricey new
[12:42:01] <t12> CV application
[12:42:04] <t12> on end effector
[12:42:05] <archivist> cheap if you are lucky
[12:42:13] <t12> doing tiny tiny moves
[12:42:26] <t12> like threading 15um holes
[12:42:49] <maxcnc> frank__: between the main pin and the pinion is a timing belt on both sides
[12:42:51] <frank__> mmm cant see it, if i think what you are thinking, i dont belive its going to work, my racks are designed to look down to the floor while the guides are on top of the frame
[12:43:01] <CaptHindsight> microorganism surveillance
[12:43:11] <frank__> excuse my english :D
[12:43:56] <t12> lol
[12:44:05] <t12> listening to their phonec alls
[12:44:38] <frank__> how do you quote my name?
[12:44:46] <maxcnc> frank__: realy easy setup and it wrks perfect
[12:44:52] <pcw_home> maxcnc: if you have sserial errors I would wait for 2.7.4, 2.7.x currently has a bug in sserial error
[12:44:53] <pcw_home> reporting that makes the error limit stuff not work
[12:45:07] <maxcnc> pcw_home: thanks
[12:45:29] <frank__> pcw your from mesa?
[12:45:33] <maxcnc> i will stay on 2.5.2
[12:45:36] <frank__> can i make you a simple question?
[12:45:43] <pcw_home> works in 2.6, bad in2.7.x
[12:46:07] <pcw_home> frank__: sure
[12:46:11] <maxcnc> frank__: just asl
[12:46:42] <maxcnc> we all love mesa and use them all over
[12:46:49] <frank__> i have a panasonic servo minas liqi series (actually 3/4 depending setup of gantry) and i want to use your cards
[12:47:15] <frank__> they are pwm, 5 serial incremental encoder
[12:47:42] <frank__> let me check more specs
[12:48:16] <frank__> 1kw ac 220-240 v
[12:48:19] <JT-Shop> pcw_home: is that bug in 2.7.2 as well?
[12:48:52] <pcw_home> I think its in all 2.7 versions but I am not sure
[12:49:09] <CaptHindsight> t12: maybe http://www.ebay.com/itm/Edmund-Optics-1-0X-Silver-Series-Telecentric-Lens-58-430-/181966245458
[12:50:07] <frank__> Pulse Differential input. Selectable by parameter. ([1]Positive/Negative pulse [2]A/B quadrature [3]Pulse/Direction)
[12:50:34] <frank__> do you need any more info? i would like to know which mesa card would be the best for my setup,
[12:51:19] <frank__> thanks in advice
[12:53:21] <pcw_home> if you want to do step/dir or quadrature (I think quadrature is the best)
[12:53:23] <pcw_home> a 5I25/7I76 will work (as will a 7I76E)
[12:54:51] <maxcnc> frank__: this will bee then a very easy setup on guntry with more motors per axis
[12:55:02] <frank__> pcw_home: awesome!
[12:55:26] <maxcnc> you even dont need guntry kinetics on this
[12:55:27] <CaptHindsight> t12: what can you use for the sensor? 1/2", 1/3"
[12:55:31] <frank__> maxcnc: you say that i should put 2 on the gantry?
[12:55:46] <maxcnc> up to you
[12:56:04] <frank__> i still havent understood how would it work with 1 motor if i cant put a ballscrew in the middle
[12:56:04] <JT-Shop> I used just one on my plasma with a jack shaft between the two sides
[12:56:10] <maxcnc> i woudt go for only 1
[12:56:24] <frank__> do you happen to have a link of a build like that?
[12:56:34] <maxcnc> as Jt said jack shaft is the google
[12:56:39] <frank__> i spent a lot of time in cnczone and never saw something like it
[12:57:09] <frank__> awesome thanks!
[12:57:48] <maxcnc> http://www.ubuilditplans.com/image1/JShaft%20kit2.jpg
[12:58:00] <maxcnc> go for timing belt on the dide
[12:58:08] <maxcnc> side
[12:58:21] <maxcnc> the gear in the middle of the pin
[12:58:38] <JT-Shop> http://gnipsel.com/images/plasma/
[12:58:58] <JT-Shop> http://gnipsel.com/shop/plasma/plasma.xhtml
[12:59:14] <frank__> yes, thats what i thought you were telling me, the thing is that the rack is a little bit hidden from the gantry they are on the sides, and connected to the gantry feet with a plate 20cm tall
[12:59:25] <maxcnc> frank__: you can do it more then one way
[13:01:08] <frank__> so no rack on one side i am correct?
[13:01:41] <frank__> am i correct?** excuse my english!
[13:03:05] <maxcnc> no rack and pinion on both sides
[13:03:17] <maxcnc> give vme 5min to get a constructin
[13:08:15] <frank__> maxcnc: no problem! thanks for the advice, althought i think i am going to go with 2 motors on the gantry to feel safer and exploit the high speeds and rigidity of my stiff gantry, besides, i kind of already bought them.. lol
[13:11:19] <frank__> actually now i am just a little worried about having the structure machined correctly, now its away from home, beeing machined.
[13:22:49] <maxcnc> done
[13:23:58] <frank__> oh, its allright, i believe i am going to go with 2 motors on the gantry to feel safer and exploit the high speeds and rigidity of my stiff gantry, besides, i kind of already bought them.. lol
[13:25:09] <frank__> i built a very heavy steel gantry, would be a shame to just use 1 motor on that thing
[13:25:46] <maxcnc> frank__: http://heimwerkermarkt-tretter.de/web/heeks.png
[13:25:58] <maxcnc> its just a 5min construction on heeks
[13:26:38] <maxcnc> im using it on a 500kg yZ
[13:26:44] <frank__> wow men
[13:26:49] <frank__> ur making it attractive
[13:26:51] <maxcnc> on 25mm rails for x
[13:27:08] <frank__> here some pics of my gantry if you want to check it
[13:27:11] <frank__> http://www.cnczone.com/forums/diy-cnc-router-table-machines/275912-cnc-cad-2.html
[13:27:15] <maxcnc> there is a timing belt for the Z differenc
[13:27:36] <maxcnc> shoudt i ad one or can you think its there
[13:28:17] <frank__> yes, the thing is that my timing bould be a little bit large, i am gonna check on the cad design
[13:28:48] <maxcnc> mine is 3meters
[13:28:52] <frank__> its allright i understand the hole idea now
[13:28:56] <frank__> wtf
[13:28:59] <frank__> sorry
[13:29:01] <frank__> lol
[13:29:05] <frank__> 3 metes of z to rack?
[13:29:27] <maxcnc> my biggest mashie is x20m y 10meters an Z 8meters
[13:29:42] <frank__> O
[13:29:42] <frank__> M
[13:29:43] <frank__> G
[13:29:57] <frank__> on what guides!?!?
[13:30:00] <maxcnc> for large brigh concrede molds
[13:30:07] <frank__> AND 1 mmotor!!?!?
[13:30:33] <maxcnc> as it only needs 1centimeter precision its all on chains
[13:30:39] <frank__> 500 kg right? sorry for the messages u got me kind of shocked
[13:30:55] <maxcnc> the plasma on that layout
[13:31:42] <frank__> where can i upload a quick picture to show you my cad ?
[13:32:00] <maxcnc> to your webpage ?
[13:32:08] <zeeshan|2> frank__: imgur.com
[13:32:28] <frank__> my webpage? i dont have one :(
[13:32:37] <CaptHindsight> imagebin.ca
[13:33:55] <frank__> your the man capthindsight
[13:34:50] <Jymmm> CaptHindsight is woman, see her roar!
[13:35:12] <maxcnc> Jymmm: my thinking its the only female here ;-)
[13:35:19] <frank__> there you have my cnc machine on cad, its a little rough (the cad)
[13:35:20] <frank__> http://imgur.com/GMd2bDg
[13:35:28] <frank__> really?
[13:35:40] <frank__> sorry capt!
[13:35:41] <maxcnc> perfect
[13:35:57] <maxcnc> remove the motor and ad a timingbelt
[13:36:05] <CaptHindsight> imgur is for posting images for other people to spend way too much time on commenting, venting, attacking and abusing the poster :)
[13:36:15] <maxcnc> use the sqare tube for transision
[13:36:32] <frank__> yes, i think it would need at elast 30cm of timing belt for each side
[13:36:48] <frank__> wouldnt that compromise rigidity?
[13:36:57] <frank__> timing belts flex
[13:37:02] <frank__> or dont?
[13:37:21] <maxcnc> yes they do
[13:37:47] <maxcnc> you can get the gear on one side
[13:37:57] <maxcnc> so its best for short elektronics
[13:38:51] <CaptHindsight> tubing is gonna twist
[13:39:02] <frank__> steel tubing?
[13:39:07] <frank__> the main gantry you mean?
[13:39:10] <maxcnc> ok im done bye Gn8 have a good xmas and party
[13:39:20] <frank__> thanks max
[13:39:25] <maxcnc> NP
[13:39:26] <frank__> ur awesome!
[13:39:53] <CaptHindsight> fill the tubing with resin/granite for quick fix
[13:40:21] <frank__> really? its 1.3mts 100x100mm tubing 6.35mm thick
[13:40:30] <frank__> really strong for wood
[13:40:32] <frank__> or not?
[13:40:38] <zeeshan|2> frank__: i think you'll be fine :P
[13:40:59] <zeeshan|2> is there a particular reason you went with such a large tube
[13:41:02] <frank__> and the z clearance its 12-15 cm for max stiffness and shor leverage
[13:41:02] <CaptHindsight> yeah, I wasn't following the application too closely :)
[13:41:42] <frank__> no, its just my first build, and really wanted to make something good,
[13:42:09] <zeeshan|2> the part i can see deflecting the most under side load
[13:42:11] <frank__> now i regret myself and think that i should have dont something easyer, like a plasma
[13:42:15] <frank__> or smaller router
[13:42:17] <zeeshan|2> is the square tubing that has been cut at a bevel
[13:42:26] <zeeshan|2> but then again, for wood i think its fine :P
[13:42:49] <CaptHindsight> he can always add some granite epoxy later
[13:43:09] <frank__> yes i thikn thats the weak point
[13:43:32] <CaptHindsight> model it and check the deflection under load
[13:43:45] <frank__> just went the easy way there, to avoid having to weld plates for the feet,
[13:44:04] <frank__> i would love to, but its really out of my territory there!
[13:44:43] <frank__> im on rhinoceros, no finite element analysis
[13:44:46] <frank__> i think
[13:45:24] <CaptHindsight> oh, then why didn't you add a t-rex in there just for fun?
[13:45:34] <Jymmm> FWIW hiwin on the rails liek that (x) has a tendancy to collect sawdust.
[13:46:13] <CaptHindsight> Jymmm: is correct, you might want to add a bellows
[13:46:23] <frank__> yep, when i welded everything i regreted not having put a single flat bar to machine on the side
[13:46:31] <frank__> much easyer, less distortion
[13:46:48] <frank__> i will be adding a collector,
[13:47:04] <frank__> maybe some discrete low quality bellows,
[13:47:21] <CaptHindsight> frank__: you'll also find that tubing isn't the straightest or flattest as you might be hoping
[13:47:29] <Jymmm> Also, give yourself added clearance on the sides to allow for edge clamping of material
[13:47:31] <ReadError> is there any 'gotchas' or downsides to the ethernet MESA gear?
[13:47:37] <frank__> its really not! i send everything to get machined
[13:48:04] <frank__> i reaally miss that piece of steel lol
[13:48:33] <CaptHindsight> ReadError: it might depend on your perspective, it's not closed source, but schematics are not open
[13:49:03] <ReadError> CaptHindsight well in terms of performance
[13:50:03] <CaptHindsight> ReadError: they suit the most common machine control applications
[13:53:44] <CaptHindsight> probably not enough room left over for integrating machine vision into the fpga or other more advanced/complicated controls
[13:58:48] <ReadError> hmm alright
[13:59:05] <ReadError> need to figure out what daughter cards are needed then I guess
[14:21:34] <frank__> hi guys, sorry to bother again but i am having a hard time figuring out my drivers need for card, it seems it needs a whooping 25 pins, i have 6 multi function digital inputs, another 4 for signal inputs, 3 multi function outputs and 7 output signal (command pulse train for a b and z phase outputs) (encoder i believe) and another 4 for grounding and power. what should i do? which mesa cards will allow me to connect so manyI/O and
[14:21:39] <frank__> thanks for the time
[14:22:48] <pcw_home> you only need a small subset of those I/O (maybe only 3 signals in the simplest case)
[14:24:19] <frank__> but wouldnt be better to give the software more information about motor, alarms etc?
[14:24:19] <frank__> i havent found an enable input
[14:24:19] <frank__> drive enable
[14:24:23] <frank__> i found a servo on input, which energizes de motor
[14:24:39] <pcw_home> ok servo-on = drive enable
[14:25:42] <frank__> another one says S-RDY and description : - Said drive signal to the output power status.
[14:27:10] <frank__> long story short, if i chose to use more i/o for better drive information i need 2 mesa daughterboards right?
[14:32:34] <pcw_home> possibly, but whether the extra information is of any real use is debatable
[14:34:26] <pcw_home> for a open loop system you probably want to monitor the drive fault output
[14:40:52] <DIYtryin> Hi guys, Anybody here? I'm burning to ask a question
[14:48:52] <malcom2073> Go get a drink of water
[14:49:05] <malcom2073> Or see a doctor, if it's *that* kind of burning :)
[14:50:09] <DIYtryin> Haha thanks:p Just got some water
[15:13:14] <malcom2073> So... burning quenched?
[15:15:04] <DIYtryin> Yeah.. Burning quenched:) Now the question: I've been challenged to create a side panel for a linuxcnc gui, and control two motors with it (arduino uno in between)
[15:16:13] <DIYtryin> I'm new to this. So I'd like to know if somebody here has experience with that kind of thing and suggestions how I should move on. It's my first day at the office and am used to web development:p
[15:16:40] <malcom2073> Well, if you're using Axis, you can make a panel with buttons and the like that run python scripts to send commands to your arduino
[15:17:32] <DIYtryin> Yeah it's axis
[15:18:08] <DIYtryin> glad to hear it's possible:p
[15:27:54] <DIYtryin> I think I'll have to choose to create the panel with gladeVCP or pyVCP. Is that right? And how do I know which one to choose?
[15:29:28] <DIYtryin> thanks for your reply earlier by the way, Malcom
[15:30:00] <malcom2073> I've never used GladeVCP, so someone else will have to pipe in on that one, I use pyVCP for mine
[15:30:26] <malcom2073> http://www.linuxcnc.org/docs/html/gui/gladevcp.html#_pyvcp_versus_gladevcp_at_a_glance
[15:30:36] <malcom2073> Pretty good list of the capability differences
[15:35:03] <zeeshan|2> gladevcp ftw
[15:44:06] <JT-Shop> well crap, I get an error about a public key missing while trying to add 2.6 back to the BP so I can make a few parts
[15:44:21] <JT-Shop> must be time to start drinking
[15:44:44] <robinsz> grootings
[15:46:26] <JT-Shop> W: GPG error: http://buildbot.linuxcnc.org lucid Release: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY EF1B07FEE0EE663E
[15:47:54] <robinsz> depending on your level of paranoia, you can just tell it to ignore
[15:49:12] <robinsz> or ... sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys EF1B07FEE0EE663E
[15:51:50] <JT-Shop> Thanks, I'll give that a try
[15:56:01] <robinsz> don't blame me if it sets on fire
[15:56:12] <robinsz> or your house explodes
[15:56:22] <robinsz> it can happen ...
[16:00:00] <JT-Shop> that can happen or I can just run it over a few times with my bulldozer
[16:00:13] <skunkworks> anger issues...
[16:01:10] <JT-Shop> frustration issues
[16:05:02] <Frank__> guys im trying to understand the closed loop servo system with linuxcnc, my drivers can be set with up to 24 pins for each one, x4 means like almost 100 pins for the motors only, should i pick 2 mesa cards to get them covered + other i/o or minimize i/o of drivers? thanks!
[16:05:08] <JT-Shop> the errors after reverting back to 2.6 and uncommenting (is that a word) the watchdog http://pastebin.com/q2uXiKxH
[16:05:57] * JT-Shop thinks it's time to evict the desk spider from behind the keyboard... one web too many
[16:08:42] <robinsz> 24 pins?
[16:10:16] <FloppyDisk> @JT - Peta will protest...
[16:10:42] <FloppyDisk> @FRank__: My guess is you don't need all those pins, or some are for encoder feedback, which would not be included in the i/o count
[16:10:57] <FloppyDisk> as encoders have their own separate inputs on the mesa 7i77.
[16:11:04] <FloppyDisk> What servo do you have?
[16:11:22] <robinsz> just buy a Mesa 5i25 and 7i77 .. job done
[16:15:12] <FloppyDisk> Oh, 2 more i/o would be your +-10v signal, so that's 7 (singled ended) or 10 pins depending on how its wired. Plus, you might have an enable and you can share that (or not), plus some fault feedback...
[16:16:08] <FloppyDisk> I guess he didn't like my answer:-( and quit... doh.
[16:18:40] <JT-Shop> guess I'll make the new powder coat rack on the drill press
[16:21:12] <FloppyDisk> @JT-shop: sorry to hear that:-(
[16:21:38] <frank__> sorry my internet connection is working awfully
[16:22:34] <FloppyDisk> To get lcnc working, on my 5i25/7i77 I put in a new (old) harddrive for 2.6.
[16:23:05] <Deejay> gn8
[16:23:09] <FloppyDisk> eitherway, it was a bit of work, I tried 3 5vdc power supplies to no avail to try to get it working.
[16:29:15] <JT-Shop> I reverted back to 2.6 but the configuration won't even start now
[16:36:34] <FloppyDisk> I'm not sure what you mean by configuration, but the INI file is different between the 2.6 and 2.7, so I have copies for each.
[16:37:14] <FloppyDisk> Also, aren't there different kernals (Preempt vs. rtai) - dunno, outa my league on that.
[16:37:50] <FloppyDisk> I think Jepler or someone had a writeup I don't remember all the details on why preempt is hte better way now, anyway, off track - sorry.
[16:41:07] <robinsz> i found a nice P&G LVDT sensor today
[16:41:17] <robinsz> that will be awesome for surface probing
[16:41:38] <robinsz> assuming i can figure out how to set up a driver for an LVDT
[16:43:41] <JT-Shop> the only difference in the configuration is the hm2 watchdog, you don't need it in 2.7
[16:49:39] <PCW> The failed to set register error is a MODBUS error
[16:50:07] <JT-Shop> from the GS2 I'd bet, thanks for looking
[16:50:47] <JT-Shop> dang, plug in the VFD!
[17:15:36] <frank__> guys, easy question, im having a servo with up to 25 pins to enable, one of them is servo ready, and this is what it does : The output transistor turns on when both the main and control power supplies are properly provided and no alarm is shown. Turns on the output transistor after absolute data are transferred, when the absolute I/F function is valid in the absolute mode-is it necessary???
[17:15:56] <Duc> What file would I need to change to stop the mesa card from outputing a 25mV signal to my servopacks. No commanded movement has been inputted
[17:17:12] <Duc> frank_ what type of servos
[17:18:32] <frank__> they are 1kw ac 220 panasonic minas liqi
[17:18:37] <frank__> incremental encoders
[17:18:49] <frank__> pwm igbt sinusoidal drive
[17:19:08] <PCW> Duc: you can use the PID offset to zero the outputs
[17:19:34] <PCW> though normally 10-20 mv is not significant
[17:19:42] <Duc> PCW: in the servo packs or in linuxcnc
[17:19:55] <PCW> in linuxcnc
[17:20:08] <Duc> PCW: the servo are slowing turning and giving f-errors so I need to offset it
[17:20:24] <Duc> PCW: Im testing the servos outside of the mill so I dont break anything
[17:20:29] <PCW> Thats not really tehe way to fix it
[17:20:55] <Duc> PCW: for some reason I need a X-ouput of .025 when I turn the system on but only on the X
[17:22:39] <PCW> That is, its more important that you disable the drives if linuxcnc is not running or has a following error
[17:23:28] <Duc> PCW: drives only turn on once I toggle on Machine power then the drive starts to move but with machine power off there is no movement
[17:24:17] <robinsz> .025 is nothing, ignore it
[17:24:42] <PCW> slow drift should be expected if the drives are enabled but the position loop is not closed
[17:24:58] <PCW> (from analog velocity mode drives)
[17:25:06] <Duc> so how does one close the position loop.
[17:25:45] <PCW> this is why the first thing that needs to be done is to get linuxcnc to enable/disable the drives
[17:26:11] <enleth> I think I'm going to be very thankful for the control-independent relay based safety circuits of my mill when I start integrating linuxcnc - all I need to do is keep using the original servo power circuit that is cut off by limit switches and activated manually with a push button that kickstarts a latching relay
[17:26:52] <Duc> linuxcnc is enabling and disabling the drives. I can see the S-on toggling in the Sigmawin software for the servos
[17:27:26] <Duc> when I turn on the machine power in axis
[17:29:50] <JT-Shop> hm2/hm2_5i25.0: Smart Serial Error: port 0 channel 0. You may see this error if the FPGA card read thread is not running.
[17:30:09] <Duc> I used PNNCONF to make the initial files and maybe that didnt close the loop. I do see feedback from the encoders when are showing a change in position in AXIS
[17:30:52] <PCW> basic steps to getting servo drives set-up:
[17:30:54] <PCW> 1. Get linuxcnc in control of the drive enables
[17:30:55] <PCW> 2. Get the encoders reading right (scale and direction)
[17:30:57] <PCW> 3. Get the feedback direction correct (expect runaways), you may need to widen the ferror limits
[17:30:58] <PCW> to see whats going on and for initial tuning
[17:31:47] <Duc> Would you suggest for me to install the drive first so it has some resistants
[17:31:56] <PCW> pncconf has a bug for servo systems, it sets the PID maxerror to .0005, this shoud be 0
[17:32:09] <Duc> let me change that quick
[17:32:38] <PCW> Its probably safer with the drives removed so you cant runaway onto the stops
[17:33:15] <Duc> ok motor went from a slow rotation to a fast rotation
[17:34:23] <PCW> so it ran away
[17:34:24] <Duc> never mind now its stample
[17:34:38] <Duc> ran away the first two times I cycled machine power
[17:35:02] <Duc> but I cycled power again and now its running away
[17:35:05] <Duc> very odd
[17:35:27] <PCW> you probably have feedback backwards
[17:36:36] <Duc> encoders or the analog signal?
[17:37:14] <PCW> you cannot reverse the encoders typically (the DRO/gcode will be reversed)
[17:37:57] <Duc> I think I can change in the servos how it behaves
[17:38:49] <PCW> did you set the following error limits so it can only runway an inch (25 mm) or so?
[17:38:50] <Duc> glad I keep my foot on the motor so it cant jump around
[17:39:03] <Duc> yea set it to 10inch so I could read data
[17:39:20] <frank__> lol
[17:39:21] <PCW> that may be a bit large....
[17:39:51] <Duc> ok
[17:40:08] <Duc> with the motor out of the system it was safe but I will change it down
[17:41:13] <PCW> this is why you get the encoders correct first
[17:41:14] <PCW> to reverse the analog output polarity you change the sign of the scalemax parameter
[17:41:21] <Duc> FERROR =1 and MIN_Ferror = .05
[17:41:51] <PCW> I would set them both to 1
[17:42:10] <PCW> and stay away from the ends of travel
[17:42:31] <Duc> motor still is outside of machine so no worries about travel
[17:42:54] <PCW> you can check the encoder scaling and direction with the drives disabled
[17:43:18] <PCW> (if you can crank the ballscrew by hand)
[17:44:03] <Duc> let me check that
[17:44:53] <Duc> I have that direction wrong
[17:45:21] <Duc> ccw moves mill neg and CCW moves axis positive
[17:45:43] <PCW> OK so the good news is when you fix the encoder direction, the feedback will be correct
[17:46:28] <Duc> ok that is fixed. changed Encoder_scale from a positive to neg
[17:46:32] <Duc> moving correct now
[17:47:26] <Duc> I think that fixed the run away now.
[17:47:34] <Duc> such simple stuff sometimes
[17:48:36] <Duc> so the computer was fighting which direction it needed to go
[17:48:41] <PCW> the 25MV was probably a runaway but limited to PID-MAXERROR * P term
[17:49:34] <PCW> since maxerror was .0005
[17:50:15] <Duc> ok The help has been amazing. Thank you
[17:50:50] <PCW> welcome
[17:51:26] <andypugh> A Pneumatic drawbar is a bit of a non-advantage when your compressor dies :-(
[17:53:12] <Duc> andypugh: had that issue when running the plasma table. Thought the plasma went belly up
[17:53:18] <Jymmmm> _methods: you in aussie land?
[17:53:26] <andypugh> I am not sure what the problem is. It’s one of the little silent compressors: This style: http://www.ebay.co.uk/itm/Jun-Air-Compressor-/252216356774?hash=item3ab9442fa6:g:9aoAAOSw1S9WcqtS
[17:53:45] <_methods> me?
[17:53:53] <frank__> pcw can i make you a little question? im figuring out my needs for mesa cards, and i am needing about 15 i/o per driver + limits, etc, which combination of 5i25 + daughterboard would work for a servo system with that many i/o requirements
[17:54:02] <Jymmmm> _methods: No, your momma!
[17:54:12] <_methods> i'm in the dirty south dawg
[17:54:26] <Jymmmm> _methods: Compton?
[17:55:09] <_methods> no the real south
[17:55:15] <_methods> with the kkk
[17:55:16] <Jymmmm> Brazil?
[17:55:34] <_methods> the ones that tried to fight off that war of northern aggression
[17:55:40] <Jymmmm> _methods: You mean those sheets wearing people?
[17:55:44] <_methods> yeah man
[17:55:58] <_methods> charleston sc
[17:56:02] <Jymmmm> _methods: Ah, I wonder if that's 400 coutn threads?
[17:56:08] <_methods> hahah
[17:56:29] <_methods> never bothered to ask
[17:56:34] <Jymmmm> I doubt it's EGYPTIAN cotton though
[17:56:41] <andypugh> frank__: What sort of IO? (Voltage, current levels?)
[17:56:56] <_methods> why you askin if i'm in AU
[17:57:10] <Jymmmm> _methods: It's their summer and you said 81 I thought
[17:57:20] <Jymmmm> on xmas day
[17:57:20] <_methods> oh that was like 2 days ago lol
[17:57:29] <_methods> yeah supposed to be 81 tomorrow
[17:57:32] <_methods> and 80 on xmas
[17:57:43] <Duc> andypugh: does the motor spin freely or is there a breaker on the unit
[17:58:08] <Jymmmm> _methods: 39/27 on Thursday, 39/26 on Friday
[17:58:15] <_methods> omg
[17:58:21] <frank__> anypugh: you mean the drivers needs? they are 1kw ac 220v panasonic
[17:58:27] <_methods> too cold
[17:58:41] <_methods> i'll be rockin the flip flops while i open presents
[17:58:41] <_methods> lol
[17:58:45] <frank__> they tech ref. manual actually has up to 24 pins to connect to each driver
[17:58:50] <frank__> the*
[17:58:52] * Jymmmm chestnuts NOT roasting on an open fire =(
[17:58:57] <_methods> tommy bahamas and margaritas
[17:59:06] <andypugh> I mean, what is the spec of the required IO?
[17:59:22] <_methods> sux i have to work 1/2 day tomorrow
[17:59:34] <Duc> frank_: do you have a link to the manual
[17:59:41] <Jymmm> _methods: Why even bother? Nobody will get anything done
[17:59:47] <Duc> seems like alot of I/O to start a drive
[18:00:17] <_methods> yeah i'm just gonna get drunk at work
[18:00:25] <_methods> gonna bring in a bottle of baileys lol
[18:00:34] <frank__> sorry for the confusion, im new to linuxcnc and figuring out my i/o neeeds for the router, gantry type
[18:01:11] <_methods> wow i'm watching the news and the minneapolis airport got shutdown by a bunch of white black lives matter protestors
[18:01:14] <_methods> lol
[18:01:15] <andypugh> Duc: It is cap-start with a current relay and thermal trip. If I spin it by hand it runs. Otherwise it just buzzes and trips the thermal trip. The current relay checks out OK. The fact it trips the thermal overload suggests that the starting winding is conductive, but could be shorted. The capacitor tests at 20uF but the case says 43uF. But that doesn’t seem like enough differene to stop it working completely.
[18:01:20] <frank__> the thing is this servo motors have to many i/o to allocate, which i have reduced to the most important ones, making them about 14/15 per driver
[18:01:23] <_methods> i didn't even know they had black people in minnesota
[18:01:59] <_methods> i guess that would explain why all the protestors were white
[18:02:01] <andypugh> frank__: Is the IO 5V logic-level, or 24V many-amps?
[18:02:06] <Duc> _methods: huge immigrant pop there from a african place
[18:02:14] <Duc> my sister used to live there
[18:02:23] <_methods> wild
[18:02:40] <frank__> there are 2 inputs which are com+ com- that require 12-24 v dc power
[18:02:45] <_methods> hey welcome to 'murrica we're gonna drop you in the middle of nowhere
[18:02:57] <Duc> andypugh: or bad caps somewhat reminds me of a RPC with the starting caps
[18:03:04] <andypugh> frank__: But, anyway, you can plug one of these extra-IO boards directly into the 7i77: http://store.mesanet.com/index.php?route=product/category&path=83_88
[18:04:07] <andypugh> And if that isn’t enough, you can add 8 more of those boards to the second header of the 5i25 through a http://store.mesanet.com/index.php?route=product/product&path=83_87&product_id=117
[18:04:30] <frank__> actually now i have none,u think i should use the 7i77+5i25?
[18:05:02] <Duc> I've had good luck with the 5i25+7i76 and now 5i25+7i77
[18:06:23] <andypugh> frank__: I can’t say what you _should_ use. And I will point out that Mesa are showing 4-6 weeks delivery delays on their web site.
[18:07:35] <frank__> yes the drivers are rs-422 and the 25 i/o pins from the driver are really small, nothing like parallel port 25 pin connctor
[18:07:54] <andypugh> All I am really saying is that if you keep adding money you can keep adding IO. Up to 432 on the smart-serial bus as well as the 48 IO lines on the 7i77
[18:08:47] <andypugh> 7i77 won’t controi RS422 drives. Do you know the protocol?
[18:10:07] <frank__> what do you mean by protocol :) ? sorry, im new, first build
[18:10:27] <andypugh> If the drives take RS422 then that’s serial data commands.
[18:10:53] <andypugh> The 7i77 supplies +/-10V analogue command voltages.
[18:11:06] <frank__> okey
[18:11:52] <andypugh> (Well, actually, it could be RS422 voltage level step-dir commands.
[18:12:43] <andypugh> So, you can’t decide on an interface until you know what command protocol the drives need.
[18:13:00] <os1r1s> andypugh Where did you see the 4-6 week lead times on mesa's website?
[18:13:22] <andypugh> (Serial data, step/dir, analogue voltage, current loop, direct PWM, three-phase PWM.. things I have forgotten)
[18:13:24] <frank__> as manual specs: Differential input; parameter-selectable (1) Positive/Negative 2) Phase A/Phase B 3) Command/Direction)
[18:13:43] <andypugh> os1r1s: On the Delivery Information page
[18:14:28] <frank__> its IGBT PWM method sine wave drive driver
[18:14:42] <os1r1s> andypugh I'll be damned, its there. I had not idea
[18:14:47] <os1r1s> no
[18:15:00] <JT-Shop> wow
[18:15:35] <andypugh> frank__: Ah, OK. Those all correlate to the different step types that the LinuxCNC, Pico and Mesa step generators can produce. So your interface hardware is likley to look a lot like a stepper system. Unless you want encoder feedback to LinuxCNC.
[18:17:31] <andypugh> frank__: Your best match is likely to be 5i25 + 7i85S (note the S): http://store.mesanet.com/index.php?route=product/product&path=83_87&product_id=124
[18:18:57] <frank__> yes, i kind of noticed too that with an (kind of) optional output called in-positon, i could close the loop at the driver, instead of linuxcnc
[18:19:49] <andypugh> frank__: Or, alternatively, in the 50-pin card family (rather than DB25 type) http://store.mesanet.com/index.php?route=product/product&path=83_86&product_id=106 and one of these in the PCI slot: http://store.mesanet.com/index.php?route=product/product&path=83_85&product_id=297
[18:20:55] <frank__> with the one mentioned before (7i85s) i would need 1 for each driver right?
[18:21:07] <andypugh> My current build is 5i24-based. Don’t get hung up on the 5i25 cards, the 5-pin cards have more pins… (23 IO rather than 17 IO per port, 3 ports per card rather than two)
[18:21:30] <andypugh> I mean 50-pin, not 5-pin
[18:22:16] <andypugh> frank__: No, the 7i85 can handle step-dir-enable to 4 drives and count encoder feedback from 4 drives.
[18:22:16] <JT-Shop> I thought I had cables for them way back when but found out I was missing a few pins on the cable lol
[18:25:01] <frank__> so i should just use the most important i/o options of the driver right? the drivers connector and options are up to 25 pins per driver, for example for positive overtravel limit, negative overtravel limit, command pulse inhibition, etc
[18:25:08] <Duc> frank_: I would be curious to see if the drives needed 14 I/O points since mine also seemed that way but alot could be eliminated
[18:25:41] <Duc> frank_: these can probably be eliminated positive overtravel limit, negative overtravel limit, command pulse inhibition, etc
[18:26:22] <andypugh> frank__: Getting a bit involved, but if you look here at a typical firmware for 5i25 + 2 x 7i85 you will see that one of the 5 RS422 channels is set aside for a smart-serial expansion card, whilst the others are step-dir generators: http://git.linuxcnc.org/gitweb?p=hostmot2-firmware.git;a=blob;f=PIN_7I85Sx2_34.vhd;h=8fb5756afbffa7e9f5a07a45479c0a04b7f5d9df;hb=HEAD
[18:27:20] <frank__> exactly, as now, i think i have gotten it to 14, com+ com- puls1,2 sing 1,2, serv-on alarm clear (inputs of driver) alarm OA (2)OB (2) OZ (2) and Cz that i dont know whats used for (open collector output for phase z)
[18:27:41] <andypugh> So, you would send high-speed step-dir directly from the 7i85 and send other IO (at only 1kHz) via a smart-serial expansion board.
[18:28:18] <andypugh> Phase Z is encodr index.
[18:29:31] <frank__> yes i have phase z, phase a, phase b, _and_ a single wire phase z Open collector output of phase Z signal that says :The emitter terminal of the transistor in the output circuit is connected to the signal ground (GND) and thus not isolated.
[18:29:35] <andypugh> You may be double-counting differential pairs there. The 7i85 has + and - for each step and dir signal. That’s 2 IO, not 4 (at both ends)
[18:30:02] <frank__> lol
[18:30:06] <frank__> newbie mistake there
[18:31:24] <andypugh> Have you looked at Pico and General Mechatronics?
[18:31:47] <frank__> not really
[18:31:49] <andypugh> http://www.pico-systems.com/univstep.html
[18:32:04] <andypugh> http://www.generalmechatronics.com/en/linuxcnc
[18:32:32] <andypugh> Just because Mesa is the only stuff _i_ know anything about doesn’t mean that it is the only choice.
[18:33:33] <andypugh> Right, I am wandering off. I will read back later.
[18:33:44] <frank__> thanks again!
[18:49:14] <JT-Shop> https://www.youtube.com/watch?v=3lKCUuyojDI
[18:50:40] <JT-Shop> https://www.youtube.com/watch?v=uO3IG-oRpis
[18:51:33] <JT-Shop> my spider left...
[19:41:30] <andypugh> You are an arachnophile then?
[19:43:23] <Duc> what type of spider?
[19:45:27] <_methods> kenny powers spider
[19:47:14] <andypugh> I used to have a pet spider in my living room. She would wander across the room at prety much the same time each evening.
[19:47:30] <andypugh> Huge, we called her “Shelob”
[19:47:55] <_methods> was she scared of bright lights lol
[19:47:57] <andypugh> (This is for UK non-lethal values of “huge”)
[19:49:08] <andypugh> Like, 3” leg span. Pretty tiny by international standards, I suppose.
[19:53:48] <_methods> spiders have never bothered me much, but ants.......
[19:53:58] <_methods> those little bastards are evil
[19:54:39] <andypugh> I quite admire ants. The only things I can’t get on with are those that lose legs easily. Like Crane Flies
[19:55:15] <_methods> ants get worse the closer you get to the equator
[19:55:26] <_methods> if you're far from it then ants are cool
[19:55:52] <renesis> theyre annoying but thats it in most of california
[19:55:53] <_methods> get close to that infernal line and ants demonic power increase exponentially
[19:56:25] <andypugh> Yes, some of the tropical ones steal cars, I hear.
[19:56:26] <renesis> we have some red ants that arent huge and dont seem to come inside
[19:56:44] <_methods> fire ants and crazy ants
[19:56:46] <_methods> F that
[19:56:54] <renesis> tropical bugs are fucked
[19:56:57] <_methods> yeah
[19:57:11] <_methods> those .50cal ants in the jungle are hell
[19:57:15] <andypugh> I imagine that ants in Australia are all instantly fatal, like everything else there.
[19:57:31] <_methods> they have some nasty ones there too
[19:58:03] <_methods> apparently the equator is to ants like the sun is to superman
[19:58:23] <_methods> hmmmm
[19:58:29] <_methods> maybe ants were from krypton
[20:00:30] <andypugh> They are famously good at carrying heavy weights
[20:00:56] <andypugh> So, that is the only possible explanation.
[20:01:32] <Duc> lets not forget the tiny black ants that are near impossible to kill
[20:02:21] <os1r1s> So how do you compare constant/peak torque parameters from a servo to a stepper?
[20:02:34] <os1r1s> It seems like the stepper only has one torque parameter
[20:03:17] <andypugh> If you look hard enough you can find torque / rpm plots for steppers.
[20:03:50] <andypugh> But steppers take max current at 0 rpm, regardless of load, so the peak torque is the only torque.
[20:06:13] <XXCoder> gonna love when plan dont get together :P
[20:06:24] <os1r1s> So would you say a 166 oz/in stepper was significantly higher than a constant torque 50 oz/in (350 oz/in peak torque) servo motor?
[20:06:29] <XXCoder> planned to tap alternator to clear derbis and make it run better
[20:06:33] <XXCoder> but its raining very hard
[20:06:38] <PCW> stepper motors spend most of their time pulling outwards on the rotor
[20:07:44] <andypugh> I tend to equate peak servo torque to stepper stall torque. Though whether that is reasonable depends on the exact application.
[20:08:29] <andypugh> So, that servo is (in my mind) twice as strong as the stepper at low rpm, but will move the machine a lot faster.
[20:09:18] <andypugh> In fact the torque @ rpm for servos is so much better that you can gear them down more, and still be faster and stronger.
[20:09:22] <os1r1s> andypugh I think these servo motors will go up to 4700 rpm, which would be 235 ipm on this machine. It would probably tear it apart :)
[20:09:35] <XXCoder> andypugh: encoders help too as it has built in feedback
[20:09:45] <XXCoder> it will stop at wrong position error
[20:10:07] <andypugh> feedback is also what allows the servo to run low current at low load.
[20:11:00] <andypugh> The new closed-loop stepper systems have muddied the waters though. Those look pretty good.
[20:11:06] <os1r1s> So with that said, why pull the encoder feedback into linuxcnc rather than use one of the closed loop drivers?
[20:11:20] <andypugh> They basically run as a 50-pole brushless motor.
[20:12:21] <andypugh> The best way is to run closed-loop velocity in the drive and position in LinuxCNC. Because the LinuxCNC PID knows more about what is needed than the drive does. For one thing, velocity feedforward.
[20:14:07] <Duc> so how much look ahead does linuxcnc have?
[20:14:39] <andypugh> Lots. But that’s a different question,
[20:14:59] <Duc> since that is normally a option for high speed milling on cncs
[20:15:07] <andypugh> The new TP in 2.7 has a configurable lookahead, but I think it defaults to 100segments.
[20:15:37] <andypugh> But that has little to do with where the PID should run.
[20:16:25] <os1r1s> andypugh Can you run a 60v servo at 48v?
[20:16:39] <Duc> yes but you reduce the torque valve
[20:16:42] <Duc> value
[20:16:43] <andypugh> Yes. But it will run proportionally slower
[20:16:53] <os1r1s> Got it
[20:17:02] <andypugh> The torque will be the same if drive-limited
[20:17:09] <os1r1s> And is a linear power supply vs switching relevant?
[20:18:15] <andypugh> os1r1s: For what?
[20:18:32] <os1r1s> A servo application
[20:18:39] <andypugh> For a servo power supply an unregulated supply is likely to be best.
[20:18:46] <os1r1s> Small mill size
[20:18:55] <andypugh> ie, just a transformer, rectifier and smoothing caps
[20:19:30] <andypugh> I run my mill (and the lathe I am currently converting) on simple rectified mains power.
[20:20:07] <os1r1s> Small mill or massive one :)
[20:20:19] <andypugh> But I have 400V drives, so the 320V that the mains gives when rectified is fine. It’s a source of current/charge not voltage, really.
[20:20:26] <andypugh> Mid-sized
[20:20:30] <_methods> he has 3ph
[20:20:42] <andypugh> Who has 3ph?
[20:20:44] <_methods> he's not on our stupid side of the pond
[20:20:46] <_methods> you
[20:20:59] <andypugh> No, I only have 240V single phase.
[20:21:02] <_methods> wut
[20:21:11] <andypugh> 3-phase would be nicer, certainly.
[20:21:12] <_methods> i thought all you euros had 3ph
[20:21:38] <_methods> well that sux
[20:21:41] <andypugh> No, not in domestic properties.
[20:21:46] <_methods> oh man
[20:21:53] <andypugh> Though a friend of mine has it.
[20:22:03] <_methods> so there is no upside to europe at all?
[20:22:05] <_methods> damn
[20:22:11] <andypugh> Because his house is a converted factory.
[20:22:14] <_methods> hahah
[20:22:18] <Duc> 3ph equipment is so much cheaper but rarely do people have it
[20:22:30] <_methods> all my stuff is 3ph
[20:22:34] <_methods> i just vfd it
[20:22:38] <_methods> or rpc
[20:22:45] <andypugh> Now that VFDs are so cheap 3-phase equipment is easy to run.
[20:22:50] <_methods> ^^
[20:23:02] <_methods> plz don't tell anyone that
[20:23:10] <Duc> a 15 hp rpc is cheap now days at around $1,000 ifyou dont buy stuff
[20:23:35] <Duc> build stuff
[20:23:49] <andypugh> And it turns out that you can fit a PC and Mesa cards in the space vacated by a single 1960s rectifier bridge
[20:24:11] <andypugh> (Have you seen those pics?)
[20:24:35] <Duc> lets see them
[20:24:58] <andypugh> Before: https://picasaweb.google.com/108164504656404380542/Holbrook#6157080037365341634
[20:24:58] <Duc> I know I removed a whole cabinet on my bridgeport mill and still have plenty of room left over
[20:25:20] <andypugh> After: https://picasaweb.google.com/108164504656404380542/Holbrook#6216384739653615938
[20:25:45] <Duc> nice what is the cap on the top right for?
[20:26:22] <os1r1s> andypugh What MB/computer is that?
[20:26:41] <andypugh> That makes the suds pump run. No point having a VFD for a suds pump
[20:27:12] <andypugh> That’s a DN2800 that I had lying about. + 7i64, 7i49, 7i44 and 7i84.
[20:28:21] <andypugh> The DN2800 has a lot of good things, like single 12V power supply and a thin form factor. But the graphics are a real struggle with Linyux.
[20:29:56] <os1r1s> andypugh Wow, looking at the stepper torque chart it drops off dramatically after 45 IPM (with 20 TPI leadscrews)
[20:30:40] <andypugh> 20TPI is probably too fine for a practical stepper machine.
[20:31:09] <andypugh> 5mm pitch ball screws would work a lot better.
[20:31:19] <os1r1s> andypugh But more appropriate for servos?
[20:31:27] <os1r1s> Or just not appropriate at all?
[20:31:48] <andypugh> Servos would work better, but swapping screws would be cheaper.
[20:33:54] <andypugh> Running ardnary screw threads at servo speeds will wear them out. Convetional threads are really meant to be tightened once.
[20:34:16] <os1r1s> andypugh Converting this machine to ballscrews I think is a pain
[20:34:38] <andypugh> What machine?
[20:34:45] <os1r1s> andypugh Taig Mill and Sherline Lathe
[20:35:00] <andypugh> Ah, yes.
[20:35:22] <os1r1s> The sherline lathe I think is a little easier
[20:35:26] <os1r1s> The taig requires modification
[20:35:36] <andypugh> So, they are actually Acme thread rather than screw thread?
[20:35:54] <_methods> yeah you'd probably be better off doing an x2 to ballscrew
[20:36:06] <os1r1s> _methods saig x2/
[20:36:11] <_methods> i used some cheap ass chinese ballscrews on mine
[20:36:20] <_methods> but that dind't work so well when i tried cutting steel
[20:36:26] <os1r1s> seif
[20:36:28] <os1r1s> seig
[20:36:32] <os1r1s> Can't spell for shit tonight
[20:36:32] <_methods> yeah
[20:36:51] <_methods> replacing chinese with roton now
[20:37:04] <_methods> the plastic caps blew out on the ballnuts lol
[20:37:19] <_methods> i tried hoggin some steel on an x2 lol
[20:37:29] <Duc> lol how much
[20:37:31] <_methods> didn't go so well for the ballscrews or the colum
[20:37:35] <_methods> column
[20:37:44] <andypugh> I have been OK with Chinese screws, except for a rotating-nut arrangement where the deflectors popped out and the balls locked up.
[20:37:49] <_methods> i was tryin to take .06" DOC lol
[20:38:01] <_methods> with a 3/8" carbide end mill
[20:38:03] <Duc> thats not bad but on a small mill...
[20:38:07] <os1r1s> andypugh I'm not sure if its acme or v-thread. Trying to look now. I'm pretty sure the sherline is v-thread
[20:38:22] <_methods> it did fine till i came off the end of the part and tried to make a 90
[20:38:35] <Duc> I know my bridgeport flipped me off when I tried to face some AR500
[20:38:45] <_methods> lol
[20:38:49] <_methods> yeah that stuff is nasty
[20:38:58] <_methods> i work with it all the time
[20:39:16] <Duc> is usually burn out targets with it but I do tomahawks on the side
[20:39:21] <_methods> weldox, armox, 46100, ar500
[20:39:24] <XXCoder> not as nasty as fiberglass
[20:39:29] <XXCoder> it eats tools
[20:39:32] <XXCoder> leave huge mess
[20:39:32] <_methods> weldox isn't too bad
[20:39:43] <os1r1s> andypugh Taig does have acme
[20:39:54] <_methods> but ar500 and 46100 are sux
[20:40:12] <_methods> machine like dog poo
[20:40:16] <andypugh> Armox is meant to be hard to make holes in. That’s its job.
[20:40:19] <Jymmm> _methods: Try AR15 instead =)
[20:40:38] <Duc> I know my last job required graphite to be machined dry. coolant would fuck up the vacuum chambers
[20:40:44] <_methods> yeah armox is better than ar500 or 46100 thouhg
[20:40:50] <_methods> the swedes make good metal
[20:41:54] <_methods> we need to be payin attention to them here in 'murrica
[20:42:06] <_methods> there is money to be made in quality metals
[20:42:16] <_methods> leave that cheap shit to china and india
[20:42:31] <andypugh> It seems that the UK has pretty much opted out of the making steel and digging coal businesses :-(
[20:42:36] <_methods> yeah
[20:42:46] <_methods> we're doing the same thing
[20:42:52] <andypugh> Last UK deep coal mine closed last week.
[20:43:02] <_methods> yeah i was reading that the other day
[20:43:26] <andypugh> Still millions of tons of coal down there, but to be honest it’s probably best not to burn it.
[20:43:42] <_methods> well i think our children will thank us for not doing it
[20:43:57] <andypugh> Not that anything other than economomics is behind it.
[20:44:00] <_methods> our generation will be widely despised for our waste i'm sure
[20:44:09] <_methods> things like racing for fun and sunday drives
[20:44:15] <andypugh> My children won’t be doing anything.
[20:44:18] <_methods> probably won't be historically loved
[20:44:30] <_methods> "children"
[20:44:39] <_methods> not necessarily yours
[20:45:00] <_methods> especially if you've chosen not to foul the planet with your seed
[20:45:27] <_methods> probably the "greenest" choice
[20:45:47] <andypugh> I wouldn’t say that the choice is mine, but I have been consistently avoided by the child-bearing sex.
[20:46:37] <_methods> well all of these hippies want to blabber on about recycling and whatever then they have 5 damn kis
[20:46:40] <_methods> kids even
[20:47:50] <_methods> no one wants to talk about the carbon footprint of their little fuck trophies
[20:47:56] <andypugh> Yeah, well, I am not going to try for any moral high ground on the not-breeding question when I have never even been in a situation where the question arose.
[20:48:17] <Duc> I have a choice but I think the wife is ok with not having kids
[20:48:34] <_methods> i think most intelligent people are opting out
[20:48:41] <_methods> for the sake of their children that might be
[20:48:57] <_methods> i'd hate to drop a child in this age of stupidity
[20:49:33] <Duc> nah I just can barely stand other humans
[20:50:06] <_methods> they're not slowing down the trailer park breeding program
[20:50:14] <andypugh> Anyway, time to log off. For the next several days I will be doing Christmas with my nephews, nieces and aging parents. Mainly aging parents, and a schedule of house modifications to make wheelchair use more practical :-(
[20:50:18] <_methods> your children will be outnumbered
[20:50:52] <andypugh> Night all
[20:50:57] <_methods> doomed to live out a masterpiece movie called "idiocracy"
[20:50:58] <_methods> lol
[20:51:46] <_methods> i keep waiting for someone to show up and install the brawndo fountain at my work
[20:51:46] <Duc> Im still confused by safe spaces
[20:54:51] <bobo__> methods can't you make a brawndo sticker for the fountion at work?
[20:56:41] <_methods> hahah i suppose so
[20:57:11] <_methods> i'm sure once trump is president the prophecy will be complete
[20:57:19] <_methods> and brawndo will be everywhere
[20:57:54] <bobo__> make some up for the neighboor's front door
[20:58:42] <_methods> heheh
[21:03:46] <bobo__> methods if you have wished for "idiocracy 2" check out https://www.youtube.com/watch?v=CAmNg651Hmg
[21:08:11] <CaptHindsight> "where's your tattoo?"