#robotics Logs

Aug 16 2015

03:53 rue_bed Triffid_Hunter, hey, long time
07:32 caveman how to choose the right propeller for a hard disk motor? spins around 7200 RPM
08:17 SpeedEvil caveman: generally - don't bother
08:18 SpeedEvil http://hobbyking.com/hobbyking/store/__2069__hexTronik_24gram_Brushless_Outrunner_1300kv.html - for example is 6 pounds.
08:19 SpeedEvil And can do 75W
08:19 SpeedEvil http://hobbyking.com/hobbyking/store/__5423__FC_28_12_Brushless_Outrunner_1534kv.html 400 pounds for 4
08:22 caveman SpeedEvil: why not bother?
08:22 SpeedEvil Because HD motor will be very heavy, produce about 10W or so, and much, much lighter motors can do better for very cheap
08:23 SpeedEvil http://www.drivecalc.de/
08:23 SpeedEvil if you actually want to work it out
08:24 caveman SpeedEvil: can this even lift itself?
08:24 SpeedEvil no
08:24 caveman i am trying to implement a motor controller for this hard disk for learning purposes
08:24 SpeedEvil It may be usable as an indoor slow-fly motor, but that would because of the weight need to be quite large
08:25 caveman that's fine with me
08:25 SpeedEvil quite large - as in - will not fly in other than an indoor tennis court
08:27 caveman i am lost. are you saying that the weight needs to be quite large? weight of what quite large? shouldn't the weight be smaller to fly better?
08:27 SpeedEvil yes.
08:27 SpeedEvil The hard disk motor is heavy, and low power output.
08:28 SpeedEvil This means very low thrust.
08:28 SpeedEvil In order to fly with very low thrust, you need a long wingspan, to cope with the weight
08:28 caveman does this assume that the torque rating is high instead?
08:28 SpeedEvil A long wingspan slow flying thing is not going to be flyable outdoors in other than unusual dead-calm conditions
08:29 SpeedEvil To a zeroth approximation, only watts matter
08:29 SpeedEvil the propeller can be picked to give gearing
08:29 SpeedEvil the 'pitch' of the propellor sets the RPM/thrust ratio
08:29 caveman pits is the sound 'zzzzzzz' it makes?
08:30 caveman pitch*
08:30 SpeedEvil no
08:30 SpeedEvil Like a screw
08:30 caveman hmm there are propellers with gearinings inside them that are tuned by screws?
08:30 SpeedEvil A 12"*6" is 12" in diameter and the prop moves through the air 6 inches for every turn it makes.
08:31 SpeedEvil the blades are steeply pitched.
08:31 SpeedEvil A 12"*3" is 12" in diameter and the prop moves through the air 3 inches for every turn it makes.
08:31 SpeedEvil To a first approximation, if you turn the second one twice as fast, you get the same result
08:35 caveman i see
08:35 SpeedEvil http://hobbyking.com/hobbyking/store/__25079__NTM_Prop_Drive_13_12_2400KV_40W.html - for example. This lists at 10W and 20000RPM, with a 3.75" diameter prop with a 3" pitch - it gets 56 grams thrust. This is probably what you're looking at maximum
08:35 SpeedEvil Perhaps at 7200RPM - 5" and 2" pitch for the same thrust
08:37 caveman SpeedEvil: how did you extract 20000RPM from the link?
08:38 caveman 2400 * 7.4?
08:39 caveman = 17,760 ≈ 20,000 --- is this what you did?
08:43 SpeedEvil yes
08:43 SpeedEvil I was doing the maths in my head roughly
08:46 caveman i am lost with their prop test data table.. for the same voltage, the amperage is different. i wonder how they measure amps?
08:46 SpeedEvil Amperage depends on load
08:46 caveman i guess a larger propeller eventually increases the resistance of the motor electrically?
08:46 SpeedEvil yes
08:46 caveman V=IR applies here? can i think of the motor + propeller as bubble and think it's a resistor with ≈ 4 ohms or whatever?
08:47 SpeedEvil Not quite.
08:47 SpeedEvil The prop load is higly non-linear with RPM
08:47 SpeedEvil power = RPM^3
08:48 caveman for these kinds of motors, does rpm only vary as i change voltage?
08:48 SpeedEvil They are controlled by a speed controller.
08:48 caveman aka motor controller?
08:48 SpeedEvil The speed controler sets the speed - up to a maximum set by the torque demands of the load and the battery voltage
08:48 SpeedEvil yes
09:15 caveman SpeedEvil: i was thinking to implement a controller by using PWM. i was thinking to only adjust the frequency of letting-current-go across the 3 wire pins, that's all. but what you said seems to me to suggest that, optimally, i should adjust my PWM signals based on the power available in the battery (i.e. battery voltage and current). am i right? thoughts?
09:16 caveman basically i was thinking to PWM to a FET transistor
12:59 rue_house >>>>> rifraf <<<<<< what did you use to generate your gear profiles?
12:59 rue_house -------------------------------------------------------------------------------
13:36 SpeedEvil caveman: no, that's not how it works at all
13:36 SpeedEvil caveman: you need to apply three waveforms, phase shifted 120 degrees to each wire.
13:36 SpeedEvil Either trapezoidal, rectangular, or square
13:40 caveman SpeedEvil: would i need a DAC to make trapezoidal waveforms by using a uC?
13:40 SpeedEvil No, it's generally simply done with power FETs on/off
13:41 caveman PWM?
13:41 SpeedEvil http://www.digikey.co.uk/en/articles/techzone/2013/mar/an-introduction-to-brushless-dc-motor-control
13:41 SpeedEvil is relevant
13:43 SpeedEvil http://scolton.blogspot.co.uk/2009/11/everything-you-ever-wanted-to-know.html
13:44 caveman i am lost. it says PWM there, but you told me "no" earlier
13:45 SpeedEvil I read what you were saying as thinking it was pure power control as to how the speed control works
13:46 SpeedEvil Whereas it needs commutation and bidirectional driving on each of the three wires, and proper timing.
13:46 SpeedEvil And sensing.
13:49 caveman sensing on each of the wires to infer the current position of the motor?
13:52 SpeedEvil yes
13:53 SpeedEvil (maybe not 'each' - exact details on how its done can vary - there are many possible ways)
14:03 caveman SpeedEvil: yeah i have expectations on doing some fun work. imo this makes a fun learning experience for a newbie like myself
14:03 caveman SpeedEvil: but back to the earlier point, did you mean that i should also sense the voltage/current state of my battery, and based on that adjust my PWM waveforms?
14:10 SpeedEvil caveman: No.
14:11 SpeedEvil caveman: you have to make three trapezoidal (or similar) waveforms, and simultaneously drive those to all three wires. The magnitude of the waveform you drive to the three wires is varied according to the needed torque at the speed.
14:11 SpeedEvil The timings of the waveforms needs adjusted so that it's precisely tracking the position of the motor, or the motor will stall
14:12 caveman i feel i'lll get lots of stalls at low rpms
14:13 caveman as it'll be harder to predict the position of the motor
14:13 caveman did you ever do this?
14:14 SpeedEvil yes, low RPMs is problematic
14:14 SpeedEvil props less so, because torque is guaranteed to be 0 at ~0 RPM
14:14 SpeedEvil Hall sensors are often used in motors that need to go very, very slowly
14:15 caveman i see
14:16 caveman do you know anyone who did this?
14:17 caveman i think i must have ADC to sense back. agree?
14:17 caveman any tricks to do it without an ADC?
14:17 caveman possibly some GIO
14:18 caveman GPIO
14:19 SpeedEvil yes, GPIO + opamp, or just GPIO in some cases, with appropriate trickery
14:20 SpeedEvil There are also open-source brushless drivers
14:21 caveman fascinating: https://www.youtube.com/watch?v=0NvKwgfjiS4
14:21 caveman simple design from nasa. i think i could do it :p
16:37 veverak why
16:37 veverak if I want to find "slip ring" on ebay
16:37 veverak 99% of the stuff is not actual ring?
16:38 veverak https://www.youtube.com/watch?v=yzk0Gqzs5RI
16:38 veverak wanted to make this, but too lazy to mess with slip contacts
17:25 rue_school that is insanely dangerous with that laser
17:40 veverak yeah
17:40 veverak use glasses!
17:41 veverak https://www.youtube.com/watch?v=sO2-tqoyGik
17:41 veverak looks better
17:43 veverak yeah
17:43 veverak I've got really big urge to make something flashy
17:44 rue_school dont use high powered lasers...
17:52 veverak yeah
17:52 veverak led ball seems much better
19:53 blockh34d veverak: i'm having fun with adafruits neopixels
19:53 blockh34d maybe you'd like them
20:17 deshipu greetings from seattle