#linuxcnc Logs
Nov 04 2023
#linuxcnc Calendar
02:44 AM Deejay: moin
05:01 AM Tom_L: morning
05:07 AM JT-Cave: morning
07:11 AM -!- #linuxcnc mode set to +v by ChanServ
07:30 AM JT-Cave: rooster must not be feeling well, all I heard was a couple of clicks from the speaker
08:32 AM lcnc-relay: <nellen3861> Hey guys just a fast question I want to hear what you think. I have 8 wire dual shaft NEMA 23 for my Y axis of my CO2 laser and was wondering what wire configuration you think it'd be best? Just a standard. Or should I do Series Configuration or parallel configuration?
09:00 AM JT-Cave: pcw--home, is it possible with mesaflash to find out what board is at -addr 10.10.10.10 without knowing the board name?
10:03 AM lcnc-relay: <jaco111#0> Hi, i am trying to convert a Chinese CNC (with a studio nc breakoutboard) to LinuxCNC with a Mesa 7i96s card. No luck yet in getting it moving. When i look at the original wiring, the breakoutboard has a P- and D- wire for each of the 3 axis. The 6 + wires are all wired together, going to the stepper drivers. Is that normal ?
10:04 AM JT-Shop: P might be pulse and D might mean direction
10:17 AM pcw--home: JT-Cave: mesaflash --device ether --addr 10.10.10.10
10:17 AM JT-Shop: thanks
10:17 AM JT-Cave: that gives a much more informative message than device not found
10:17 AM JT-Cave: making mesact check the device before flashing now
10:19 AM JT-Shop: so the check for ethernet card could fill in the board and address?
10:19 AM Tom_L: sure could
10:20 AM Tom_L: you could test the 192 addy if you wanted but that's up to you
10:20 AM Tom_L: both are known
10:20 AM Tom_L: unless someone used some odd 10.xx addy
10:21 AM Tom_L: later..
10:21 AM JT-Shop: so a find ethernet board button?
10:21 AM Tom_L: or just do it on startup, that's up to you
10:22 AM Tom_L: 'no board found'... no harm done
10:22 AM roycroft: assuming that pcw has a documented range of ethernet addresses to assign his boards, you could arp for mac addresses within that range and report them
10:22 AM Tom_L: there would be lots of times no board would be connected yet
10:23 AM JT-Shop: usually it's just 10.10.10.10 and 192.168.1.121
10:23 AM Tom_L: just something to think about
10:23 AM JT-Shop: and > or
10:24 AM Tom_L: hup, both are defaults
10:24 AM Tom_L: yup
10:24 AM JT-Shop: the main focus is to not flash a 7i92 with 7i92t firmware :)
10:24 AM * Tom_L puts his glasses back on
10:24 AM Tom_L: they should identify different
10:24 AM Tom_L: and
10:25 AM Tom_L: the T shouldn't allow a bit file but rather a bin file
10:25 AM Tom_L: you can test that too
10:25 AM JT-Shop: if a 7i92t is the selected board only firmware for that board is available to flash
10:26 AM JT-Shop: well actually I "think" the flash would fail if the wrong board is selected but I wan't to give the user a meaningful reason
10:27 AM * JT-Shop gets back to servicing chainsaws and clearing the work table for a fork
10:31 AM pcw--home: I just changed mesaflash a bit to make it harder to use the wrong name after a forum issue I think mesaflash should also print the list of valid names if an invalid one is used
10:33 AM pcw--home: That is, people try mesaflash --device 7i92tm and the response is "No 7i92tm card found"
10:34 AM JT-Shop: cool
10:34 AM JT-Shop: just when you think you have outsmarted new bee's you get surprised :)
10:46 AM pcw--home: Easiest thing would be to print the entire card name list on an invalid card name
10:47 AM pcw--home: nicer would be to munge 7i92tx to 7i92t etc but that's quite tricky to get right in the general case
12:25 PM LUNADA_DESKTOP: pcw--home i ended up downloading the servo drive software and changing the encoder output to a smaller value. seems to have done the trick
12:28 PM pcw--home: Those were pretty crazy numbers even for a mill
12:37 PM LUNADA_DESKTOP: agreed. no idea why they had it like that
02:43 PM LUNADA_DESKTOP: this z axis is kicking my ass. can't seem to get it tuned. there's an air cyl on it to counteract the weight, but it doesnt seem to move air fast enough and causes a huge spike in the following error
02:43 PM LUNADA_DESKTOP: no idea how they were able to get it tuned with the old controller. guessing it wasnt very good
02:47 PM XXCoder: its bit amusing that most ,achinist channels make tools for shop
02:50 PM lcnc-relay: <jaco111#0> So, struggling with my cnc conversion 😦 I had thought getting it moving would be the easy part but it is not moving at all. I have a Mesa 7i96s board, and DMA860H stepper drivers. I used pncconf. and went through all the steps to configure the 3 axis. No errors when i start LinuxCNC but it just does not move, even tho on the screen the coordinates are moving when i jog. The machine is not moving. No special sou
02:57 PM JT-Shop: are the steppers locked up when power is on or do they free turn?
03:00 PM lcnc-relay: <jaco111#0> THey are locked when power is on
03:03 PM JT-Shop: when you change direction does it move one step
03:04 PM JT-Shop: and of course you might want to try my Mesa Configuration Tool II
03:05 PM lcnc-relay: <jaco111#0> Hmm, if what you mean is jog minus and then jog positive, it makes no move or sound
03:07 PM JT-Shop: yeah just trying to see if the step and direction are reversed
03:07 PM JT-Shop: are you using the Axis GUI?
03:08 PM lcnc-relay: <jaco111#0> Yes
03:10 PM JT-Shop: Machine > Show Hal Configuration to see the pins
03:10 PM JT-Shop: oh yes does the DRO and the backplot change when you jog?
03:12 PM lcnc-relay: <jaco111#0> yes, DRO changes according to my jogging
03:13 PM lcnc-relay: <jaco111#0> Which pins exactly are we interested in the Show Hal Config ?
03:20 PM JT-Shop: well if the DRO changes then I assume the stepgen is changing
03:21 PM JT-Shop: but you can put stepgen in the watch window and see if it moves
03:22 PM lcnc-relay: <jaco111#0> yes there are values that are changing in stepgen
03:26 PM JT-Shop: sounds like a wiring issue now
03:30 PM pcw--home: Common causes of no steps are:
03:31 PM pcw--home: 1. Too short step lengths (double the drives minimum time to check)
03:31 PM JT-Shop: I was just going to say that... back to making a fork seal driver
03:31 PM pcw--home: 2. wiring issues (dont use differential drive with OPTO coupled inputs)
03:32 PM pcw--home: 3. Latency issues so bad you get a watchdog timeout
03:34 PM pcw--home: 4. step drive enables wired wrong (these are typically _disable_ inputs )
03:35 PM pcw--home: that is, they disable the drive if current flows into these pins
03:35 PM lcnc-relay: <jaco111#0> ENA are not wired on the stepper drivers
03:36 PM lcnc-relay: <jaco111#0> Hmm, too shorts step lengths ? So that is a config issue, right ?
03:37 PM lcnc-relay: <jaco111#0> That is possible. I know from the manufacturer that for X and Y i need 0.011976 mm/pulse
03:38 PM pcw--home: That's scaling. The step length is drive interface timing
03:39 PM lcnc-relay: <jaco111#0> Is that something that i need to configure ?
03:39 PM pcw--home: yes
03:40 PM lcnc-relay: <jaco111#0> I just install Mesa CT 2
03:41 PM lcnc-relay: <jaco111#0> Is it the StepGen settings ? With Custom (as my DMA860H is not listed)
03:43 PM pcw--home: I would use double the Drives spec (drive says 2.5 usec so use 5 usec or 5000)
03:44 PM pcw--home: Also use single ended wiring (PUL+,DIR+ to +5V. PUL- to 7I96S Step-, DIR- to 7I96S Dir-)
03:45 PM JT-Shop: use the TB6600
03:45 PM lcnc-relay: <jaco111#0> We are talking about this right ? https://wiki.linuxcnc.org/cgi-bin/wiki.pl?Stepper_Drive_Timing
03:47 PM lcnc-relay: <jaco111#0> So only 2 wires per stepperdriver to the mesa board ? And all the 6 Positives to a Wago that is connected to the input 5V ?
03:49 PM lcnc-relay: <jaco111#0> Leadshine USA Analog M860H 80VAC 7.2A 1500 1500 8000 5000 Rising Edge
03:49 PM lcnc-relay: <jaco111#0> Step time and Step space is there set as 1500
03:50 PM pcw--home: Each 6 pin step/dir connector on the 7I96S has +5V so its usually best to use those (to allow easy swapping)
03:50 PM pcw--home: I would change the 1500 to 5000
03:52 PM lcnc-relay: <jaco111#0> ok
03:53 PM lcnc-relay: <jaco111#0> Ok, So the PUL+ and DIR+ wires do not go to these labeled ports but to the 5V port
03:54 PM pcw--home: The per 6 pin connector 5V (that's there for this purpose)
03:55 PM -!- #linuxcnc mode set to +v by ChanServ
03:56 PM Tom_L: pcw--home, do you recall when --device ether was introduced to mesaflash?
03:56 PM Tom_dev: ETH device 7I96S at ip=10.10.10.10
03:56 PM Tom_dev: pretty cool
03:57 PM Tom_dev: No ETHER board found
03:57 PM pcw--home: Its been there a long time
03:57 PM Tom_dev: JT-Cave, pretty easy to add that to your tool i think
03:57 PM JT-Shop: yup
03:57 PM Tom_L: i thought it had, just never used it much
03:57 PM pcw--home: I just noticed I busted card naming in 3.5.2 so best to use 3.5.3
03:58 PM Tom_L: just updated to 3.5.3
03:58 PM Tom_L: since you'd mentioned adding board names
03:58 PM pcw--home: (the stricter name checking I added depended on a list that was missing some cards)
03:59 PM pcw--home: it worked before even it the card was not in the list
03:59 PM pcw--home: added 7i97t 7i94t 7i76eu
03:59 PM LUNADA_DESKTOP: so in my z axis, the following error looks ok at lower speeds, but it seems like it cant keep up in higher velocities
04:00 PM LUNADA_DESKTOP: the following error increases with velocity. not cnaging directions or anything, just a start and stop
04:00 PM LUNADA_DESKTOP: accel settings are pretty slow
04:00 PM LUNADA_DESKTOP: changing*
04:01 PM LUNADA_DESKTOP: any ideas?
04:01 PM pcw--home: Have you tuned FF1? the the most important parameter for velocity mode drives
04:01 PM LUNADA_DESKTOP: i'm talking a huge error, like .1" or more
04:01 PM LUNADA_DESKTOP: yea
04:01 PM LUNADA_DESKTOP: it's pretty high
04:02 PM lcnc-relay: <jaco111#0> Peter you rock !!
04:02 PM lcnc-relay: <jaco111#0> Its moving 🙂
04:02 PM Tom_L: pcw--home, so it shows the difference say between 7i92 & 7i92t so jt could pair the firmware in his tool
04:03 PM pcw--home: yes, if you parse the results of mesaflash --device ether --addr someipaddr you can get the name
04:04 PM pcw--home: I guess it could put the name in quotes or something to make it easier to parse
04:04 PM Tom_L: if you keep the format it shouldn't be hard to find
04:04 PM Tom_L: ETH device xxxx
04:05 PM pcw--home: Yeah
04:06 PM pcw--home: "the following error increases with velocity" sounds like FF1 is wrong
04:06 PM Tom_L: so in the background on mesact startup he could interrogate both the 192.xx addy as well as the 10.xx addy searching for a card then auto fill some parameters
04:07 PM Tom_L: or however he chooses to do it
04:07 PM LUNADA_DESKTOP: i'll keep trying.
04:08 PM pcw--home: Too much FF1 means the motion will be ahead of the command, too little means the motion will lag the command
04:48 PM LUNADA_DESKTOP: https://pasteboard.co/IPLldOvNr95B.jpg
04:48 PM LUNADA_DESKTOP: cant tell if the downward spike is during decel or accel
04:48 PM LUNADA_DESKTOP: to me it looks like too much ff1
04:59 PM pcw-home: it looks ok until it starts to decelerate is this a Z down move?
05:02 PM LUNADA_DESKTOP: yes
05:02 PM LUNADA_DESKTOP: i can put the air cyl back on
05:11 PM pcw-home: It look more like the drives torque capabilities are being exceeded that a tuning issue
05:20 PM LUNADA_DESKTOP: it's possible, but man... it's slow
05:20 PM LUNADA_DESKTOP: but it's always been pretty slow
05:24 PM pcw-home: maybe needs different gearing...
05:29 PM LUNADA_DESKTOP: yea i'd rather have the accel than the velocity
05:47 PM pcw-home: If this is a router like design I would expect Z has a small travel so accel would be more important
06:07 PM LUNADA_DESKTOP: it's a router but it has like a 30+" z
06:08 PM LUNADA_DESKTOP: pcw-home i'm beginning to think the rpms are limited in the servo software. i dont recall if the z was speed limited with the old controller
06:09 PM LUNADA_DESKTOP: the accel was so terrible that it never really saw max velocity in most cases
06:09 PM LUNADA_DESKTOP: it's fine until i exceed 500ipm
06:10 PM LUNADA_DESKTOP: then it cant keep up. it's a hard cutoff at 500ipm
06:28 PM XXCoder: yay joe is back to minure machoine machining
06:28 PM XXCoder: https://youtu.be/trtdzeKK1qU
06:39 PM LUNADA_DESKTOP: weird, i set my max velocity for the z axis in the ini file but it doesnt seem to be limiting it
06:47 PM LUNADA_DESKTOP: im an idiot. fixed.
07:06 PM javabean_ is now known as javabean
07:30 PM LUNADA_DESKTOP: i got the following error pretty good on the whole machine but it's a little shaky when i jog
07:31 PM LUNADA_DESKTOP: lowering p doesnt seem to fix it
08:07 PM Tom_L: LUNADA_DESKTOP, what about [TRAJ] MAX_LINEAR_VELOCITY & MAX_LINEAR_ACCELERATION
08:09 PM Tom_L: also in the [DISPLAY] section
08:09 PM Tom_L: DISPLAY may be where jogs are controlled
08:10 PM Tom_L: also, looks like 2.9 is listed as current stable now
08:10 PM Tom_L: woot!
08:13 PM Tom_L: also mesaflash 3.5.3 is out
08:54 PM rene-dev: andypugh good job!
08:55 PM Tom_L: we will never know the work behind the numbers
11:39 PM solarwind: PSA: these carbide inserts are LEGIT: https://www.aliexpress.com/item/1005005486105971.html
11:40 PM solarwind: that is, all the DESKAR ones
11:40 PM solarwind: I think they're the original manufacturer and their prices are 1/5th to 1/10th of the big brand names but they hold their tolerances and last a long time
11:41 PM solarwind: They actually have some specs and cutting parameter data and they indeed perform as advertised. No visible wear so far running at 180m/min on a face mill
11:45 PM Connor: Silly question, Does LinuxCNC invert the direction of a stepper/servo when used in tandem mode? (Or perhaps PnP conf did it for me?) I was expecting to have to invert the direction of the servo, but, it was already going the right direction.