#hazzy Logs

Aug 15 2020

12:07 AM Not-61bf: [02qtpyvcp] 07adargel synchronize pull request 03#48: ENH: Add conversational drilling GCode generator. - 13https://git.io/JJ5rc
04:19 AM Joco[m]: hazzy: got a test done. Cutting the perimeter of a rectangle. X and Y dimensions are within 0.02mm
04:25 AM Joco[m]: It's not perfect and I need to spend some more time on tuning. Not convinced I have an optimal PID yet.
09:06 AM JT-Cave: TurBoss, got sunrise and sunset started on the DotStar https://github.com/jethornton/dotstar
09:13 AM TurBoss: on the pi?
09:13 AM TurBoss: great
09:29 AM JT-Cave: yes on the Rpi
09:30 AM JT-Cave: checked it with a LUX meter and it is very linear in brightness
09:35 AM TurBoss: nice
09:40 AM JT-Cave: now I need to calculate the delay time based on the duration of dawn to sunrise and the number of leds in the string
11:50 AM roguish[m]: JT-Cave: what about over cast days ????
11:53 AM roguish[m]: JT-Cave: get a light sensor, and adjust your lighting according to ambient. can bound by time of day.
12:11 PM TurBoss: JT-Cave: some systems start/stop 3 min latter each day to compensate date
12:12 PM TurBoss: my system was used to raise canaries
12:18 PM TurBoss: https://github.com/TurBoss/LainBot/blob/master/main.py#L80
12:18 PM TurBoss: check this python scheduler
12:19 PM TurBoss: this is a bot the main atraction of #lain:matrix.org
12:19 PM TurBoss: it post random Lain picture once a day
01:15 PM JT-Cave: roguish[m], not sure I understand why to adjust inside lights to outside?
01:16 PM JT-Cave: TurBoss, the RPi gets astral information updated each new day at 24:05 local time
01:16 PM TurBoss: amazing!
01:17 PM JT-Cave: def updateSun(self):
01:17 PM JT-Cave: self.todaySun = sun(PBMO.observer, date=datetime.today(), tzinfo=PBTZ)
01:17 PM JT-Cave: PBTZ = pytz.timezone('US/Central')
01:17 PM JT-Cave: PBMO = LocationInfo("Poplar Bluff", "Midwest", "US/Central", 36.873512, -90.488008)
01:18 PM JT-Cave: # show todays sunlight
01:18 PM JT-Cave: todaySun = daylight(PBMO.observer, date=datetime.today(), tzinfo=PBTZ)
01:18 PM JT-Cave: todayLight = todaySun[1] - todaySun[0]
04:37 PM roguish[m]: another toastie day. 106 right now.............
05:06 PM hazzy-m: roguish: that's real toasty!
06:24 PM roguish[m]: hazzy: topped out at 109.3 on my weather station..............think it's a record here..................
07:16 PM Joco[m]: this is screwy. I am now getting a pretty consistent 0.12mm to large on X. Y is always within 0.005mm
07:17 PM * Joco[m] is feeling a bit stumped
07:38 PM Joco[m]: especially since the graphing of the following error does not show this level of deviation
08:24 PM hazzy-m: could there be play in the spindle housing?
08:24 PM hazzy-m: or the z?
08:25 PM Joco[m]: wouldn't that impact Y as well?
08:25 PM Joco[m]: and its REALLY consistent. 0.12mm bang on each time
08:29 PM hazzy-m: yes, it would probably impact Y at least a little
08:31 PM Joco[m]: I'm taking light cuts for the finishing pass. 0.25mm cut in aluminium
08:32 PM hazzy-m: are you using any backlash comp now?
08:32 PM hazzy-m: If the comp effects the cmd'ed pos then it might be looking for the comp value to show up in the linear scale feedback...
08:33 PM Joco[m]: no - turned off all backlash and screw mapping
08:34 PM Joco[m]: even went back to a nominal scale for steps/mm
08:34 PM Joco[m]: Y is really good and X is that 0.12mm over size
08:35 PM Joco[m]: lol - i even checked the gcode for a bug but that calc's as the correct length
09:01 PM Joco[m]: ok - back to basics. Did a travel distance test over 100mm. Using a DTI. With the scales it should have delivered 100mm within 0.01mm or better. Its consistently travelling 0.07-0.08mm to far.
09:03 PM Joco[m]: think I need to hook into the encoder position and see what it is saying.
11:31 PM Joco[m]: Might have solved the “to large on X” problem
11:36 PM Not-61bf: [02qtpyvcp] 07KurtJacobson synchronize pull request 03#47: ENH: Add support for virtual keyboard and numpad - 13https://git.io/JJ9sp
11:42 PM Not-61bf: [02qtpyvcp] 07KurtJacobson commented on pull request 03#47: ENH: Add support for virtual keyboard and numpad - 13https://git.io/JJdlQ
11:43 PM hazzy-m: Joco: what was it?
11:44 PM Joco[m]: hazzy: f’ing glass scales are not delivering the correct counts per mm.
11:45 PM hazzy-m: well that would explain it!
11:45 PM Joco[m]: Nominally there should be 200 counts per mm. There is really 199.xxx (long decimal number)
11:46 PM Joco[m]: So have been playing with the encoder scale to get a better fit.
11:46 PM hazzy-m: well that's easier to fix than mechanical issues