#linuxcnc | Logs for 2016-11-06

Back
[01:18:01] <terkaa> hi all
[01:34:43] <XXCoder> well it finally happened
[01:34:50] <XXCoder> izzy swan went crazy. https://www.youtube.com/watch?v=Z37g5NZL8u4
[02:31:04] <Deejay> moin
[03:05:02] <linux_vampire> morning
[03:05:12] <XXCoder> good morning madam
[03:05:43] <linux_vampire> lol
[03:05:51] <XXCoder> heh well whats up
[03:06:25] <linux_vampire> i'm trying to display my current feed rate
[03:07:40] <linux_vampire> i saw this post https://forum.linuxcnc.org/forum/21-axis/30416-display-current-feedrate
[03:10:04] <linux_vampire> i clicked the "show velocity" on the menu but i don't see any change on the screen
[03:53:43] <XXCoder> dunno
[03:53:59] <XXCoder> wtf
[03:54:04] <XXCoder> ah yes fall back
[03:54:18] <XXCoder> I wondered how I spoke much later but was earlier than you linux_vampire
[04:19:51] <linux_vampire> what do you mean? XXCoder
[04:20:38] <XXCoder> you last said something at 1:41, and 40 minutes later I said something at 1:25 lol
[04:23:14] <linux_vampire> i need to add the current feed rate, and make 2 digital outputs
[04:24:25] <jthornton> morning
[04:25:15] <XXCoder> jt I tweaked printed so hard, and printed 3d benchy
[04:25:26] <XXCoder> still has an issue but much better
[04:30:48] <linux_vampire> nice,
[04:31:16] <linux_vampire> I'm go for a few.. cooking time
[04:31:21] <XXCoder> have fun
[04:31:37] <linux_vampire> thanks
[04:34:36] <jthornton> I found the print fans a huge difference on the bench test
[04:35:18] <XXCoder> I bet. mine had slight issue with heat but luckly worked fine. no my printer most major problem is damned zits and lines agaibn
[04:41:55] <jthornton> I need to take the Z axis apart and replace the bearings with the double long ones and tighten up the rod pockets in the acrylic parts
[04:42:04] <jthornton> but first time to check the coffee
[04:44:02] <XXCoder> ah yea
[04:44:31] <XXCoder> still unsure on some stuff for my printer lol like how setup fans. im waiting on induction probe before I plan fans
[04:52:15] <jthornton> you just plug them in and they work
[04:52:45] <XXCoder> oh I know, I meant where to place em, so on
[05:02:38] <jthornton> did you see the photo of where I put mine?
[05:04:23] <XXCoder> cant recall heh
[05:08:37] <XXCoder> jthornton: got pics? argh my printer is still making those damned lines
[05:08:42] <jthornton> http://gnipsel.com/images/prusa-i3-clone/print-fans-01.jpg
[05:08:46] <XXCoder> its really annoying
[05:09:12] <jthornton> I just made a L bracket to connect the print fans to the hot end fan
[05:09:16] <XXCoder> I still find it funny how different ours are now.
[05:09:43] <jthornton> see the tape shim on the bottom of the right Z rod?
[05:09:53] <XXCoder> yeah
[05:10:01] <XXCoder> I guess its to prevent movements there
[05:10:30] <jthornton> that took out some of the slop in the rod which translated to the print head
[05:10:58] <XXCoder> is your x assembly rods rubbing on screws?
[05:12:29] <XXCoder> https://www.simplify3d.com/wp-content/uploads/2015/09/Blobs-And-Zits.jpg
[05:12:34] <XXCoder> thats problem my printer has
[05:12:37] <XXCoder> really annoying
[05:16:52] <jthornton> I got them when I had random start points checked off
[05:17:16] <XXCoder> I wonder if its slic3r bug
[05:17:33] <jthornton> could be
[05:17:41] <XXCoder> im trying 2 mm retract and -2mm return :P meaning it dont bother to come back
[05:17:51] <jthornton> I don't understand the question about my X assembly rods
[05:17:51] <XXCoder> just see what the heck going on
[05:18:08] <XXCoder> oh my machine, the X rods rub on Z screws
[05:18:16] <XXCoder> leaving the screw wobble
[05:18:21] <_methods> that's usually cause by an issue with your extruder not feeding consistently?
[05:18:27] <_methods> are you using a bowden setup?
[05:18:31] <XXCoder> (not z screw wobble, which is confusing lol)
[05:18:49] <XXCoder> met bowden yeah
[05:18:59] <_methods> yeah get rid of that and you'll fix that
[05:19:16] <_methods> or try and remove slack/movement from the bowden setup
[05:19:25] <jthornton> yea mine broke through and I had to put some shaft collars on the upper X rod
[05:19:37] <jthornton> get a photo after while
[05:19:39] <XXCoder> mine never had any lol
[05:19:53] <XXCoder> so it always rubbed on screw
[05:21:24] <XXCoder> shift collars interesting lemme look up
[05:21:42] <XXCoder> that is nice
[05:21:59] <XXCoder> 8mm
[05:24:52] <XXCoder> jth: trying aligned now not ranodm
[05:25:24] <XXCoder> where did you buy that shaft collar?
[05:25:28] <jthornton> yea I had to order some from mcmaster carr
[05:25:36] <XXCoder> _methods: movement you meant tube movements or?
[05:26:05] <_methods> yeah tube movement
[05:26:15] <_methods> try and get that out of the system if possible
[05:26:16] <XXCoder> it affects print?
[05:26:22] <_methods> better yet get rid of the damn bowden setup
[05:26:27] <_methods> of course
[05:26:33] <jthornton> I too some copper wire and wrapped the fittings under the collar to keep them from moving during a retract
[05:26:37] <_methods> you have a tube constricting your filament movement
[05:26:37] <XXCoder> how because if I restrain it it will rip out if it fgets out of range
[05:26:55] <_methods> don't restrain it
[05:27:06] <_methods> try and set it up so it has as little motion in it as possible
[05:27:14] <XXCoder> ah yea lemme look a sec
[05:27:24] <_methods> you want to remove any source of impingement in the filament path in the tube
[05:28:15] <XXCoder> mine curves pretty smoothly from upwards curve down to hotend assembly
[05:28:37] <_methods> obviously not
[05:29:08] <XXCoder> been figuring how to cut in least 200mm off
[05:29:23] <XXCoder> and still no harash bend
[05:30:17] <XXCoder> https://www.amazon.com/Ruland-MSC-8-F-Screw-Collar-Metric/dp/B0063KWI42/
[05:31:13] <XXCoder> LOL!! https://www.amazon.com/clamp-collar-single-split-1-0414-burnished/dp/B00872AG20/
[05:32:51] <archivist> burnished my arse, that looks drilled/turned
[05:33:03] <XXCoder> it gets annoying after 10th item with $50 shipping
[05:34:05] <XXCoder> jthornton: aligned the lines disappear, so there is only screw rub wobble
[05:34:21] <XXCoder> can see screw length on part lol
[05:34:39] <XXCoder> (thread thickness I meant)
[05:35:18] <XXCoder> https://www.amazon.com/Climax-Metal-M1C-08-S-Collar-Stainless/dp/B000IL5GPK/ finally no gold plated shipping method
[05:35:58] <jthornton> XXCoder: you don't want a screw collar you want a split collar
[05:36:31] <XXCoder> kind that come in 2 parts?
[05:37:00] <jthornton> yea so you don't have to take it apart to put them on
[05:37:14] <XXCoder> ah that makes sense yeah
[05:37:27] <jthornton> https://www.mcmaster.com/#shaft-collars/=14x4yb7
[05:37:52] <jthornton> https://www.mcmaster.com/#6063k14/=14x4yiq
[05:38:11] <jthornton> damn time change been up since 4am now
[05:38:11] <XXCoder> omg
[05:38:24] <XXCoder> just spotted bolt hole
[05:38:33] <XXCoder> lemme try that
[05:44:16] <XXCoder> worked.
[05:44:32] <XXCoder> rod still touches screw a little but not pressing on it anymore
[05:51:39] <XXCoder> wow that sure messed with leveling
[06:00:02] <jthornton> anything you do messes with leveling lol
[06:00:22] <XXCoder> including looking at it wrong way
[06:00:42] <XXCoder> glad I never removed that dial dindictor adoptor LOL
[06:00:47] <XXCoder> sometimes lazyness wins :P
[06:08:47] <XXCoder> I had to hit zstop to stop it from trying to go down when maxed down, so now its bit messed up now :P
[06:08:56] <XXCoder> print smooshes a little on first layer
[06:15:47] <XXCoder> jthornton: no more screw rub wobble
[06:16:01] <XXCoder> but very bad first layer lol gonna realign z later
[06:18:05] <jthornton> cool
[06:18:35] <XXCoder> so thats 2 kinds of z related wobble thats gone lol
[06:23:06] <jthornton> you just stuck a 2.5mm screw in the hole on the left Z thingy?
[06:23:16] <XXCoder> screwed down to it
[06:23:33] <XXCoder> made sure rod is kinda out of way
[07:04:04] <linux_vampire> hii
[07:10:37] <XXCoder> wb
[07:41:08] <linux_vampire> is there a way to reload pyvcp without close and open the whole linux cnc?
[07:42:30] <archivist> not really as the connections need to be made and checked
[07:43:07] <linux_vampire> :(
[07:44:17] <linux_vampire> also i want to change the font size
[07:52:15] <archivist> for all of axis or just the panel
[07:53:00] <archivist> or just titles/whatever
[08:03:37] <andypugh> linux_vampire: The font size is a parameter in the XML file for the PyVCP
[08:04:05] <andypugh> (And you might find GladeVCP to be a better option, before you go too far down this line)
[08:04:49] <pink_vampire> lets see
[08:07:13] <pink_vampire> yes, it look much better
[08:08:09] <pink_vampire> how do i add it to the linux cnc?
[08:10:23] <pink_vampire> andypugh:
[08:15:14] <pink_vampire> it's like the extensions is firefox?
[08:15:47] <pink_vampire> i can't find stuff like that in the menu
[08:16:13] <archivist> read the integrator document
[08:18:14] <pink_vampire> ?
[08:18:32] <pink_vampire> where i fund it?
[08:20:56] <pink_vampire> GladeVCP is Gmoccapy ?
[08:30:44] <jthornton> in the docs
[08:42:32] <pink_vampire> there is no GladeVCP in the "user interface" section
[08:44:23] <pink_vampire> http://i.imgur.com/a8fU1bL.png
[08:44:32] <pink_vampire> http://linuxcnc.org/docs/html/
[08:45:45] <pink_vampire> maybe something missing?
[08:46:17] <archivist> did you google GladeVCP
[08:46:46] <pink_vampire> yes
[08:46:52] <pink_vampire> 08:52:34] <pink_vampire> GladeVCP is Gmoccapy ?
[08:47:39] <archivist> wrong question
[08:48:12] <archivist> you should have ended up at http://linuxcnc.org/docs/html/gui/gladevcp.html
[08:52:16] <pink_vampire> i have no idea even what I'm looking for, actually i can't understand why in 2016 (almost 2017) there is no decent user interface and convenient way to change settings and install plugins.
[08:56:33] <pink_vampire> maybe the linuxcnc core is grate, but all the surround stuff should be tied down, in a presentable way.
[08:57:27] <Kevin`> what do you mean? 2016 is the year of the half-assed web ui
[08:57:55] <pink_vampire> maybe is the female touch that missing...
[09:00:38] <pink_vampire> i mean that the user interface and the settings need to be more slick, easy and intuitive to edit.
[09:02:17] <pink_vampire> like installing an extension in firefox, you have a list, you select it, and you have it.
[09:03:23] <pink_vampire> edit text files is not the way to go.
[09:09:12] <nubcake> anyone here using the geeetech GT2560 Controller Board for their 3D printer?
[09:09:45] <pink_vampire> wrong channel
[09:10:00] <nubcake> well i do know that, but maybe someone could point me to the right channel
[09:10:13] <nubcake> that's why i was asking
[09:10:16] <pink_vampire> #reprap
[09:10:24] <nubcake> thanks :)
[09:13:03] <JT-Shop> because LinuxCNC is a machine controller not a game
[09:17:33] <archivist> pink_vampire, you can write a slick pointy clicky gui maker if you want
[09:18:11] <archivist> everyone are volunteers, they write what they want at the time
[09:19:29] <pink_vampire> also the toolpath display look so old
[09:20:50] <JT-Shop> the cool thing about open source software is YOUR free to change it but don't expect someone else to do your work
[09:22:37] <pink_vampire> i know.. but i'm not the a programmer, i'm the user.
[09:24:05] <JT-Shop> then hire a programmer
[09:24:19] <archivist> other option is to go without or pay someone to implement what you want, which is how the path pilot interface came about
[09:24:40] <JT-Shop> think of it like this my yard is a mess and it's almost 2117 YOU need to come fix it because I don't know how
[09:24:53] <JT-Shop> exactly
[09:25:14] <JT-Shop> complaining about something never gets you anywhere
[09:26:11] <pcw_home> Also you are not a really just a user, but also a system integrator
[09:26:12] <pcw_home> if you buy a Tormach, you are a user
[09:28:04] <pink_vampire> i'm using solidworks, and add / remove icons in the menu take 10 sec, here i need to become "system integrator" just to add a button to menu.
[09:29:02] * JT-Shop wanders away
[09:29:08] <archivist> unreasonable expectations
[09:29:11] <BeachBumPete> yup so you are on your own unless you can afford to pay someone to do the programming for you. I mean hey not everyone is a programmer type after all....
[09:31:22] <pink_vampire> http://i.imgur.com/BqIX5cK.png
[09:31:55] <pink_vampire> why linux cnc can't be that easy to change the interface
[09:31:58] <pink_vampire> ???
[09:33:22] <pcw_home> Because they dont have 20 full time programmers writing user interface code?
[09:33:38] <pink_vampire> it work like that for almost any program, so i'm sure someone already invent the wheel.
[09:33:55] <archivist> not that simple
[09:35:27] <archivist> try writing something to understand the complexity and effort, then try to understand it is a hobby for most
[09:36:52] <archivist> adding an icon is very different to adding code and functionality
[09:38:00] <pink_vampire> but this is the #1 issue with linux cnc, it's a controller that not design to a specific machine, it is critical to be able to get a customizable interface to the machine that each user have.
[09:38:45] <archivist> you have already learned that you can with pyvcp!
[09:38:58] <JT-i3> why do you just continue this useless bitching?
[09:39:22] <CaptHindsight> archivist: someone here has discovered that the squeaky wheel gets oiled first, whining has worked so well so far
[09:41:13] <pink_vampire> all i'm saying that yeah it's a hobby, but hobby stuff not have to look outdated , so maybe we can push linux cnc forward to become very good and slick cnc controller.
[09:41:42] <pcw_home> We look forward to you contributions
[09:41:50] <CaptHindsight> exactly
[09:42:05] <archivist> http://thebestyoumagazine.co/i-want-doesnt-get-by-mel-carnegie/
[09:42:51] <pink_vampire> maybe one of you as a programmers can estimate what need to be done to reach that goal.
[09:43:29] <archivist> faster to edit one of the panels than write something new
[09:44:49] <archivist> and remember a pointy clicky misses the point of true programmability
[09:45:07] <pcw_home> Also whos goal is it?
[09:45:15] <pink_vampire> maybe, but it's not the right way to do something.
[09:46:00] <tiwake> oh man, I made some really kick butt stir fry last night
[09:46:28] <pink_vampire> if you design something you need to have a path to go inside and a goal.. otherwise you go lost. like linux cnc go lost.
[09:46:35] <pcw_home> it may be the right way if the configuration space is too large to adequately represent with menu choices
[09:48:41] <pink_vampire> the way i see it, the user interface is the face of ant product, and without good interface you are going to spend a lot of time for no reason.
[09:48:59] <archivist> I dont think people understand the limitless choice problem
[09:49:18] <CaptHindsight> tiwake: thanks, now I'm hungry :)
[09:49:22] <pink_vampire> what do you mean?
[09:49:33] <pink_vampire> archivist: ^
[09:49:43] <archivist> an icon cannot represent all the choices
[09:50:00] <archivist> icons are terrible for that
[09:50:34] <archivist> "modern looks" does not mean works well
[09:50:51] <pink_vampire> you can drag and drop whatever you want,
[09:50:53] <tiwake> CaptHindsight: its literally the best stir fry I've ever made, on par with my dads stir fry, who is a professional chef and custom caterer for 45 years
[09:51:02] <tiwake> so good
[09:51:27] <archivist> pink_vampire, it is what is behind under the hood not the drag operation
[09:52:41] <IchGucksLive> hi
[09:52:51] <pink_vampire> i believe that all the stuff that we do now by edit text files we can do easier and faster.
[09:52:56] <pink_vampire> hi IchGucksLive
[09:53:22] <IchGucksLive> did you get you ports to work
[09:53:36] <archivist> not true, there is no sensible programming in icon dragging
[09:54:40] <pink_vampire> IchGucksLive: i'm trying to solve much more important issue - the poor user interface.
[09:54:41] <archivist> there is no way to create a hobbing machine interface by drag and drop, there is maths behind it
[09:54:53] <tiwake> archivist: yeah, though for the most part that only works for people who are doing it all the time, us lowly peons who do not do that stuff for a living can not do it enough to streamline such a process and have it worth while :P
[09:54:58] <IchGucksLive> most of us like it
[09:55:15] <IchGucksLive> as yo can add your buttons and widgets to sidepanel
[09:55:30] <IchGucksLive> what are yoiu usingf
[09:55:37] <IchGucksLive> gmocapy ?
[09:55:42] <IchGucksLive> axis
[09:55:59] <IchGucksLive> there are so many examples in the sim area
[09:56:13] <IchGucksLive> noone neads more to work with
[09:56:15] <pink_vampire> IchGucksLive: now i'm using the thing on the side
[09:56:29] <IchGucksLive> pyvcp
[09:56:34] <pink_vampire> yes
[09:56:47] <IchGucksLive> cool there you can do so many things
[09:57:01] <IchGucksLive> logo buttons for mdi
[09:57:11] <IchGucksLive> infos to the user
[09:57:26] <IchGucksLive> easy and workable
[09:57:45] <pink_vampire> i need to add probing option, current feedrate, and few buttons for macros
[09:57:57] <IchGucksLive> then axis has all this nice subprogramms
[09:58:02] <Tom_L> pink_vampire you're not solving anything
[09:58:07] <Tom_L> you're just bitching
[09:58:47] <JT-i3> the reason I'm ignoring
[09:59:25] <_methods> Yeah I don't know why I helped her get her card working
[10:00:07] * Tom_L kick bans _methods
[10:00:32] <_methods> Sorry
[10:00:34] <Tom_L> cause you're just that kinda guy
[10:00:51] <JT-i3> Tom_L: I've solved all the problems to make the bot like I like it now to put it all together]
[10:01:05] <Tom_L> good stuff
[10:01:14] <CaptHindsight> _methods: did you have problems with the wch ch382l in the past?
[10:01:45] <Tom_L> my smartphone is confused today
[10:01:53] <Tom_L> it's showing 2 different times
[10:02:24] <_methods> I just had issues getting it working
[10:02:57] <_methods> Not being recognized
[10:03:12] <CaptHindsight> huh, just works every time I try
[10:03:21] <_methods> Like what was happening to stink vampire
[10:03:42] <CaptHindsight> does it act that way when there is more than one LPT port?
[10:04:20] <CaptHindsight> e.g. integrated LPT plus ch382 card
[10:04:34] <_methods> Can't remember I got it working agrees ago
[10:08:28] <_methods> Ages even
[10:09:28] <pcw_home> Looks like wheezy's RTAI kernel doesn't recognize it
[10:09:30] <pcw_home> CaptHindsight: did you use another OS/kernel?
[10:10:43] <CaptHindsight> pcw_home: I believe it was Wheezy from around summer of '15. I'll check for sure later this week.
[10:10:59] <CaptHindsight> definitely nothing newer
[10:11:10] <andypugh> pink_vampire: GladeVCP is an alternative to PyVCP.
[10:11:37] <andypugh> pink_vampire: If you want to make your own pretty UI, then Gscreen was designed to be easy to edit.
[10:13:04] <pink_vampire> andypugh: but how i get them
[10:13:04] <pink_vampire> ?
[10:13:51] <pink_vampire> what bring them to linux cnc? it's a plugin? extension? i need to apt-get install something?
[10:13:58] <archivist> have you read the integrators manual yet, GET is generally not needed
[10:14:54] <pink_vampire> GET 200?
[10:15:05] <pink_vampire> ok
[10:15:06] <archivist> a minor edit instead
[10:23:59] <JT-i3> umm archivist we don't have an integrators manual any more
[10:24:32] <archivist> why the fuck not, what is it called now
[10:24:41] <JT-i3> the manual
[10:24:47] <JT-i3> Documents
[10:25:15] <JT-i3> everything is in one manual so you can link to anywhere in the PDF's
[10:25:36] <archivist> please dont make it impossible for people to help users
[10:26:02] <archivist> I hate pdf with a passion, I use online docs
[10:26:35] <JT-i3> no one said you have to look at the PDF
[10:27:40] <pink_vampire> all i want is to get that to work https://forum.linuxcnc.org/49-basic-configuration/29187-work-with-probe
[10:27:44] <Roguish> JT-i3: the newish manual structure is fine. some ways better. html and pdfs are both good.
[10:28:08] <pink_vampire> html is better
[10:28:09] <sector_0> hey
[10:28:40] <JT-i3> everyone must be on the rag today
[10:28:43] <pink_vampire> enable a way to comment about pages
[10:28:46] <sector_0> If I'm gonna be using linux cnc, what needs to sit in between my stepper motor drivers, and the pc?
[10:29:00] <archivist> a break out board
[10:29:45] <sector_0> archivist, can a dev board, like an STM32F4 work?
[10:29:54] <Tom_L> JT-i3 hah
[10:30:11] <archivist> sector_0, er wot that is over the top
[10:30:41] <sector_0> archivist, well I suspect so but it's what I have lying around
[10:31:10] <sector_0> I've previously used it as the sole controller for the steppers, but I want to switch to linuxcnc
[10:31:14] <archivist> the break out board has opto isolators and drivers
[10:31:48] <sector_0> archivist, it has drivers?
[10:32:03] <sector_0> so the steppers connect directly to the breakout?
[10:32:30] <archivist> no ttl/open collector drivers
[10:32:37] <JT-i3> no, you need stepper drives
[10:32:54] <pink_vampire> http://www.christoph-lauer.de/wp-content/uploads/2015/05/STM32F4-DISCOVERY_BoardContent1.jpg
[10:32:58] <archivist> they drive thee optos in stepper drives
[10:32:59] <pink_vampire> sector_0: ^
[10:33:05] <sector_0> oh ok
[10:33:19] <JT-i3> TB6600 or something like that is the cheapest IIRC
[10:33:32] <sector_0> pink_vampire, yeah that's what I have
[10:33:48] <pink_vampire> and you have drives?
[10:34:13] <nubcake> wow.. #reprap really is helpful.. (not) :D
[10:34:25] <sector_0> pink_vampire, yeah L298N
[10:34:31] <sector_0> for NEMA17s
[10:35:52] <JT-i3> nubcake: what are you seeking? and yes #repcrap can be a bitch
[10:35:55] <linux_vampire> .
[10:37:20] <_methods> time to make some chocolate chip cookies
[10:37:25] <nubcake> JT-i3: my Geeetech Prusa i3 Pro-B (GT2560) didn't like that i flashed the Marlin RC Nov. 03 and went all gibberish display output.
[10:37:47] <sector_0> pink_vampire, so will my current setup work?
[10:37:51] <JT-i3> what did you change?
[10:38:02] <Roguish> time for my ride..............
[10:38:10] <JT-i3> enjoy Roguish
[10:38:31] <nubcake> i just made the adjustments for board_type and auto-bed leveling coordinates
[10:38:49] <archivist> JT-i3, see http://linuxcnc.org/docs/2.7/html/ go down to user interfaces section, no overview section for the vampire, ie concept of edit ini to choose display
[10:39:11] <nubcake> just took the values from the original firmware and put them into the new one
[10:39:25] <JT-i3> did you set up the lcd too?
[10:39:36] <nubcake> whoops, yes i did
[10:39:42] <nubcake> forgot to mention
[10:39:45] <JT-i3> archivist: nope it's in the ini section
[10:40:05] <archivist> that is why noobs get lost!
[10:40:08] <JT-i3> might have made a mistake?
[10:40:32] <nubcake> i'll doublecheck again ^^
[10:40:40] <JT-i3> hmm I want edit the ini file so I look in how to use a gui no wonder I get lost all the time
[10:40:59] <nubcake> oh wow..
[10:41:02] <JT-i3> nubcake: yea sounds like a typo or something wrong in the lcd secion
[10:41:10] <nubcake> weird...
[10:41:24] <nubcake> i flashed a basic arduino example (blinking)
[10:41:37] <nubcake> then started arduino ide 1.5 and flashed the old firmware - worked
[10:41:43] <JT-i3> that can happen too
[10:42:37] <nubcake> well, now i have the display being normal and stuff back, but it's the old firmware :/ brb checking for typos.. thanks for helping btw.
[10:42:45] <archivist> ini section is after the user interface section no wonder she is lost
[10:43:58] <archivist> I think the ini is more important than the important section
[10:44:41] <archivist> does a noob really need to know about the TP
[10:45:07] <JT-i3> you can blame the manuals on her getting lost all day long...
[10:59:16] <archivist> http://linuxcnc.org/docs/2.7/html/user/user-intro.html also does not tell the user how they are chosen/used
[10:59:29] <archivist> the guis
[11:28:58] <Jymmm> Any time you are looking at cad drawings for a furnace and you there is mention of marshmallows, it's always a good thing =)
[11:31:02] <Jymmm> Hmmm, didnt realize salt melted at 1474F (801C), I thought it was much higher than that.
[11:39:33] <roycroft> salt is an ancient (and still commonly used) glaze on ceramics
[11:40:20] <roycroft> think of the german bier stein - the big grey ones with the bluish glaze
[11:40:23] <roycroft> that is salt glaze
[11:40:38] <Jymmm> nice, I was gonna say LICK IT =)
[11:40:59] <Jymmm> roycroft: So once glazed it is no longer soluable in water?
[11:41:52] <pink_vampire> why type of salt?
[11:42:42] <roycroft> correct
[11:43:13] <roycroft> potters would toss a handful of salt into the kiln on second firing and the salt would vaporise, then combine with the clay to form a coating of sodium silicate
[11:43:52] <CaptHindsight> the sodium chloride vaporizes and reacts with steam to form hydrogen chloride and soda
[11:44:12] <Jymmm> CaptHindsight: HCL?
[11:44:30] <CaptHindsight> https://en.wikipedia.org/wiki/Salt_glaze_pottery look at the process section
[11:45:08] <Jymmm> Ah "...iron impurities can help produce good salt glaze"
[11:46:16] <CaptHindsight> that's why you should not burn your McDonalds fries
[11:46:39] <roycroft> because they will become petrified?
[11:46:57] <CaptHindsight> or rather salt sticks with mild potato flavor
[11:47:08] <Jymmm> I like crispy fries
[11:47:25] <Jymmm> the soaked up oil will burns for days?
[11:48:01] * roycroft has never seen moldy mcdonald's food, and suspects the reason is not that it's so good *someone* will eat it no matter the conditions of disposal
[11:48:12] <CaptHindsight> lol
[11:48:57] <Jymmm> roycroft: It has to be made from organic material to actually mold.
[11:49:12] <CaptHindsight> http://aht.seriouseats.com/archives/2010/11/the-burger-lab-revisiting-the-myth-of-the-12-year-old-burger-testing-results.html
[11:51:57] <Jymmm> Hmmm, salt heat storage
[11:54:48] <Jymmm> Either I melt salt and store heat in it's molten state for 24 hours, or less time in it's solid state. Either way, use as sorta of a storage device and release gradually as needed
[11:58:44] <SpeedEvil> 'salt' in this case is rarely simply naCl
[11:59:16] <Jymmm> SpeedEvil: No, but it's $12/block http://www.tscstores.com/20KG-PLAIN-SALT-BLOCK-P3557.aspx
[12:01:31] <Jymmm> And I already have a place that only sells salt that I've been buying from for years.
[12:08:06] <SpeedEvil> yes, but you probably want to optimise for a lower temperature storage salt in most cases
[12:08:21] <SpeedEvil> and perhaps a storage temperature range, and other factors
[12:08:28] <Jymmm> http://www.sciencedirect.com/science/article/pii/0017931084900346
[12:08:33] <Jymmm> Darnit ^^^
[12:13:40] <Jymmm> Not gonna pay $10/page
[12:13:54] <zeeshan> for
[12:15:25] <IchGucksLive> hi
[12:16:29] <IchGucksLive> linux_vampire: ?
[12:21:00] <pink_vampire> yes
[12:21:42] <IchGucksLive> mashine ?
[12:23:46] <archivist> pink_vampire, read the ini section of the manual, there you find edit the display section to change axis to one of the others
[12:34:54] <sector_0> can I connect the parallel port directly to my a4988?
[12:35:50] <IchGucksLive> sector_0: what BOB is on the mashine
[12:38:37] <archivist> sector_0, I would not recommend that,
[12:41:44] <IchGucksLive> sector_0: no way for a arduino direct
[12:41:56] <IchGucksLive> hi rob_h
[12:42:17] <archivist> one avoids direct stepper grounds connecting to PC ground
[12:43:43] <rob_h> hi
[12:45:04] <archivist> ho
[12:46:19] <sector_0> IchGucksLive, BOB?
[12:46:43] <IchGucksLive> BreakoutBoard not related to Arduino
[12:46:56] <IchGucksLive> but go for a pice like
[12:47:12] <sector_0> oh ok
[12:47:23] <archivist> sector_0, http://www.ebay.co.uk/itm/5-Axis-CNC-Breakout-Board-Adapter-Stepper-Motor-Driver-Mach3-USB-Cable-Tools-/252307419456
[12:47:28] <sector_0> well I was referring to not having a breakout board
[12:47:31] <IchGucksLive> http://www.ebay.de/itm/111925695212
[12:48:12] <archivist> the usb cable is just for 5v
[12:48:30] <CaptHindsight> sector_0: why would you want to remove protection from your PC used to control the drivers?
[12:49:11] <CaptHindsight> save up the extra $10 and buy a BOB
[12:49:28] <IchGucksLive> archivist thats the cool part
[12:49:52] <IchGucksLive> the plugs on the board are xh-3s
[12:49:58] <archivist> I have 3 of those BOBs now
[12:50:10] <IchGucksLive> i did mount 3 yesterday
[12:50:10] <sector_0> well I have a dev board (not a breakout board), but I just assumed it would be redundant to simple route the step, dir, etc signal through a microcontroller
[12:50:26] <IchGucksLive> and today i did unpaxk 4 and testet on Oszi
[12:50:49] <sector_0> what exactly does the breakout board do?
[12:51:14] <archivist> buffers the signal to drive the stepper driver optos
[12:51:33] <CaptHindsight> sector_0: are you controlling motors in the real world or just poking around on the bench?
[12:51:35] <archivist> and has optos for inputs
[12:52:09] <sector_0> CaptHindsight, I have a small robotic project I want to start working on
[12:53:03] <sector_0> archivist, what are the optos?
[12:53:16] <sector_0> is that the same as an H-bridge controller?
[12:53:40] <CaptHindsight> sector_0: https://en.wikipedia.org/wiki/Opto-isolator
[12:53:55] <sector_0> oh ok
[12:54:31] <sector_0> so let me see if I understand the setup...
[12:54:36] <archivist> they save the smoke getting out
[12:54:57] <archivist> and stop the interference getting in
[12:55:25] <Jymmm> SpeedEvil: Hey, KNO2 is salt too, Could heat that to it's melting point as well.... PERFECTLY safe ;)
[12:55:40] <sector_0> the pc sends the step and dir signals down to the parallel port, then the breakout board buffers the signals and sends them to the optos which then drives the steppers?
[12:55:52] <andypugh> Most stepper drivers have opto-inputs. I (personally) have no qualms about connecting directly to FPGA pins or parallel port pins.
[12:56:39] <archivist> parallel ports have insufficient drive for some inputs
[12:56:51] <sector_0> ohh
[12:57:46] <sector_0> forgive me if I seem skeptical, I'm just trying to get clarification, because i've never had these concerns with my 3d printer setup
[12:58:30] <Jymmm> sector_0: and you are driving your 3d printer rom a parallel port or usb?
[12:58:36] <Jymmm> from*
[12:58:46] <Jymmm> apples/oranges
[12:59:00] <andypugh> Jymmm: It is fairly common to use molten salt for heat-treatment. The eutectic mixture of pottasium and sodium nitrate is common.
[12:59:12] <andypugh> (metls at 250C)
[12:59:29] <sector_0> I had the a4988 connected directly to the stm32f4, which drove the LM298N (h-bridge controller), which drove the NEMA 17
[12:59:58] <IchGucksLive> the LM is isolated
[13:00:05] <andypugh> FLinNak is rather less nice: https://en.wikipedia.org/wiki/FLiNaK
[13:00:08] <sector_0> so now that I'm going to use the pc, i thought i'd simply replace the stm32f4 with the pc
[13:00:55] <Jymmm> andypugh: I was looking at slat as a form of (overnight) thermal "storage", much like the mojave solar plant did (which is what gave me the idea) https://en.wikipedia.org/wiki/Mojave_Solar_Project
[13:01:00] <Jymmm> salt*
[13:01:00] <andypugh> I worked on a heat-treatment process that used molten Sodium Hydroxide. And have a very small scar to prove it.
[13:02:10] <archivist> IchGucksLive, the LM is an H bridge not isolated
[13:02:59] <IchGucksLive> oh i thought it is the 297/298 combi he uses
[13:03:38] <archivist> sector_0, what can be "fun" is the stepper driver switching noise causing problems, this is where the optos help as well
[13:06:00] <archivist> I did have one of my steppers step with noise from the VFD, the BOB fixed that
[13:06:16] <archivist> I was working direct :)
[13:07:05] <IchGucksLive> this boards are cracy good at that price
[13:07:35] <IchGucksLive> and without PWM or 0-10 or relay only 5V is the limit
[13:07:48] <IchGucksLive> no need vor higher supply
[13:10:46] <IchGucksLive> im off Gn8
[13:10:57] <zeeshan> http://i0.wp.com/TheAutomationBlog.com/wp-content/uploads/2016/01/ControlLogix-Hardware.png
[13:11:03] <zeeshan> has anyone worked with this type of plc before?
[13:11:13] <zeeshan> or really a "PAC"
[13:11:17] <zeeshan> programmable automation controller
[13:16:13] <pcw_home> andypugh: As a teenager I once made sodium metal from molten sodium hydroxide, fairy nasty...
[13:16:57] <andypugh> Impressive ;-)
[13:17:40] <pcw_home> I'm lucky to still have all fingers
[13:18:08] <linux_vampire> i'm getting error "joint 2 following error" and the z axis stop move"
[13:18:40] <linux_vampire> and all the controls become gray
[13:19:44] <pcw_home> a following error usually means a stepgen tuning issue (or bad latency)
[13:19:57] <linux_vampire> http://i.imgur.com/4V348P2.png
[13:20:31] <andypugh> I had a small chemistry set, and that included sulphur and potassium permanganate. Those, in chemistry-set quantities were vaguely inteesting. But I was using it in the cellar, where the gardening chemicals were, including a big box of Flowers of Sulphur and a big brown bottle of Potassium Permanganate. I succeeded in making the air in the cellar 100% unbreathable, but got out.
[13:21:15] <archivist> naughty boy :)
[13:21:17] <pcw_home> :-)
[13:22:05] <linux_vampire> it happen only in the z axis
[13:23:11] <linux_vampire> on slower jog speed it work find
[13:23:21] <linux_vampire> fine*
[13:23:34] <andypugh> linux_vampire: How many steps per second at the peak ini-file speed?
[13:23:49] <andypugh> And how does that compare to your base-thread rate?
[13:24:48] <linux_vampire> i know i need at least 22khz
[13:25:43] <linux_vampire> http://paste.debian.net/893068/
[13:25:51] <linux_vampire> this is my ini file
[13:28:01] <andypugh> 1638.4 steps per mm and 9.5 max velocity is 15kHz.
[13:28:09] <andypugh> What is your base-thread rate?
[13:28:45] <andypugh> (Sorry, found it: BASE_PERIOD = 100000 )
[13:29:35] <andypugh> So, you have a base thread of 100,000 nS which is 10kHz and you have a system that needs 15kHz to reach full speed.
[13:30:25] <andypugh> If you set the BASE_PERIOD to 50000 you should be OK, if the machine latency is low enough.
[13:30:51] <linux_vampire> the x y axis are 80 steps per mm max feed 5000 mm/min, the z axis is 1638.4 steps per mm, max feed 600 mm / min,
[13:31:11] <linux_vampire> 400000+982800+400000
[13:31:20] <linux_vampire> 1782800/60.0
[13:31:39] <linux_vampire> 29713 hz
[13:32:22] <linux_vampire> andypugh, now i need to convert that to nano secs?
[13:32:42] <andypugh> I don’t understand your maths above
[13:33:03] <andypugh> You don’t need to add the axes up
[13:33:43] <linux_vampire> 80 step/ mm *5000 mm/min = 400000
[13:34:05] <linux_vampire> 1638*600=982800
[13:34:08] <andypugh> Yes, but what is the 400000+982800+400000 ?
[13:34:41] <linux_vampire> 400000 x + 400000 y + 982800 z = 1782800
[13:35:05] <linux_vampire> then 1782800 steps / min / to 60 secs
[13:35:13] <linux_vampire> = 29713 hz
[13:35:20] <archivist> wrong to add them
[13:35:22] <andypugh> Yes, I see the numbers and I understand what the plus sign means. What I don’t understand is why you are adding the numbers together
[13:35:42] <archivist> they are in parallel
[13:35:54] <andypugh> What numbers did the latency test give you?
[13:36:18] <linux_vampire> lets see
[13:37:34] <linux_vampire> about 6500
[13:38:23] <archivist> for how long did you test
[13:38:24] <andypugh> OK, so then you should be able to set BASE_PERIOD in the INI to 30000 without any worries, and the following error will go away.
[13:38:27] <linux_vampire> with youtube it jump to 20,000
[13:40:00] <linux_vampire> what is the correct way to calculate the max speed of all the axes together?
[13:40:47] <linux_vampire> BASE_PERIOD = 100000
[13:40:47] <linux_vampire> SERVO_PERIOD = 1000000
[13:40:54] <linux_vampire> i have 2 option
[13:41:23] <linux_vampire> to change both of them to 30,000?
[13:41:29] <pcw_home> no
[13:41:46] <linux_vampire> just the base
[13:41:55] <pcw_home> yep change only the base period to 30000
[13:41:59] <linux_vampire> i'm using servos
[13:42:08] <pcw_home> doesnt matter
[13:42:30] <pcw_home> (if they are step/dir driven)
[13:42:41] <linux_vampire> yes, it's the g320x
[13:44:04] <linux_vampire> here where did you gut the 9.5? <andypugh> 1638.4 steps per mm and 9.5 max velocity is 15kHz
[13:44:21] <linux_vampire> what the 9.5 mean?
[13:44:34] <pcw_home> A 100000 nS base period will limit you to 10 KHz max step rate (using reset or just 5KHz without reset))
[13:45:01] <andypugh> linux_vampire: I was looking in your INI file, and the Z axis max velocity is 9.55 mm/sec
[13:45:18] <linux_vampire> i never put it there.
[13:45:31] <linux_vampire> maybe it calculate it ?
[13:45:41] <andypugh> Well, 9.50002432006 to be exact
[13:46:15] <linux_vampire> on the x, y it's 30
[13:46:19] <andypugh> Which is 570mm/min (which is about what you saud)
[13:46:45] <linux_vampire> but SCALE = 80.0
[13:46:51] <linux_vampire> yes.
[13:46:58] <linux_vampire> correct.
[13:47:51] <linux_vampire> my question, what is the correct way to calculate my step base number?
[13:48:37] <archivist> slower than the latency with a safety margin
[13:48:51] <linux_vampire> i know..
[13:49:11] <archivist> that then sets speed limits
[13:49:26] <linux_vampire> but here <andypugh> 1638.4 steps per mm and 9.5 max velocity is 15kHz.
[13:49:38] <linux_vampire> what is the math ?
[13:50:24] <linux_vampire> 1638.4*9.5 = 15564.80?
[13:50:49] <andypugh> 1638.4 steps/sec * 9.5 steps/mm = 15565/sec
[13:51:05] <pcw_home> mm/S * steps/mm = steps/S
[13:51:22] <andypugh> Sorry. Yes, got my denominators wrong
[13:52:31] <linux_vampire> and i can change the settings on the x y axes to be also 1638.4 instead of 80?
[13:52:51] <linux_vampire> and the base number will be the same?
[13:53:03] <linux_vampire> or i need to add all of them?
[13:53:24] <linux_vampire> i know mach3 add all the signals.
[13:53:31] <archivist> the are being stepped at the same time, dont add them
[13:53:38] <archivist> they
[13:54:22] <linux_vampire> that grate.
[13:56:07] <zeeshan> do servo drives usually need filtered power?
[13:56:09] <zeeshan> or unfiltered power
[13:56:31] <zeeshan> can i just simply do 1ph 240v xfmr -> rectifier -> servo drive?
[13:56:52] <zeeshan> like for the be25a20 for example
[13:57:01] <linux_vampire> if i want to get smaller number on the latency test, what i need to upgrade? cpu? ram? video card?
[13:57:50] <archivist> video card is sometimes a problem, onboard usually worse than a plug in card
[13:58:24] <linux_vampire> any recommended card?
[13:58:50] <archivist> last week my onboard gave 90000, a plug in card 15000
[13:58:56] <nubcake> i'm using a gtx 580 Ti Phantom, never had any problems with that one
[13:59:35] <nubcake> think i had less than 15000 but not a 100% certain about that.
[14:00:36] <linux_vampire> so the gaming cards are better then the stuff like the tesla?
[14:00:54] <nubcake> no, i don't know about the teslas
[14:00:55] <pcw_home> zeeshan: you need a filter capacitor
[14:01:05] <nubcake> they're probably even better than the gaming stuff
[14:01:29] <nubcake> but i'm not an expert with that
[14:01:31] <zeeshan> pcw_home: im trying to power up my fanuc drives
[14:01:47] <zeeshan> i'm thinking they can take a unregulated supply
[14:02:01] <zeeshan> unregulated means what?
[14:02:02] <pcw_home> you dont have the Fanuc power supply?
[14:02:05] <zeeshan> i do
[14:02:06] <zeeshan> but its 3 phase
[14:02:14] <zeeshan> so kinda useless for me
[14:02:28] <zeeshan> the drives take in dc
[14:02:37] <zeeshan> so its really just a matter of finding the correct supply for them
[14:02:39] <pcw_home> 1 phase just has more filter capacitance
[14:03:02] <zeeshan> thge fanuc power supply is doing a lot more than just providing power to servo drives
[14:03:11] <zeeshan> its got some rails at different voltages which i'm not really using
[14:03:18] <zeeshan> (they went to the controllers)
[14:03:29] <zeeshan> i was really hoping i could just use a toriod or something
[14:03:33] <zeeshan> and get to the voltage i want
[14:03:39] <pcw_home> the servo power will normally be unregulated (but filtered)
[14:04:31] <zeeshan> i hope i'm making the right decision by keeping the fanuc drives...
[14:04:35] <zeeshan> i don't want to do what i did with the mil
[14:04:39] <zeeshan> and have not enough decel :{
[14:04:43] <pcw_home> much fancier drives have a cycloinverter to dump power back into the line when braking but the standard Fanuc drives are simpler
[14:04:50] <archivist> reservoir capacitor being its other name
[14:04:51] <linux_vampire> the psu that go to my servos is just diode bride and large capsitors
[14:06:06] * zeeshan will read the fanuc drive manuals today while machining
[14:10:53] <pcw_home> what model number are your drives?
[14:12:27] <pcw_home> anything after xxx6057 is a digital (PWM or FSSB) drive
[15:09:47] <enleth> zeeshan: 48V telecom power supplies work well for DC servo drives, if your servos take a more or less round multiple of 48V
[15:10:50] <enleth> zeeshan: I've bought a few Tebechop 2.7kW single phase modules that turn on with no chassis connection/external control
[15:10:59] <enleth> they can be connected in series, within reason
[15:11:18] <enleth> and similar units seem to be abundant on auction sites
[15:32:33] <tiwake> is it difficult/expensive to titanium nitride coat aluminum?
[15:32:40] <andypugh> coin-op machines (fruit machines etc) have some fairly hefty PSUs. This one is 5A @ 48V + 7.5A @ 34V: http://www.ebay.co.uk/itm/Barcrest-Mpu-4-Power-Supply-Unit-Fully-Working-/262666392063?hash=item3d282311ff:g:eycAAOSw8w1X-nkC
[15:37:22] <CaptHindsight> had to loop up "fruit machine" http://www.reelfruits.co.uk/375_500_csupload_35874970.jpg?u=3904810643
[15:40:06] <_methods> i thought he was talkin about this
[15:40:07] <_methods> https://s10.postimg.org/ea7o6m4o9/tumblr_m9t4oxx_N0n1re5uw8o1_500.jpg
[15:41:16] <CaptHindsight> heh, what I initially pictured http://images-cdn.9gag.com/photo/awrYNYD_700b.jpg
[16:02:06] <sync> tiwake: expensive, probably, difficult not so much iirc
[16:02:15] <zeeshan> enleth: servos are at 180 v
[16:03:38] <tiwake> sync: compared to coating steel?
[16:04:05] <sync> andypugh: I was pretty confused by reciprocal space maps, just remembering you did xray stuff before
[16:04:27] <sync> tiwake: well, it probably depends if your coating shop has experience and/or will risk trying it
[16:05:10] <tiwake> sync: I am kind of thinking of things that would increase the life of hydraulic cylinders
[16:05:49] <tiwake> TiN coating seems like it would make it rather badass
[16:05:59] <enleth> zeeshan: many 48V supplies can be cranked up to 55-65V with no loss of maximum amperage, three of those might do
[16:06:05] <zeeshan> pcw_home: drives are a06b-6047-h003
[16:06:20] <sync> well, usually they wear the seals down
[16:06:29] <sync> and the ram is already chrome hardplated
[16:06:32] <enleth> zeeshan: they typically output 40 tp 60A, which should be good for 3 servos in a reasonably big machine
[16:06:57] <zeeshan> enleth: are they switch mode suppliers
[16:07:00] <zeeshan> supplies, or linear
[16:07:08] <andypugh> sync: I have touched on that sort of thing a long time ago, but it was nver critical to my area.
[16:07:22] <enleth> zeeshan: switchmode, of course
[16:07:31] <sync> ah, too bad andypugh
[16:07:32] <enleth> almost no one uses linear power supplies for high power anymore
[16:07:35] <zeeshan> i personally dont like switch mode for servo drives
[16:07:44] <zeeshan> i dunno about that!
[16:07:45] <andypugh> I did once find myself reading a paper which had a “reciprocal effective mass tensor”
[16:07:52] <zeeshan> the way i see servo drives powered on original controls
[16:07:54] <zeeshan> fanuc/mazak etc
[16:08:07] <zeeshan> is a monster transfer that goes into a rectifier and some other module
[16:08:20] <sync> it is pretty neat tho, andypugh
[16:08:44] <enleth> zeeshan: those telecom supplies often power RF transmitters, which are quite sensitive to supply voltage noise and capable of very sharp and sudden changes in power draw
[16:08:51] <sync> I can get stress and concentration info out of one measurement in crystal lattices
[16:08:55] <zeeshan> ah
[16:09:01] <sync> (although one measurement takes like 60h)
[16:09:20] <tiwake> alright, gotta run
[16:09:51] <andypugh> Well, the “reciprocal effective mass tensor” was describing the way that the effective mass of photons changed with the direction they were moving in reciprocal space. I still don’t know if it was really a useful concept.
[16:12:54] <enleth> zeeshan: I'll let you know what the results are on mine, so far I've got one axis wired and it seems to be doing OK, but I haven't done any serious voltage monitoring on the DC bus
[16:13:47] <enleth> zeeshan: I'm rather inclined to trust those supplies, though - there's no fundamental reason for switchers to be unsuitable to power servo drives
[16:14:12] <zeeshan> on my lathe
[16:14:19] <zeeshan> i fried my switchers pretty quick :P
[16:14:20] <zeeshan> old lathe
[16:14:29] <enleth> what were those?
[16:14:31] <zeeshan> so i have had a bad experience with them. doesn't mean theyre not good
[16:14:35] <zeeshan> some crap from cnc4pc
[16:14:42] <enleth> MeanWell?
[16:14:46] <zeeshan> no
[16:16:49] <pcw_home> Switching supplies are not a great idea for DC buss servo power
[16:17:20] <zeeshan> man i lost the manual for this damn velocity control unit
[16:17:27] <pcw_home> because they are not designed to absorb energy
[16:18:00] <sync> that is why you have a braking unit or braking resistors in the drives
[16:18:03] <enleth> pcw_home: any better ideas to get several kW at much more than 100V, that don't involve a quarter ton of laminated iron and copper?
[16:18:23] <sync> a 4 quadrant bridge is the neater idea
[16:18:24] <pcw_home> You still want large output capacitance
[16:18:47] <sync> yaskawa does that
[16:18:48] <pcw_home> so you can dump excess energy back into the supply
[16:18:55] <sync> they push power back into the grid
[16:19:09] <pcw_home> yes fancier drives do that
[16:19:19] <pcw_home> (usually larger drives)
[16:19:51] <sync> 14kW iirc
[16:21:06] <pcw_home> enleth: line operated (no xfrmr)
[16:21:37] <enleth> yeah, the old servo drive on the bridgeport did that
[16:21:46] <enleth> I can't really say I was sold on the idea
[16:21:58] <pcw_home> most newer drives do as well
[16:22:00] <andypugh> I run both my machines on rectified line voltage. That’s 300+V here.
[16:22:07] <pcw_home> (like a VFD)
[16:22:43] <zeeshan> andypugh: makes the most sense to me to do that
[16:22:52] <andypugh> You could fit an isolation transformer if you want to be slghtly safer. But I am not sure how much it really helps.
[16:22:57] <zeeshan> i was thinking of dropping the ac voltage to whatever i need
[16:22:59] <zeeshan> to get 180v out
[16:23:12] <sync> andypugh: it is pretty safe tho
[16:23:21] <enleth> pcw_home: VFDs I've seen cracked open seemed to be fully isolated, although I've never messed with a unit bigger than 4kW
[16:23:23] <sync> if something fails it will do so to ground
[16:23:46] <sync> zeeshan: just run 230V from the wall to the motors
[16:23:48] <enleth> pcw_home: and they were, in fact, just fancy switchmode power supplies
[16:23:53] <sync> and limit max rpm in software
[16:24:01] <pcw_home> the problem with normal switching supplies is that the main filter capacitors are on the primary side
[16:24:03] <pcw_home> so dont do any good for absorbing power
[16:24:25] <zeeshan> sync 230vac from wall
[16:24:29] <zeeshan> is like 380vdc isnt it
[16:24:30] <zeeshan> i forget.
[16:24:34] <sync> you don't want to absorb power in the supply, you just need a braking resistor
[16:24:39] <enleth> pcw_home: the telecom supplies I was talking about earlier have much more capacitance on the output side, FWIW
[16:24:40] <sync> *sqrt(2)
[16:24:47] <andypugh> I actually have a HAL component that operates my PSU. It switches in and out the charge and discharge relays, and monitors bus voltage (via the 8i20 drives, which transmit that data back)
[16:24:57] <zeeshan> 340vdc
[16:24:57] <zeeshan> jeez
[16:24:58] <zeeshan> :p
[16:25:16] <zeeshan> 120vac is pretty much what i need going into the transfer :P
[16:25:19] <zeeshan> *transformer
[16:25:37] <enleth> pcw_home: those units cost around $1500 a piece (for 48V at 2.7kW) when they were new, 15-20 years ago
[16:25:44] <andypugh> zeeshan: So use one of your weak-ass USA 110V outlets ;-)
[16:25:49] <zeeshan> lol
[16:26:31] <pcw_home> You dont want your brake resistor switched in every millisecond of you want any kind of
[16:26:32] <pcw_home> efficiency the power supply must absorb some reverse energy
[16:26:34] <linux_vampire> what is the max jitter for 1638.4 steps per mm, and MAX_VELOCITY = 30.0
[16:26:59] <andypugh> Over here there are lots of big 240 to 110 V transformers for building-site use, where the power tools tend to be 110V.
[16:27:02] <enleth> now I'm getting them for $40, because a local mobile network replaced several thousand BTSes at once, including power supplies
[16:27:10] <linux_vampire> if i multiply them i'm getting 49152
[16:27:21] <linux_vampire> 50kz?
[16:27:25] <sync> you can push some power into them without issue, I used to run a machine with the braking resistor set to Vbus+2V
[16:27:28] <andypugh> linux_vampire: Yes, 50kHz
[16:27:29] <sync> nothing ever failed
[16:27:42] <sync> huh andypugh, safety tools run at 42V here
[16:27:43] <linux_vampire> now i need to convert it to ns?
[16:28:15] <andypugh> linux_vampire: 1S / 50kHz
[16:28:33] <andypugh> 0.00002 seconds
[16:28:34] <linux_vampire> 20000 ns
[16:29:12] <andypugh> Should be OK as long as you don’t use YouTube
[16:29:32] <andypugh> But… Think about what the next lower speed is…
[16:29:34] <Deejay> gn8
[16:29:58] <linux_vampire> what do you mean the next lower speed?
[16:30:14] <nubcake> n8 everyone
[16:30:38] <linux_vampire> andypugh,
[16:30:41] <andypugh> It’s 50kHz with one pulse every 20,000nS. But 25kHz with one pulse ever second base-period. Might be OK with your servo system, but a stepper would not be able to follow that.
[16:31:58] <linux_vampire> this is why i'm using the g320x
[16:32:31] <linux_vampire> what is the maximum that i can push in to the parallel port?
[16:32:55] <andypugh> 50kHz is really pushing it with a parport
[16:34:09] <linux_vampire> maybe i need to keep it slower?
[16:34:17] <andypugh> Can’t you just set the G320x to fewer pulses per rev?
[16:34:31] <linux_vampire> i can..
[16:34:35] <JT-i3> most gecko are 10 micro stepping
[16:34:44] <linux_vampire> yes..
[16:34:49] <JT-i3> one which I forget is settable
[16:35:14] <linux_vampire> now the encoder is 4000 ppr but the g320x is on x10
[16:35:27] <linux_vampire> so the computer see just 400 ppr
[16:35:43] <linux_vampire> on the z axis i left it on 8192 ppr
[16:35:57] <andypugh> It isn’t like you need to adjust Z height to half-microns.
[16:36:40] <linux_vampire> i found it run much smooth this way.
[16:37:21] <linux_vampire> my goal eventually is to get better rails and high speed spindle
[16:37:51] <andypugh> You could look at a Mesa 5i25. That can step at a few MHz
[16:38:24] <linux_vampire> but i need much much lower jitter
[16:38:57] <andypugh> Not with a 5i25. In fact those can work with 500,000 jitter
[16:39:26] <linux_vampire> how??
[16:39:34] <andypugh> The 6500 you have now really is quite low. Not a lot of PCs go that low.
[16:40:18] <linux_vampire> it's q9300 with 4gb ram
[16:40:32] <andypugh> linux_vampire: It gets sent a step-rate every 1mS and runs a dpll to adjust for update jitter from the host. The steps are generated by an FPGA that isn’t doing anything else.
[16:41:55] <linux_vampire> the fpga is like ssd?
[16:42:10] <linux_vampire> it's gan burn off over time?
[16:43:13] <sync> yes and no
[16:43:19] <sync> all semis are due to fail at some point
[16:43:34] <sync> but a fpga will take its time
[16:44:31] <andypugh> linux_vampire: No, the FPGA isn’t being read and written continuously, it is like a very specialised little computer doing only one thing (makin steps, running servos, etc)
[16:45:02] <andypugh> The FPGA should last as long as the CPU in the PC.
[16:45:13] <andypugh> Probably longer, it doesn’t get hot.
[16:45:35] <linux_vampire> i see now
[16:46:03] <sync> getting hot doesn't mean it is not stressed tho :P
[16:46:10] <sync> or rather not hot
[16:47:15] <linux_vampire> i know arduino can burn off after about 10,000 times that you flash it.
[16:47:33] <XXCoder> andypugh: basically computer just tell 5i25 "do this" and it handles the signaling and stuff?
[16:47:43] <XXCoder> so there is no computer latency issues
[16:47:53] <Magnifikus> not really
[16:48:02] <Magnifikus> if the computer cant tell the 5i25 in time what todo
[16:49:40] <newbie5000> hi andy how can I use gmoccapy with classic ladder thc?
[16:49:46] <andypugh> The FPGA will give up and stop if the next servo-thread update data doesn’t arrive (though this is a programming decision, it could just carry on….)
[16:49:56] <Magnifikus> im abit hyped for hostmot for zynqs :)
[16:50:06] <Magnifikus> when the fpga lives in your address space
[16:50:43] <andypugh> newbie5000: Well done! You have picked three things that i know nothing about :-) (I have never used classic ladder, or a plasma, or Gmocccapy)
[16:50:58] <newbie5000> :d ok
[16:51:03] <linux_vampire> The 5I25 use a Xilinx Spartan6 ~400k gate FPGA
[16:51:19] <linux_vampire> what the 400k gate mean?
[16:51:30] <Magnifikus> how much logic you can put into the fpga
[16:51:36] <andypugh> 400,000 indicvidual logic gates
[16:51:47] <zeeshan> linux_vampire: a fgpa is basically a shit load of switches
[16:51:57] <Magnifikus> lookup tables :P
[16:51:58] <zeeshan> which you can use to make gates
[16:52:08] <andypugh> newbie5000: You probably don’t need CL to use a THC, there are HAL components that already exist for the task
[16:52:12] <zeeshan> gates are things like and or nor xor etc
[16:52:36] <zeeshan> man i remember the 1 year i did EE
[16:52:41] <newbie5000> really?! how to use them? what is the neccessary hardware?
[16:52:41] <zeeshan> all we were doing was truth tables
[16:52:50] <zeeshan> that's when idecided, EE isn't for me
[16:52:54] <zeeshan> even though i did well in that class
[16:52:57] <zeeshan> =P
[16:53:01] <andypugh> newbie5000: http://linuxcnc.org/docs/2.7/html/man/man9/thc.9.html
[16:53:12] <Magnifikus> zeeshan, fpgas are basicly giant truth tables + flip flops :D
[16:53:15] <newbie5000> thanks alot Andy
[16:53:20] <zeeshan> :)
[16:53:29] <Magnifikus> oh and interconnects ^^
[16:53:43] <linux_vampire> so it's like the ram on the computer.. not like ssd
[16:54:24] <Magnifikus> the logic gates?
[16:54:37] <Magnifikus> its more like how big your cpu is
[16:54:40] <andypugh> That’s aimed at using the Mesa THC hardware, which puts out a pulse-rate proportional to arc voltage. LinuxCNC calculates the pulse rate using a HAL encoder counter, and then the THC component moves the Z up and down.
[16:54:46] <Magnifikus> cause fpga do anything parallel
[16:54:49] <andypugh> That’s the same in any UI
[16:55:23] <newbie5000> In English, please. : d I'm just a begginer.
[16:55:47] <Magnifikus> in fpgas you have no mhz or ram
[16:56:00] <andypugh> newbie5000: http://store.mesanet.com/index.php?route=product/product&product_id=128
[16:56:17] <Magnifikus> you define what they do and they do it all the time like static hardware logic
[16:56:31] <Magnifikus> and the gate count limits how much they can do
[16:57:16] <Magnifikus> think about a giant pcb with logic on
[16:57:29] <andypugh> Magnifikus: Well, that’s not 100% true any more. Many FPGA implementations actually have a little CPU running inside them and they end up acting more like a uP than a pipleine.
[16:57:52] <Magnifikus> sure you do that if you need cpu like stuff
[16:57:53] <andypugh> (it’s a soft-CPU, made of gates)
[16:58:04] <Magnifikus> cause fpgas suck if its getting not boring
[16:58:51] <Magnifikus> and the newer socs are even better cause you get arm cpus integrated into the fpga
[16:59:10] <Magnifikus> and which run linuxcnc with ease
[16:59:12] <newbie5000> I can't buy the mesa Thc, unless they can ship to Egypt.
[17:00:08] <andypugh> I am sure that they can, but portugal is closer: http://eusurplus.com/index.php?route=product/product&product_id=613&search=THCAD&description=true&sub_category=true
[17:01:21] <andypugh> newbie5000: There is another option, http://linuxcnc.org/docs/2.7/html/man/man9/thcud.9.html
[17:01:38] <newbie5000> ok
[17:01:58] <andypugh> newbie5000: The THCUD component just needs a “Higher, no, lower” input from some external hardware.
[17:02:06] <pcw_home> linux_vampire: Yes the FPGAs used on the 5I25 are volatile static RAM based, they are loaded at power up by a flash memory chip
[17:02:08] <pcw_home> the flash memory chip has a limited cycle life but you are not likely to change the firmware 10,000 times
[17:02:42] <newbie5000> what do you mean by Higher, no, lower” input from some external hardware?
[17:02:46] <andypugh> pcw_home: (I wouldn’t put it past her)
[17:03:06] <sync> I don't really see the need for that Magnifikus
[17:03:13] <sync> you can just run it on a regular arm
[17:03:16] <andypugh> newbie5000: Just two logic signals, one for higher and one for lower
[17:03:22] <sync> they are clocked so high now
[17:03:32] <newbie5000> I got it .
[17:03:45] <JT-Shop> newbie5000: is there a problem shipping to Egypt?
[17:03:55] <JT-Shop> I ship all over the world
[17:04:05] <Magnifikus> but you miss the tight coupling
[17:04:18] <Magnifikus> and they are cheap
[17:05:08] <newbie5000> Can you ship it?
[17:05:31] <newbie5000> what shipping method do you accept?
[17:05:40] <newbie5000> payment*
[17:06:53] <linux_vampire> so each time that i'm using the machine it's flash a new firmware in to the fpga?
[17:07:18] <pcw_home> no, only once
[17:07:21] <Roguish> linux_vampire: not necessarily. depends upon which card you have
[17:07:34] <linux_vampire> 5i25
[17:07:42] <JT-Shop> newbie5000: I use paypal to take credit cards or paypal
[17:08:33] <pcw_home> on the 5I25 you only flash it if you want different firmware
[17:09:08] <newbie5000> send me the paypal link
[17:09:13] <newbie5000> please
[17:10:26] <Magnifikus> btw sync any good idea for 24v sync ac servos? :D
[17:10:50] <Magnifikus> i got the controller working with linuxcnc but the 8ms update rate on canopen suxx
[17:10:51] <JT-Shop> http://mesaus.com/
[17:11:06] <JT-Shop> newbie5000: just go to my store and place the order
[17:11:09] <XXCoder> https://www.youtube.com/watch?feature=player_embedded&v=nkn2ZXWDl6k
[17:11:11] <XXCoder> enjoy
[17:11:21] <XXCoder> it moves video so its earth that moves not stars
[17:11:25] <JT-Shop> newbie5000: http://mesaus.com/index.php?route=product/category&path=65
[17:11:35] <sync> Magnifikus: the stmbl will work
[17:11:55] <JT-Shop> hey XXCoder I know what it sounds like when the hot end fan gets unplugged
[17:12:04] <Magnifikus> oh nice thought is more for higher voltage
[17:12:23] <sync> it is, but that does not keep it from working at 24V
[17:12:29] <XXCoder> jt heh attempt to discribe it? lol
[17:12:36] <Magnifikus> kewl project for next year :)
[17:12:53] <XXCoder> https://www.youtube.com/watch?annotation_id=annotation_173144&feature=iv&src_vid=nkn2ZXWDl6k&v=h714VOr-6nY nice
[17:13:05] <Magnifikus> before i realize the resolvers suck
[17:13:14] <sync> probably not
[17:14:18] <Magnifikus> but i found quite beefy dsp's in the servos: http://cache.nxp.com/files/dsp/doc/data_sheet/DSP56F807.pdf?pspll=1
[17:14:32] <Magnifikus> too bad nxp is selling the toolchai for 500€+
[17:17:27] <JT-Shop> XXCoder: loud clicking sounds with the extruder jumping
[17:17:41] <XXCoder> ouch
[17:17:49] <JT-Shop> then you take the hot end apart and clear the jam lol
[17:17:52] <XXCoder> did you manage to clear interior of hotend?
[17:18:01] <JT-Shop> yea it's pretty easy
[17:18:08] <XXCoder> not bad
[17:18:13] <Magnifikus> did you burn your fingers?
[17:18:25] <JT-Shop> unscrew the cold end from the cold break and push the melted bit otu
[17:18:28] <JT-Shop> out even
[17:18:43] <JT-Shop> at 200C? naw
[17:19:32] * JT-Shop got the 325lb 3 in 1 machine moved off the table and into the new shop... glad I've been working out
[17:20:15] <pcw_home> did you get your welding table?
[17:20:34] <JT-Shop> yes, trying to make room to assemble it now
[17:20:56] <JT-Shop> one box was 149lbs and the UPS guy was happy I was here to help
[17:21:02] <linux_vampire> if i'm getting just the 5i25 and the flat cable, it can work as a direct replacement to the dual lpt?
[17:21:13] <JT-Shop> yes
[17:21:30] <pcw_home> Different hal file setup though
[17:21:40] <JT-Shop> if by flat cable you mean http://mesaus.com/index.php?route=product/product&path=71&product_id=70
[17:21:57] <linux_vampire> correct
[17:23:20] <XXCoder> JT-Shop: btw test print after I rised z a little - it was so damn smooth
[17:23:26] <linux_vampire> how many axes i will be able to connect to a single 5i25?
[17:23:51] <XXCoder> no wobble, only remaining stuff I can see is TINY movement error ones as X is little too loose now
[17:24:08] <linux_vampire> to each DB 25 connector
[17:24:17] <pcw_home> axes are pretty heavy so I dont know...
[17:24:39] <linux_vampire> lol
[17:25:01] <linux_vampire> i will tig weld them..
[17:26:03] <newbie5000> thanks andy and jt, I'll place the order.
[17:26:21] <JT-Shop> XXCoder: nice
[17:27:06] <JT-Shop> I put the LM8LMM bearings in place of the LL8MM bearings in the Z axis
[17:27:23] <XXCoder> yeah? hows it
[17:27:56] <JT-Shop> Z is better but there is still a tiny bit of slop on the X axis rods
[17:28:09] <linux_vampire> JT-Shop, ^
[17:28:49] <JT-Shop> ?
[17:28:56] <Nick-Shop> JT-how is that 3d printer working out - any good?
[17:29:40] <XXCoder> jt that affects surpising very little
[17:29:47] <JT-Shop> for a $218 complete printer I can't complain, I have had to make several minor upgrades but overall it's pretty good
[17:29:48] <linux_vampire> JT-Shop, how many axes i will be able to connect to a single 5i25?
[17:30:02] <XXCoder> in fact your screws being not stright affects much more, thats why I designed and use anti-wobble
[17:30:08] <XXCoder> it also makes printer impact-proof
[17:30:15] <Nick-Shop> Like what upgrades?
[17:30:19] <XXCoder> it cant hit surface for enough damage
[17:30:24] <JT-Shop> you have two ports that you can connect your bob's to so the limit is your bob's
[17:30:46] <JT-Shop> the biggest was adding a couple of 40mm print fans for $8
[17:31:01] <XXCoder> mine is legs lol man that made tons difference
[17:31:07] <JT-Shop> making the rods fit the holes with shipping tape shims etc
[17:31:13] <XXCoder> makes it easier to move, no wobble when sitting on table
[17:32:02] <linux_vampire> JT-Shop, is there any advantage to the daughtercard over a bob?
[17:32:43] <JT-Shop> oh I printed some feets with o-rings at the bottom to make it skid resistant http://www.thingiverse.com/thing:1821456
[17:32:47] <XXCoder> jt 4 of those and any 8 mm thread funture footing https://www.aliexpress.com/item/2020-aluminum-extrusion-profile-european-standard-white-length-50mm-industrial-aluminum-profile-workbench-OB2020-50-1pcs/32706264190.html
[17:32:53] <newbie5000> JT-Shop:Do I need any other boards with the THCAD-300?
[17:33:04] <XXCoder> jt yeah I like your idea too :)
[17:33:28] <Nick-Shop> JT - I'm putting one together now any it's very lacking - have to lock down the x axis rods so the belt tention dosen't shift things as it's rising on z-
[17:33:35] <JT-Shop> linux_vampire: the 7i76 has 5 axes, spindle, 32 inputs, 16 outputs
[17:33:49] <XXCoder> FINALLY! stuff I need for next upgrade is shipped.
[17:34:31] <JT-Shop> newbie5000: not for a torch height control, but if you want to upgrade your electronics you could get more
[17:34:42] <JT-Shop> XXCoder: what's on the way?
[17:34:59] <XXCoder> washers nuts and 2 types of resistors
[17:35:02] <newbie5000> get more like?
[17:35:08] <XXCoder> nuts are same kind we use for 2020 extensions
[17:35:20] <XXCoder> my psu still isnt back on machine uet
[17:35:34] <JT-Shop> I moved my psu to the back side
[17:35:47] <XXCoder> define "back end"?
[17:36:00] <JT-Shop> left side behind vertical
[17:36:09] <JT-Shop> more out of the way there
[17:36:11] <linux_vampire> JT-Shop, so if i take a simple bob i can get 6 axes per bob but no extra outputs?
[17:36:11] <XXCoder> oh thats my psu orginial position
[17:36:22] <XXCoder> its just layibg on table now, disconnected to steal nuts
[17:36:34] <XXCoder> needed more nuts to attach legs
[17:36:44] <JT-Shop> linux_vampire: right, whatever your bob's have is it
[17:36:59] <linux_vampire> also why i can't use the there 32 outputs in the 7i76 for step/dir?
[17:37:01] <JT-Shop> I need to order some nuts from aliexpress
[17:37:07] <XXCoder> I plan to figure and move circuit boards under the machine, covered too. so theres no fobs staning on way
[17:37:13] <pcw_home> 6 axis BOBs are pretty rare
[17:37:32] <JT-Shop> rare as chicken lips?
[17:37:33] <XXCoder> once I finally remove that damned z switch limit, I plan to add another 90 degree 2020 brace
[17:38:02] <XXCoder> eat chicken with respect - they ahas t-rex as ascendary
[17:38:11] <linux_vampire> every parallel port is 12 outputs, divide that by 2 and you get 6 axes
[17:38:22] <pcw_home> but you could drive 8 step/dir axis from one connector if you had the right breakout or drive direct
[17:38:28] <newbie5000> Jt, What boards can I use, if I want to upgrade?
[17:38:29] <XXCoder> pink sure but not all BOB expose all ports
[17:38:48] <JT-Shop> newbie5000: steppers?
[17:38:56] <newbie5000> no
[17:38:59] <newbie5000> yes
[17:39:01] <newbie5000> yes
[17:39:07] <newbie5000> I use stepper motors
[17:39:24] <JT-Shop> I use this for my plasma http://mesaus.com/index.php?route=product/product&path=69&product_id=78
[17:39:28] <linux_vampire> pcw_home, how do you get 8 x 2 from parallel port?
[17:39:30] <JT-Shop> + the THCAD
[17:40:09] <pcw_home> not talking about parallel ports, the 5I25 can chose In or out for all pins
[17:40:11] <JT-Shop> newbie5000: do you have pci or pcie slot on your motherboard?
[17:40:41] <andypugh> Magnifikus: Resolvers are much better than encoders for most things.
[17:41:08] <newbie5000> I think the motherboard can use both. It's Dell optiplex 360
[17:41:21] <pcw_home> resolvers are certainly a lot tougher than encoders
[17:41:22] <newbie5000> I set it to PCI
[17:41:27] <linux_vampire> i have 2 c10 installed.
[17:41:39] <andypugh> linux_vampire: The 7i76 GPIO pins are serially-addresed, and only update once every 2mS
[17:41:45] <andypugh> (sorry, 1mS)
[17:42:19] <andypugh> So, you _could_ use them for step/dir outputs, but only up to 500Hz (1mS on, 1mS off)
[17:43:03] <linux_vampire> that actually not bad at all !!
[17:43:51] <sync> andypugh: idk, resolvers are usually lower resolution the only thing they have going for them is the robustness
[17:44:20] <andypugh> Resolvers don’t have a resolution.
[17:44:58] <JT-Shop> newbie5000: the 5i25 (PCI) cost a bit less than the 6i25 (PCIe)
[17:45:58] <andypugh> The resolver interface I am using gives 16-bits of position information, though it’s probably only useful to 12 bits.
[17:46:25] <JT-Shop> my mom gave me a worn out copy of The Foxfire Book a few years before she passed just now reading it... pretty interesting the old mountain ways
[17:47:23] <sync> andypugh: well yeah, but due to noise and shits you get a finite amount of resolution
[17:47:37] <pcw_home> I think the 7I49 gives you 32 bits (but most is noise)
[17:47:38] <pcw_home> depending on resolver you can get around 15+ bits of resloution
[17:47:44] <sync> and yeah 16 or 12 bit is lower than most regular encoders
[17:47:57] <sync> yeah 13-14 is reasonable usually
[17:48:46] <andypugh> Unlike an encoder, though, oversampling vives you more accurate position
[17:49:17] <andypugh> ie, low-pass filter the data and you get more resolution, because it is analogue.
[17:50:14] <pcw_home> high res encoders are a bit like that also since they use analog interpolation
[17:50:34] <sync> sure but you lowpass your position, which is eh
[17:50:52] <andypugh> So, a single-turn resolver ends up about equivalent to a 4000 count or 8000 count encoder. But if you happen to have a 2-tun then its 16000 count, and a 4-turn is 32000 and pretty quickly you are into good resolution.
[17:51:01] <sync> and 21bit abs data is usually enough
[17:51:28] <pcw_home> but of course 21 bit encoders are analog also
[17:51:50] <andypugh> Also, with a resolver, there is no danger of accumulating count errors.
[17:51:59] <linux_vampire> 21bit encoder...
[17:52:14] <sync> why would you accumulate count errors?
[17:52:16] <pcw_home> I have a 24 bit encoder
[17:52:36] <sync> yeah, they are not too fancy anymore
[17:52:39] <andypugh> sync: If you get a mis-count with an encoder, that error stays there for ever.
[17:53:05] <sync> not if you get actual angular values instead of A/B signals
[17:53:06] <andypugh> (well, until you next home or do an index reset anyway)
[17:53:29] <pcw_home> the new Avago absolute encoders are 30 some bits (but not very accurate)
[17:53:29] <sync> but getting a miscount is relatively unlikely
[17:53:50] <sync> yeah, idk what the deal with them is
[17:54:01] <sync> probably to make deriving the velocity easier but uh
[17:54:08] <andypugh> Renishaw Resolute ir pretty neat.
[17:54:13] <pcw_home> They are quite neat for multiturn without a battery or gears
[17:54:24] <pcw_home> (the Avago ones)
[17:54:46] <sync> the rls things are also quite neat
[17:55:26] <pcw_home> which reminds me I think theres a bug or missing feature in the HostMot2 SSI/all Absolute encoder interfaces
[17:55:39] <pcw_home> the M option
[17:56:04] <andypugh> I haven’t even looked at that for years. I have never had any of the hardware.
[17:58:08] <andypugh> pcw_home: What would the M-option be?
[17:58:27] <pcw_home> The bug is this: for SSI encoders at least (maybe Fanuc and BISS as well) the counter detects wraps in the absolute count
[17:58:28] <pcw_home> which is fine for single turn absolute encoders, but a bad idea for multiturn
[17:58:44] <sync> uhhh
[17:58:46] <sync> yeah
[17:59:00] <sync> but you have to have an axis that can turn more than 4096 revs
[17:59:09] <sync> or whatever multiturns it supports
[17:59:59] <pcw_home> so if you get transient wrong data from a multiturn, it wraps and gives you wrong data forever
[18:00:01] <andypugh> pcw_home: How should it handle wrap?
[18:00:23] <andypugh> Ah, ok, so that’s a problem with transient bad data?
[18:00:38] <pcw_home> M would mean never wrap
[18:01:00] <andypugh> Shouldn’t it unwrap with the next un-bad data loooking like an unfeasible jump the other way?
[18:02:06] <pcw_home> just saying that the normal use case of multi-turn absolute (or linear absolute) encoders is that you never wrap
[18:02:40] <sync> well you can have that situation when a jackass puts the motor on in any old way
[18:02:49] <sync> and your wrap point is in the middle of the travel
[18:03:12] <pcw_home> thats a sign change, not a wrap
[18:04:02] <sync> well it goes from say 0 turns to 4096, but yeah it just changes sign I guess
[18:05:55] <andypugh> pcw_home: It’s not an easy change, as the SSI, BSPI rct encoders all use the same code as the smart-serial encoders.
[18:05:56] <andypugh> https://github.com/LinuxCNC/linuxcnc/blob/master/src/hal/drivers/mesa-hostmot2/sserial.c#L1451
[18:06:36] <pcw_home> for single turn absolute (like resolvers or magnetic encoders) you need to wrap, for multi-turn it causes trouble
[18:07:30] <andypugh> Surely it causes just as much trouble with single-turn if bad data gives an erroneous wrap?
[18:08:35] <pcw_home> its not a big thing but for example if you take a linear 26 bit SSI absolute encoder and unplug it then plug it in again the count is permanently off by 64 million
[18:08:50] <andypugh> Though I guess that it is a _much_ bigger error with a 30-bit absolute than a 10 bit incremental.
[18:10:18] <andypugh> We would need to create a new LBP_ENCODER_NOWRAP type I think.
[18:11:30] <andypugh> Hmm, maybe not
[18:12:50] <pcw_home> you can lose position information on 1 turn absolute (or quadrature) encoders but ideally multiturn/linear absolute encoder should be immune
[18:13:39] <andypugh> pcw_home: Can you put this in an email to make it harder for me to forget?
[18:17:42] <andypugh> I can add an “allow wrap” flag in the data fields. Note that smart-serial will get the faciltity too, it’s all handled by the same code.
[18:18:32] <pcw_home> as long as its an option it should be fine (and wrap is default)
[18:18:55] <pcw_home> (so "no-wrap" flag)
[18:19:16] <linux_vampire> the probing menu become white, ant not responding.
[18:19:46] <andypugh> Which probing menu?
[18:20:20] <linux_vampire> one sec i will take a print screen
[18:22:20] <linux_vampire> andypugh, http://i.imgur.com/pBNwey0.png
[18:23:22] <andypugh> Ah, right, the one from Fern?
[18:24:05] <andypugh> Looks like a missing subroutine file, reading the text on the screen
[18:24:43] <linux_vampire> what do you mean ?
[18:25:11] <linux_vampire> http://i.imgur.com/p8esLB7.png
[18:25:17] <linux_vampire> this is the bug
[18:25:30] <andypugh> EOF in file … seeking o-word o<xplus>
[18:26:15] <andypugh> Hmm, does it come back if you chage tabs?
[18:26:42] <linux_vampire> no
[18:26:53] <andypugh> It looks to be running a program
[18:27:13] <andypugh> What happens if you press program-stop?
[18:27:26] <linux_vampire> it back to life
[18:28:14] <andypugh> Maybe it is meant to become unresponsive when probing is not possible due to a program running?
[18:28:30] <andypugh> Or was the program that was runnning actually a probe routine?
[18:28:42] <linux_vampire> no. it's not consistent
[18:29:22] <linux_vampire> and it actually not smart at all to hide it under tab
[18:29:37] <linux_vampire> i can't see anything
[18:33:05] <linux_vampire> why it move the z axis if i press the x+ probing button??
[18:34:57] <andypugh> No idea. I have not used those routines yet.
[18:35:57] <linux_vampire> it look nice and fancy but the nothing more then that.
[18:36:25] <andypugh> You were the one demanding nice and fancy earlier!
[18:36:50] <linux_vampire> the buttons look very nice.
[18:37:11] <linux_vampire> but if i press the x+ i want the x to move not the x and z
[18:38:50] <andypugh> Something funny going on, because the x-plus routine only contains X moves: https://github.com/verser-git/probe_screen/blob/master/macros/xplus.ngc
[18:40:24] <linux_vampire> i see those macros, and it's look like something combined them together to the actual path
[18:41:44] <andypugh> Yes, the routine for x-plus is here: https://github.com/verser-git/probe_screen/blob/master/python/probe_screen.py#L540
[18:42:37] <andypugh> So, there is a Z-move from the safe-height to the probe-height and back.
[18:43:05] <linux_vampire> it is not safe!
[18:43:30] <andypugh> Without reading the instructions, I don’t know which setting is which.
[18:43:43] <linux_vampire> if you are using large ball in a slot, and the z go up you crash the ball
[18:44:41] <_methods> you always have to be more careful when you're working with big balls
[18:45:06] <linux_vampire> !!!!!!!
[18:45:17] <andypugh> In that case yuu would need to program the safe-Z to be the same as the probe-Z
[18:46:08] <linux_vampire> z clearance?
[18:47:15] <andypugh> There is a setting, somwhere on the screen, that defines how far up and down the probe moves during a routine.
[18:47:22] <andypugh> Set that to zero.
[18:49:24] <linux_vampire> i'm trying now
[18:49:37] <linux_vampire> one sec
[18:56:20] <andypugh> Glade is driving me to distraction!
[18:56:59] <andypugh> It has a habit of changing the width of the table as I change the settings of the widgets.
[18:57:58] <andypugh> Worst is when I change the axis number of the HAL DRO, it takes a default value that is a 120 digit number(!) and makes every cell in the table wide enough to hold the same thing.
[18:58:10] <andypugh> Which can’t be recovered from….
[18:58:12] <linux_vampire> is there a way to make a button that i can press and it will run some g-code?
[18:59:05] <linux_vampire> i mean button in pyvcp
[19:04:40] <linux_vampire> andypugh,
[19:07:04] <andypugh> Yes
[19:07:18] <andypugh> It’s a bit clunky
[19:07:32] <andypugh> It’s _much_ easier in GladeVCP
[19:07:51] <andypugh> GladeVCP has buttons that run G-code built-in
[19:08:43] <andypugh> With PyVCP you need to define button with a HAL pin, then connece that HAL pin to one of the MDI_COMMAN HAL pins, then put the G-Code in the INI file.
[19:09:18] <andypugh> http://linuxcnc.org/docs/2.7/html/config/ini-config.html#_halui_section
[19:09:29] <andypugh> Though that basically just says the same as I said
[19:10:12] <jdh> JT has examples in the wiki
[19:14:05] <linux_vampire> all i need is 7 buttons x+ ,x- ,y+ ,y- ,z- ,probe circle in, probe circle out
[19:16:51] <linux_vampire> i need to get some sleep
[19:32:46] <rene-dev> linux_vampire why dont you just use the probe screen?
[19:38:52] <andypugh> rene-dev: She doesn’t like some of the features, like the z-retract setting.
[19:39:53] <rene-dev> andypugh I dont like some of them... you can edit the gcode
[19:42:08] <andypugh> I think you need to edit the Python to disable the z-retract, but that’s not any harder than editing the G-code
[19:44:52] <dioz> yo
[19:45:29] <dioz> yo
[19:45:35] <dioz> so here i am
[19:46:10] <andypugh> ?
[19:46:39] <dioz> andypugh: how much diameter on a rod does a person need on a metal lathe?
[19:47:00] <andypugh> Depends what you are making, I guess.
[19:47:12] <dioz> shafts and struts for RC car
[19:47:16] <andypugh> It’s hard to make a 2” part from a 1” rod
[19:47:38] <dioz> i wonder how thin i could lathe a piece of rebar down
[19:47:40] <dioz> before it broke
[19:47:54] <dioz> or just bent
[19:48:08] <andypugh> I wouldn’t use rebar for anything on an RC car.
[19:48:18] <jdh> rebar is good for concrete
[19:49:24] <andypugh> But if you machine down a 1mm length to 0.5mm, then do the next mm, you can go very small. It comes down to the lenght-to-diameter ratio much more than the actual diameter.
[19:49:47] <andypugh> So, machine it in short lengths.
[19:51:55] <SpeedEvil> at some point it'll start whipping around too
[19:52:07] <SpeedEvil> so you need some form of ever more exotic support
[19:53:14] <SpeedEvil> i suppose saying it's a lathe if it's got a hollow headstock and tailstock, even if those happen to have dies in is cheating.
[20:05:45] <Jymmm> SpeedEvil: "... more exotic support", TWO FFF bras, at the same time?
[20:10:48] <SpeedEvil> multiple steady-rests
[20:11:15] <SpeedEvil> and possibly a powered tailstock
[20:11:15] <Jymmm> multiple how? over the distance of the lathe?
[20:11:35] <Jymmm> stacked?
[20:11:41] <SpeedEvil> along the length of the work
[20:11:45] <Jymmm> ah
[20:11:54] <SpeedEvil> if the work is too narrow, it will whip
[20:12:05] <Jymmm> ah, gotcha
[20:13:01] <Jymmm> SpeedEvil: have you used the forced air propane heates before?
[20:13:06] <Jymmm> heaters
[20:13:42] <SpeedEvil> not if you mean the shop heating type ones
[20:13:58] <Jymmm> Like these: http://www.homedepot.com/p/Dyna-Glo-Pro-60K-BTU-Forced-Air-Propane-Portable-Heater-RMC-FA60DGP/202223041
[20:17:57] <Frank_10> hi
[21:06:43] <XXCoder> I AM ALLIVE!
[21:08:38] <SpeedEvil> Jymmm: what about them?
[21:12:53] <Jymmm> SpeedEvil: any drawbacks? annoyances?
[21:15:28] <SpeedEvil> Jymmm: it's a gas heater.
[21:15:52] <SpeedEvil> it makes CO, if the safety goes fucky, it makes lots of water, which may cause condensation on various parts of the structure.
[21:16:34] <SpeedEvil> it requires you to purchase gas. Large temperature swings are going to invite condensation anyway, if you leave it off for times
[21:27:52] <sector_0> hey, I was reading about stepper motor controlling on the reprap wiki, and from what it's saying, an H-bridge controller is better than a stepper controller chip (e.g. a4988) at higher speeds
[21:27:55] <sector_0> is this true?
[21:29:37] <SpeedEvil> it's stupid.
[21:29:46] <SpeedEvil> As many stepper controller chips are h-bridges
[21:30:19] <SpeedEvil> Stepper controller chips generally regulate current better, so torque stays constant over RPM
[21:30:26] <SpeedEvil> and the stepper doesn't get hot.
[21:30:53] <SpeedEvil> With a stepper controller chip with adequate supply voltage, it can be lots better than a simple h-bridge unregulated as it can run the stepper at much higher voltage
[21:31:09] <SpeedEvil> As to the impliemtnation in reprap - that is a while nother question.
[21:31:51] <sector_0> http://reprap.org/wiki/Stepper_motor_driver
[21:31:59] <sector_0> http://reprap.org/wiki/Stepper_motor_driver#Microcontroller-based_Stepper_Drivers
[21:32:06] <sector_0> maybe i understood it wrong
[21:33:13] <sector_0> also is it possible to use an a4988 with power MOSFETs on the output to drive let's say a NEMA 34?
[21:34:25] <sector_0> well then again, I know it's possible, but is it advised?
[21:36:53] <Jymmm> SpeedEvil: thanks for the info on the condensation
[21:41:41] <eeriegeek> hmnn, they make an A4989 just for driving external MOSFETs
[21:56:36] <SpeedEvil> sector_0: it's poorly written in several aspects
[22:00:22] <XXCoder> guys
[22:00:36] <XXCoder> what do I use to protect 123 blocks but keep it clean and well 123 sized heh
[22:01:45] <Tom_L> store them at room temperature in an oxygen free environment
[22:02:17] <XXCoder> oh was talking about coating
[22:02:20] <SpeedEvil> Vacuum is better
[22:02:30] <XXCoder> work uses this oil but dunno what name it is
[22:02:33] <SpeedEvil> a light oil
[22:02:55] <Tom_L> any light oil should be ok
[22:03:40] <XXCoder> got example of such?
[22:03:41] <SpeedEvil> if your shop is condensing, you have other problems.
[22:03:52] <SpeedEvil> any '3 in 1' oil even
[22:03:57] <SpeedEvil> it's very non-critical
[22:04:31] <Tom_L> they usually ship with a light coat of cosmolene
[22:04:37] <XXCoder> its my room actually just beem a long while since I last oiled it
[22:04:56] <XXCoder> maybe just bring to work spray it a little wipe off and bring home
[22:05:33] <Tom_L> http://www.cosmolinedirect.com/
[22:05:43] <Tom_L> been used for years in some form or another
[22:06:46] <SpeedEvil> i recommend a light spray of WD40.
[22:06:59] <SpeedEvil> i have a constant recirculating shower setup of wd40 in my shop.
[22:07:14] <SpeedEvil> Works great for keeping off the rust and discouraging smoking.
[22:19:55] * Jymmm sends SpeedEvil a randomized tesla generator
[22:22:08] <XXCoder> KL73 interesting
[22:32:24] <XXCoder> I wish theres Al 123 blocks or nickel so its good for some uses (those probably strunk or expand too much to be as useful as steel)
[22:32:46] <XXCoder> and yes there is 3d print 123 block lol sure thats so precise :P
[22:34:30] <XXCoder> http://www.popularwoodworking.com/woodworking-blogs/editors-blog/arts-mysteries-blogs/rust-prevention-for-woodworkers
[22:34:33] <XXCoder> interesting tips
[22:35:16] <XXCoder> Starrett precision tool oil (mineral oil) "
[22:35:22] <enleth> XXCoder: some places use hot wax dipping to protect tools when not in use, but that's not very practical for an individual
[22:35:46] <XXCoder> 16 bucks not too bad on amazon
[22:36:48] <enleth> although it sounds quite practical for a place with a bunch of dedicated tool room guys who strip the wax before handing the tools out and wax them again upon return
[22:37:48] <enleth> unless you use some tool like once a year, in which case it might just be the easiest thing to do short of dunking it in a bucket of oil for storage
[22:40:11] <XXCoder> enleth: one thing I need to do is dehumidifer
[22:40:28] <XXCoder> my room and worse, my nephew's have water problem, expecially during fall and winter
[22:40:37] <XXCoder> cracked fountin maybe
[22:40:48] <Tom_L> dehumidifier
[22:40:58] <XXCoder> tom thats what i said
[22:41:09] <Tom_L> oh i just came back
[22:41:17] <XXCoder> lol np
[22:41:27] <XXCoder> not sure what is good though
[22:41:55] <enleth> Tom_L: you win today's "Captain Obvious" badge, wear it with pride
[22:42:19] <Tom_L> Y thank you !
[22:42:52] <Jymmm> Is that anything like a CaptHindsight badge?
[22:43:24] <Tom_L> or bag them with some desiccant bags
[22:43:35] <CaptHindsight> it's more of a discoloration than a badge
[22:45:18] <Jymmm> heh
[22:45:32] <zeeshan> so means
[22:45:33] <zeeshan> :]
[22:45:51] <XXCoder> meaningful you mean
[23:02:41] <XXCoder> I love it when they use photoshop
[23:02:50] <XXCoder> they always get scale wrong
[23:02:56] <XXCoder> https://www.amazon.com/KEDSUM-Approved-Thermo-Electric-Dehumidifier-Basement/dp/B01KTAF4I8/
[23:03:02] <XXCoder> look at one where they show closet
[23:46:27] <SpeedEvil> XXCoder: work out air changes. It can take a _LOT_ of dehumidification to remove the water
[23:46:41] <SpeedEvil> https://en.wikipedia.org/wiki/Psychrometrics#/media/File:PsychrometricChart.SeaLevel.SI.svg is awesome
[23:48:11] <XXCoder> how do I read the chart
[23:49:52] <SpeedEvil> Temperature is on the horizontal axis, water content is on the vertical.
[23:50:49] <XXCoder> i see humidity curves
[23:50:52] <SpeedEvil> the top curve is the 100% water saturation point, so at 20C, go up to the top red line with 100%, and then to the right, and you get 100% water saturation is at about 14 grams of water a kilo
[23:51:46] <SpeedEvil> This means that if you take air at 20C, 100% humidity and heat the space by 10C, it is at about 55% humidity
[23:52:09] <SpeedEvil> Or if you cool it to 10C, about half the water condenses out
[23:52:32] <SpeedEvil> the curved lines are lines of constant humidity
[23:53:33] <XXCoder> ahh
[23:53:35] <XXCoder> thanks
[23:53:42] <XXCoder> my room is at 25c approx
[23:54:18] <XXCoder> approx 50%
[23:54:31] <XXCoder> thats around 0.01
[23:54:50] <XXCoder> not too sure how to translate that to grams of water
[23:55:10] <SpeedEvil> The density of air is about 1.3kg/m^3 at STP
[23:55:31] <SpeedEvil> So multiply to get kilos of water per m^3
[23:56:32] <XXCoder> 165 oz
[23:56:47] <XXCoder> dunno sounds huge
[23:57:16] <XXCoder> I assume room is 15'x15'x20' then converted to m^3 got 127.4
[23:57:37] <SpeedEvil> you have a 20 foot tall room?
[23:57:57] <XXCoder> not really lol
[23:58:55] <XXCoder> 82.81 corrected oz