#linuxcnc Logs
Aug 03 2017
#linuxcnc Calendar
12:08 AM Crom: damn VA,, taking thier sweet time..
12:45 AM CaptHindsight: shouldn't ENCODER_SCALE be a positive whole number?
12:46 AM CaptHindsight: his X axis has a negative value of -20000
12:46 AM CaptHindsight: hus Y and Z are both positive
12:52 AM CaptHindsight: he inverted it since he thought he was getting encoder signals going the opposite direction
12:54 AM CaptHindsight: I'd change to velocity mode and get it all working first then see if you really want torque mode
12:58 AM CaptHindsight: http://www.dmm-tech.com/Files/CTADYN4FT-1214E0_P1.pdf according to this doc his motors have absolute encoders
01:02 AM CaptHindsight: http://www.dmm-tech.com/Files/DYN4MS-ZM1-A17a.pdf the manual 76 pages
01:04 AM CaptHindsight: his drive takes the absolute encoder input from the motors and outputs quadrature encoder outputs
01:06 AM CaptHindsight: get the setting in the drive wrong and the encoder output signals to the 7i77 are wrong
01:10 AM IchGucksLive: morning from Germany
01:13 AM IchGucksLive: im off bad weather outside today
02:14 AM CaptHindsight: Adamj12b-mobile: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Tuning_EMC2/HAL_PID_Loops
02:15 AM CaptHindsight: plus you need to be sure your servo drivers are properly configured as well
02:31 AM Deejay: moin
05:00 AM selroc2: test
05:00 AM XXCoder: Error: Irc channel not found
05:00 AM selroc3: Test
05:00 AM XXCoder: Error: Irc channel not found
05:35 AM jthornton: morning
05:36 AM XXCoder: hey
05:38 AM XXCoder: laughed so much on this one,http://cdn.smosh.com/sites/default/files/2017/08/coloring-corruption-kama-sutra.jpg
05:50 AM Deejay: hey
05:50 AM XXCoder: yo
06:48 AM IchGucksLive: evgenii: du wolltest mit einem deutschen reden
06:50 AM evgenii: MAGIG CNC ? :)
07:30 AM IchGucksLive: hi Adamj12b-mobile
07:41 AM IchGucksLive: im off Till later ;-)
08:11 AM adamj12b-web: Good Morning!
08:15 AM JT-Shop: morning
08:33 AM codepython777: Does anyone have any suggestions on how I can pull this motor out of the wheel: https://i.ytimg.com/vi/Xv4UIHgjna0/maxresdefault.jpg ?
08:34 AM archivist: the wheel is the motor
08:34 AM archivist: remove tire
08:35 AM codepython777: archivist: I want to pull out this: https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTs0SYm8-XkpHAgbo-BN-Xr7yoInxyG0pfzdMpf9orAmEBYhtOE
08:35 AM codepython777: The winding core from the wheel - can't get it to pull out...
08:35 AM archivist: that is just the rotor
08:36 AM archivist: the magnetism is quite high, will take some force, I have played with a similar LG washing machine motor
08:37 AM archivist: the bearings may be an interference fit, press out, can be tons of force needed
08:37 AM archivist: not enough detail in pics to see
08:39 AM codepython777: one sec
08:41 AM codepython777: archivist http://imgur.com/a/nctdH
08:41 AM archivist: I have a 3 legged puller that I would use or a press
08:42 AM gonzo_: you're pulling my leg
08:42 AM archivist: why take it apart anyway
08:42 AM Adamj12b-mobile: codepython777 How hard did you have to work to remove the front cover?
08:43 AM Adamj12b-mobile: Also, can you post a picture of the backside?
08:44 AM codepython777: archivist: Can you show me a picture of that tool?
08:44 AM codepython777: Adamj12b-mobile: https://images-na.ssl-images-amazon.com/images/I/71MShNW4wgL._SY355_.jpg
08:45 AM archivist: https://www.machinemart.co.uk/c/pullers-and-seperators/
08:45 AM codepython777: archivist: I want to run some tests on the motor winding. Understand the mechanical assembly.
08:46 AM archivist: looks pressed, may not be removable without damage
08:46 AM Adamj12b-mobile: I would have to guess that under that back cover is a snap ring that is holding the shaft through the bearing.
08:46 AM archivist: you can test the winding in place
08:46 AM codepython777: archivist: it is pressed - https://www.machinemart.co.uk/p/cht227-hydraulic-puller-set/ - how does one attach the middle part to the wheel?
08:47 AM codepython777: archivist: I want it away from the magnets
08:47 AM archivist: in your case drill hole in wheel and that peg pushes it out
08:47 AM archivist: is that star thing a nut or just bling
08:48 AM codepython777: just bling - outside is one piece
08:48 AM codepython777: I could make a hole, I was hoping to not kill the wheel
08:49 AM archivist: the other option is the claws under the rotor spokes
08:49 AM codepython777: The outside wheel is plastic and metal - so heat is not an option either
08:50 AM archivist: push the center, remove cable first if possible
08:50 AM codepython777: we need to pull the center black cylinder, not push, right?
08:50 AM codepython777: unless of course we destroy the wheel
08:50 AM adamj12b-web: If the backside does not come off, then the shaft/bearing on that side would have a light press fit. Use some heat from the back outside to quickly heat the aluminum without heating the bearing too much. Thermal expansion will lessen the interference.
08:51 AM adamj12b-web: Dont want to get too hot or you can kill the magnets
08:51 AM archivist: pulling the rotor off the shaft is the only non damaging option, make tooling to pull on the spokes
08:52 AM archivist: not designed for repair
08:53 AM archivist: what is the distance spokes to bearing spigot below
08:55 AM codepython777: adamj12b-web: there is plastic near that alum piece - hence the heat problem
08:55 AM codepython777: I think archivist is right - pulling the right way is the only option
08:55 AM codepython777: but that needs a new tool
08:56 AM FinboySlick: archivist: Did you catch that cute tiny slanted bed CNC lathe with dual ATC I linked yesterday?
08:56 AM archivist: FinboySlick, no
08:57 AM adamj12b-web: codepython777: I think you are correct. Have to pull it out.
08:57 AM FinboySlick: archivist https://www.youtube.com/watch?v=2jmOk86-jgM
08:57 AM codepython777: https://www.machinemart.co.uk/p/cht227-hydraulic-puller-set/ - this does not help
08:57 AM archivist: pfft, yt, cannot see on this box
08:58 AM FinboySlick: archivist: Oh, right :P
08:58 AM archivist: codepython777, there are various legs, plus make your own
08:59 AM archivist: codepython777, Sykes Pickavant http://www.sykes-pickavant.com/products/pullers-and-separators
09:00 AM archivist: if you were local I have them :)
09:00 AM adamj12b-web: codepython777: I think you will need to make one. You will want to pull aginst the wheel and not the center shaft.
09:00 AM codepython777: There is a bearing on the end of this black cylinder
09:02 AM archivist: there is an inner ring on the rotor that may be accessible if the feet are thin enough
09:03 AM archivist: there was a set of pullers at the army surplus all sorts of interesting shapes
09:04 AM CaptHindsight: adamj12b-web: any luck with the servos today?
09:05 AM adamj12b-web: Hi CaptHindsight YES!!
09:05 AM CaptHindsight: what was the issue/s?
09:05 AM adamj12b-web: PID loop
09:05 AM adamj12b-web: ill explain
09:06 AM CaptHindsight: <--- 3 hrs sleep
09:06 AM adamj12b-web: I tried the 10000 for ferror with no difference. also tried inverting the encoder. all the same results
09:07 AM adamj12b-web: then I was looking at the pid output and error and it never tried to correct max output was +/-0.25mV
09:07 AM CaptHindsight: ah hah
09:07 AM adamj12b-web: so I set the P=500 and it would spin looking for the encoder. if you manually turned it past 0, it would spin the other direction.
09:08 AM adamj12b-web: so i corrected the inverted encoder back to -20000 and upped the gain to 5000
09:08 AM adamj12b-web: this time I get violent oscillation on the servo trying to hold position.
09:09 AM CaptHindsight: adamj12b-web: torque mode is trikier to tune from having no starting place
09:09 AM adamj12b-web: I know its WAY overshooting, but its really trying to lock on. but PROGRESS!
09:09 AM CaptHindsight: is there a reason why you need torque mode?
09:11 AM adamj12b-web: I am totally open to try velocity(Dmm calls it speed) mode. Throughout my searching. I cannot find a good answer what real machines use. it seems 50/50/ So when I saw that Torque was easier to tune, I figured I would start there.
09:12 AM CaptHindsight: it's more difficult to tune than velocity mode
09:15 AM CaptHindsight: http://gnipsel.com/linuxcnc/tuning/servo.html
09:15 AM CaptHindsight: velocity mode ^^
09:17 AM adamj12b-web: Yes. I was reading that last night. I will try velocity tonight as now I have the system "working" and know where the issue was.
09:26 AM CaptHindsight: with the axes on a mill you're more concerned with position and velocity
09:26 AM CaptHindsight: feed rate and position
09:26 AM CaptHindsight: easily solved with velocity mode PID
09:28 AM adamj12b-web: Out of curiosity what would you use torque mode for?
09:29 AM CaptHindsight: a gripper
09:31 AM CaptHindsight: https://forum.linuxcnc.org/12-milling/27653-analog-servos-torque-mode-velocity-mode-both?limitstart=0
09:34 AM CaptHindsight: http://www.wellking.com/default/en/article/AppNotes/msk/msk_11.pdf
09:35 AM CaptHindsight: TORQUE CONTROL vs. SPEED CONTROL ^^
12:02 PM IchGucksLive: hi
12:48 PM adamj12b-web: Is it possible to calibrate an open loop spindle using a comp file like on an axis?
12:53 PM IchGucksLive: why calibrating a soiindle
12:53 PM IchGucksLive: spindle
12:53 PM IchGucksLive: if 8k rpm or 8.2
12:53 PM IchGucksLive: try to get your best match
12:53 PM IchGucksLive: on pwm
12:53 PM adamj12b-web: Im just asking.
12:54 PM adamj12b-web: I noticed during my spindle testing that I was off a few hertz on the drive at different levels. It was not linear
12:57 PM pcw_mesa: if you want to get fancy you can linearize/match with the lincurve component
12:57 PM pcw_mesa: but in general spindle speeds are not that critical
12:58 PM Tom_L: are there examples of 'orient' or M19 somewhere?
12:58 PM Tom_L: i didn't have alot of luck finding any
01:09 PM adamj12b-web: pcw_mesa: I figure you might be the one to ask this. Do you know if it is possible to use the "use_serial_numbers" when configuring a 7i84 with pncconf?
01:10 PM pcw_mesa: You cannot use that feature currently with a 7I77
01:11 PM adamj12b-web: Well that clears that up quick.
01:11 PM adamj12b-web: I thin I will be adding a 7i74 eventually with some more serial i/o.
01:13 PM pcw_mesa: The 7I77 is "funny" because it has 2 sserial devices (analog and digital I/O) both named 7I77 (but on different channels)
01:13 PM pcw_mesa: This causes a non unique serial number issue (probably should be fixed by some kind of hashing on the serial numbers)
01:15 PM adamj12b-web: what about the serial expansion channel on the card?
01:16 PM adamj12b-web: TB6
01:17 PM pcw_mesa: Normally that's sserial port 0 channel 2
01:17 PM pcw_mesa: if you add a 7I74 thats normally sseririal port 1 channels 0..7
01:28 PM CaptHindsight: http://www.compumotor.com/catalog/cataloga/A72-A96.pdf various servo applications and how they were solved
01:30 PM CaptHindsight: servo applications with applying a known torque http://irtfweb.ifa.hawaii.edu/~tcs3/tcs3/vendor_info/Baldor/AN00156-apply_known_torque/AN00156-001%20-%20Applying%20a%20Known%20Torque.pdf
01:52 PM IchGucksLive: Gn8
01:53 PM CaptHindsight: http://linuxgizmos.com/tis-99-dlp-pico-projector-board-connects-to-beaglebone-black/ a low cost DLP dev board $99
01:53 PM CaptHindsight: useful for SLA printing
02:07 PM FinboySlick: CaptHindsight: You aware of anyone duplicating the 'oxygen float layer' that carbon3d uses to prevent the sla print from sticking to the bed?
02:08 PM CaptHindsight: yes, how they fooled so many investors is beyond me
02:08 PM FinboySlick: What are they using?
02:08 PM CaptHindsight: plus it only applies to free radical cure vs cationic
02:09 PM CaptHindsight: well top projection solves the non problem
02:09 PM FinboySlick: I'm not sure what that means.
02:09 PM CaptHindsight: the type of curing chemistry used
02:09 PM FinboySlick: It introduces the need for a wiper though.
02:10 PM CaptHindsight: not if you are continuously curing
02:11 PM FinboySlick: A 'curing below the surface' kind of deal then?
02:11 PM CaptHindsight: wiper is used with higher viscosity resins when printing in layers
02:11 PM CaptHindsight: it's automatic with free radical cure
02:11 PM CaptHindsight: oxygen is a polymerization inhibitor with free radical
02:12 PM CaptHindsight: anyway have to run, bbl
02:12 PM FinboySlick: Ok
02:16 PM CaptHindsight: https://www.mitsuichemicals.com/opulent_cha.htm
02:16 PM CaptHindsight: TPX has one of the highest oxygen permeable ratings
02:17 PM CaptHindsight: https://www.mitsuichemicals.com/tpx.htm
02:19 PM FinboySlick: You're going a bit fast for me ;) I understand that some resins won't cure when in contact with oxygen. What magic material do they use to keep the oxygen later below the puddle? A porous thing that is exposed to pressurized O2?
02:21 PM FinboySlick: :s/later/layer/
02:29 PM CaptHindsight: tpx is porous and optically clear
02:29 PM CaptHindsight: they seem to have invented something similar
02:30 PM CaptHindsight: resins that use free radical curing will not cure at the surface where it is in contact with oxygen
02:30 PM CaptHindsight: bbl
04:47 PM Deejay: gn8
05:33 PM JT-Shop: phiplii: you have gained an i
05:44 PM phiplii: perhaps methinks, because I inelegantly departed this room
05:48 PM Cromaglious_: heh
05:59 PM Cromaglious_: I'm in a holding pattern until i get more money.... Damn VA!
06:00 PM phiplii: how impatient are you feeling? Anything you can sell on ebay / make and sell?
06:01 PM phiplii: My brother sells audio effects when he is a few quid short
06:04 PM Cromaglious_: There is a bunch of C.R.A.P (Curios, Rare and Precious) I can sell
06:06 PM Cromaglious_: ahhh... My wifes replacement phone has a decent camera! kewl! phiplii Thanks.. I didn't have a decent camera to take pictures of the C.R.A.P, now I do...
06:08 PM phiplii: Don't for get to thank her!
06:08 PM Cromaglious_: heh... yeah...
07:13 PM Tom_L: JT-Shop around?
07:37 PM Adamj12b-Web: BeachBumPete: Hey, Made some progress.
07:41 PM BeachBumPete: Adamj12b-Web thats great man glad to hear it...
07:42 PM Adamj12b-Web: The pid was too weak. once I cranked it up, it came alive.
07:47 PM BeachBumPete: Awesome
07:47 PM BeachBumPete: I know tuning my servos was kind of a stab in the dark
07:47 PM BeachBumPete: but once I got close it became a lot easier to get it better
07:51 PM Cromaglious_: I just received my tuner for my iHSS57-36-20 closed loop steppers. The current settings are not what the manual says they are.
08:59 PM Adamj12b-mobile: When tuning, what should the acceleration for the axis be set to?
09:04 PM CaptHindsight: something less than or equal to that you will use in it's final settings
09:04 PM CaptHindsight: I start low and slow and work my way up
09:06 PM Adamj12b-mobile: I thibk im too low currently at 2in/sec as i get more ringing as the motion goes on. Evens out somewhere around decel
09:08 PM CaptHindsight: no such thing as too slow, it should be stable at low speeds
09:10 PM CaptHindsight: homing will be done at low speed and accel
10:03 PM CaptHindsight: http://www.bbc.com/news/av/technology-40447883/the-house-being-built-by-robots-and-3d-printers
10:04 PM CaptHindsight: a mobile robot that can tack weld is handy
10:56 PM misha is now known as Guest50366
10:57 PM Guest50366: hello?
11:00 PM CaptHindsight: howdy