#hazzy Logs
Jul 10 2020
#hazzy Calendar
12:03 AM -!- #hazzy mode set to +o by ChanServ
03:29 AM Joco[m]: <hazzy-m "the cool thing about using exter"> Been thinking about this some more and reading up on eoffsets. If I undertand things correctly if I can get a running delta between scales and the position the steppers think it is at then could use that delta to drive the eoffsets
03:32 AM Joco[m]: Which would mean - again in theory - that holding a 0.02mm or better tolerance could be VERY achievable,
03:33 AM Joco[m]: I have already checked the response rate on the scales. 60m/min should be plenty fast enough for anything I'm doing on this mill
05:05 AM JT-Cave: morning
08:09 AM hands[m]: serious ouch!
02:08 PM TurBoss: Hello
02:13 PM hands[m]: howdy TurBoss
02:13 PM TurBoss: Hi hands
02:14 PM TurBoss: I don't know if vertical probebasic is finished yet
02:15 PM hands[m]: would be happy to help... But the docs suggested it was "built in" but I couldn't find a way to kick it off hence the ask..
02:16 PM TurBoss: let me check
02:18 PM TurBoss: current status unusable
02:19 PM TurBoss: did you git clone probe_basic from github?
02:22 PM TurBoss: we have to organice it currently there is a VCP name probe_basic_vertical
02:22 PM TurBoss: some files in probe_basic
02:23 PM TurBoss: a mess
02:23 PM TurBoss: i don't know wich ones are the latest
02:24 PM TurBoss: Lcvette: may know
02:24 PM Joco[m]: hazzy: how on earth did that shaft crack?
02:25 PM TurBoss: :D
02:27 PM hands[m]: I did an install from 0.28
02:28 PM hands[m]: its in a VM for testing and validation.. I'll uninstall 0.28 and do an install for dev off a fork..
02:29 PM TurBoss: cool
02:29 PM TurBoss: if you need some help please just ask
02:29 PM hands[m]: I'll have to come up with a build and deploy cycle for my machine so as I learn and correct things I can push them over to the actual machine I am building/retrofitting
02:30 PM * hands[m] uploaded an image: IMG_3099.jpeg (1840KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/rpVMgrifWugyvabJlARGoAjA >
02:30 PM TurBoss: niceee!!!
02:30 PM hands[m]: should be fun 🙂
02:30 PM TurBoss: sure!
02:31 PM hands[m]: thanks for checking and pointing me in the right direction
02:31 PM TurBoss: you welcome
02:32 PM Joco[m]: hands: nice setup. What is the machine you are retro fitting?
02:38 PM TurBoss: inst a skyfire
02:40 PM hands[m]: yah a Skyfire.. gave up on the Adtech controller.. just nothing but frustration
02:40 PM hands[m]: machine has great bones but the controller is trash..
02:41 PM roguish[m]: hands: hey, nice looking machine. what is it ?
02:41 PM hands[m]: Skyfire SVM-2.. (think 1/2 way between a Tormach 770 and 1100)
02:46 PM roguish[m]: cool
02:46 PM roguish[m]: are you aware that Tormach uses Linuxcnc? with there own GUI.
02:47 PM roguish[m]: and MESANET originated boards.....
02:47 PM Joco[m]: Just looked them up. Heh the new ones can be optioned with mesa controllers and linuxcnc.
02:47 PM hands[m]: Oh yah, known that for quite a while.. My background is more in Mach 3/4 & ESS.. so this is a touch new..
02:47 PM Joco[m]: Specs read well
02:48 PM hands[m]: worst choice I made was NOT picking the LinuxCNC controller.. but I thought the Industrial controller would be "better"...
02:48 PM Joco[m]: hands: i like things i can hack. :-)
02:48 PM hands[m]: Delivery is SUPER slow and quality is not great.. It is more machine than anyone else for the price, but you got to be willing to get your hands dirty tuning it... Mine came WAY out of spec on alignment..
02:49 PM Joco[m]: Sounds familiar. Alot of the “budget” stuff from china is really a kitset.
02:50 PM hands[m]: @joco yah, 20 years in Enterprise Custom App dev.. It was the need to hack that caused my frustration and final giving up on the controller... Their implementation of Macro-b was trash.. and the firmware updates to fix the bugs I found broke the spindle...
02:51 PM hands[m]: kitset, good word...
02:52 PM Joco[m]: hands: lol. Similar drivers. But i couldnt justify the approx 20,000 nzd that usd price turns into. So i converted a BF20
02:52 PM Joco[m]: Well - still converting. I think the mill is the hobby
03:13 PM * TurBoss uploaded an image: Captura de pantalla de 2020-07-10 22-03-22.png (55KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/llYwKXDdWMWWndJYtAhINdsj >
03:51 PM * TurBoss uploaded a video: linear_atc.webm (1157KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/SGAHUPOKAnIHmIUvtOTFuxsN >
03:55 PM TurBoss: the most important part got cut
03:55 PM TurBoss: lol
03:58 PM * TurBoss uploaded an image: linear_atc.gif (33KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/NYHzewPAGCDVjXROUHLBKgRr >
04:02 PM Joco[m]: hazzy: for when you are about, any reason the linuxcnc.command stat attribute is in the STATUS exclusion list?
05:05 PM hands[m]: What am I missing regarding the conventions for working on a template? I can edit probe_basic and probe_basic_lathe but I can't figure out why I can't load probe_basic_vertical.. I thought it was a naming thing so I renamed all the files in the folder... But I feel like editvcp is looking for some kind of convension for file naming that I am failing to duplicate..
05:10 PM TurBoss: al files in probe_basic_viertical need a refactor
05:10 PM TurBoss: * all files in probe_basic_viertical need a refactor
05:11 PM TurBoss: they are called probe_basic.* and should be named probe_lasic_vertical.*
05:11 PM TurBoss: I can work on that
05:11 PM TurBoss: do you have a fork?
05:16 PM hands[m]: not yet, I just did a pull, I was going to make sure I had things mentally worked out then make a fork... I figured I would be blowing things a way a few times..
05:16 PM TurBoss: vertical is not useable
05:16 PM TurBoss: yet
05:16 PM TurBoss: xD
05:17 PM TurBoss: ok I'll create a branch for that
05:17 PM hands[m]: I did rename everything probe_basic_vertical.* . and went into probe_basic_vertical.yaml and repointed ui and qss
05:17 PM hands[m]: figured kind of classic porting process, step one get stuff to load, then start in with the fixing..
05:18 PM TurBoss: in the probe_basic cirectory there are vertical ui files
05:18 PM TurBoss: I'm not sure wich is the lates
05:18 PM TurBoss: * I'm not sure wich is the latest
05:18 PM TurBoss: I'm checking if more fines need changes
05:19 PM TurBoss: just did this for probe_basic_latc yesterday
05:19 PM TurBoss: * I'm checking if more files need changes
05:20 PM hands[m]: is it just a naming convention that the folder and ui files all have to have the same name for editvcp to pick it up.. or is there something I need to do with local entry points or something to get it to be "found" (like in ~/.local/bin/)
05:21 PM TurBoss: you may need to run pip install -e . in the probe_basic root
05:21 PM TurBoss: ```
05:21 PM TurBoss: python -m pip install --ugrade -e .
05:21 PM hands[m]: ahh yes.. I did see the need to add the entry point to the setup file
05:21 PM TurBoss: * ```
05:21 PM TurBoss: python -m pip install --upgrade -e .
05:21 PM TurBoss: ```
05:21 PM hands[m]: an upgrade may be what I am missing to wire that up
05:21 PM TurBoss: yes
05:22 PM TurBoss: cool
05:22 PM hands[m]: boom that was it
05:22 PM TurBoss: it does something?
05:22 PM TurBoss: \o/
05:22 PM hands[m]: so entrypoints in the setup.py and then reinstall
05:23 PM TurBoss: probe_basic_vertical
05:23 PM TurBoss: you can also rename the main class
05:23 PM TurBoss: ProbeBasic
05:23 PM TurBoss: to ProbeBasicVertical
05:24 PM TurBoss: I have some things commited but I need Lcvette to review before pushing to master
05:24 PM TurBoss: the same you did but for linear atc
05:28 PM hands[m]: I think at this point I have enough figured out that I need to make my fork and that way I can send pull request to you guys as I get things worked out... Thanks 🙂
05:28 PM TurBoss: :)
05:28 PM TurBoss: we use a branch for feature
05:29 PM TurBoss: ej make vertical work can be a branch
05:30 PM hands[m]: sure thing I'll create a branch on my end vertical_ui a good name for it?
05:30 PM TurBoss: yup
05:30 PM TurBoss: thaaaanks!
05:32 PM hands[m]: thank you guys.. I was mentally planning to learn how to write all this from scratch, glad I learned about the project :-)...
06:47 PM hazzy-m: Joco: not sure how that shaft got cracked, the shaft on the other side of the machine was broken clean in two
06:47 PM * hazzy-m posted a file: 20200708_100836.jpg (3745KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/lEFQwBLQNHsbWSrgAPioSBUW >
06:48 PM * hazzy-m uploaded an image: ima_3d23242.jpeg (3374KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/fWyiaRIFqwOqqDOjTQzwrzbg >
06:52 PM hazzy-m: originally I thought it had just fatigued, but we replaced the shafts, servos and gearboxes and put everything back together, and now one of the servo drives is acting up, slamming the beam around very violently, to the bad drive might have just twisted the shafts to failure
06:58 PM Joco[m]: there must be some serious torque in the mix. I have had that acting up for a bit to cause a fatigue failure
06:58 PM Joco[m]: ?
07:00 PM Joco[m]: about to start testing how to make the MPG input read a glass scale for screw mapping. this could be easy of a monumental failure.
07:05 PM Joco[m]: F'ing AWESOME. I'm getting raw counts on halshow!!!
07:05 PM hazzy-m: Joco: SWEET!!
07:06 PM Joco[m]: scale is a bit borked so need to solve that I guess. But just the fact it is reading the TTL feed is a big step
07:06 PM * Joco[m] does a wee jig
07:16 PM hazzy-m: Joco: `linuxcnc.command` is ignored as it is the command interface, not a status item
07:19 PM Joco[m]: oh - just the way the docs read is it returns the currently executing command. Not it "sends/executes" a command. Doc's not too clear in that case,
07:20 PM Joco[m]: ```
07:20 PM Joco[m]: command
07:20 PM Joco[m]: (returns string) - currently executing command.
07:21 PM Joco[m]: this being in linuxcnc.stat
07:21 PM Joco[m]: which is different to linuxcnc.command
07:22 PM Joco[m]: unless I have completely lost the plot - quite possible by the way
07:24 PM Joco[m]: <hazzy-m "Joco: `linuxcnc.command` is igno"> snap - I'm referring to linuxcnc.stat.command.
07:26 PM Joco[m]: is that included in the status channel? it looked on first look that when the status channel setup is scanning through the artributes under linuxcnc.stat that command was being ignored. Am i reading things wrong?
07:45 PM hazzy-m: Joco: good eye!
07:46 PM hazzy-m: I don't know why I excluded command
07:46 PM hazzy-m: fix coming up!
07:46 PM hazzy-m: it even seems to show commands running in a remap
08:03 PM Joco[m]: cool - cose this could be a useful little hook
08:04 PM Joco[m]: <hazzy-m "I don't know why I excluded comm"> you were probably in the zone around linuxcnc.command
08:06 PM hazzy-m: must have been
08:06 PM hazzy-m: I did the same with debug
08:07 PM Joco[m]: i'm running 5mu scales so I I guess my scale needs to be 50 to get sensible readings off position
08:08 PM Joco[m]: I wonder how I can tell how accurate they are. Hmmmm
08:10 PM Joco[m]: probably just going to have to mount it and run some measurements using 123 blocks and a DTI
08:10 PM Joco[m]: right ... next to mount X scale.
08:37 PM Joco[m]: crap - old mounting approach wont work anymore due to some changes made when converting to CNC. Gonna have to come up with a new system
09:38 PM * hands[m] sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/wnAmehRmKUQlmrMHLMUoQdJB >
09:39 PM hands[m]: (so in qtdesigner, double click on a control.. or right click the control and select "Edit Widget Rules")
09:39 PM Joco[m]: erk - that's a hazzy job I think. Or TurBoss
09:42 PM TurBoss: does that only happend on vertical?
09:43 PM hands[m]: happens on normal as well.. thought it might be something in vertical so I loaded up probe_basic to see the difference in the config and poof.. (that error is actually from probe_basic)
09:43 PM hazzy-m: hands: my bad, I bushed a bad commit
09:43 PM hazzy-m: should be fixed on latest master
09:44 PM hands[m]: ok I'll pull and resync my fork.. and give it a try.. thanks
09:44 PM hazzy-m: hands: https://github.com/kcjengr/qtpyvcp/commit/f93e2cb633ae08738fff70052c056aaa4358397d
09:44 PM hazzy-m: sorry about that
09:45 PM hands[m]: all good man.. lots of moving parts
09:46 PM hazzy-m: and too little sleep ... lol
09:48 PM hands[m]: hey widget rules!... thanks hazzy!
09:48 PM hazzy-m: sweet!
09:50 PM TurBoss: great
09:55 PM TurBoss: hands: so it boots?
09:55 PM TurBoss: I mean with linuxcnc?
10:00 PM hands[m]: oh I haven't gotten that far 😉
10:00 PM hands[m]: I have about 20 issues that are being thrown by QTdesigner in the commandline
10:01 PM hands[m]: I want to clear those before I try to get it to boot in linuxcnc
10:02 PM TurBoss: cool
10:02 PM * hands[m] sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/PmHomPkRWrOJropHduWSvTuf >
10:03 PM TurBoss: yes that stuff can be old
10:03 PM TurBoss: :=
10:04 PM hands[m]: Its all just XML so once I figure out the pattern of what is "right" I can knock it out in VScode...
10:14 PM hands[m]: ahhh its all coming clear... DROWidget replaces all of these old ways to display the DRO... I guess you use to have rules to set the DRO and you encapsulated that... Sorry for the running commentary while I learn... feel free to tell me to shut-up..
10:14 PM TurBoss: :-)
10:18 PM Joco[m]: nah - you going way faster then me. I was all like "what the fuck?" lol Actually I still am. :-p
10:20 PM Joco[m]: hows this for a #8 wire setup ..
10:20 PM hazzy-m: hands: yes! The new DRO widgets are a lot cleaner. Feel free to replace the old status labels with them
10:21 PM * Joco[m] uploaded an image: ima_959db99.jpeg (379KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/SdgNbbYfqLxLkaCoJYEODRDU >
10:21 PM Joco[m]: note the gclamps to mount the scale - lol
10:22 PM Joco[m]: but on the plus side ...
10:22 PM * Joco[m] uploaded an image: ima_fbfa9e8.jpeg (146KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/HaqYWQITsWCIoymdnrVTkOYj >
10:22 PM Joco[m]: 300mm move and its 0.04mm over and ...
10:22 PM * Joco[m] uploaded an image: ima_4875b14.jpeg (180KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/kMPXbATatuvweRPuhhETrDGV >
10:22 PM hazzy-m: Joco: that works!
10:23 PM Joco[m]: even though sign is backwards 300.04mm on the encoder reading from the scale
10:23 PM Joco[m]: * even though sign is backwards 300.04mm on the encoder reading from the scale. The "position" pin
10:24 PM Joco[m]: so that would suggest we have the correct scale :-)
10:24 PM hazzy-m: Joco: your field voltage is a bit off ...
10:25 PM Joco[m]: lol - 23.99949 for a nominal 24V? Not tooo bad
10:25 PM hazzy-m: haha, that's a lot of decimal places
10:25 PM Joco[m]: But I could tweak the trim pot on the PSU
10:26 PM Joco[m]: personally I'm still in shock that I got that scale to actually work via the MPG
10:26 PM hazzy-m: That has to be rounding error, that's got to be only an 8bit ADC
10:26 PM Joco[m]: probably
10:27 PM Joco[m]: I'm gonna need to get an add on card to give me some proper encoders.
10:27 PM hazzy-m: it's just a linear encoder , not you need to hook it up as an MPG, and you can control your machine by moving the head lol
10:27 PM hazzy-m: oh, is that not a hardware encoder input?
10:27 PM Joco[m]: oh - just meant was via the ecnoders intended for MPG
10:27 PM Joco[m]: no
10:28 PM Joco[m]: its via the mode 2 MPG encoders on the 7i76
10:29 PM hazzy-m: I see, so a little slow
10:29 PM Joco[m]: I'm stretching things a bit here. First plan is to use this pattern to get screw maps. But want to get proper h/w encoders to support scales on all axis
10:29 PM Joco[m]: with h/w encoders I can start thinking about application of eoffsets
10:30 PM Joco[m]: in the mean time I can use this setup to map screws with a LOT of data points and accuracy.
10:31 PM Joco[m]: Getting all the scales back in place and done in such way that they will not drown in a flood coolant setup
10:38 PM hazzy-m: better plugin docs: https://www.qtpyvcp.com/plugins/
10:38 PM hazzy-m: the spaminator is broken :(
10:42 PM -!- #hazzy mode set to +o by ChanServ
10:42 PM hazzy-m: or maybe the bridge is down ..
10:45 PM Joco[m]: quick - save the spaminator!!
10:45 PM hands[m]: last question for the night.. How does the rule on the Offsets know which axis it is pulling the status on (specifically the tool offset rule).. all 5 controls look identical, I am not following how they know which entity in the tulip is "their" axis offset?
10:45 PM Joco[m]: "All be back"
10:45 PM Joco[m]: ROFL
10:46 PM Joco[m]: * "I'll be back"
10:46 PM hands[m]: I mean I assume it works, but I would have expected an axis signifier in the rule or the control expression or something.
10:47 PM Joco[m]: are you look at code?
10:47 PM Joco[m]: or widget rule?
10:47 PM hazzy-m: hands: it's probably specified in the rules expression, something like `ch[0][<axis_num>]`
10:48 PM Joco[m]: the first rule, is look for a widget rule
10:48 PM hands[m]: ahhhhh ok yah so 0 = X, 1 = Y and so forth
10:48 PM Joco[m]: yup
10:49 PM hands[m]: Joco I was looking at the widget rules... I just didn't notice that ch[][] in the formatting section.
10:49 PM Joco[m]: gotcha
10:51 PM * hazzy-m hits the hay
10:51 PM hazzy-m: gn8
10:51 PM Joco[m]: nite
11:54 PM * TurBoss uploaded an image: Captura de pantalla de 2020-07-11 06-44-56.png (42KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/WIzfpeItbAJfbBGlrxnIObPI >
11:55 PM TurBoss: hands: the super should call ProbeBasicVertcal too