#linuxcnc Logs
Jan 03 2023
#linuxcnc Calendar
12:13 AM -!- #linuxcnc mode set to +v by ChanServ
12:27 AM isaacggw[m]: This is on a pi4
12:40 AM solarwind: > This is Yolanda from STEPPERONLINE and thank you for contacting us.
12:40 AM solarwind: Really? Is your name _really_ Yolanda?
12:42 AM solarwind: Moving on, nice to see that they can customize the servo and swap the magnetic encoder for an optical encoder
12:42 AM sensille: not "Jane"? half of chinese companies have Jane in support
12:43 AM solarwind: There was a really old George Carlin show and one of his random jokes was "Not too many Chinese guys named Rusty"
12:43 AM solarwind: That was from 1980 or so. I find that to still be true. Haven't met a Chinese guy named Rusty yet.
12:44 AM sensille: but have you met jane?
12:45 AM roycroft: i've run into a number of support folks named "cindy"
01:25 AM Deejay: moin
02:28 AM randy: morning
04:01 AM He[m]1: What kind of machine are you running? Steppers/servo mesa card?
04:05 AM He[m]1: Morning btw
04:27 AM Tom_L: morning
04:39 AM -!- #linuxcnc mode set to +v by ChanServ
04:39 AM Tom_L: isaacggw[m], http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Following_Error
04:54 AM pere: Tom_L: the wiki page seem to lack the fe2.png graph.
05:02 AM JT-Cave: morning
05:38 AM He[m]1: Well, still got the 5ms lag between step pulse being generated and linuxcnc getting the encoder pulse.
05:38 AM He[m]1: dang
05:38 AM He[m]1: Guess I'll try the other drives and see if they have the same lag
05:41 AM He[m]1: Then I guess I'll try hooking one of the drives up to the +-10V output of the 7i76 and see if lag decreases at all
06:29 AM Tom_L: JT-Cave, new debs are up
06:29 AM JT-Cave: thanks
06:30 AM Tom_L: updated the build date/version
06:43 AM jpa-: He[m]1: have you calculated the expected delay from motor inertial load and torque?
06:48 AM jpa-: though i guess 5 ms is unrealistically long for any practical setup - probably poorly tuned PID on the driver then
07:09 AM He[m]1: Inertial load should be very low as it's just a motor that's not connected to anything yet
07:12 AM He[m]1: The scope inside the drive seems to follow the commanded velocity/position really well too
07:19 AM * travis_farmer[m] wonders how to display spindle temperature on qtdragon_hd, or if a remote Arduino powered display would be better...
07:45 AM Tom_L: JT-Cave, did you get that error i posted?
07:56 AM JT-Cave: was that from trying to view a pdf from the info tab?
08:02 AM Tom_L: i think so yes
08:23 AM JT-Cave: I deleted that tab this morning
08:29 AM Tom_L: no more links to doc.pdfs?
08:29 AM Tom_L: maybe just as well
08:30 AM JT-Cave: in the menu under Downloads
08:30 AM Tom_L: i'll check later on
08:31 AM JT-Cave: now I'm hell bent on reducing the gui down to 800 x 600 so it fits on my rpi screen
08:31 AM Tom_L: heh
08:32 AM Tom_L: hope nothing gets lost in the move
08:34 AM Tom_L: bbl
09:04 AM solarwind: Did you guys see the new EL8 series servo drives from Leadshine? https://www.leadshine.com/upfiles/downloads/26a49980c0b3d63cfe0cd8cb597d83fd_1670408956734.pdf
09:04 AM solarwind: Ethercat + analog input, can accept second encoder like a glass scale
09:04 AM roguish[m]: good morning. a bit chilly here. about 40F, and cloudy. tomorrow big storm
09:25 AM solarwind: SimpleFOC is an excellent project https://docs.simplefoc.com/position_sensors
09:25 AM solarwind: A wealth of information + actual working code there
09:25 AM solarwind: and it works _really_ well
09:29 AM markgonz: hello everyone
09:30 AM markgonz: I have linuxcnc 2.9.0, I am trying to make a pncconfig for a mesa 7i96s
09:31 AM markgonz: however the board doesn't appear on the list, closest is 7i96
09:35 AM JT-Shop_: https://gnipsel.com/linuxcnc/mesact/index.html
09:42 AM markgonz: thank you I will look into it
10:11 AM isaacggw[m]: I am running a raspberry pi four connected to a 7i96s
10:37 AM travis_farmer[m]: i think i have gotten too OCD with my HAL file... 441 lines, so i could create various lockouts for safety. like for instance, the manual tool release button on the spindle, when pressed, the spindle must be stopped (for obvious reasons), and works on a toggle HAL component, so i press once to release, and again to lock the next tool in. if nothing else, i think i am learning a lot about HAL. :-)
10:45 AM He[m]1: I have read somewhere that you might want to increase the servo thread to 2ms on a pi, let's see if I can find the thread
10:46 AM sensille: travis_farmer[m]: did you spindle already arrive?
10:47 AM ZincBoy[CAON][m]: Wow that is a small hal file! I have 1266 lines in my hal files for my mill 😛
10:48 AM travis_farmer[m]: sensille: not for a while, but i am working on getting the HAL file ready, as well as updating the schematic so i know how it should all connect up
10:48 AM travis_farmer[m]: ZincBoy[CA, ON]: it's not the size, it is how you use it ;-)
10:48 AM ZincBoy[CAON][m]: lol
10:49 AM sensille: comments and empty lines don't count
10:50 AM travis_farmer[m]: sensille: :-P
10:50 AM travis_farmer[m]: ;-)
10:50 AM JT-Cave: He[m]1, https://gnipsel.com/linuxcnc/rpi4-11.html
10:50 AM JT-Cave: https://gnipsel.com/linuxcnc/rpi4-11-emc.html
10:51 AM JT-Cave: latency +- 50us
10:51 AM ZincBoy[CAON][m]: I put all of the logic in custom comp files and that shrunk down the hal file size quite a bit. I actually ran out of HAL memory at one point and had to optimize things.
10:53 AM travis_farmer[m]: Hmm, there is an idea, it would clean things up a lot if i put all the logic in a comp... :-)
10:53 AM ZincBoy[CAON][m]: It is much easier to implement logic in the comp file. Doing it HAL is painful.
10:53 AM travis_farmer[m]: yeah, and makes the hal file harder to read...
10:57 AM travis_farmer[m]: well, before i make the logic into a comp file, i will wait for my spindle to show up and be connected, before i break my idea of how it should work when converting... best to have something to test with ;-)
10:59 AM travis_farmer[m]: oh, hey, look at the time... gotta make some lunch, before my fatness starves to death ;-)
11:38 AM skunkworks[m]: Jt-Cave: what image are you using for the pi now?
11:46 AM sensille: i just thought it would be nice to mount a 3d probe permanently to the spindle, lowered by some actuator. does anybody do that? (no toolchanger)
11:51 AM skunkworks[m]: sensille it would have to be very accurate.. certainly do-able. The offset vs spindle could be in the tool table..
11:53 AM sensille: changing the collets back and forth just to mount the probe is quite the effort
11:55 AM travis_farmer[m]: anyone know what the input type is on qtdragon_hd for HAL input "qtdragon.spindle-fault"? just wondering if i can use it to display spindle temperature in the GUI field?
11:55 AM sensille: even a manual switch to push it down would be fine
11:55 AM sensille: or a kinematc mount :)
11:57 AM He[m]1: that somehow reminded me to do the no coalesce thing for the intel LAN chip and uh, it appears that the lag for the encoder feedback is now 2ms instead of 5-6ms
11:58 AM * He[m]1 uploaded an image: (93KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/jKFoGCMUTINaZwynFDrcFuPm/image.png >
11:58 AM He[m]1: Also set the netmask to 255.255.255.0 as suggested
11:59 AM He[m]1: I haven't a clue why that made things better but it seems to have done so regardless
12:01 PM He[m]1: What is an acceptable latency value for that anyway?
12:07 PM CaptHindsight[m]: He²: well under 1mS is the goal
12:08 PM CaptHindsight[m]: For the servo thread latency
12:20 PM He[m]1: You mean like jitter in the latency test?
12:58 PM sensille: i'd like to integrate a probe screen into axis. is verser-git/probe_screen_v2 the way to go?
12:58 PM sensille: and what to use for the tool length sensor?
01:01 PM boozeman[m]: Hi, Mesact 1.2.1 version and build date on main window is version 1.2.0 12/31/2022 🙂
01:09 PM travis_farmer[m]: so, went out to my shop to run my HAL file changes... had a few errors... and then i noticed that encoder inputs, even if not used, are still encoder inputs... in other words, i can't use encoders 04 and 05, and leave the I/O that resides on the lover encoder inputs, just because encoders 04 and 05 were convenient pins... so i have to change my HAL file and move the I/O off the lower encoder inputs so i can use those, and change my
01:09 PM travis_farmer[m]: hardware to match... PITA, but likely my own fault, thinking i would never use encoder inputs...
01:11 PM bjorkintosh: Is there a freely available database of tools and insert images?
01:15 PM zephyr9900: travis_farmer I think you can still use un-used encoders for gpio
01:16 PM travis_farmer[m]: no, i got HAL errors
01:17 PM ZincBoy[CAON][m]: Not without changes to the FPGA code AFAIK
01:17 PM zephyr9900: what error?
01:19 PM travis_farmer[m]: "./my_LinuxCNC_router.hal:125: parameter or pin 'hm2_7i80.0.gpio.003.invert_output' not found" the pin is not found as it belongs to a encoder pin
01:20 PM zephyr9900: hmm did you set .is_output TRUE?
01:23 PM skunkworks[m]: you have all the encoders enabled.. You need to tell the card you only want a few..
01:24 PM travis_farmer[m]: yes, i know that.... now
01:24 PM travis_farmer[m]: ;-)
01:26 PM skunkworks[m]: loadrt hm2_eth board_ip="10.10.10.10" config="num_encoders=1 num_pwmgens=1 num_stepgens=3 sserial_port_0=0xxxxx"
01:26 PM skunkworks[m]: something like that..
01:31 PM Tom_L: boozeman[m], that has been fixed
01:33 PM travis_farmer[m]: hmmm, if i keep using up I/O pins, i may have to pair up my first 7i80 with my second one... (kidding, hope i don't get to that point) :-)
01:33 PM Tom_L: sserial
01:33 PM travis_farmer[m]: ?
01:34 PM Tom_L: for more io
01:34 PM Tom_L: but my god who needs 144 io?
01:34 PM travis_farmer[m]: Hmmm, interesting idea, but i hope i don't need that much I/O LOL
01:35 PM Tom_L: i ended up using a few off the 2nd header
01:35 PM travis_farmer[m]: using all 3 headers on mine
01:36 PM ZincBoy[CAON][m]: You can hit that pretty easily implementing a machining center with tool changer. I am over 100io on my machine and it is a small one.
01:36 PM travis_farmer[m]: and i don't have the tool changer built yet, or configured ;-)
01:37 PM Tom_L: i know they can go quick but i'd think you could do a tool changer fairly efficiently
01:38 PM * JT[m]1 uploaded an image: (349KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/fPgOozGxiyNIcdFpgiWMghZR/20230103_133735_1133142041779151247.jpg >
01:38 PM Tom_L: JT[m]1, build or wait?
01:38 PM travis_farmer[m]: for some reason i never doo things efficiently ;-)
01:39 PM * Tom_L feels JT[m]1 is knee deep in some old project
01:39 PM XXCoder: i saw video of someone making toolchanger that dont use any electrics nor pressured air
01:39 PM XXCoder: it was semi failure but just needs some adjustment. it was interesting
01:40 PM JT-Shop_ is now known as JT-Shop
01:40 PM -!- #linuxcnc mode set to +v by ChanServ
01:40 PM travis_farmer[m]: what's all the valves, JT?
01:41 PM JT-Shop: Tom_L, wait till in the morning to build
01:41 PM Tom_L: ok
01:41 PM sensille: XXCoder: https://hackaday.com/2016/06/20/hackaday-prize-entry-diy-automatic-tool-changer/
01:41 PM JT-Shop: they are off of a machine I'm converting from pneumatic logic to PLC logic
01:41 PM XXCoder: oh yeah been a while since i saw thatm but no
01:41 PM XXCoder: its on top of mill
01:42 PM XXCoder: using double levers to get enough pressure to open
01:42 PM JT-Shop: I had to study that one valve to figure out how to take it apart lol
01:54 PM travis_farmer[m]: there, changed my HAL to move a bunch of pins... will change the hardware to match, tomorrow. what a fun PITA :-)
02:02 PM travis_farmer[m]: according to my schematic ( https://github.com/travis-farmer/linuxcnc/blob/master/7i80hd-16_connections%20v77.pdf ) with the 7i80 headers along the left, i should still have enough I/O to build a automatic changer. :-) i even have a stepgen available :-) (btw, schematic still has errors...)
02:05 PM travis_farmer[m]: still not sure how i will build the changer. will either be a linear changer, or something... i don't have a lot of tools (bits) at the moment, but i want it to be future-proof, if possible :-)
02:23 PM Tom_L: build a 100 position chain drive with prefetch
02:26 PM travis_farmer[m]: i thought of a chain drive, but it looked like it could be more expensive than i can handle ;-)
02:29 PM roycroft: well that was a very frustrating phone call
02:29 PM roycroft: i just called shure shot sprayers to ask about using denatured alcohol in their sprayers
02:30 PM roycroft: they said that "model a and m have chemical resistant seals, and the model b has the standard viton seals"
02:30 PM roycroft: i asked what "chemical resisant means", as ethanol damages seals that most other solvents don't
02:31 PM roycroft: and i was told "denatured alcohol is a chemical, and model a and m have chemical resistant seals"
02:31 PM roycroft: so i asked if there was any rubber in the seals and i was told "the are chemical resistant"
02:32 PM Tom_L: and you expected intelligence on the other end?
02:32 PM roycroft: i did not expect it
02:32 PM roycroft: but i hoped for it
02:32 PM travis_farmer[m]: sounds like it could turn into a revolving door conversation
02:32 PM Tom_L: those folks don't answer the phones
02:32 PM roycroft: it turned circular immediately
02:33 PM roycroft: but at this point i think i can order one and see how it works, and i'll have a valid reason to return it if it fails
02:33 PM Tom_L: call back and ask what the seals in A & M sprayers have
02:33 PM roycroft: i should take a look at my existing ones
02:33 PM Tom_L: what material
02:33 PM roycroft: to see if the seals look like rubber
02:33 PM roycroft: i've already been told what material they're made of - they're made of "chemical resistant"
02:33 PM Tom_L: viton won't do?
02:34 PM roycroft: viton should be fine, but i don't want the model b
02:35 PM roycroft: speaking of circular things
02:35 PM roycroft: the house of representatives are now on their third round of voting for a speaker, with the same results each time :)
02:35 PM roycroft: this is the first time in 100 years it's gone more than one round
02:35 PM travis_farmer[m]: lol
02:35 PM XXCoder: chemical resistant lol
02:36 PM roycroft: i'm thinking of emailing them
02:36 PM roycroft: but it's a small company
02:36 PM roycroft: there's a good chance the person who answers the email is the same person who answers the phone
02:38 PM travis_farmer[m]: and you would be lucky if they spoke english...
02:38 PM XXCoder: if company is small, I doubt they would not hire supports outside usa
02:39 PM XXCoder: and problem lays on company leaders, not workers. anyone can learn facts, its down to complete lack of training from companies
02:39 PM XXCoder: and they may be given bad materials for support
02:40 PM roycroft: the person who answered the phone is likely the bookkeeper as well
02:40 PM XXCoder: so its not always the case that support person is one thats problemic
02:40 PM roycroft: and possibly the janitor and hr manager
02:40 PM travis_farmer[m]: lol
02:40 PM roycroft: and shipping clerk
02:40 PM XXCoder: and ceo?
02:42 PM XXCoder: "recreational math" lol gonna love pracical engineering yt
02:43 PM travis_farmer[m]: yeah, do a few lines of math, see if you get a "buzz" on ;-)
02:44 PM XXCoder: hah
02:44 PM XXCoder: https://youtu.be/Ry_3icwd68o btw. fun
04:08 PM Tom_L: roycroft, would you rather deal with small companies or china?
04:08 PM * JT[m]1 uploaded an image: (255KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/GuSobPTXQvAbydCzpBJmbxur/20230103_160646.jpg >
04:09 PM Tom_L: uh oh, new project in the works
04:10 PM roycroft: small companies
04:10 PM roycroft: the closer to me the better
04:10 PM * JT[m]1 uploaded an image: (265KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/SsCwUqmuQmYJYTBrVRNqYXDU/20230103_161010.jpg >
04:10 PM roycroft: we're on round four now
04:10 PM roycroft: and mccarthy is losing votes
04:11 PM john__: the actual project, the first one was just a practice bowl
04:12 PM XXCoder: nice clamps
04:13 PM XXCoder: nice clamps jt
04:15 PM JT-Woodshop: thanks, you can never have too many clamps...
04:15 PM Tom_L: XXCoder, we have a river that originally used a rubber dam
04:15 PM XXCoder: really? how did that dam work
04:16 PM Tom_L: inflatable but i suspect it got replaced due to debri hitting it
04:16 PM travis_farmer[m]: damn good...
04:16 PM * JT[m]1 uploaded an image: (360KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/plIRHtvHrRhZIOLYLYxatgBY/20230103_161542_1768081531150269095.jpg >
04:16 PM travis_farmer[m]: ;-)
04:16 PM * JT[m]1 uploaded an image: (330KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/DbNJRRweVXRjlXlzaPFNfhAR/20230103_161611_4391319880996486062.jpg >
04:17 PM Tom_L: now it had 3-4 doors with large hydraulic rams holding them up
04:17 PM * JT[m]1 uploaded an image: (403KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/sqtPYZTyezUXqwEdZYLlYSrT/20230103_161703_7795846075863827289.jpg >
04:17 PM Tom_L: https://lh3.googleusercontent.com/p/AF1QipNNA3gqW0EckJjud19-3yxyZpYIBrPBzk4-oaqS=s680-w680-h510
04:17 PM * travis_farmer[m] is very jealous of JT's shop...
04:17 PM Tom_L: XXCoder, ^^
04:18 PM Tom_L: you can see the yellow rams
04:18 PM JT-Woodshop: the cylinders just raised and lowered the rubber thing
04:18 PM Tom_L: to the side is a fish spillway so they can swim thru
04:18 PM Tom_L: JT-Woodshop, those are solid now, not rubber
04:19 PM Tom_L: this replaced the rubber dam
04:19 PM JT-Woodshop: ah
04:19 PM Tom_L: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcS7UulB50oSczziq-QKesjCaDRhlodbSaqUBJ8ooaM0Zw&s
04:19 PM Tom_L: the fish spillway
04:19 PM JT-Woodshop: cool
04:20 PM roycroft: roycroft.us/Clamp-Racks
04:20 PM roycroft: those clamp racks are all full now
04:20 PM roycroft: i need to build some more
04:20 PM Tom_L: happens to be right down the street from me
04:20 PM roycroft: but i don't have room for any more, so i'm not sure what to do
04:20 PM JT-Woodshop: send the excess clamps to JT
04:20 PM roycroft: i don't have any excess clamps
04:21 PM roycroft: i have a deficiency of clamp storate
04:21 PM JT-Woodshop: lol
04:21 PM roycroft: storage
04:21 PM * travis_farmer[m] uploaded an image: (336KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/znJmNPBtxwjXJmDFFAdVIFDR/IMG_0312.png >
04:21 PM travis_farmer[m]: This is why i am jealous of everybody else's shop...
04:21 PM JT-Woodshop: I need to get some of those parallel clamps
04:21 PM Tom_L: wow, they have a whole page about it http://www.kansastravel.org/wichita/lincolnstreetdam.htm
04:21 PM JT-Woodshop: that's way bigger than my first shop
04:21 PM roycroft: i love the parallel clamps
04:21 PM travis_farmer[m]: low headroom though :-)
04:22 PM roycroft: but what i really need now are more of the bessey tradesman clamps
04:22 PM roycroft: longer ones
04:23 PM travis_farmer[m]: i use old style pipe clamps...
04:24 PM JT-Woodshop: has anyone guessed what is in that last photo of the pipe clamps?
04:24 PM JT-Woodshop: roycroft, which clamp are you speaking of?
04:25 PM travis_farmer[m]: crosbow?
04:25 PM JT-Woodshop: pretty close
04:25 PM JT-Woodshop: a ballista
04:25 PM JT-Woodshop: https://gnipsel.com/ballista/ballista.xhtml
04:25 PM travis_farmer[m]: yeah, the spring-mechanical version. i couldn't remember the name
04:25 PM JT-Woodshop: but I cheated and used a truck spring instead of a rope torsion box
04:26 PM * JT-Woodshop wonders why he never uploaded any more photos of the ballista
04:26 PM travis_farmer[m]: nice, go hunting with it much? ;-)
04:27 PM travis_farmer[m]: my crossbow ;-)
04:27 PM * travis_farmer[m] uploaded an image: (5154KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/JhxXuEidWhVoEQeCOBybfiDR/20211030_134539.jpg >
04:27 PM roycroft: and the house have given up
04:28 PM roycroft: they just adjourned without electing a speaker
04:28 PM JT-Woodshop: a ballista is not a hunting weapon, it's a siege weapon
04:28 PM JT-Woodshop: nice cross bow
04:28 PM travis_farmer[m]: i know, i was making a joke ;-)
04:28 PM travis_farmer[m]: thanks :-)
04:29 PM * JT-Woodshop only hunts for great restaurants to eat at...
04:29 PM travis_farmer[m]: https://www.cabelas.com/shop/en/centerpoint-sniper-xt390-crossbow-package
04:30 PM * JT-Woodshop looks for some wood for the next segment ring
05:20 PM roguish[m]: JT-Woodshop: JT are you still using your mill6 on the BP ???
05:21 PM JT-Woodshop: yes or maybe mill 7 I forget
05:21 PM roguish[m]: in your git
05:22 PM roguish[m]: mind if I take a look at it, and maybe use it?
05:22 PM JT-Woodshop: it's there for anyone to use as they like
05:23 PM roguish[m]: thanks.
07:07 PM CaptHindsight[m]: travis_farmer: I wish I had a mobile shop
07:40 PM roycroft: capthindsight: check with your local u-haul/ryder rental places
07:40 PM roycroft: they regularly sell their used box trucks for pretty low prices
07:41 PM XXCoder: I used to work for vending company
07:41 PM XXCoder: boss will never buy uhaul
07:41 PM roycroft: the vehicles tend to have high mileage, but are usually quite suitable for conversion to a mobile shop
07:41 PM XXCoder: theyre fixed at cheapest shops possible and held together by wires and spit by time they sell it off
07:41 PM XXCoder: buy pesoli trucks, they have centerized shop with good workers
07:42 PM XXCoder: penske
07:44 PM XXCoder: while its something he said, he have been in vending business for over 20 years and I dont doubt his words on trucks
07:46 PM XXCoder: nyway business vans, 3 of em is penski, and one is larger white box truck
08:28 PM CaptHindsight[m]: roycroft: I don't even like renting u-haul
08:29 PM CaptHindsight[m]: I'm waiting for the u-haul data breach since they take a pic of your license on their u-haul smartphones
08:30 PM CaptHindsight[m]: I'm sure their security is top notch 🙃
08:30 PM roycroft: i haven't rented a u-haul since the end of the last century
08:32 PM CaptHindsight[m]: but isn't a camper shop great for mobile machining on demand
08:33 PM CaptHindsight[m]: like those mobile vets and pet groomers
08:34 PM CaptHindsight[m]: people already pay 100% markup on restaurant food delivery
08:35 PM CaptHindsight[m]: sounds a new show for the discovery channel
08:36 PM CaptHindsight[m]: Appalachian Highway Machinists
08:45 PM roycroft: where your mechanic is your sister, and also your wife, and also your cousin
08:47 PM XXCoder: ugh
08:47 PM XXCoder: 2 blocks is a killer
08:58 PM -!- #linuxcnc mode set to +v by ChanServ
09:42 PM Bleepshop: pcw---home: You alive/awake/aware? 5V/1A should be enough to run my 7i96s with nothing else connected right?
09:47 PM * Bleepshop is chasing the 'hm2/hm2_7i96s.0: error finishing read! iter = x' error.
09:50 PM pcw---home: 1A is fine, beware long skinny wires though
09:50 PM Tom_L: Bleepshop, i use a chopped off usb cord
09:56 PM pcw---home: What Kernel are you using an what MAC?
10:02 PM Bleepshop: Linux accu-cut 4.19.0-21-rt-amd64 #1 SMP PREEMPT RT Debian 4.19.249-2 (2022-06-30) x86_64
10:03 PM Bleepshop: Network on that machine is inoperable due to a borked config at the moment.
10:04 PM Bleepshop: Intel 82574 chipset. [Checked via Sneakernet]
10:06 PM Bleepshop: It's got a Realtek baked into the mobo and I'm trying to get another intel or a dual port intel to kill it off.
10:08 PM Bleepshop: Looking at one of these. https://www.amazon.com/10Gtek-Gigabit-Ethernet-Converged-Network/dp/B01LXTF48X
10:08 PM Bleepshop: Or is that too fancy?
10:09 PM pcw---home: Do you have irq-coalescing disabled on the Intel MAC?
10:11 PM Bleepshop: 'hardware-irq-coalesce-rx-usecs 0' is the last line in /etc/network/interfaces
10:12 PM pcw---home: What kind of ping times do you get?
10:13 PM Bleepshop: 64 bytes from 10.10.10.10: icmp_seq=1 ttl=64 time=0.319 ms
10:14 PM Bleepshop: 0.314-0.319 with 'ping -c 10 10.10.10.10'
10:15 PM Bleepshop: 0.187--0.232 pinging back to this machine. [network finally came up]
10:17 PM pcw---home: that's really slow
10:18 PM Bleepshop: pcw---home: /etc/network/interfaces. https://ctxt.io/2/AAAQe862EQ
10:19 PM pcw---home: I would expect the first ping to be slow as it incorporates and ARP transaction
10:19 PM Bleepshop: https://ctxt.io/2/AAAQvc8YFQ
10:20 PM Bleepshop: Ping returns
10:20 PM pcw---home: You might need a slower servo thread with such long ping times
10:21 PM pcw---home: what host PC?
10:22 PM pcw---home: I get this on my G800:64 bytes from 10.10.10.10: icmp_seq=1 ttl=64 time=0.179 ms
10:22 PM pcw---home: 64 bytes from 10.10.10.10: icmp_seq=2 ttl=64 time=0.122 ms
10:22 PM pcw---home: 64 bytes from 10.10.10.10: icmp_seq=3 ttl=64 time=0.131 ms
10:23 PM pcw---home: (the first in longer because of the ARP)
10:23 PM pcw---home: is longer
10:24 PM Bleepshop: Dell Optiplex 390 w/ 12gb RAM, 120gb SSD.
10:25 PM Bleepshop: Intel(R) Core(TM) i3-2120 CPU @ 3.30GHz
10:26 PM Bleepshop: Running as 2 cores, VT/VM/HT off, no isolcpus.
10:27 PM pcw---home: Does is possibly have AMT or other "Ethernet spyware" features?
10:27 PM pcw---home: (often present in corporate desktops)
10:28 PM pcw---home: I know HP 8300s have this and disabling it drops about 150 usec from the ping times
10:31 PM Bleepshop: pcw---home: Not that I can find. NIC options are Disabled, Enabled, Enabled with PXE, Enabled with ImageServer. I've got it at just Enabled.
10:34 PM pcw---home: makes you wonder what its doing for a couple 100 usec...
10:35 PM Bleepshop: That is why I was thinking of a dual port card. Kill of the Realtek anc the internal NIC and just run PCIe.
10:36 PM pcw---home: You could try a lower servo thread rate (and take a look at the tmax values)
10:39 PM Bleepshop: Current is <property name="servoperiod" type="float" value="50000.0"/>
10:41 PM pcw---home: OK well that will not work, try 1000000
10:42 PM Bleepshop: pcw---home: I've got time. My silicone wiring kit won't be here till the 12'th or so and I'm going to wire it right instead of the bodge job I did with the BOB..
10:44 PM pcw---home: A 20KHZ servo thread will not work with almost any hardware 1 ms (1000000 ns) is standard
10:46 PM Bleepshop: pcw---home: BTW Those generic breakouts are labeled 'DB25-1205', I picked them up from Longs Motor or Leadshine years ago.
10:50 PM Bleepshop: pcw---home: Well. Now I feel like an idiot. I had 25,000 instead of 250,000 in my configs. LOL
10:51 PM roycroft: operator error is always the preferred problem, because it can usually be fixed
11:02 PM Bleepshop: pcw---home: Got a 30' x 10 color silicone wire kit and a 100 pack of axial lead ferrite beads for EMI/RFI supression on all the step/dir/input leads/etc. Gonna do it right. LOL