#linuxcnc Logs
Jan 19 2021
#linuxcnc Calendar
12:24 AM roycroft: yeasties are very resiliant
12:30 AM perry_j1987: yep
01:46 AM -!- #linuxcnc mode set to +v by ChanServ
01:46 AM -!- #linuxcnc mode set to +v by ChanServ
01:49 AM CaptHindsight: perry_j1987: which BOB ended up working?
03:23 AM -!- #linuxcnc mode set to +v by ChanServ
03:23 AM -!- #linuxcnc mode set to +v by ChanServ
03:28 AM Deejay: moin
04:16 AM Loetmichel: *sigh* what do they say? "the kids of the shoemaker go barefoot"? ... just got a call from my sister who is admin in a medium sized company. Her APC UPS at home died. Battery error... "when was the last time you changed the batteries?" "oh, about 5 years?" "there is your problem lady..." :)
04:35 AM rs[m]: Loetmichel: so your kids are cutting with a cheese knife?
04:35 AM Loetmichel: i have no kids. thankfully
05:36 AM Loetmichel: Hmm... 7*20 M3 threads cut in alu sheet in less than 5 minutes. Not bad! And now my righ arm cramps from holding the battery drill stiff and switching directions 140 times... :)
05:42 AM alex_joni: get a CNC :)
05:45 AM Loetmichel: i have a CNC
05:45 AM Loetmichel: just no tool changer
05:46 AM Loetmichel: so its easier to do it by hand than to realign the workpiece and switch to the tap
05:53 AM JT-Cave: morning
06:06 AM XXCoder: yo
06:14 AM Loetmichel: *note to self:* when the finger that spreads the alcohol onto the alu for easier deburring starts to burn the burrs have worn thru the skin. And its cold on top! :)
06:16 AM Tom_L: morning
06:16 AM Loetmichel: mornin'
06:17 AM -!- #linuxcnc mode set to +v by ChanServ
06:41 AM XXCoder: hey
09:02 AM jymmmm: Morning
09:07 AM alex_joni: evening for some
09:07 AM * alex_joni just finished work for today
09:08 AM gregcnc: good(or nat) non specific time of day to you as well
09:09 AM jymmmm: hi alex
09:09 AM jymmmm: hi greg
09:09 AM gregcnc: where you been jymmmm?
09:10 AM jymmmm: life gets in the way sometimes
09:10 AM gregcnc: yeah
10:35 AM unterhaus: the fact that usb c can supply different voltages seems pretty neat
10:43 AM flyback: my pololu order shipped
10:43 AM flyback: BMCC
11:30 AM unterhaus: I'm really curious why you can't find kennametal indexed tooling on their site https://www.ebay.com/itm/NEW-KENNAMETAL-A10CRGNL3-BORING-BAR/173373358703?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m1438.l2649
11:33 AM JT-Shop: text to customer slight delay of game, ran out of materials last night but McMaster will have them to me by lunch
11:33 AM JT-Shop: customer can we get 4 or 5 installed today?
11:33 AM JT-Shop: LOL
11:34 AM Thorhian: Oh boy lol
11:35 AM JT-Shop: I should have told him sure but they won't do anything until I bring them back and finish them
11:37 AM Thorhian: You know, I've heard both good and bad things about McMaster Car. They seem to be of good quality and good customer service (at least for US customers?), but terrible prices on some stuff.
11:37 AM norias: It's a speed / money thing
11:37 AM norias: If you recognize that, they are great.
11:38 AM norias: You will almost always pay more, but you will almost always get it delivered in a reliably short time
11:38 AM Thorhian: Hmmm, I see.
11:38 AM norias: They also have what I think is a insanely easy to use web site
11:39 AM JT-Shop: it really depends on the item as far as the price, for example don't order structural steel from them it's twice as high
11:39 AM norias: and all the little details like billing, etc are done well
11:39 AM norias: for B2B they are super easy to deal with
11:40 AM Thorhian: I really wish Misumi would update their crud compared to McMaster. I've had good interactions with them, but their login/billing/shipping setup is kind of ancient in terms of websites.
11:41 AM norias: Misumi is definitely interesting
11:42 AM Thorhian: Lol, what do you mean?
11:42 AM norias: They have some interesting product offerings.
11:42 AM norias: Fun stuff.
11:43 AM Thorhian: Indeed. That's actually where I got my timing belt and pulleys for my mill's spindle. My father got a decent deal on 20mm profile t-slot extrusions too a while back.
11:45 AM roycroft: i don't find mcmaster's prices to be out of line on the things i buy from them
11:45 AM Thorhian: I also have one of their paper catalogs. Makes it a lot easier to browse their offerings.
11:45 AM roycroft: and they dispatch out of fresno or reno 99+% of the time, so i get things 2 days ground
11:45 AM roycroft: and yes, their website is easy to navigate
11:45 AM roycroft: and it's so easy to order from them
11:46 AM roycroft: my main criticism is that they don't have a good quote/project system
11:46 AM roycroft: one can make an order, and save it instead of processing it
11:46 AM roycroft: but it's a little clunky
11:46 AM Thorhian: You are in Oregon right? Also, that would be nice.
11:46 AM mrec: does anyone know is there a register that can be read to get the current tool offset?
11:46 AM roycroft: yes
11:47 AM mrec: G43.1 ZX ... I'd like to read X
11:47 AM roycroft: it would be nice to be able to build a bom for a project over time on their site
11:47 AM roycroft: save the bom
11:47 AM roycroft: and just order it when you need to build that project
11:47 AM norias: I've mostly used McMaster for business.
11:47 AM roycroft: which kind of can be done, but with mcmaster you convert the saved order to an active order
11:47 AM roycroft: and then you're done once you order it
11:47 AM norias: So, when it's time to do expense reports
11:48 AM norias: i can download all the invoices fairly quickly, etc
11:48 AM roycroft: yes, that is nice
11:48 AM norias: most things i can get next day
11:48 AM roycroft: also, my other minor nit is that they don't calculate shipping when you place an order
11:48 AM norias: and i get the tracking info and get a text notification for delivery
11:48 AM roycroft: you only find out when they send you the invoice, which is usually after you get the order
11:48 AM norias: because all my stuff is prototyping so hours / days count
11:49 AM norias: well, was
11:49 AM Thorhian: That's kind of obnoxious roy.
11:49 AM norias: haha, doesn't seem like i'll be machining anytime soon
11:49 AM roycroft: it's not really obnoxious, imo
11:49 AM roycroft: they don't use shipping as a profit center, so their shipping charges are minimal
11:50 AM roycroft: but it would be nice to know at the time of the oder what the total delivered price will be
11:50 AM norias: yeah, i usually have my own shipping account they bill
11:51 AM roycroft: i have shipping accounts with ups and fedex, but i don't get the same discounts that mcmaster do
11:51 AM norias: which is handy when folks ask why stuff costs so much
11:51 AM roycroft: it's always cheaper to have them bill me for shipping
11:51 AM norias: most of the overnight shipping i put in the "Cost of Speed" category
11:51 AM norias: like Cost of Quality accounting
11:51 AM Thorhian: XD
11:51 AM roycroft: do you add a nuisance fee when the customer asks too many questions? :)
11:52 AM norias: Then I can say, "We do next day for most things. We can do 3 day for x% less."
11:52 AM norias: well, i'm always an internal shop reporting these things to management
11:52 AM norias: so i turn it in to their decision
11:52 AM Thorhian: XD
11:52 AM norias: "Do you want the standard to be 1 day or 3 day? I'll do it either way, but this is what 1 day costs."
11:53 AM roycroft: i'm rarely in such a hurry for things
11:53 AM norias: I'm not, but the people that pay me are
11:53 AM roycroft: which is a big reason i have no interest in amazon prime
11:54 AM norias: usually because they have mismanaged a project
11:54 AM roycroft: they need to simmah down :)
11:54 AM roycroft: jit is not the same as omg
11:54 AM roycroft: the former entails planning; the latter only panic
11:54 AM norias: yeah, a lot of my job is turning other people's panic in to planning
11:55 AM norias: you can only tell people they planned poorly
11:55 AM norias: so many times
11:55 AM Thorhian: Having purchased a lot of things on Amazon, pretty sure I saved a good chunk of money on shipping with prime, plus prime video is nice *shrug*
11:55 AM norias: before you learn it's not an effective strategy
11:56 AM roycroft: i cannot remember the last time i paid shipping on something from amazon
11:56 AM roycroft: i use their free super-saver shipping on everything
11:56 AM roycroft: and if it takes a couple weeks to get here, so what
11:57 AM roycroft: i'm pretty good at planning
11:57 AM roycroft: panic mode does not fit my personality well
11:57 AM Thorhian: Now I just need to wait for that bandsaw to come in stock locally here at Harbor Freight.
11:58 AM gregcnc: unterhaus which number did you search?
11:58 AM roycroft: besides, super-saver shipping rarely takes a couple weeks these days - it's usually a matter of 4-5 days vs. 2 days for prime
12:00 PM unterhaus: gregcnc many variations of A10CRGNL3 A10-crg etc. Also with the N15
12:00 PM unterhaus: it's not clear that the box goes with the bar
12:02 PM gregcnc: the bar it'self is ancient, which is why you can't find it
12:02 PM unterhaus: that' what I thought, but I thought they probably still want to sell inserts
12:02 PM gregcnc: with crap images of the pocket, I'd move on
12:03 PM gregcnc: unless you need that specific part number, but then you'd know what it is
12:12 PM unterhaus: says new in box and looks to be new. But if I can't find any info, it's not worth buying
12:17 PM unterhaus: a10crgnl3 is an ansi-iso part number. Kennametal puts it on the tool but not in their catalog in a searchable form
12:35 PM Tom_L: unterhaus, try A20UDDPNR4KC3
12:41 PM unterhaus: Tom_L, I don't think that's the right part number, I think the boxes got switched
12:43 PM unterhaus: it's really hard to tell because there is nothing for scale
12:43 PM Tom_L: yeah
12:57 PM gregcnc: do you have a bunch of round insets to consume?
12:58 PM Thorhian: I've noticed that round milling inserts are kind of hard to get unless you order from China. Am I just missing something?
12:59 PM gregcnc: the world has moved on I think
12:59 PM Thorhian: Moved on round inserts?
12:59 PM Thorhian: from round inserts*
01:00 PM gregcnc: looka t soem tool catalogs to see what's being used today
01:00 PM gregcnc: typing extra poor today
01:00 PM Tom_L: gotta keep ya buying new tooling
01:01 PM Thorhian: I've just read that round inserts have good geometry for cutting more material efficiently. It just seems like people sell tons of the other stuff, but very little round ones. Kennametal still has a few, but I ain't got money for kennametal tooling lol
01:02 PM gregcnc: what do you need?
01:03 PM Thorhian: I don't "need" anything, I'm still a hobbyist. Figured it might be an interesting thing to try out for my lower powered spindle (BT30, 1.8KW).
01:03 PM gregcnc: do you have milling head that uses them?
01:04 PM Thorhian: Not yet, didn't want to bother buying one if it would be expensive/buying chinesium all the time.
01:04 PM gregcnc: if not chinaco, and new from dealers, it will be expensive
01:05 PM gregcnc: ebay if you want a better price
01:05 PM gregcnc: milling heads are tricky as they come in so many flavors
01:07 PM Thorhian: Pretty much everything on ebay is from china if it isn't used most likely. The only ones I considered there were a more reputable looking Chinese company in terms of quality. Wish Tegara in Taiwan made some lol
01:07 PM Thorhian: SGCO I believe it was.
01:08 PM gregcnc: no I mean real tools, not random cheap garbage
01:09 PM gregcnc: I buy Sandvik, Iscar, etc. from ebay. Not 10USD per box inserts even if they carry these names
01:10 PM Tom_L: gregcnc, do you have a shop or mostly for hobby?
01:11 PM gregcnc: somewhere in between
01:11 PM Tom_L: kinda what i figured
01:11 PM gregcnc: i read far too many catalogs
01:11 PM Thorhian: https://www.ebay.com/itm/202266568068 This is the holder I was thinking about. Problem is finding decent insert. I'll try searching for brandname ones and see what comes up
01:12 PM Tom_L: stopped by to see my friend today and he wasn't doing anything
01:12 PM Tom_L: pretty slow around here right now
01:12 PM Tom_L: had a part for a mutual friend's racecar in the manual lathe and that was it
01:14 PM Thorhian: That's cool, the lathe part Tom
01:14 PM gregcnc: that head seems small for facing. with 10mm inserts you have 16mm finish width
01:14 PM Tom_L: i didn't bother looking to see what it was
01:16 PM Thorhian: I have a small machine greg, unless you think I could do better for facing steel greg?
01:16 PM Tom_L: what's the thread size on that?
01:16 PM Thorhian: I still need to fill my column up with E/G for more weight and dampening lol. Don't want to go crazy.
01:16 PM Thorhian: M12 Tom_L
01:16 PM gregcnc: do the math for cutting power. That's how I settled on 40mm heads
01:17 PM Thorhian: I have seen the equations, but I haven't been able to do it since I didn't know the watt constants for the materials, but my new machinery handbook I believe does :)
01:17 PM Tom_L: what rev?
01:18 PM Tom_L: mine is old
01:18 PM Thorhian: 31, got it for Christmas :D It's my first one lol
01:18 PM gregcnc: you'll never buy another
01:18 PM Tom_L: mine's mid 20 something
01:19 PM Tom_L: i got a ready reference with mine
01:20 PM Tom_L: https://www.amazon.com/Machinists-Ready-Reference-10th-Weingartner-ebook/dp/B01CO7APAQ
01:20 PM Tom_L: different cover but looks the same
01:20 PM Thorhian: Nice
01:20 PM Thorhian: It's just packed with information. I'm glad these later editions have stats in metric and imperial.
01:23 PM Thorhian: But gregcnc, for the smaller one, I figured it could be used as a roughing end mill as well, especially since my machine is configured for lower RPM, with my spindle motor only goes 3000 RPM max, and I only have 1.59 multiplication in RPM to the spindle with my pulley situation. I figured I could use larger diameter tooling to my advantage.
01:29 PM * JT-Shop wires up the last PLC
01:30 PM Thorhian: A lot of what I will be cutting is aluminum in general, so maybe I am overthinking it lol,
01:30 PM Thorhian: Also, nice JT :D
01:44 PM roycroft: it's lunch time - time to apply the last coat of paint to the section of the bandsaw stand i'm currently working on
01:44 PM * roycroft will be able to do some actual assembly tomorrow
01:45 PM unterhaus: I imagine chinaco inserts can be pretty good
01:48 PM roycroft: i don't use much insert tooling for my metalworking machines, but i do use them on some woodworking machines, and they perform very well
01:49 PM roycroft: i do have a 3" facing mill for my milling machine that uses triangular inserts, and those also work well, and are from china
01:49 PM roycroft: the chinese know how to make high quality tooling
01:50 PM roycroft: and you can get it, but not at places like harbor freight
01:50 PM roycroft: what the chinese don't do is make both high quality and shite quality tooling of the same type and sell both for the same price
01:51 PM roycroft: and if you do want true shite quality tooling for a song, increasingly you have to look to india for that
01:52 PM roycroft: the chinese are improving their quality across the board
01:52 PM roycroft: after wwii, it was the japanese who dumped cheap crap on the us market
01:52 PM roycroft: they they started making good stuff, and it was taiwan who made the crap
01:52 PM roycroft: then taiwan upped its game, and mainland china made the crap
01:52 PM roycroft: now the crap is made in india and vietnam
01:53 PM roycroft: and when those countries start making higher quality stuff, the crap market will probably shift to the african continent
01:53 PM gregcnc: Thorian, you need to look at your needs and capability. You're just starting and will find what works.
01:53 PM roycroft: i don't know what we'll do when we run out of countries to make cheap crap
01:53 PM gregcnc: you'll have plenty of tools you never touch again along the way
01:54 PM roycroft: don't buy high quality tooling at first
01:54 PM roycroft: don't buy high quality anything consumable at first
01:54 PM roycroft: until you figure out what you actually use
01:55 PM gregcnc: I didn't and wouldn't do that
01:56 PM unterhaus: I usually look on ebay for something like what I want and then look at the app notes. Even a short description can make a big difference in selecting the right insert
01:57 PM gregcnc: catalogs will tell you everything, but some give away more than others
01:57 PM unterhaus: but most hobby machines need inserts that are a bit unusual due to low speeds and feeds
01:58 PM gregcnc: you just need to apply them correctly
01:58 PM unterhaus: Red cross once again reminds me why I hate giving blood thru them, their website stinks
01:59 PM unterhaus: sure, but a lot of inserts are made for huge DOC that most of us can't pull off
01:59 PM gregcnc: don't use those grades?
01:59 PM gregcnc: most hobbyists should be using medium to finishing grades
01:59 PM unterhaus: I hope you understand that we are agreeing?
01:59 PM gregcnc: yes
02:00 PM gregcnc: there is a significant learning curve to inserts
02:00 PM gregcnc: selecting inserts
02:01 PM gregcnc: and everything is heading to faster and more power, so you need to pay attention
02:07 PM gregcnc: machinery's handbook should split the antique info and create a modern version
02:12 PM norias: yeah, that'd be interesting
02:12 PM gregcnc: the metric "machinery's handbook" isbn 9783808519141
02:13 PM norias: really?
02:13 PM gregcnc: nothign to do with machinery's handbook
02:16 PM norias: ok
02:16 PM norias: Yeah I thought that there was already a ton of metric data in there
02:16 PM norias: uh, a short ton, to be precise
02:17 PM gregcnc: not enough
02:17 PM Thorhian: Thanks greg. I also hesitate to buy chinese in general. It's unavoidable, but I don't like slave labor either, so I do what I can.
02:17 PM roycroft: just ask abom what inserts to use
02:17 PM roycroft: then buy machines as big as his :)
02:18 PM roycroft: i have that book, gregcnc
02:18 PM norias: ya know, sometimes it's not just size or HP
02:18 PM roycroft: you're correct - it has nothing to do with machinery's handbook, but it's an excellent book
02:18 PM norias: the one place i worked, there was a 90HP lathe that beat the 150HP machines every time
02:18 PM roycroft: when it comes to tooling it kind of is, norias
02:19 PM norias: roughing cast iron
02:19 PM roycroft: abom hogs off 0.200 or more at a time on his lathe
02:19 PM norias: and it was physically smaller
02:19 PM gregcnc: 0.2mm?
02:19 PM roycroft: inserts that work for that won't work well when you're only able to turn 0.050" at a time
02:19 PM roycroft: sorry
02:19 PM roycroft: 0.200"
02:19 PM gregcnc: .200"
02:20 PM roycroft: i'm not saying that big machines are better
02:20 PM norias: who is abom?
02:20 PM gregcnc: real and rated HP is nice to know
02:20 PM roycroft: he's a well-known youtube machinst
02:20 PM gregcnc: YT machinist
02:20 PM roycroft: abom79
02:20 PM norias: gregnc: fare, but it was sorta apples to apples both being the rated
02:20 PM roycroft: he knows his inserts very well
02:21 PM norias: on the other hand the 90HP was belt drive, and the 150's were gear
02:21 PM roycroft: but he knows what works for him, and he has medium size to medium plus size machines
02:21 PM roycroft: i saw him try to operate a 12x36 atlas lathe once
02:21 PM roycroft: it was ridiculous
02:22 PM roycroft: he had no clue how to use it
02:22 PM roycroft: he was laughing the whole time
02:22 PM norias: i think the inserts for those lathes were around $120 / ea
02:23 PM Thorhian: norias: That's a lot of lost power in gears compared to belt drive.
02:23 PM roycroft: sure, but you get to gain a whole lot of noise!
02:23 PM Thorhian: XD
02:23 PM norias: Thorhian: i don't think they were losing it all in the gears
02:24 PM norias: it was also just the geometry of the machine
02:24 PM norias: the radial resultant force in the smaller machine went straighter down to the ground
02:24 PM roycroft: i am done with this round of painting, but given that it's winter, i'm going to wait at least a full day before doing any assembly
02:24 PM norias: i.e. through the frame
02:25 PM norias: the bigger ones, it angled out a bit, so it was always trying to cantilever the carriage
02:26 PM norias: i'm sure there's some design tradeoff there
02:27 PM norias: incidentally, switching between vim and a word processor is rough
02:27 PM norias: my document has all these i's and A's in it
02:28 PM roycroft: i mostly use vi for all my writing and editing
02:28 PM roycroft: then i switch to a word processor for formatting, if i'm no t using a markup language for the formatting
02:29 PM roycroft: but i don't switch over the word processor until the editing is all done
02:29 PM norias: makes sense
02:29 PM roycroft: i still use troff for a lot of the documents i write, if i don't have to share the source
02:29 PM norias: i should just learn latex
02:29 PM roycroft: not that i'm unwilling to share troff markup documents
02:29 PM roycroft: but most people don't know how to deal with them
02:30 PM roycroft: i used to do some stuff with tex
02:30 PM roycroft: tex has a superior page model to troff, there's no doubt about that
02:30 PM roycroft: but even using a good macro set like latex i find the markup more tedious with tex than with troff
02:31 PM norias: i never tried troff
02:31 PM norias: maybe i should
02:31 PM roycroft: i've been using troff since the late '70s, and wrote my first troff macro package in the early '80s
02:31 PM roycroft: tex was not an option back then
02:32 PM norias: how well does troff handle mathematical symbols?
02:32 PM roycroft: and btw, back in the '80s microsoft word was a modal editor
02:32 PM roycroft: before the gui days
02:33 PM roycroft: when it ran on ms-dos
02:33 PM roycroft: not even remotely as well as tex
02:33 PM roycroft: there's an eqn macro package that does maths kind of ok
02:33 PM roycroft: and there's tbl, which formats tables
02:33 PM roycroft: that actually works really well
02:34 PM roycroft: but tex runs circles around troff/eqn when it comes to maths
02:34 PM norias: fair
02:34 PM norias: that's really what i'm after
02:34 PM roycroft: then latex would be a great choice
02:34 PM norias: cool
02:34 PM roycroft: i was doing data-driven stuff
02:35 PM roycroft: directories, cookbooks
02:35 PM roycroft: i used to write software documentation
02:35 PM roycroft: i never had a great need for mathematical formulae
02:35 PM Tom_L: and i thought wordperfect was the shizzle back then
02:35 PM roycroft: i never used it much
02:35 PM roycroft: although i did have a xenix version of wp for a while
02:36 PM roycroft: and actually, i had a linux version of it during the brief period of time they supporte linux
02:36 PM Tom_L: and all the paper keyboard layouts...
02:37 PM roycroft: but i had years and years of experience doing composition -> editing -> formatting as three distinct sequences
02:37 PM roycroft: an all-encompassing word processer that did it all at the same time didn't make much sense to me
02:37 PM roycroft: when i'm writing i don't think about font size or exactly where the illustration fits on the page
02:38 PM roycroft: i think about expressing ideas in words
02:39 PM Thorhian: I use latex and emacs org mode usually these days. One of my classes wanted a .docx file, not even a pdf for submission. I was miffed. Turns out org mode has an ODT converter, so I sent it to openoffice and converted it to .docx.
02:39 PM Thorhian: Doom Emacs is my editor of choice. Evil Mode (Vim mode) and good configurations :D
02:40 PM roycroft: the nice thing about open document standards is that i can use a workflow and set of tools that works well for me, and you can use a workflow and set of tools that works well for you, and we can still exchange information without forcing one to conform to the other's methods
02:48 PM Thorhian: Stuff like Word and LibreWriter just seem so bloated for what they do, along with trying to make sure you can do everything and all of that needs some gui. I much prefer working with plain text and jumping around via Vim.
02:52 PM unterhaus: I like latex, but it's a lot of firepower. It's such a pain if you don't use it all the time though
02:53 PM unterhaus: there was a dos tex clone I bought for my masters thesis. didn't have latex, so the file was really raw code
02:53 PM unterhaus: looked nice though, except for being printed on a dot matrix
03:05 PM perry_j1987: hows it going guys
03:17 PM unterhaus: ha, booted myself
03:19 PM unterhaus: anyone know about the 50 pin to 25 pin converter for mesa?
03:22 PM unterhaus: okay, its the HD-DB ADAPT
03:23 PM perry_j1987: trying to figure out what could possibly be in this gcode that made this cut .22mm longer in Z than it should have been
03:24 PM perry_j1987: tested the z movement numerous times/ways and seems to be moving dead on now
03:24 PM unterhaus: that suggests a reference problem to me
03:24 PM perry_j1987: reference?
03:24 PM unterhaus: touch off, tool length, summen like that
03:25 PM perry_j1987: x is dead on, z it faces and then makes a groove a little ways in
03:25 PM perry_j1987: with the same tool
03:25 PM perry_j1987: so shouldnt possibly be reference issue
03:26 PM gregcnc: grooving tool?
03:26 PM perry_j1987: ya
03:27 PM gregcnc: tool width correct?
03:27 PM perry_j1987: it should be
03:27 PM gregcnc: "should"
03:28 PM perry_j1987: https://www.amazon.com/gp/product/B0824QCB56/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1 one of these
03:28 PM perry_j1987: unless the inserts wildely vary in sizes
03:29 PM gregcnc: is the front facing cut at Z0. or Z2.
03:30 PM perry_j1987: it faces and then moves over and then makes a groove
03:31 PM perry_j1987: i do a few plunges with the grove tool and then a finish pass to smooth it all out and leave a nice finish
03:31 PM CaptHindsight: perry_j1987: what ended up being the problem with the switch inputs to the 2nd BOB?
03:31 PM perry_j1987: bad BOB
03:31 PM perry_j1987: second one diff make in a row lol
03:31 PM CaptHindsight: 3rd BOB's a charm?
03:31 PM gregcnc: buy 3, 1 will work
03:31 PM perry_j1987: put one of my known good BOBS from the old lathe on and works fine
03:32 PM perry_j1987: tomorrow a replacement arrives they exchanged it so i'll try that one
03:32 PM perry_j1987: what sucks is the BOB on there now doesnt have the 0-10v
03:32 PM gregcnc: did you measure the tool?
03:32 PM perry_j1987: so using a dc-dc board to give voltage for the VFD atm
03:34 PM perry_j1987: ok im looking at the tool table in fusion360
03:34 PM perry_j1987: it has 1.89 mm for the width..
03:34 PM perry_j1987: they are suppose to be 2mm
03:34 PM gregcnc: oops
03:35 PM perry_j1987: ah groove width set to 1.9
03:35 PM perry_j1987: and 1.89 is in there for the holder
03:36 PM perry_j1987: but that doesnt line up
03:36 PM perry_j1987: if the gcode thinks the tool is 1.9 instead of 2
03:36 PM perry_j1987: it'd make a smaller grove vs one thats .22 over
03:36 PM gregcnc: not how that works
03:37 PM perry_j1987: ohhh wait im getting there
03:37 PM perry_j1987: visualizing it now
03:37 PM perry_j1987: the actual tool was wider than what it thought it was making it cut more
03:43 PM perry_j1987: https://amzn.to/3o3FVaV are these better or worse groove tools
03:44 PM gregcnc: what problem are you having?
03:44 PM Thorhian: Hey gregcnc, I decided to practice calculating power requirement estimates using the info in my machinery's handbook.
03:44 PM Thorhian: https://www.shars.com/products/indexable-cutting/indexable-milling/bt30-25mm-modular-90-degree-indexable-end-mill-apkt-insert
03:44 PM perry_j1987: looks like this has a little more reach
03:46 PM Thorhian: I calculated around 1.3kW for facemilling hard chromoloy steel at the full diameter of the tool with a 1mm depth of cut.
03:46 PM Thorhian: I figured that would be well within my spindle's power constraints (1.8kW).
03:47 PM unterhaus: Are you sure the length is consistent? I am not a big fan of those tools since there is barely any registration
03:48 PM unterhaus: ^talking about the lathe groove tool
03:49 PM perry_j1987: no clue
03:49 PM perry_j1987: looks like its similar to the one i got just registers against back wall and clamp it down
03:50 PM gregcnc: Thorian sounds pretty good, the only other thing to consider is whether the machine makes the torque required at the RPM you've chosen
03:52 PM gregcnc: if it's running at rated frequency, this isn't a concern
03:53 PM Thorhian: That I don't know how to do. It should be running within it's rated frequency. Spindle should be going at 1941 RPM. With the pulley situation it can get up to 4700 RPM max.
03:55 PM unterhaus: what kind of spindle motor?
03:57 PM Tom_L: 35hp brute force monster
03:57 PM Tom_L: the lights dim when it starts
03:57 PM Thorhian: AC servo motor. 3 Phase, 240V. Supposedly has a holding torque of 6NM but I don't know about when it's spinning.
03:58 PM Thorhian: Only 1.8KW.
03:58 PM Thorhian: Maximum RPM is 3000RPM.
03:59 PM Tom_L: for a mill?
03:59 PM gregcnc: oh right do you have a link
04:00 PM sinned6915: has anyone tried using a magnetic encoder like an AS5047 or AS5147 to sense shaft rotation.
04:00 PM sinned6915: https://ams.com/as5047d
04:00 PM Thorhian: https://www.ebay.com/itm/184285171784
04:01 PM Tom_L: sinned6915, several here have used hall sensors
04:01 PM sinned6915: that one in particular?
04:02 PM Tom_L: not that i'm aware of
04:02 PM sinned6915: that one is suppost to be a drop in replacment for optical shaft encoder
04:05 PM CaptHindsight: I use their AS5311 for linear measurement
04:06 PM CaptHindsight: they have several for rotary measurement like the 5047
04:08 PM gregcnc: 6Nm at 3000 is rated power of 1.8kW. It says it will do 18Nm so you have a lot of headroom, but what you get will depend on drive settings
04:09 PM Thorhian: Sounds good. For 1941 RPM the drive will be running at 1220 RPM
04:09 PM CaptHindsight: sinned6915: they mostly vary by type of interface, output types and bit resolution
04:09 PM gregcnc: So you can figure torque available from your ratios
04:10 PM gregcnc: for that type of motor you have full torque from 0 to rated RPM
04:10 PM Thorhian: Oh okay.
04:11 PM Thorhian: So the step down for torque would be 3.77 NM at the standard 6NM.
04:12 PM Thorhian: Is there a good way of calculating needed torque based on the power required?
04:13 PM gregcnc: https://www.engineeringtoolbox.com/work-torque-d_1377.html
04:24 PM unterhaus: wow, the national instruments website is horrible
04:25 PM unterhaus: they somehow decided I was in the uk
04:25 PM gregcnc: did it go to a phone style site like most fast food places?
04:57 PM perry_j1987: i dunno about this whole servo thing lol
04:58 PM perry_j1987: just fired up the machine and reran the same gcode and now its off in both x and z heh
05:02 PM CaptHindsight: perry_j1987: isn't your setup just like a stepper with just Step and Dir?
05:02 PM perry_j1987: ya
05:03 PM CaptHindsight: you should not have to do any LCNC servo tuning
05:03 PM perry_j1987: im just complaining cause things are not working and im not sure whats going on yet heh
05:04 PM CaptHindsight: do your X, Y and Z encoders only go back to the motor driver?
05:05 PM CaptHindsight: are you losing steps in X and Z?
05:05 PM CaptHindsight: be conservative with the accel until you know all the scaling is correct
05:07 PM perry_j1987: ya only to the g320x
05:07 PM CaptHindsight: be sure that your max step rate is below what the machine can handle
05:08 PM CaptHindsight: max jog speed, max feed rates etc
05:09 PM CaptHindsight: this is all in the config file
05:09 PM perry_j1987: the jitter test gave me number around 15,000
05:09 PM perry_j1987: its a ga350n gigabyte board
05:09 PM perry_j1987: amd e350 proc with 16 gigs of ram
05:11 PM perry_j1987: stepconf had dropdown option for gecko 320 that i used
05:11 PM perry_j1987: 3500 step time, step space 500
05:11 PM CaptHindsight: that gives you a starting config
05:12 PM perry_j1987: i dont mind the thing being set conservitively
05:12 PM CaptHindsight: you need to set all the Max_Velocity, Max_Acceleration, Max for jog etc etc
05:12 PM perry_j1987: i'd rather reliable operation
05:12 PM perry_j1987: those are set to 25 and 750
05:12 PM CaptHindsight: by default those are probably too high right now
05:14 PM CaptHindsight: "just fired up the machine and reran the same gcode and now its off in both x and z "
05:14 PM CaptHindsight: How is it off?
05:15 PM CaptHindsight: you need to find out how and why and make the proper adjustments to the config settings
05:15 PM perry_j1987: i just lowered vel to 15 and accel to 250
05:17 PM CaptHindsight: they are in the Display section as well as each Axis section
05:21 PM gregcnc: How are you homing?
05:22 PM gregcnc: gecko 320 don't lose steps
05:26 PM perry_j1987: trying some new gcode
05:26 PM perry_j1987: i have microswitches
05:26 PM Tom_L: sure it's not the G90 G91 thing all over again?
05:27 PM perry_j1987: i think its something fusion360 did
05:28 PM Tom_L: no big surprise there
05:28 PM gregcnc: same code can't be different?
05:29 PM gregcnc: if you have space on the BOB, home to index
05:31 PM perry_j1987: ok so my old gcode works but its making the grove slightly too wide
05:31 PM perry_j1987: i could edit the gcode and fix that
05:31 PM perry_j1987: but i thought well i'll just fire up fusion360 and regenerate the gcode from the same thing after editing the proper tool width
05:32 PM perry_j1987: something happened and its a totaly different tool path generated now vs what i was using before
05:32 PM perry_j1987: fusion updates i guess..
05:32 PM gregcnc: lol
05:32 PM perry_j1987: going to have to spend some time mucking around with it figure out what they changed in CAM
05:32 PM CaptHindsight: also check for typos
05:33 PM perry_j1987: how often do i need to redownload the linuxcnc turning post processor
05:33 PM CaptHindsight: pesky decimal points in the wrong place etc etc
05:33 PM perry_j1987: as fusion is updating constantly etc
05:33 PM gregcnc: good luck
05:33 PM perry_j1987: well dont make me sound doomed gregcnc haha
05:33 PM gregcnc: that is the stupidest thing fusion does
05:34 PM Tom_L: i'd think a tool would regenerate the same tool path if nothing changed
05:34 PM Tom_L: what a piece of work...
05:34 PM gregcnc: 2 axis lathe isn't hard to hand code
05:34 PM Tom_L: no
05:34 PM perry_j1987: nothing i've ever done from scratch
05:34 PM Tom_L: we generally mdi'd most of the parts we did
05:35 PM Tom_L: i _do_ however have a lathe cam package, just never use it
05:35 PM CaptHindsight: If Fusion360 doesn't work for you then try AI :)
05:36 PM Tom_L: who is this AI guy anyway?
05:36 PM Tom_L: sure like to meet him
05:36 PM JT-Shop: WOW I'm done
05:36 PM CaptHindsight: ^5
05:36 PM Tom_L: yay!
05:37 PM Tom_L: now hurry up and get em installed... we got a deadline here ya know!
05:38 PM JT-Shop: take me about 5 minutes each to install\
05:38 PM Tom_L: you mean unless he tries to help?
05:39 PM JT-Shop: he has to help hold them in place while I screw them to to wall
05:39 PM unterhaus: yay, my 7i48 is on the way
05:39 PM JT-Shop: then I'm done, the electrician has to bring power in
05:40 PM Tom_L: JT-Shop, do you keep any of those single analog boards in stock?
05:40 PM Tom_L: for spindle
05:40 PM unterhaus: The SPIN?
05:40 PM Tom_L: yeah
05:41 PM Tom_L: just curious if they were popular
05:41 PM unterhaus: mesa has 55 spinx1
05:41 PM Tom_L: i know
05:42 PM JT-Shop: Tom_L, which one?
05:43 PM unterhaus: where is JT's mesa page?
05:43 PM Tom_L: mesaus
05:43 PM unterhaus: thx
05:43 PM Tom_L: JT-Shop, the spinx1 i think
05:43 PM JT-Shop: yea I have them
05:44 PM unterhaus: 3 in stock
05:44 PM unterhaus: doesn't have ency's though
05:44 PM JT-Shop: they are being built
06:03 PM perry_j1987: interesting
06:04 PM perry_j1987: there's too much movement in this cross slide
06:04 PM perry_j1987: lifting up on the apron it moves up off the ways
06:05 PM perry_j1987: the bottom of the casting is down far making the strip not able to contact the bottom of the ways
06:11 PM CaptHindsight: gravity assist cross slide ways
06:12 PM CaptHindsight: fine if the cross slide is >1,000lbs (450Kg)
06:13 PM perry_j1987: this isnt made out of neutron star for sure heh
06:13 PM perry_j1987: just poking around the thing, havnt found a dissassembly method yet
06:13 PM perry_j1987: hidden screws under the cross slide hold the strips on
06:14 PM CaptHindsight: so the ways just sort of provide a bit of linear constraint
06:14 PM CaptHindsight: like a suggestion
06:14 PM perry_j1987: ya
06:14 PM perry_j1987: im going to have to fix this
06:22 PM CaptHindsight: not much cutting force required for cheese, might be useful as a dairy lathe
06:22 PM perry_j1987: there's a setscrew on bottom of the ballscrew nut holder
06:22 PM perry_j1987: no flange on this ballscrew
06:23 PM perry_j1987: appears to just slide in this machined hole and held with setscrew
06:23 PM perry_j1987: doesnt budge with setscrew loose though heh
06:23 PM perry_j1987: hmm
06:27 PM Tom_L: it's not a double setscrew is it?
06:29 PM CaptHindsight: yeah sneaky one hiding under the top one
06:30 PM perry_j1987: ah didnt test that theory
06:30 PM perry_j1987: i'll check
06:30 PM perry_j1987: doesnt appear so
06:30 PM perry_j1987: not enough meat just pulled the setscrew out
06:31 PM CaptHindsight: sounds like you'll have some fixin to do, check for other surprises
06:32 PM perry_j1987: ya
06:32 PM perry_j1987: project just got bigger heh
06:32 PM Tom_L: you needed something to do
06:33 PM perry_j1987: ya but not this lol
06:34 PM perry_j1987: the screws that hold the entire bracket on the back are some magical size allen key that i dont have lol
06:36 PM perry_j1987: oh wait
06:36 PM perry_j1987: 3/16
06:36 PM perry_j1987: got it
06:49 PM asdfasd: anyone tried the rtai kernel for buster?
06:53 PM perry_j1987: tada
06:53 PM perry_j1987: got it all apart
06:54 PM Centurion-Dan2 is now known as Centurion_Dan
07:01 PM CaptHindsight: asdfasd: do you NEED the RTAI kernel?
07:01 PM CaptHindsight: perry_j1987: what was the secret to take the nut out?
07:02 PM perry_j1987: i left it in place heh
07:02 PM perry_j1987: took the servo off and then the bracket that held the servo and screw and cross slide
07:02 PM perry_j1987: took the gib out of cross slide and slid the whole thing off the back of the lathe hehe
07:02 PM perry_j1987: i got the bare casting for the cross slide here now
07:03 PM perry_j1987: there's little bit of paint on the spot where the rear gib bolts down
07:03 PM perry_j1987: and its very rough casting
07:03 PM perry_j1987: both the front and rear gib surfaces
07:04 PM perry_j1987: so when it was all together i could pull up on the apron and get some upwards movement on the cross slide with the gibs tightened as much as they could
07:04 PM perry_j1987: so im going to have to file down the casting a bit to give me some adjustment room it looks like
07:04 PM asdfasd: I have it installed, but it seems something is wrong
07:04 PM asdfasd: very poor latency
07:06 PM CaptHindsight: asdfasd: which processor?
07:06 PM asdfasd: intel E5800
07:06 PM CaptHindsight: asdfasd: did you install the LCNC RTAI package?
07:07 PM asdfasd: yes, as per instructions
07:07 PM asdfasd: first buster rt and then rtai
07:07 PM CaptHindsight: heh, instructions for what? there are all sorts for using the packages or for building it yourself
07:07 PM roycroft: you can't use a preempt kernel?
07:08 PM asdfasd: its not very good I think, compared with wheezy rtai
07:08 PM roycroft: and yay, mail arrived today for once
07:08 PM roycroft: but only the letters
07:08 PM asdfasd: as I need it for LPT control
07:09 PM roycroft: the 5 parcels that have been backlogged for delivery were all marked out for delivery, but did not arrive
07:09 PM CaptHindsight: asdfasd: preempt_rt can also work well on the LPT
07:09 PM asdfasd: jitter more than 10000 is it good for LPT?
07:10 PM CaptHindsight: asdfasd: if you are installing the packages then all you can do it try them and see which works best for you
07:10 PM CaptHindsight: how much more than 10,000?
07:10 PM asdfasd: some time jump to 15000
07:10 PM CaptHindsight: run the latency test and open a browser, play a youtube, start and stop some applications
07:11 PM CaptHindsight: 15,000 is really good
07:11 PM asdfasd: good? on my old machine wich is 10 years older, its 6000
07:11 PM gregcnc: 6000 is excellent
07:11 PM asdfasd: on the new machine with wheezy is around 5000
07:11 PM asdfasd: but buster go to 15000
07:11 PM roycroft: on a modern machine you're lucky if you even get a parallel port
07:12 PM asdfasd: whell, I call it new, but its not new actually haha
07:12 PM CaptHindsight: the RTAI kernel and LCNC have to be built against each other
07:12 PM roycroft: i would posit that in the amount of time you've spent trying to get this working you could have generated enough revenue to more than pay for a mesa card for your machine
07:12 PM roycroft: :)
07:13 PM CaptHindsight: any kernel can run on any Linux distro
07:13 PM gregcnc: really depends how fast you need to step. if you need 6000, 15,000 doesn't work
07:14 PM CaptHindsight: asdfasd: how many steps per second do you need to run your machine?
07:14 PM asdfasd: http://linuxcnc.org/docs/2.8/html/getting-started/getting-linuxcnc.html#cha:Installing-RTAI
07:14 PM CaptHindsight: you need to know this first
07:14 PM asdfasd: let me check
07:15 PM CaptHindsight: asdfasd: yeah that RTAI kernel is just experimental, RTAI is still broken
07:16 PM CaptHindsight: asdfasd: how fast do your rapids need to go? Then you work from there back though the screws, pulleys and motors to see what step rate is required
07:16 PM perry_j1987: gah sorry i just realized i kept saying cross slide gibs i meant saddle gibs
07:17 PM perry_j1987: so bottom of the saddle is too low so the gibs when fully tightended dont even make contact with the lathe bed
07:17 PM CaptHindsight: perry_j1987: why the deal on the lathe looked so good?
07:18 PM perry_j1987: its still a good deal heh
07:18 PM asdfasd: MAX_VELOCITY = 230.0
07:18 PM asdfasd: MAX_ACCELERATION = 800.0
07:18 PM asdfasd: STEPGEN_MAXACCEL = 1000.0
07:18 PM asdfasd: SCALE = 156.86274
07:18 PM perry_j1987: im having fun still
07:18 PM asdfasd: BASE_PERIOD = 26000
07:19 PM CaptHindsight: asdfasd: so what is that in Steps/second?
07:19 PM asdfasd: these are the current settings on the old computer
07:19 PM CaptHindsight: asdfasd: so using preempt_rt you can use the same settings
07:19 PM CaptHindsight: on the new computer
07:21 PM asdfasd: I cant remember how steps/second are calculated, searching now
07:23 PM asdfasd: I think it is a bit more than 30khz
07:24 PM asdfasd: may be up to 35 khz should be enough
07:25 PM asdfasd: yes I will try to use the same settings but the jitter is twice worst
07:25 PM asdfasd: where the old computer with 6000 is on its limits
07:26 PM asdfasd: on the new one is jumping to 15000ns
07:30 PM CaptHindsight: just change BASE_PERIOD = 26000 to BASE_PERIOD = 31000
07:31 PM asdfasd: yes I can tweak these, but Im worried about the jitter, as it is worst compared with 10 years older pc
07:31 PM roycroft: just feed 600 volts to the computer :)
07:31 PM roycroft: *vroom*
07:31 PM asdfasd: yeah, of course :)))
07:32 PM gregcnc: gosh I had to go back and look it's been so long since I did my parport
07:32 PM gregcnc: now I'm trying to understand if I even had a clue when I did it
07:33 PM roycroft: i need to pull my linuxcnc machine out and do some work on it - i just got some m2.5 screws and can finally install the wifi card in it
07:33 PM CaptHindsight: asdfasd: different kernels act differently on different processor
07:34 PM gregcnc: but I see the Gecko drives I used are fast
07:34 PM asdfasd: <CaptHindsight> the RTAI kernel and LCNC have to be built against each other
07:35 PM asdfasd: could you please explain that
07:35 PM asdfasd: is there better way than apt-get install ?
07:36 PM CaptHindsight: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?TweakingSoftwareStepGeneration
07:36 PM CaptHindsight: asdfasd: this is not for you, this for those that can build kernels and applications from scratch
07:37 PM CaptHindsight: asdfasd: installing packages is the BEST way for you
07:37 PM CaptHindsight: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?TweakingSoftwareStepGeneration
07:38 PM CaptHindsight: choose a BASE_PERIOD that is the sum of the longest timing requirement of your drive, and the worst case latency of your computer.
07:38 PM CaptHindsight: If you are running a Gecko with a 20uS hold time requirement, and your latency test said you have a maximum latency of 11uS, then if you set the BASE_PERIOD to 20+11 = 31uS (31000 nano-seconds in the ini file),
07:40 PM asdfasd: in that case the new PC wont do the job with debian buster
07:41 PM CaptHindsight: http://wiki.linuxcnc.org/uploads/StepTimingCalculatorPlus.ods
07:41 PM Centurion-Dan2 is now known as Centurion_Dan
07:41 PM perry_j1987: there we go
07:41 PM perry_j1987: fixed
07:41 PM perry_j1987: no longer any jossling sloshing around heh
07:42 PM CaptHindsight: asdfasd: why not?
07:42 PM asdfasd: because the old machine with 6000 jitter is on its limits with 26000 base period
07:43 PM gregcnc: what limit?
07:43 PM asdfasd: of maximum speed
07:44 PM gregcnc: due to drives that need long periods?
07:44 PM asdfasd: no, motors can go faster, but not the PC
07:45 PM gregcnc: with 6000 jitter base period could be much smaller it seems?
07:45 PM CaptHindsight: how did you compute the base thread for the older system?
07:46 PM asdfasd: I didnt, try-error
07:46 PM asdfasd: I did lower it until stop working
07:46 PM CaptHindsight: if max jitter was 6000 on the old system and 15000 on the new system then add 9000 to the base period for the new system
07:47 PM asdfasd: yeah, unfortunately
07:47 PM CaptHindsight: or buy a Mesa card and step even faster than you did before
07:47 PM CaptHindsight: or get a different PC with lower latency
07:48 PM asdfasd: I got few more PCs but trying to understand where the problems are and how to test properly
07:48 PM asdfasd: so when I plug the other PC I need to know what and how to test
07:49 PM gregcnc: that's the time vs. cost of a card comes in
07:49 PM roycroft: honestly, if you can just get a mesa 5i25 and don't even think about latency ever again
07:50 PM asdfasd: well... it wasnt a problem before
07:51 PM roycroft: or be stubborn ...
07:51 PM roycroft: :)
07:51 PM asdfasd: it become a problem with new computers
07:51 PM asdfasd: Im stubborn haha
07:51 PM roycroft: and two of us have told you repeatedly how to make it not a problem
07:51 PM gregcnc: you got lucky with a 6000 jitter few machines are that fast
07:52 PM asdfasd: 6000 is on so old computer that cant even boot from usb
07:52 PM gregcnc: sure
07:52 PM roycroft: i tend to be long-winded, but i'll to keep this story short
07:53 PM roycroft: a few years ago i wanted to build a computer to host a linuxcnc installation
07:53 PM roycroft: i did a bunch of research
07:53 PM roycroft: ended up getting some good information from this very channel on irc about an ibm system board that had low latency and worked well for linuxcnc
07:53 PM roycroft: i bought that board
07:54 PM roycroft: i don't recall the exact numbers i got on the jitter test, but they were really good
07:54 PM roycroft: shortly after that, the board was discontinued
07:54 PM gregcnc: beating the latency game is a challenge....
07:54 PM roycroft: i had a brand new one, and i had good numbers
07:54 PM roycroft: but i realised that i was lucky to have gotten a good one
07:54 PM roycroft: and decided to not worry about it and bought a 5i25
07:55 PM roycroft: which is installed on that same board
07:55 PM roycroft: but i am confident that when that machine dies, i can get any random system board, plug the 5i25 into it, and go happly on my way
07:55 PM roycroft: happily
07:56 PM roycroft: i just don't see the point of trying to use a legacy parallel port for real time applications any more
07:56 PM asdfasd: the funny part is that the old computer work perfectly fine, and I really dont want to change it
07:56 PM roycroft: it's all but impossible to find a printer with a parallel port these days
07:56 PM roycroft: no, it's not funny
07:56 PM roycroft: it's perfectly logical
07:56 PM asdfasd: I just decided that wheezy is too old and have to upgrade
07:57 PM asdfasd: and because few pcs are going in the bin soon, decided to pick one for the upgrade
07:57 PM roycroft: when your older system board was made, a significant number of people still used parallel ports for printers
07:57 PM roycroft: now it's almost impossible to find a printer with a parallel port
07:58 PM roycroft: as i said earlier, if you buy a brand new system board today you'll be lucky to have a parallel port on it
07:58 PM asdfasd: I know
07:58 PM asdfasd: that is the reason I kept few old PCs until now
07:59 PM roycroft: then why are you surprised that your legacy hardware supports a deprecated interface better than newer hardware?
07:59 PM asdfasd: no, Im not surprised, Im just trying to find what can be done to be improved
08:00 PM asdfasd: ah and forgot to mention - wheezy on the new machine work just perfect - 5000
08:00 PM roycroft: the solution is to eliminate the legacy interface, and that is a ~$100 solution
08:00 PM asdfasd: may be
08:00 PM gregcnc: you'll make millions by maximizing rapids
08:00 PM asdfasd: will see
08:01 PM roycroft: you've easily spent $100 worth of the channel's collective time discussing it already :)
08:01 PM asdfasd: really?
08:01 PM asdfasd: I have not seen these millions haha
08:01 PM gregcnc: oh course you be done before you start
08:01 PM roycroft: no, of course you haven't
08:01 PM roycroft: because you're still using your parrallel port
08:01 PM asdfasd: haha, logix
08:01 PM roycroft: that is the bottleneck that is keeping the money tree from raining on your head
08:01 PM asdfasd: logic
08:02 PM asdfasd: of course haha
08:02 PM gregcnc: the machine will print $$ faster the the fed
08:02 PM asdfasd: ok, sorry guys, wont waste your time anymore
08:02 PM roycroft: you're not wasting our time
08:02 PM roycroft: *if* you learn somethign
08:02 PM gregcnc: the time is wasted no matter who is here
08:02 PM asdfasd: keep making money
08:02 PM roycroft: time is relative
08:02 PM asdfasd: Im not on that path currently
08:03 PM roycroft: sadly, my relatives are mostly pretty cheap
08:03 PM gregcnc: time never waits
08:03 PM roycroft: and i'm not
08:03 PM roycroft: i'm frugal
08:03 PM roycroft: which is way different to cheap
08:04 PM roycroft: full disclosure: pcw is a friend of this channel, and is mesanet
08:04 PM roycroft: jt is also a friend of this channel, and is a mesanet reseller
08:05 PM asdfasd: ok guys, thanks for your help, its bed time here
08:05 PM asdfasd: all the best
08:05 PM roycroft: that said, i still strongly recommend the mesanet hardware, relationships with those two notwithstanding
08:05 PM roycroft: it just ends your worries quickly and cheaply
08:06 PM asdfasd: I know, but my question was about jitter
08:06 PM asdfasd: anyway
08:06 PM asdfasd: thanks
08:06 PM gregcnc: bios settings, multiple cores, etc. sometimes you can fix it other times you can't
08:20 PM roycroft: looks like he got too jittery
08:23 PM perry_j1987: hmm one of the servos is making a noise
08:24 PM perry_j1987: like errrrrhhmmmmmm .... errrrrrrhhhmmmmm ... errrrrrhhhmmmm
08:24 PM perry_j1987: repeating
08:24 PM roycroft: get better earplugs :)
08:24 PM perry_j1987: about 2seconds long
08:35 PM perry_j1987: the pots are identical on that driver vs the other driver same servo
08:35 PM perry_j1987: hmm
08:37 PM Tom_L: what a waste of time :/
08:38 PM Tom_L: roycroft,
08:38 PM Tom_L: he's wanting to upgrade because of windows 10
08:39 PM Tom_L: file sharing
09:13 PM perry_j1987: ok that servo that is making the noise all of a sudden
09:13 PM perry_j1987: is warm
09:13 PM perry_j1987: the other servo on z is cool
09:13 PM perry_j1987: it wasnt doing this earlier today
09:16 PM roycroft: sounds like time to swap in a spare
09:18 PM flyback: you can get pci-e parallel ports
09:18 PM flyback: the only ting that might not work is dma which almost nothing ever used anyways
09:18 PM perry_j1987: i dont have a spare
09:19 PM flyback: perry_j1987, time to scrap it since you can't re-bearing steppers or servos without weaking the magnetic field
09:19 PM flyback: whcih is a really stupid design btw
09:19 PM roycroft: time to order one
09:19 PM flyback: whoever came up with this design limitation should be shot DEAD
09:20 PM perry_j1987: its functioning
09:20 PM roycroft: murder/execution is never appropriate
09:20 PM perry_j1987: just when its idle sitting there its making a humm reapeating
09:20 PM flyback: roycroft, yes it is
09:20 PM roycroft: for now, perry_j1987
09:20 PM roycroft: i'd order a spare
09:20 PM roycroft: it's good to have one on hand
09:20 PM perry_j1987: aye
09:21 PM roycroft: and that servo may fail at any moment
09:21 PM roycroft: i'm a big advocate of sparing
09:23 PM perry_j1987: i took that servo off when i was fixing the gibs
09:24 PM perry_j1987: put it back on and its making the noise
09:46 PM sorki is now known as srk
10:47 PM mrec: linuxcnc becomes scary..
10:47 PM mrec: https://imgur.com/a/UgWBJRK
10:51 PM mrec: that issue must be derived from some preceding calculation
10:52 PM W1N9Zr0: do your axis limits match your joint limits?
10:53 PM mrec: there are accuracy issues in linuxcnc
10:53 PM mrec: double has an accuracy of around 14/15 digits
10:53 PM mrec: everything below it is not guaranteed (Depending on the architecture)
10:53 PM mrec: and linuxcnc is not taking care about that
10:54 PM mrec: it's a bug
10:54 PM mrec: but to make it worse I think all the linuxcnc code has to be checked for it..
10:55 PM mrec: I have found 3 places in the code where it happens (and yes it happens to my setup unfortunately)
10:56 PM W1N9Zr0: before you blame it on bugs, what are your axis and joint limits? i've never seen this happen on my machines
10:56 PM W1N9Zr0: post your ini and hal
10:57 PM mrec: 80 of course, it's a code bug I fixed it on my side with updates.
10:57 PM mrec: just read about the double accuracy and you'll see.
10:59 PM W1N9Zr0: i know about double accuracy thank you very much, but the chances of you finding this bug in a 10+ year old project that nobody else ran into before are slim
11:00 PM W1N9Zr0: your accelerations might be misconfigured and your axis cannot slow down before you run past the axis limit
11:01 PM mrec: it's an open loop setup
11:01 PM mrec: the physical setup should be irrelevant
11:02 PM mrec: open loop with closed loop servos as their own unit but without feedback to emc/linuxcnc
11:02 PM W1N9Zr0: did you read this part:
11:02 PM W1N9Zr0: STEPGEN_MAXACCEL = 21.0 - Acceleration limit for the step generator.This should be 1% to 10%larger than the joint MAX_ACCELERATION. This value improves the tuningof stepgen’s "position loop". If you have added backlash compensationto an joint then this should be 1.5 to 2 times greater thanMAX_ACCELERATION.
11:03 PM mrec: I did not set up backlash compensation
11:04 PM W1N9Zr0: if your STEPGEN_MAXACCEL is less than or equal to MAX_ACCELERATION, this difference will make your open loop axis decelerate slower than motion planner expects, and you will overrun your soft limits
11:04 PM mrec: in any case the values in the screenshot are part of the double inaccuracy - not guaranteed to be correct
11:06 PM mrec: http://paste.debian.net/1181858/
11:07 PM mrec: MAX_ACCELERATION = 1500.0
11:07 PM mrec: STEPGEN_MAXACCEL = 1875.0
11:08 PM mrec: and just look how small the value in the screenshot is, for sanity this shouldn't happen.
11:11 PM mrec: that mentioned the machine is small I'm really using the limits on a bigger machine I wouldn't have to go that far
11:11 PM mrec: I guess the linuxcnc userbase is not really that big for small machines
11:12 PM CaptHindsight: most are over 4ft tall
11:12 PM CaptHindsight: ~1.3m
11:12 PM roycroft: are you saying that hobbittses do not like to machine?
11:13 PM mrec: few hundred mini mills maybe.
11:14 PM mrec: roycroft: I just segment the userbase, and I'm hitting the limit with a small machine
11:14 PM mrec: other users probably have bigger machines
11:14 PM CaptHindsight: well 20 years ago EMC for short people didn't sell well
11:14 PM mrec: and don't depend on the limit
11:14 PM roycroft: the size of the appropriate machine is more a function of the parts than the operator, no?
11:15 PM CaptHindsight: I have LCNC on some machines with only 1cm of travel
11:15 PM mrec: but most probably don't have to retract to the max limit (also it takes time on a bigger machine to fully retract the head)
11:16 PM mrec: I'd prefer to not retract completely, but I have a rotary head installed at the moment so I need to
11:16 PM CaptHindsight: small machine avg size user
11:17 PM CaptHindsight: max velocity 240m/min
11:18 PM CaptHindsight: 4m/sec pullin some G's there
11:18 PM roycroft: but who uses lego mill? chipmunks?
11:18 PM roycroft: oh, i know
11:18 PM roycroft: the brain does
11:20 PM mrec: for small parts it makes sense :-)
11:23 PM CaptHindsight: BASE_PERIOD = 45000
11:26 PM mrec: it should be around 25k some old settings.
11:26 PM mrec: possibly more like 15k since I'm using isolcpus
11:28 PM mrec: the error value is on digit #14
11:28 PM mrec: the value is ridiculous small
11:29 PM CaptHindsight: how fast the machine really move?
11:29 PM CaptHindsight: does the
11:29 PM mrec: it might be some derived issue
11:29 PM mrec: does it matter in the open loop setup?
11:29 PM CaptHindsight: yes
11:29 PM mrec: the moving part is all fine
11:30 PM mrec: https://www.jmc-motor.com/product/901.html
11:30 PM CaptHindsight: kewl
11:30 PM mrec: I'm using those motors
11:31 PM mrec: anyway it's irrelevant. I guess you mean the settings?
11:32 PM CaptHindsight: no I was wondering about the actual velocity of the machine
11:32 PM W1N9Zr0: if this was a float issue, wouldn't people see it all the time with axis configured with a limit near 0? that's a pretty common config
11:33 PM CaptHindsight: sure, you can reproduce his settings and check
11:34 PM CaptHindsight: mrec: where did you post your fixes/patches?
11:34 PM mrec: I only posted one but since there are multiple places it should be "reworked"
11:35 PM mrec: https://github.com/LinuxCNC/linuxcnc/pull/1047
11:35 PM mrec: is there any "standard" accuracy that linuxcnc is known to work with? eg. 5 digits, 6 digits?
11:37 PM mrec: Exceed NEGATIVE soft limit %.5f not very helpful if the value is smaller. (control.c)
11:37 PM mrec: and it's far smaller.
11:37 PM CaptHindsight: ask in #linuxcnc-devel
11:38 PM mrec: by the way I'm happy with linuxcnc I'm not criticizing it :-)
11:38 PM CaptHindsight: Andy and Dewey have looked at it
11:39 PM CaptHindsight: I haven't run into the issue and I avoid coding
11:39 PM W1N9Zr0: note that the limits are exclusive, not inclusive. so if you want your machine to accept commands to move to 80, your limit should be >80
11:39 PM W1N9Zr0: most configs i've seen have limits set to something like 80.001 for this reason
11:40 PM mrec: W1N9Zr0: is that documented somewhere and not your personal statement?
11:40 PM mrec: actually it's inclusive the checks show that
11:40 PM mrec: foo > limit -> throw error
11:40 PM mrec: it's not foo >= limit -> throw error
11:41 PM mrec: just the bug has made your statement a habbit
11:43 PM CaptHindsight: mrec: or post on the forums https://forum.linuxcnc.org/
11:43 PM mrec: the developer has seen it I think it should be fine
11:43 PM CaptHindsight: the devs seem to be aware, they just aren't in here
11:47 PM CaptHindsight: https://imgur.com/b4zPwKo I used a similar motor here
11:48 PM CaptHindsight: though that pic was taken with a different motor and drive
11:51 PM mrec: I guess I will sell my mill once I'm ready with the first parts and get a bigger one. The small one is perfect for learning and getting into it
11:51 PM W1N9Zr0: i guess it's not documented, but it is consistently an exclusive limit
11:52 PM W1N9Zr0: you're only getting the joint error because you patched the axis limit and not the joint limit, right?
11:52 PM mrec: W1N9Zr0: I have applied the patch at 3 places so far and it's fine now
11:52 PM mrec: the screenshot was the last one.
11:53 PM mrec: the checks are clear.. everything outside the limit should throw the error
11:54 PM mrec: I guess the limit check should be the only obvious critical part where accuracy matters
11:54 PM mrec: but the double issues could also be in other areas in the code..
11:55 PM mrec: and no real machine would notice such small "errors"
11:58 PM W1N9Zr0: it's probably an issue on all machines where float parsing rounds the numbers bigger, just people assumed it's an exclusive limit and have been working around it for years by adding 0.001mm to their limits :P
11:59 PM CaptHindsight: sub angstrom position errors