#linuxcnc Logs

Jan 29 2025

01:17 AM lcnc-relay: <meisterdippel@> moin
02:03 AM Deejay: moin
02:25 AM memleak: you guys would make excellent navy seals if i had to go off your sleep schedules alone.
02:27 AM memleak: always well rested, waking up naturally just in time for mission briefing for some fancy night op, just gotta grab NODs and you're good to go. you guys like swimmin? :D
02:28 AM * memleak from now on will see Deejay as a navy seal
02:31 AM lcnc-relay: <vibram@> zmrdko@ do you manage to move forward on the index home ethercat?
02:33 AM memleak: plus for machining you need to know trigonometry and stuff which is typically required for special ops. knowing how the coriolis effect works for sniping etc. so basically, once you're experienced with machining, you already have all the skills to be a navy seal.
02:34 AM memleak: and you guys wouldn't ever run out of ammo, just bring your equipment and CNC more bullets.
02:35 AM memleak: you'd just have to figure out with your squad if you're going hollow point or full-metal jacket.
02:41 AM Deejay: hey memleak
02:42 AM Deejay: i hope you can help me
02:42 AM memleak: i'm surprised you're not the one telling me i need help! :)
02:42 AM memleak: what's the issue?
02:43 AM Deejay: just bought a new intel n100 itx board. running debian 12 bookworm works fine. using kernel 5.4.278, x-server doesnt start
02:43 AM Deejay: i think the i915 kernel driver may be to old?
02:44 AM memleak: what does /var/log/Xorg.0.log report? can you post to dpaste/pastebin?
02:44 AM lcnc-relay: <zmrdko@> vibram@: Hello, i am slowly chewing through the lcec_deasda.c code.
02:44 AM memleak: you can try force probing it if is too old
02:46 AM Deejay: memleak: https://pastebin.com/nsgskDgB
02:46 AM Deejay: i tried to append i915.force_probe=xxx to grub
02:46 AM Deejay: but doesnt change anything
02:47 AM memleak: now post it with the 5.4.278 kernel
02:48 AM Deejay: that is with 5.4.278 kernel
02:48 AM memleak: nope that's with 6.1.0-30-rt-amd64
02:49 AM Deejay: oops. sry. give me a second
02:49 AM lcnc-relay: <vibram@> zmrdko@: OK
02:50 AM memleak: you can: `cp /var/log/Xorg.0.log /home/<your user>/xorg_log.txt` to save it, then reboot into the 6.1 kernel again
02:50 AM lcnc-relay: <vibram@> I would like to use servo but I'm afraid about this part, finding all infos through the interface
02:50 AM Deejay: memlean: https://pastebin.com/xuWNeCwc
02:51 AM Deejay: *memleak
02:51 AM memleak: this is the issue: [ 2876.621] (EE) open /dev/dri/card0: No such file or directory
02:51 AM memleak: now i need `dmesg` output
02:52 AM lcnc-relay: <zmrdko@> vibram@: Check DMs
02:52 AM Deejay: first thing i notice: there is no /dev/dri
02:52 AM Deejay: jep.
02:52 AM Deejay: ok, dmesg. one sec
02:52 AM memleak: yeah the drm driver isn't doing it's thing
02:54 AM Deejay: dmesg: https://pastebin.com/5Q6AuMah
02:56 AM memleak: yeah i915 cant see your card
02:57 AM memleak: it should be able to use the framebuffer though
02:57 AM memleak: do you have a /dev/fb*?
02:57 AM Deejay: from which lines can you see that?
02:57 AM Deejay: fb0 is there
02:57 AM memleak: ok what happens if you try to run `sudo startx`
02:58 AM Deejay: it fails, tried before
02:58 AM Deejay: thats where the Xorg.log comes from
02:58 AM memleak: oh ok
02:59 AM Deejay: so the i915 module is older than for 6.1.x kernel?
02:59 AM Deejay: or what is the difference?
03:00 AM memleak: yeah your hardware is apparently missing from the 5.4 kernel
03:00 AM Deejay: d'oh
03:00 AM memleak: did you specify i915.force_probe=<your actual id> ?
03:00 AM memleak: or was it a random one?
03:00 AM Deejay: it was 46d1
03:01 AM Deejay: i think it should be the correct one
03:01 AM Deejay: but perhaps you can confirm that :)
03:01 AM memleak: i'm not sure... try `lspci -nnvv`
03:03 AM Deejay: output lspci: https://pastebin.com/Ei6cPJ72
03:03 AM Deejay: first line: 8086:46d1?
03:03 AM Deejay: i thought this would be the number...
03:04 AM Deejay: (i removed all other devices, the list is loooooong)
03:04 AM memleak: yeah that looks right
03:04 AM Deejay: ok, let me try that again
03:04 AM memleak: i915.force_probe=46d1
03:05 AM Deejay: just to make sure. when booting, i choose edit in grub, append the line you posted to the "linux ..." line, right?
03:06 AM Deejay: somewhere before ro quiet?
03:06 AM memleak: yes
03:07 AM Deejay: no x starting :/
03:07 AM Deejay: what should i post, dmesg, xorg.log?
03:08 AM memleak: dmesg
03:08 AM memleak: how are you posting to pastebin so fast without X btw?
03:09 AM Deejay: https://pastebin.com/AG2f5wi5
03:09 AM Deejay: dmesg > foo.txt
03:09 AM Deejay: copy it by ssh to my normal computer
03:09 AM Deejay: :)
03:10 AM Deejay: have some years linux experience and also compiled kernels modules for myself, but thats years ago
03:10 AM Deejay: so if that might be an option, i am not afraid of trying that
03:11 AM memleak: ahhhh
03:11 AM memleak: building a custom kernel won't help, it's just too old
03:11 AM memleak: you should still be able to use the framebuffer though but that's not working either
03:11 AM memleak: simple-framebuffer simple-framebuffer.0: fb0: simplefb registered!
03:14 AM memleak: i have to get sleep, ambien kicking in
03:14 AM Deejay: any suggestions what i can try while u sleep? ;)
03:15 AM memleak: i dont know what to try.. `nomodeset` maybe?
03:15 AM memleak: on kernel cli?
03:15 AM Deejay: ok, i will try
03:15 AM Deejay: thanks a lot so far for your help!
03:15 AM memleak: maybe because drm is trying to work and fail, it needs to be off completely, not sure
03:16 AM memleak: good night!
03:16 AM Deejay: gn8, sleep well!
03:22 AM lcnc-relay: <zmrdko@> Does anybody have any info about scottlaird?
03:22 AM lcnc-relay: <zmrdko@> He maintains linuxcnc-ethercat repository
03:22 AM lcnc-relay: <zmrdko@> Last commit is from september
04:00 AM -!- #linuxcnc mode set to +v by ChanServ
04:08 AM -!- #linuxcnc mode set to +v by ChanServ
04:23 AM -!- #linuxcnc mode set to +v by ChanServ
04:46 AM Tom_L: morning
06:18 AM rdtsc: Morning. OT: so there is a new AI in town, and it's making waves... here I ask the 7-billion-parameter model (tiny compared to ChatGPT) a question that requires reasoning: https://paste.debian.net/1347322/
06:20 AM rdtsc: Then I asked it questions like 'what happens if you hit a firecracker with a hammer' (says it will explode), and 'how do I synthesize carbonic acid' - that one produced hundreds of pages of output, where it tried hundreds of scenarios - settling on three practical and valid ways.
07:55 AM -!- #linuxcnc mode set to +v by ChanServ
08:44 AM lcnc-relay: <nutlinux@> Lol
09:12 AM lcnc-relay: <nutlinux@> Deejay, maybe Debian doesn't like using simplefb for X.. Gentoo can bring up X without DRM at all. I was thinking about adding RTAI support as an optional feature for my Gentoo image.
09:13 AM lcnc-relay: <nutlinux@> That way X always works regardless of GPU and you'll practically be guaranteed low latency.
09:14 AM lcnc-relay: <nutlinux@> Only downside is that EXT4 will have to be the only filesystem. XFS filesystems created with newer kernels and utils will not boot on the 5.4 kernels and if you're playing with btrfs or f2fs on an old kernel like 5.4 you're playing with fire.
09:14 AM lcnc-relay: <nutlinux@> I could do it though
09:15 AM lcnc-relay: <nutlinux@> I can just force EXT4 in the installer if RTAI is enabled.
09:16 AM lcnc-relay: <nutlinux@> Debian is probably missing something for simplefb
09:17 AM lcnc-relay: <nutlinux@> You should never need the GPU hardware to be initialized for X to work, it should function just enough to bring up the display but everything else is done on the CPU.
09:17 AM lcnc-relay: <nutlinux@> I'd need a Debian system to troubleshoot
09:18 AM lcnc-relay: <nutlinux@> There's probably some hack I could do for it to use the UEFI framebuffer or VESA or something
09:19 AM lcnc-relay: <nutlinux@> The brilliant thing with X is the practically endless ways to configure it. If one thing doesn't work, there's a million other switches to play with.
09:19 AM lcnc-relay: <nutlinux@> And there's just so many different ways to work the framebuffer with X too.
09:24 AM lcnc-relay: <nutlinux@> rdtsc: https://imgur.com/a/S510zw5
09:58 AM Deejay: nutlinux, thanks, but at the moment, even some framebuffer-stuff doesn't run. right now, i am just trying the 6.1.x*-preempt-rt-kernel to see what the latency is
09:59 AM Deejay: tried around with isolcpus and cstates and so on, but not getting good enough latency values for running steppers via parallel port
10:44 AM memleak: ok, sounds like i need a gentoo rtai image
10:44 AM memleak: or a debian system
10:46 AM memleak: can you run this script on your hardware? https://github.com/NTULINUX/gentoo_backup/blob/linuxcnc/verify_psabi.sh
10:47 AM memleak: git clone --single-branch -b linuxcnc --depth=1 https://github.com/NTULINUX/gentoo_backup ; cd gentoo_backup ; ./verify_psabi.sh
10:47 AM memleak: just want to make sure your cpu is new enough to run it
10:51 AM memleak: can you post your `lscpu` output as well?
10:52 AM memleak: i want to look up that cpu model and see what microarchitecture it is and what specific intel gpu it is
10:54 AM memleak: i'm going to try and dig through the kernel source and see when that gpu was added
10:54 AM memleak: we know it was sometime between 5.4 and 6.1
11:02 AM memleak: got it: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/include/drm/i915_pciids.h?h=v6.1&id=7e28d0b26759846485978ada860ef4a427e06c8f
11:02 AM memleak: 6.1 is the first LTS kernel to have it
11:03 AM rdtsc-w: could this be as simple as installing the appropriate driver?
11:04 AM memleak: the driver (i915) is installed but 5.4 is missing support for it
11:05 AM memleak: i could technically try backporting support for alderlake
11:05 AM memleak: he's on ADL-N which is a sub-platform of ADL-P which is also missing
11:06 AM memleak: so i need to grab both of those and throw it into the 5.4 kernel
11:06 AM memleak: proper fix is to build a sane linux image with a working framebuffer
11:07 AM memleak: debian is too dumb to figure out how to use simplefb
11:07 AM rdtsc-w: intel gpu support is a "hole-y" mess ;)
11:07 AM memleak: gentoo starts X just fine with simplefb if you build it right
11:07 AM memleak: brb
11:47 AM lcnc-relay: <nutlinux@> Might be able to make a quick fix for debian
11:47 AM lcnc-relay: <nutlinux@> Almost done
11:47 AM Deejay: how?
11:48 AM lcnc-relay: <nutlinux@> Testing before I give you advice xD
11:48 AM Deejay: :D
11:49 AM lcnc-relay: <nutlinux@> Ok got vesa working
11:50 AM lcnc-relay: <nutlinux@> So X actually doesn't work with simplefb I thought it did...
11:50 AM lcnc-relay: <nutlinux@> So we need to kill /dev/fb0 to get vesa up
11:50 AM lcnc-relay: <nutlinux@> Vesa won't start if there's a framebuffer in the way
11:51 AM Deejay: ah okay
11:52 AM lcnc-relay: <nutlinux@> Can you try this just to see if it works
11:52 AM lcnc-relay: <nutlinux@> "sudo rm -rf /dev/fb0 && sleep 1 && startx"
11:54 AM lcnc-relay: <nutlinux@> On the RTAI kernel
11:56 AM Deejay: nope, not working: https://pastebin.com/V45meHkc
11:57 AM Deejay: :(
11:58 AM lcnc-relay: <nutlinux@> Wtf.. V_BIOS address 0x0 out of range
12:00 PM lcnc-relay: <nutlinux@> Ok, reboot and try running: "sudo apt purge xserver-xorg-video-fbdev"
12:00 PM lcnc-relay: <nutlinux@> Then startx
12:01 PM lcnc-relay: <nutlinux@> I have vesa working here
12:03 PM lcnc-relay: <nutlinux@> Oh crap
12:03 PM lcnc-relay: <nutlinux@> Don't do that
12:03 PM lcnc-relay: <nutlinux@> They make the xorg server a dependency of that package
12:03 PM Deejay: too lake
12:03 PM Deejay: *late
12:04 PM Deejay: the error is the same
12:04 PM lcnc-relay: <nutlinux@> Ok
12:04 PM Deejay: i will now re-install that package, ok?
12:04 PM lcnc-relay: <nutlinux@> Ok make sure xserver-xorg-core is still installed
12:04 PM Deejay: :D
12:04 PM lcnc-relay: <nutlinux@> Didn't mean to pull that out if it did
12:05 PM Deejay: dpkg.log tells me what was removed
12:05 PM lcnc-relay: <nutlinux@> You got the V_BIOS error after rebooting and uninstalling the fbdev driver?
12:06 PM Deejay: yep, same as before, adress 0x0 out of range
12:07 PM Deejay: +d
12:07 PM Deejay: should i pastebin it?
12:07 PM lcnc-relay: <nutlinux@> No that's ok
12:07 PM lcnc-relay: <nutlinux@> Weird as hell though, vesa should always be able to find a memory address
12:07 PM Deejay: [ 27.943] (EE) VESA(0): V_BIOS address 0x0 out of range
12:09 PM lcnc-relay: <nutlinux@> Ok maybe startx is broken in this instance
12:09 PM lcnc-relay: <nutlinux@> Someone on arch Linux had the same problem
12:09 PM Deejay: hmhmh
12:09 PM lcnc-relay: <nutlinux@> Uninstall that package again, then instead of using startx, reboot
12:10 PM lcnc-relay: <nutlinux@> That will let the desktop manager start X instead of startx
12:10 PM Deejay: eh, this is what i did
12:11 PM lcnc-relay: <nutlinux@> You already did that?
12:11 PM Deejay: wait, i am not sure. i am doing it agai. to make sure
12:11 PM lcnc-relay: <nutlinux@> You rebooted after removing that package?
12:14 PM Deejay: jep, its the same
12:14 PM Deejay: here is the log: https://pastebin.com/Yr1HFimq
12:15 PM Deejay: it says: oh no, fbdev module not found. but in the end, its the same error
12:15 PM Deejay: and: /dev/fb0 is still there
12:15 PM Deejay: wb memleak
12:15 PM memleak: that's fine, vesa doesn't care if /dev/fb0 is there as long as the fbdev driver is gone apparently
12:15 PM memleak: this is a really weird problem.. going to ask in #dri-devel
12:15 PM Deejay: memleak, would it really be possible to add the new i915 support to the 5.4.xxx kernel?
12:15 PM memleak: OFTC
12:16 PM memleak: iiiiiiiiiiiiiiii don't know
12:16 PM memleak: i could definitely try and see how far i get lmao
12:16 PM Deejay: you are supposes to answer with yes ;)
12:16 PM Deejay: *supposed
12:17 PM Deejay: plan B is to buy a mesa card and no need for RTAI any more :/
12:18 PM Deejay: but then i also need another mini-itx-case and perhaps power supply
12:18 PM Deejay: so not my preferred solution
12:19 PM Deejay: but i even dunno if the n100 will give good results with rtai... so thats also not for sure...
12:20 PM memleak: as a sanity check, can you make sure /dev/mem is there?
12:20 PM memleak: maybe debian disabled it
12:20 PM Deejay: crw-r----- 1 root kmem 1, 1 29. Jan 19:03 /dev/mem
12:20 PM memleak: /dev/mem is really only useful for the vesa driver for pretty much anything remotely relevant in today's age
12:20 PM memleak: ok so it's there...
12:22 PM Deejay: memleak: btw, lscpu: https://pastebin.com/Ugw28gBH and verify_psabi: https://pastebin.com/hV2bPbdn
12:23 PM memleak: mmxext might not show up on intel platforms i guess
12:23 PM memleak: your cpu definitely has the full MMX instruction set
12:23 PM memleak: what does `lsmod` report?
12:24 PM Deejay: lsmod: https://pastebin.com/PMEyiAQu
12:26 PM -!- #linuxcnc mode set to +v by ChanServ
12:26 PM memleak: the output of `lsmod` was run on the 5.4 rtai kernel?
12:27 PM Deejay: deejay@cnc:~$ uname -a
12:27 PM Deejay: Linux cnc 5.4.279-rtai-amd64 #1 SMP Tue Jul 9 19:17:37 UTC 2024 x86_64 GNU/Linux
12:27 PM Deejay: jep
12:27 PM memleak: ok, shot in the dark, reboot and try setting `nomodeset` again with the fbdev driver uninstalled
12:27 PM memleak: i915 is loaded, maybe there's some weird memory management going on
12:30 PM memleak: one way to test RTAI latency btw is by running the testsuite. `sudo /usr/realtime*/testsuite/run` i think should work
12:30 PM Deejay: heh, pastebin asks me to solve a captcha. i think i post too much in too little time ;)
12:30 PM Deejay: Xorg.log.0 https://pastebin.com/
12:31 PM Deejay: oh no
12:31 PM Deejay: cannot paste any more
12:31 PM memleak: try dpaste or pastebin.ca
12:33 PM Deejay: https://quadrokopter.net/temp/Xorg.0.log
12:34 PM Deejay: my own webspace now
12:34 PM memleak: yeah same problem. vesa just doesn't want to go
12:34 PM Deejay: yep
12:37 PM memleak: i know what might be easier than to backport the i915 code
12:38 PM Deejay: let me ask a (perhaps stupid) question: i would need a matching i915.ko (kernel module) that is new enough to support my hardware and is compiled to match 5.4.278-rtai, right?
12:39 PM Deejay: memleak, what do you have in mind?
12:40 PM memleak: i'm going to try backported simpledrm to 5.4
12:40 PM memleak: *backporting
12:40 PM Deejay: okay
12:50 PM memleak: compile testing
12:57 PM memleak: nope not feasible
12:58 PM Deejay: :/
12:58 PM memleak: what's the latency anyway with the rtai testsuite?
12:58 PM Deejay: öh, not tested yet. can i run it without x?
12:58 PM memleak: yep no x needed
12:58 PM Deejay: oh, great news
12:59 PM Deejay: do you have a command for me? :)
12:59 PM memleak: `sudo /usr/realtime*/testsuite/run` should work if my memory is right
01:00 PM memleak: and yes you are right about the i915 module
01:00 PM memleak: but that doesn't exist
01:04 PM Deejay: https://quadrokopter.net/temp/run_output
01:05 PM Deejay: please explain me the values. what is lat max and ovl max?
01:06 PM Deejay: ovl_max is the maximum latency?
01:06 PM memleak: lat max is the maximum latency of the thread when it was running, ovl max is the highest latency ever recorded since the beginning of the test suite
01:06 PM Deejay: ah okay
01:07 PM memleak: ovl max is the equivalent of latency-test or latency-histogram
01:07 PM Deejay: ok, i understand
01:07 PM memleak: so just under 15 microseconds :)
01:07 PM memleak: not bad
01:08 PM Deejay: yeah, but i did not put much stress on the machine
01:08 PM Deejay: just little cpu-load
01:08 PM Deejay: but cannot glxgears or youtube-stuff... since there is no x :(
01:08 PM memleak: you can use stress-ng
01:09 PM memleak: also as a simple stress test, while true ; do echo 1 > /dev/null ; done
01:09 PM Deejay: yeah, did 'yes > /dev/null' ;)
01:09 PM memleak: ah lol :)
01:09 PM Deejay: found it somewhere on the web
01:10 PM memleak: well rtai does have hope on your system at least
01:11 PM memleak: problem is X
01:11 PM memleak: do you have a spare gpu lying around?
01:11 PM Deejay: nope. there is no slot for it
01:11 PM memleak: ohh damn
01:11 PM Deejay: only pcie-1 (1 lane?)
01:12 PM Deejay: really shooort pcie-slot
01:12 PM memleak: sounds like you need gentoo lol
01:12 PM Deejay: like for mesa-card 6i25
01:13 PM memleak: i'll make sure vesa works with the 5.4 kernel here before shipping, when do you need this by?
01:13 PM Deejay: is there a rtai-kernel for gentoo and working with linuxcnc?
01:13 PM memleak: i can build one
01:14 PM memleak: i'll ship out the preempt_rt build today (hopefully)
01:14 PM memleak: then do one for rtai
01:14 PM memleak: i've been saying that for awhile though... really need to get it sent out and tested
01:14 PM memleak: can you remove the mmxext line from verify_psabi.sh and re-run it?
01:15 PM Deejay: without mmxext:
01:15 PM Deejay: Done. Your processor is x86-64-v3 or newer.
01:15 PM Deejay: You may safely use the Gentoo image for LinuxCNC.
01:15 PM Deejay: whoopsie ^^
01:15 PM memleak: ok awesome, i'll pull that out
01:18 PM Deejay: may i ask whats your job / your profession ("in real life")?
01:20 PM memleak: i started dabbling in linux systems around 12 yrs old, i like being a delivery driver though lol
01:21 PM Deejay: oh, okay :D
01:22 PM * Deejay had first linux contact with ~16 years
01:22 PM memleak: i held a few senior software engineering roles over the years
01:22 PM memleak: i have a stupid patent with my name on it too..
01:23 PM Deejay: a patent? that sounds nice
01:23 PM memleak: i hate patents lol
01:23 PM Deejay: hehe, but having one yourself... why not :)
01:23 PM Deejay: can bring you money
01:24 PM memleak: i'm not with them anymore (thank god)
01:24 PM bjorkintosh: Deejay: I think of it like writing: you just need paper and pencil. easily purchased. will anyone care about your story?
01:24 PM bjorkintosh: well.
01:25 PM Deejay: hmm yes, you need the right person / company / whatever who wants/needs you patent / licence...
01:27 PM bjorkintosh: 'xactly.
01:28 PM Deejay: okay, i think i will shut my computer down now. tested around the whole day (its 8pm here)
01:28 PM Deejay: memleak & nutlinux, thank you a lot for your help so far!
01:29 PM memleak: i'll be working on an rtai image for gentoo soon
01:29 PM memleak: andypugh asked about it too
01:30 PM Deejay: andy also has an intel N100 board if i am right
01:30 PM bjorkintosh: Deejay: still no gui?
01:30 PM Deejay: bjorkintosh: x server is not working
01:30 PM Deejay: (with rtai-kernel)
01:31 PM andypugh: I was awarded a patent on Christmas day :-)
01:31 PM Deejay: ah, there here is :)
01:31 PM Deejay: *he
01:31 PM Deejay: sorry for my english :/
01:31 PM memleak: if all goes well, i'll have one in a few days
01:32 PM memleak: rtai on gentoo
01:32 PM Deejay: okay, gentlemen, good bye. see you in some hours! and thanks again
01:32 PM Deejay: memleak, that would be great!
01:32 PM Deejay: you make my day!
01:32 PM memleak: need to re-work my installer is all
01:32 PM memleak: take care!
01:33 PM Deejay: have a nice day while i fall asleep ;)
01:33 PM Deejay: cya
01:34 PM bjorkintosh: andypugh: for what?
01:34 PM andypugh: An “unpickable” lock.
01:34 PM bjorkintosh: :-)
01:35 PM bjorkintosh: has the lock picking lawyer had a go at it?
01:36 PM xxcoder: I got big wire cutter that says its unlockable
01:36 PM andypugh: No, but a lovkpicker who is prepared to make a video about _not_ picking a lock has. https://www.youtube.com/watch?v=T9u3g-0uYko
01:37 PM andypugh: (LPL didn’t reply to emails, McNally asked for one, then didn’t make a video, or give it back)
01:41 PM bjorkintosh: fascinating! how does one learn how to make locks in the first place, andypugh?
01:42 PM andypugh: Well, firstly by picking them, then taking them apart, then having an idea how to make one harder to pick.
01:45 PM bjorkintosh: sounds like that time consuming thing called ... Experience!
01:45 PM bjorkintosh: it's a good thing too.
01:50 PM xxcoder: geez one that didnt give it back
02:04 PM memleak: tarring up the gentoo image and deploying on USB for testing, fingers crossed, brb
02:13 PM lcnc-relay: <skunkworks8841@> andypugh@ congratulations! (patents are a pain for joe blow public)
02:16 PM lcnc-relay: <skunkworks8841@> andypugh@ so - is there not a way on the mesa card to invert the enable logic? Does it have to be external?
02:16 PM lcnc-relay: <skunkworks8841@> (I will play with it tonight - I haven't gotten back to it yet)
02:17 PM lcnc-relay: <skunkworks8841@> (for the pwmgen)
02:17 PM xxcoder: oh missed that. congats andy!
02:19 PM lcnc-relay: <andypugh@> skunkworks8841@ I am only guessing that inverting the enable is what is making things not work. Have you confirmed that changing that fixes the behaviour/
02:20 PM lcnc-relay: <skunkworks8841@> andypugh@ I assume so.. but I will work on it tonight.
02:20 PM lcnc-relay: <skunkworks8841@> (because the rest of the pwmgens seem to be working correctly and I am not inverting the enable.)
02:22 PM lcnc-relay: <skunkworks8841@> (but i just might be reading the pin wrong.. lol) I just assumed 'setp hm2_7i97.0.pwmgen.04.enable.invert_output true' inverted the enable logic
02:36 PM lcnc-relay: <andypugh@> I think that is a reasonable assumption. There may be a flaw in the driver, or the VHDL
02:37 PM lcnc-relay: <nutlinux@> Tarring up and testing before the big push today 😄
02:38 PM lcnc-relay: <nutlinux@> THIS IS IT
02:38 PM lcnc-relay: <nutlinux@> A fresh PREEMPT_RT Gentoo image ready to go for LinuxCNC
02:38 PM lcnc-relay: <nutlinux@> So excited!
02:39 PM lcnc-relay: <nutlinux@> Up next: RTAI on Gentoo
02:40 PM lcnc-relay: <roguish> andypugh@: thanks for that last release....
03:04 PM lcnc-relay: <nutlinux@> Yes, thank you!
04:20 PM memleak: running off gentoo usb now
04:26 PM memleak: brb
04:38 PM memleak: caught a few bugs, all fixed!
04:39 PM JT-Shop: debugging used to mean cleaning the bugs out
04:41 PM memleak: need break, image almost ready to push
04:41 PM memleak: also linuxcnc is finally showing up in the science category in the menu
04:42 PM xxcoder: first computing bug was caused by actual bug getting crushed by gears and jamming
04:43 PM xxcoder: I wonder what we would call debugging if that didnt happen
04:43 PM JT-Shop: a computer used to be a person that computed numbers
04:43 PM JT-Shop: like for offset tables on ship drawings
04:44 PM xxcoder: robot used to mean forced labor
04:45 PM xxcoder: if we get unbder robot overlords, thats why ;)
04:52 PM Tom_L: JT-Shop, how's the upgrade at the grill plant goin?
04:52 PM JT-Shop: connecting the dots between the remotes and the main panel is done
04:53 PM JT-Shop: working on the 480 feed box now
04:56 PM JT-Shop: changing the oil and filter on the c3 this weekend and hope to do a compression test
04:57 PM JT-Shop: it still has a vibration at 65 and I think it's the tires
04:57 PM Tom_L: the manual probably doesn't give a compression test number
04:58 PM JT-Shop: I didn't see any but didn't look real hard
04:59 PM Tom_L: takin the kid's truck in for a new windshield tomorrow
05:01 PM Tom_L: had that car been sitting a while/
05:01 PM Tom_L: ?
05:01 PM JT-Shop: you guys go through a lot of windshields lol
05:01 PM JT-Shop: I have no clue
05:01 PM Tom_L: yeah we do
05:01 PM JT-Shop: tires for the c3 are cheap compared to the c8
05:01 PM Tom_L: i'm due for one too but not that worried about it
05:02 PM Tom_L: i bet
05:02 PM JT-Shop: $800 for the front and $950 for the rear on the c8
05:02 PM Tom_L: we'll probably take his truck for the trip so i'm getting that done now
05:03 PM JT-Shop: $185 plus tax and install for each on the c3
05:03 PM Tom_L: tires have gone up quite a bit since my last set
05:03 PM JT-Shop: https://www.discounttire.com/buy-tires/bfgoodrich-radial-t-a/p/38709
05:03 PM Tom_L: you buy em and have em shipped?
05:04 PM JT-Shop: I just tell them what I want at the discount tire store
05:04 PM Tom_L: i didn't know what you had there for that
05:04 PM JT-Shop: it's in Farmington which is 60 miles awasy
05:04 PM JT-Shop: so we make a day of it
05:05 PM Tom_L: good night for a steak i think
05:06 PM JT-Shop: taco night here
05:06 PM Tom_L: her birthday
05:06 PM JT-Shop: of course it's steak night then
05:57 PM lcnc-relay: <skunkworks8841@> andypugh@ yes - the invert enable is the issue. If you set invert enable to true - the output goes to -10v.
05:57 PM lcnc-relay: <skunkworks8841@> when you enable the pwm - the output goes to 10v
05:57 PM lcnc-relay: <skunkworks8841@> it does invert the enable though
05:57 PM rdtsc: interesting
06:17 PM lcnc-relay: <skunkworks8841@> peter says you can't inver the enable on the 7I97/7I97T/7I36 as it effects the analog cirucitry
06:38 PM Tom_L: skunkworks, would a bit of external circuit work to invert it?
06:38 PM Tom_L: say a transistor etc
06:50 PM lcnc-relay: <skunkworks8841@> I just read found the manual for the drive
06:51 PM lcnc-relay: <skunkworks8841@> This TTL level input signal turns off all power devices of
06:51 PM lcnc-relay: the “H” bridge when pulled to ground with SW1-6=ON. If
06:51 PM lcnc-relay: SW1-6 = OFF pulling this pin to ground will enable the
06:51 PM lcnc-relay: amplifier. This inhibit will cause a fault condition and a red
06:51 PM lcnc-relay: LED.
06:51 PM lcnc-relay: <skunkworks8841@> so - I think I can just flip a switch lol
06:51 PM Tom_L: led me to find single gate inverters which i didn't know exhisted
06:51 PM Tom_L: https://www.mouser.com/ProductDetail/Nexperia/74AHC1G04GM115?qs=5C9Q4QJFsuMUMRaev4gY8A%3D%3D
06:52 PM Tom_L: of course the first one i clicked on is out of stock
06:54 PM rdtsc: lets hope it's that easy... if not, can add a transistor and resistor to "flip" it
06:55 PM Tom_L: yeah i've done that but these single invert chips are kinda intriguing
06:59 PM Tom_L: either way would take about the same realestate
07:00 PM lcnc-relay: <skunkworks8841@> lol - started with this.. https://photos.app.goo.gl/Ly4k9HVYbW4nec89A
07:00 PM lcnc-relay: <skunkworks8841@> but mesa is just so much nicer
07:00 PM Tom_L: true that
07:00 PM Tom_L: nothing against your handywork...
07:00 PM lcnc-relay: <skunkworks8841@> none taken
07:01 PM Tom_L: i used a single diode on a centertap to get my 24v supply and it's worked fine
07:02 PM Tom_L: hidden in some 'inline' shrinkwrap
07:02 PM Tom_L: 1/4 wave
07:02 PM lcnc-relay: <skunkworks8841@> nice 🙂 hopefully you remember that..
07:03 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/rue/Control2.jpg
07:03 PM Tom_L: on the black wire on the center blue cap
07:04 PM Tom_L: the smps regulators weren't rated for 48v input
07:05 PM Tom_L: not pretty but it's worked flawless for quite a while
07:07 PM memleak: so glad i'm not a full-time software engineer anymore. i get burned out and can't look at linux for the day.
07:10 PM rdtsc: Tom those transformers are making guitar amps jealous :)
07:10 PM Tom_L: i'm quite sure i didn't need all of those... one probably would have been plenty but they were surplus and the guy had 3
07:11 PM Tom_L: same with all the caps
07:13 PM Tom_L: switching supply would have been nicer but those were practically free
07:21 PM lcnc-relay: <skunkworks8841@> That fixed it
07:21 PM lcnc-relay: <skunkworks8841@> the smaller amc drives don't have that option (changing the enable with a switch)
07:21 PM Tom_L: awesome
07:22 PM lcnc-relay: <skunkworks8841@> now making it all work 🙂
07:23 PM Tom_L: you'll have fun tinkering
07:23 PM lcnc-relay: <skunkworks8841@> yes - it should be fun. You know - only the original monarch 10EE drive works... nothing else will
08:47 PM memleak: hello rdtsc