#linuxcnc Logs

Aug 11 2025

12:05 AM lcnc-relay: <meisterdippel@> moin
02:03 AM lcnc-relay: <lucid.nonsense@> In terms of mesa setup what would you folks go for if converting a turret mill today and ripping out everything? Planning on putting new servos and drives to replace some old dc servos and probably getting a chinese 5th axis. Off to Portugal on Sunday and thought I'd stop by and see Filipe at EUsurplus.
02:31 AM lcnc-relay: <lucid.nonsense@> Cool, so are they more advanced than say the 7i95t or 7i96s? Mesa website isn't exactly easy to make sense of for comparisons
02:33 AM lcnc-relay: <meisterdippel@> lucid.nonsense@: I think it's just a new FPGA because the Spartan6 is no longer being manufactured.
02:34 AM lcnc-relay: <lucid.nonsense@> Ok, just found this, sorta interesting, https://albersx.com/static/93fe234741f0b3cf3ed1885b07009231/d8104/mesa_cards.png
02:35 AM lcnc-relay: <meisterdippel@> It's best to wait until Tom arrives; he knows all about Mesa cards.
02:39 AM lcnc-relay: <meisterdippel@> lucid.nonsense@: not really complete
02:52 AM lcnc-relay: <lucid.nonsense@> meisterdippel@: Cool, thanks, yeah, see there's UDP based boards now
03:03 AM lcnc-relay: <meisterdippel@> How cool is that 🙂
03:03 AM lcnc-relay: I asked EUsurplus if they would give me a discount, as I want to port their board to RIO and mention them in the video/forum. They did 🙂
03:15 AM lcnc-relay: <lucid.nonsense@> Just happy they're still selling the cards, last time I visited a few years back, Filipe showed me around and it was clear that it had become a very small part of their business.
03:17 AM lcnc-relay: <meisterdippel@> I think there are only two distributors in Europe,
03:17 AM lcnc-relay: so offering them will definitely not be a loss-making business.
04:26 AM Tom_L: morning
04:29 AM Tom_L: they switched to the efinix chips during the chip shortage. the 7i96s is a good all around card
04:30 AM Tom_L: 26 pin header and sserial for expansion
04:31 AM Tom_L: it's geared more for steppers though i think
04:32 AM Tom_L: you might find it a bit easier to navigate jt's site: mesaus.com
04:34 AM Tom_L: https://mesaus.com/product/7i97t/
04:34 AM Tom_L: servo card
04:34 AM Tom_L: ethernet interface
05:33 AM lcnc-relay: <lucid.nonsense@> Cool, thanks!
05:44 AM Deejay: hi
06:02 AM lcnc-relay: <.le_potato@> i really have to replace my pc with rpi5?
06:03 AM lcnc-relay: <.le_potato@> how is rpi4?
06:03 AM lcnc-relay: <.le_potato@> i have those with 2gb ram :l
06:42 AM lcnc-relay: <meisterdippel@> .le_potato@: no, but you can use your pc with the ethernet cards , but RPI4 works also with ethernet and SPI cards
06:43 AM lcnc-relay: <meisterdippel@> 2gb is ok
06:43 AM lcnc-relay: <.le_potato@> meisterdippel@: I have HP mini pc now with ethernet but its realtek and using 2ms time
06:43 AM lcnc-relay: <.le_potato@> maybe that 2ms isnt a problem
07:13 AM rdtsc: Can use rpi4 + 7c8x, but it will run slower than a traditional PC. Rpi5 is much faster.
08:25 AM BorgPipe is now known as NetPipe
11:14 AM JT-Cave: new refrigerator "could" be here in as little as 45 minutes...
11:19 AM roycroft: nice
11:19 AM roycroft: i assume you're not holding your breath, though
11:23 AM lcnc-relay: <.le_potato@> sorry for a dumb question
11:23 AM lcnc-relay: <.le_potato@> how can i monitor input pin?
11:24 AM roycroft: halscope?
11:26 AM JT-Shop: show hal watch window as well
11:26 AM JT-Shop: no, not holding my breath it's been many months without one
11:27 AM roycroft: is there an snmp interface to linuxcnc? and if so, what mibs does it use?
11:27 AM * roycroft has never looked into that
11:27 AM JT-Shop: what is snmp?
11:28 AM roycroft: simple network monitoring protocol
11:28 AM roycroft: it's been the industry standard for monitoring for decades
11:28 AM roycroft: it would be neat to have for linuxcnc
11:28 AM JT-Shop: what would you monitor in linuxcnc?
11:28 AM roycroft: the "simple" part really is simple - there are three operations one can do with snmp
11:28 AM roycroft: the monitoring system can issue a "put" or a "get"
11:29 AM roycroft: and the monitored system can send a "trap"
11:29 AM lcnc-relay: <.le_potato@> ty
11:30 AM roycroft: if a machine needs a tool change, for example, a trap could be sent and the monitoring system could take action based on that, such as send an email or ring a bell or flash a light
11:30 AM JT-Shop: interesting
11:30 AM roycroft: if linuxcnc does not support it, the underlying os does, and scripts could be written to send the traps
11:31 AM JT-Shop: you can have python code in linuxcnc
11:32 AM roycroft: home assistant can do snmp monitoring
11:32 AM roycroft: as can virtually all network monitor systems such as nagios or zabbix or whatever
11:33 AM roycroft: if it were built into linuxcnc it would be trivial
11:33 AM roycroft: if it requires a few scripts it would be close to trivial
11:33 AM roycroft: anyway, it's probably kind of a niche thing, but it would be nice
11:35 AM JT-Shop: Easy SNMP is a fork of Net-SNMP Python Bindings that attempts to bring a more
11:35 AM JT-Shop: Pythonic interface to the library. This module provides a full-featured SNMP
11:35 AM JT-Shop: client API and supports all dialects of the SNMP protocol.
11:39 AM rdtsc-w: remember seeing some sort of network interface for linuxcnc, but unsure if that was snmp
11:39 AM roycroft: it's the mib (management information database) that's the problem
11:39 AM roycroft: that has to be written and an agent run that can query linuxcnc
11:40 AM lcnc-relay: <.le_potato@> i wired the switch to INCOM and to IN0
11:40 AM lcnc-relay: <.le_potato@> i monitoring the 0.gpio.000.in
11:40 AM lcnc-relay: <.le_potato@> and its not changing
11:42 AM lcnc-relay: <.le_potato@> so for detecting there should be a pull-up resistor but there is nothing about the schematic i have no clue
11:42 AM rdtsc-w: does the pull-up resistor from +V to IN0 make it go on?
11:43 AM lcnc-relay: <skunkworks8841@> and are the gpio read/write functions running?
11:44 AM lcnc-relay: <.le_potato@> at I/O map i dont see any IN function
11:45 AM lcnc-relay: <.le_potato@> P1-01/DB25-01 41 IOPort None
11:46 AM lcnc-relay: <.le_potato@> i watching wrong input
11:46 AM lcnc-relay: <.le_potato@> :l
11:47 AM lcnc-relay: <skunkworks8841@> I don't remember how the gpio mapping works.. I remember it being odd
11:47 AM lcnc-relay: <skunkworks8841@> https://linuxcnc.org/docs/stable/html/drivers/hal_gpio.html
11:47 AM lcnc-relay: <.le_potato@> still zero
11:49 AM lcnc-relay: <.le_potato@> this help document for different
11:50 AM lcnc-relay: <.le_potato@> ok there is an input led for it
11:50 AM lcnc-relay: <.le_potato@> to check it
11:50 AM lcnc-relay: <.le_potato@> but it stay off, but with multimeter its okay
11:50 AM lcnc-relay: <.le_potato@> its like want to switch 5V instead of GND
11:52 AM lcnc-relay: <.le_potato@> why there is no correct manual of the board 🙁
11:53 AM lcnc-relay: <.le_potato@> Terminal block TB6 is one the 7I95Ts isolated I/O connectors. This has 16 inputs
11:53 AM lcnc-relay: (0 through 15)
11:53 AM lcnc-relay: <.le_potato@> The 7I95T has 24 isolated inputs and 6 isolated outputs. All 24 Isolated inputs have
11:53 AM lcnc-relay: a common pin per input pair. This common pin must be connected to ground for active
11:53 AM lcnc-relay: high inputs and connected to the I/O power for active low inputs. The 6 isolated outputs
11:53 AM lcnc-relay: are completely floating switches so can be use for pull-up/pull-down and mixed voltage
11:53 AM lcnc-relay: switching
11:55 AM lcnc-relay: <.le_potato@> a switch is NPN
11:55 AM lcnc-relay: <.le_potato@> For NPN type sensors or switches with a common ground, the input common for a
11:55 AM lcnc-relay: given input pair is connected to +5 to +36V and the input pins are grounded to activate an
11:55 AM lcnc-relay: input.
11:56 AM JT-Shop: hm2_7i95.0.inmux.00.input-00 is an input pin for the 7i95t not gpio
11:57 AM lcnc-relay: <.le_potato@> ah ty
11:58 AM lcnc-relay: <.le_potato@> ok, for search for input is not exist
11:58 AM lcnc-relay: <.le_potato@> so in hal oscilloscope the search is broken :3
11:58 AM JT-Shop: prob not broken
11:59 AM lcnc-relay: <.le_potato@> that one doesnt work as well, lol
11:59 AM JT-Shop: what is a "that one"?
11:59 AM lcnc-relay: <.le_potato@> input-04
12:02 PM JT-Shop: driver just called and he's about 30 minutes away
12:03 PM * roycroft is going to attempt a debian upgrade to 13.0 on a non-critical machine
12:04 PM roycroft: after taking a snapshot of the 12.11 install, of course
12:13 PM JT-Shop: it's been 3 months without a full size refrigerator upstairs
12:27 PM roycroft: trixie alters some network device names, so one should be cautious about upgrading - your network config
12:28 PM roycroft: there are several other precautions that must be taken as well - it is not as simple an upgrade as the previous few have been
12:28 PM xxcoder: wonder why names was changed
12:32 PM roycroft: no idea
12:32 PM roycroft: and the release notes say nothing about postgresql, but i got a warning that it was running when i attempted the ugprade the first time, with an option to continue anyway
12:33 PM roycroft: i did so, and the upgrade of libc6 failed
12:33 PM roycroft: and even apt -y --fix-dependencies failed
12:33 PM roycroft: i restored the snapshot, started over, stopped postgresql, and now it's proceeding past the previous point of failure
12:33 PM lcnc-relay: <skunkworks8841@> I had some ubuntu systems I have upgraded through the years.. They also did a totally different network stack that took me a while to figure out.
12:36 PM roycroft: this is a personal machine, btw, and the database is not something critical that is updated in real time - it's just an indexer for my archived emails
12:37 PM roycroft: i only keep 5 years of emails on my mail server, as it's a leased machine with limited stoage
12:37 PM roycroft: but i have emails going back to the '80s on my archive server
12:45 PM lcnc-relay: <skunkworks8841@> I have only kept work emails back to 00
12:50 PM roycroft: since my archive server has virtually unlimited storage and since emails are pretty small, especially in the pre-mime days, it's not a problem to just keep all the old ones
12:53 PM roycroft: it's fun looking at old emails with uucp adressing
01:43 PM lcnc-relay: <.le_potato@> somehow it doesnt work
01:43 PM lcnc-relay: <.le_potato@> have to switch 5V on it?
01:43 PM lcnc-relay: <.le_potato@> why with common ground doesnt work?
01:45 PM lcnc-relay: <.le_potato@> Input_0_4 = Joint 1 Home
01:46 PM rdtsc-w: maybe some of these can be helpful - see how others are doing it https://duckduckgo.com/?q=how+to+test+inputs+and+outputs+linuxcnc+mesa
01:46 PM lcnc-relay: <.le_potato@> yeah, input doesnt work
01:46 PM lcnc-relay: <.le_potato@> led is OFF
01:47 PM lcnc-relay: <.le_potato@> why no schematic thats show electronic inside
01:47 PM rdtsc-w: 7i90?
01:47 PM lcnc-relay: <.le_potato@> 7i95t
01:53 PM rdtsc-w: inputs are likely bidirectional LED optocouplers. put one side of battery to TB6 INCOM0,1, other side to INPUT0, should light (always check your manual to see what voltages are acceptable - here it can be 4 to 36 Volts)
01:55 PM lcnc-relay: <.le_potato@> redtsc-w: how i wire a switch?
01:56 PM pcw---home: The isolated inputs need a voltage difference between the input common and the input pins
01:56 PM pcw---home: as explained in the manual
01:57 PM pcw---home: so if you have a switch to ground, the input common must be at a potential 4V to 36V different from ground
02:00 PM pcw---home: I would suggest higher voltages (say 24V) with mechanical switches for reliability
02:03 PM pcw---home: So with a sinking sensor (NPN or switch to ground) the input common is normally connected to +5 to +36V
02:04 PM pcw---home: And with a sourcing sensor (PNP or switch to +5V to +36V) the input common would be ground
02:08 PM rdtsc-w: and if you need to invert the output state, that can be done in .hal
02:09 PM rdtsc-w: *output meaning what linuxcnc "sees"
02:10 PM pcw---home: Yes, physical inputs are hal output pins
02:18 PM roycroft: these old phones i have are going for $10 each on ebay
02:18 PM roycroft: i think it's donate time - i can't afford to list them for that price
02:26 PM * roycroft has a box full of ip phones that needs to go away
02:42 PM Scopeuk: can you run a job lot?
02:46 PM roycroft: i'll make more money donating them
02:46 PM roycroft: with less hassle
02:50 PM xxcoder: tax writeoffs I guess
02:52 PM roycroft: yes
02:52 PM roycroft: i already have to invest the time doing factory resets on them all
02:53 PM roycroft: i've decided that the only further time investment i can afford is to put them in a box and put the box in my car
02:53 PM Tom_L: just smash em with a hammer and go on with your day
02:54 PM pcw---home: might blow-up...
02:54 PM Tom_L: could happen
02:54 PM Tom_L: pcw---home, since you seem to be here, https://github.com/jethornton/mesact/blob/master/mesact/src/libmesact/flash.py#L300
02:54 PM roycroft: if i did that i'd have to pay money at the electronics recycling center
02:55 PM roycroft: if they are intact i can recycle them for free
02:55 PM Tom_L: is that what --reload is in mesaflash?
02:55 PM roycroft: but since they still work i'd rather donate them to st vinnie's and get a tax writeoff
02:55 PM Tom_L: he's working on those in mesact
02:56 PM pcw---home: --reload allows you to reload the FPGA from the flash EEPROM without a power cycle
02:56 PM Tom_L: right
02:56 PM Tom_L: in the other channel he wanted me to grep the logs but it returned way too many results
02:57 PM Tom_L: you issue that right after a --write
02:57 PM pcw---home: You can
02:58 PM pcw---home: On xilinx cards you can also do tricks like --reload --fallback
02:58 PM pcw---home: That may work on Efinix, I forget
02:58 PM Tom_L: :)
03:00 PM pcw---home: the --reload --fallback trick can be used to test iffy firmware (loaded with --write --fallback)
03:00 PM Tom_L: what does that do? just reload the fallback section?
03:00 PM pcw---home: Yes
03:01 PM Tom_L: how many more tricks are up that sleeve?
03:02 PM pcw---home: Lots of tricks with Xilinx since you can dynamically change the FLash load address
03:02 PM xxcoder: all of them
03:02 PM Tom_L: that could get you in trouble quick
03:02 PM pcw---home: Th Efinix chips are more basic with their startup options
03:05 PM pcw---home: (a --write --addr and --reload --addr option might be nice if you wanted to be able to switch between multiple images without changing the flash EEPROM}
03:06 PM Tom_L: i did that on the 7i90 using the jumper
03:06 PM Tom_L: and the 2 eeprom
03:07 PM Tom_L: i have spi loaded on the secondary one and epp on the primary
03:07 PM Tom_L: ya just gotta remember ya did that
03:08 PM Tom_L: is there room in the eeprom for more than one with a fallback?
03:09 PM roycroft: so the phones are reset and ready to go, and the last box of books, which was underneath the phone box, is unpacked and the books shelved
03:09 PM roycroft: those were the last two moving boxes in my office
03:10 PM Tom_L: --addr would conflict with eth and spi --addr
03:14 PM pcw---home: with 32Mbit you could have 4 images
03:15 PM Tom_L: might be handy for testing
03:15 PM pcw---home: Yeah so maybe --offset or some such
03:15 PM Tom_L: yeah
03:16 PM Tom_L: i'd rather see a --read before --write --offset came about
03:16 PM Unterhaus__ is now known as Unterhausen
03:32 PM lcnc-relay: <.le_potato@> pcw---home: so, switch 5V with the switch?
03:40 PM rdtsc-w: is this switch a physical switch, or a proximity sensor? model number would be helpful. 2-wire or 3-wire?
03:41 PM lcnc-relay: <.le_potato@> found out
03:44 PM lcnc-relay: <.le_potato@> so its comparing the COM with INPUT
03:44 PM lcnc-relay: <.le_potato@> so if i wire 24V to Input then i have to ground the Com
03:44 PM lcnc-relay: <.le_potato@> if i put 24V on the Com then i can control the input with Ground
03:46 PM roycroft: correct, .le_potato
03:51 PM lcnc-relay: <.le_potato@> thank you!
03:52 PM roycroft: and iirc, if you tie com to earth, you can use different voltages on different input pins and they will all work
03:52 PM roycroft: within the allowed voltage range, of course
03:53 PM * roycroft needs to go outside to retrieve a parcel
03:54 PM roycroft: i was hoping to avoid outside until it cools off tonight
03:54 PM xxcoder: its very hot here, but also very very windy
03:54 PM xxcoder: so I had no problem getting my soylent package.
03:55 PM xxcoder: I was literally air-cooled ;)
03:57 PM roycroft: no wind here today, thank goodness
03:57 PM xxcoder: no air coolant for you
03:57 PM xxcoder: but same time, winds dont need help
03:57 PM xxcoder: err fires
03:57 PM roycroft: high heat and high wind make extreme fire danger
03:57 PM roycroft: yes
03:59 PM lcnc-relay: <.le_potato@> roycroft its have an opto comparator
03:59 PM lcnc-relay: <.le_potato@> but its not writing about it, its have fancy writing about it
03:59 PM lcnc-relay: <.le_potato@> not just saying for an electrican, hey its a comparator
04:22 PM lcnc-relay: <.le_potato@> if i set home offset to 200, why is it want to go back to 0?
04:22 PM lcnc-relay: <.le_potato@> just set final to 0?
04:34 PM lcnc-relay: <.le_potato@> home = 200
04:38 PM lcnc-relay: <big_kevin420@> home = 200 offset =200
04:39 PM lcnc-relay: <big_kevin420@> should make nothing move that way
04:40 PM Tom_L: HOME = 0.000
04:40 PM Tom_L: HOME_OFFSET = 0.050
04:40 PM Tom_L: i set mine with a little offset to back off the switch when homing
04:41 PM Tom_L: HOME_SEARCH_VEL = 2.50
04:41 PM Tom_L: HOME_LATCH_VEL = -0.20
04:41 PM Tom_L: latch is negative so it will reverse and stop
04:43 PM lcnc-relay: <.le_potato@> my home offset is 200 coz i homing to positive
05:14 PM * JT-Shop waits for the new fridge to get cool
05:15 PM JT-Shop: left over pizza night
05:15 PM xxcoder: wouldnt be best to wait 24 hours before turn on?
05:15 PM JT-Shop: it might take 24 hours for the first ice to be made
05:15 PM JT-Shop: but no, just plug it in
05:16 PM xxcoder: hmm ok
05:16 PM JT-Shop: interesting the fridge cools down before the freezer
05:17 PM JT-Shop: years ago if it was on it's side you need to wait a bit before turning it on
05:17 PM JT-Shop: but these are shipped and stored upright
05:17 PM xxcoder: I recall it as any large movements
05:17 PM JT-Shop: I had to move it from the road to my deck with my truck, the delivery truck couldn't make the turn into my driveway
05:18 PM _unreal_: that was fun
05:18 PM _unreal_: just got home
05:25 PM JT-Shop: maybe in the old days you had to wait after moving a fridge
05:28 PM xxcoder: maybe
05:28 PM xxcoder: its not like I keep up with ref tech :D
05:29 PM JT-Shop: me neither, the manual just says plug it in :)
05:35 PM xxcoder: cool :D
05:35 PM xxcoder: both definitions
05:46 PM roycroft: i was always told to let a refrigerator/freezer stand for several hours after moving it if it's been tipped
05:46 PM roycroft: not period
05:47 PM Tom_L: so the liquid doesn't acumulate in the compressor
05:49 PM Tom_L: (oil and refrigerant)
06:12 PM lcnc-relay: <ccatlett1984@> several hours is important
06:13 PM lcnc-relay: <ccatlett1984@> at least 8 hours, if possible
06:21 PM lcnc-relay: <skunkworks8841@> I can see why... you want the liquid freon in the correct place before starting it.. But i hav done both with no issues...
06:34 PM lcnc-relay: <skunkworks8841@> https://www.bbc.com/news/live/c20114gel3lt
07:45 PM Tom_L: skunkworks, http://tom-itx.no-ip.biz:443/~webpage/cnc/Rpi5/pics/RPI5_spi_cable1.jpg
07:45 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/Rpi5/pics/RPI5_spi_cable2.jpg
07:45 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/Rpi5/pics/RPI5_spi_cable3.jpg
07:52 PM lcnc-relay: <.le_potato@> Tom_L: IDE Cable is 1:1 fit?
07:53 PM Tom_L: yes
07:53 PM lcnc-relay: <.le_potato@> i didnt know tha
07:53 PM lcnc-relay: <.le_potato@> t
07:53 PM Tom_L: one plug is flipped over for convenience
07:53 PM Tom_L: just on the 7c8x boards
07:53 PM Tom_L: 7i90 is different
07:56 PM Tom_L: http://tom-itx.no-ip.biz:443/~webpage/cnc/Rpi5/pics/RPI5_spi_cable4.jpg
08:55 PM lcnc-relay: <.le_potato@> thats nice
08:56 PM xxcoder: finally crossed the heat peak
09:24 PM Tom_L: for testing, i didn't want to mount the rpi
10:41 PM uffywolflf is now known as fluffywolf