#hazzy Logs

Jul 09 2020

12:53 AM Joco[m]: SMB? Sounds fiddly.
01:34 AM TurBoss: surface mount components
01:35 AM * TurBoss uploaded an image: Captura de pantalla de 2020-07-09 08-25-22.png (14KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/SMDJhmXAXEtPEEQdQCtuCdHw >
05:18 AM JT-Cave: morning
05:22 AM Joco[m]: hi JT
06:03 AM hazzy-m: morning
06:04 AM Joco[m]: Morning.
06:06 AM JT-Cave: going to be another hot day here in Swamp East Missouri
06:07 AM Joco[m]: Not having that issue here. 9C
06:09 AM JT-Cave: 24°C right now at sunrise
06:10 AM JT-Cave: time to clean chicken poop
06:10 AM Joco[m]: Oh - finally got screw compensation working on X axis. Had to use a type 0 file. Type 1 just did not seem to work. Now I need to move from 6 samples to 200 odd
06:21 AM * TurBoss uploaded an image: VectorImage_2020-07-09_011145.jpg (353KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/uklYdkFXUcvnBvFpQikWqDxP >
06:21 AM TurBoss: new ribbon for my avr programer
06:21 AM TurBoss: cat eat previous one
06:22 AM Joco[m]: Evil cat. Lol
11:55 AM * hazzy-m uploaded an image: ima_8fc6a60.jpeg (81KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/ztRSzLgRgwfAxxehsklljqWH >
11:56 AM hazzy-m: One of this mornings projects. Got it done while doing phone support
12:02 PM * JT-Cave has a Sercos module that can't talk to the drives :(
12:43 PM -!- #hazzy mode set to +o by ChanServ
03:44 PM Joco[m]: hazzy: what does the project do?
03:44 PM Joco[m]: some kind of control interface?
03:45 PM TurBoss: it read plc stats from an api
03:45 PM TurBoss: or looks like
03:50 PM TurBoss: just finished polishing some corners of mine
03:51 PM * TurBoss uploaded an image: Captura de pantalla de 2020-07-09 22-40-39.png (38KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/fvaFcRduIetDDeZrmJjyFDBE >
03:52 PM * TurBoss uploaded an image: Captura de pantalla de 2020-07-09 22-42-22.png (52KB) < https://matrix.org/_matrix/media/r0/download/jauriarts.org/KIJDZihKPDqiuUFsfZANQGoq >
03:53 PM TurBoss: > * <@freenode_JT-Cave:matrix.org> has a Sercos module that can't talk to the drives :(
03:53 PM TurBoss: how is that?
03:54 PM TurBoss: does it have any model number?
04:05 PM hazzy-m: Joco: yes, TurBoss was spot on. Reads machine state info from the PLC on our folders
04:13 PM Joco[m]: cool
04:54 PM JT-Cave: TurBoss, http://www.plctalk.net/qanda/showthread.php?t=125522
08:14 PM Joco[m]: cool - I get to try and make my glass scale talk to linuxcnc this weekend.
08:16 PM Joco[m]: Or get a rotary encoder with a wheel on an arm to measure table travel. Will see what ends up being easier. I do have a half formed idea about using the 5um scales to provide some form of feedback on actual position to support some error adjustment factor in how the steppers operate.
08:16 PM Joco[m]: Once the screw is fully mapped that might not be necessary I guess
08:17 PM Joco[m]: * Once the screws are fully mapped that might not be necessary I guess
09:25 PM hazzy-m: Joco: you could just leave the scales on and use them for feedback ...
09:26 PM Joco[m]: I have started to think about that. Need to get an add on card for the encoders.
09:26 PM hazzy-m: not sure how hard that would be to get tuned to you would not get oscillations
09:31 PM Joco[m]: I think it comes down to how much impact you let the scale feedback have on a PID loop. I have been reading how others have done it.
09:35 PM hazzy-m: yes, since you are using steppers you don't need fast feedback, so I think it should not be very difficult to just have it correct for long term deviation from CMD pos, cause by lost steps or lead error
09:36 PM hazzy-m: maybe could even apply the pos error as an external offset
09:36 PM Joco[m]: Yup. And then if you have a detailed screw map that will help as well
09:37 PM hazzy-m: correct
09:37 PM hazzy-m: you will have the most accurate little mill on the planet lol
09:38 PM Joco[m]: And I suspect an external offset approach might be the least intrusive and least jerky approach to applying the error correction
09:40 PM Joco[m]: Heck I could see the potential in the screw compensation from just 6 position samples
09:44 PM hazzy-m: the cool thing about using external offsets is you could enable/disable lead error comp from gcode, and only use it for finishing passes
10:27 PM Joco[m]: <hazzy-m "the cool thing about using exter"> Now that is SWISH.
10:36 PM Joco[m]: I might well be sold. Maybe this weekend will be re engineering some things to allow me to get my scales back on. Even thought only 5mu under the model we have been discussing that will be just fine with steppers. If o get 0.02mm accuracy on parts I will be a very happy camper.
11:26 PM -!- #hazzy mode set to +o by orwell.freenode.net