#robotics Logs

Apr 01 2015

05:37 malinus I never really did anyhting moving before. 4 cheap dc motors from ebay (no datasheet), gear them 1:200, small wheels with good grip. That should move 2kg, right?
05:38 Triffid_Hunter one way to find out
05:39 malinus Make all the calculations :P?
05:40 deshipu theoretically, on a flat surface you can move any weight with any torque
05:40 deshipu in practice, the wheels and the floor bend a little, so it's never perfectly flat, and they stick to the floor too, so you need some force to move
05:43 malinus deshipu: that's only for things on wheels, right?
05:43 deshipu malinus: or tracks
05:45 deshipu malinus: actually, it's the same for things without wheels too
05:45 deshipu malinus: only the sticking is stronger :)
05:45 malinus ah
05:46 deshipu but if you had a pefectly flat perfectly hard non-stick surface, you could move anything on it with minimal force
05:46 deshipu of course, inertia means it would take some time to get to speed
05:46 malinus like they do in space, or you can kinda do on ice
05:46 deshipu right
05:47 malinus Back to my thing though - would you predict such a setup would work?
05:47 deshipu no data
05:48 malinus Yeah. That's what you get for $4 dc motors
05:48 deshipu you'd also need to know the friction of the wheels, etc.
05:48 deshipu and it would depend on the surface
05:49 malinus So what do you do in practice?
05:49 malinus (I'm assuming you won't take all that into account))
05:49 orlock Bolt wheels to old cordless drills.
05:49 orlock torque to spare
05:49 malinus haha
05:49 malinus yeah, that would work
05:49 orlock i need to get that bot running again
05:49 deshipu malinus: I build walking robots, problem solved, they always have to be able to lift their own weight :)
05:49 orlock i have a lawn now
05:49 orlock and it has a spare unidirectional motor output
05:50 deshipu but seriously, try it and replace the motors if they are too weak, I guess
05:51 malinus How big difference is there between motors that are the same size, anyway (assuming no gearing)?
05:51 deshipu huge
05:51 malinus How os?
05:52 deshipu it all depends on how much power they draw
05:53 deshipu you can make two motors that are identical mechanically, but have different coils and draw different power
05:54 malinus Yeah, I just figured everyone went for as much coil as could fit in the design :)
05:54 deshipu more coil actually draws less power
05:55 deshipu anyways
05:55 deshipu by the way, I found a great source of cheap electric motors
05:55 deshipu a bought a box of very old (the plastic went yellow already) dildos from a local sex shop
05:56 malinus :|
05:56 deshipu a little bit of sawing and I have the motors
05:56 malinus deshipu: mind explaining how a higher impedance results in less power?
05:57 deshipu malinus: when it doesn't move ;)
05:57 malinus deshipu: I was thinking, more coild = higher impedance = higher current.
05:57 deshipu you are right, sorry
05:57 deshipu I was thinking more coil = higher resistance = lower current
05:57 malinus Oh okay, you got me confused there.
05:57 deshipu sorry, I was wrong
06:03 Triffid_Hunter malinus: P=V²/R so more R = less P for the same V
06:04 Triffid_Hunter as far as I understand the math, the area available for putting the coil into defines the peak power, and from there the motor can be wound in various ways to deliver that amount of power at various voltages
06:05 Triffid_Hunter assuming things like core material and magnet strength are the same of course, they have a significant impact on peak power
06:05 deshipu they also have to dissipate the heat
06:05 Triffid_Hunter yeah that's a big concern
06:05 deshipu so, for instance, motors in metal chassis have a chance of handling more power
06:06 deshipu than the plastic ones
06:06 Triffid_Hunter interesting thing about the heat is that it's proportional to current, so if you feed the motor from a high voltage, current-limited supply you can easily exceed its power rating without exceeding its current rating
06:07 Triffid_Hunter ie you could destroy the bearings or fragment the rotor without melting any windings
06:07 Triffid_Hunter sorry, Δt is proportional to square of current
08:49 rue_bed Triffid_Hunter, your formula is for resistance, not inductance
08:50 rue_bed so it will tell you the heat, but not torque
08:50 rue_bed torque would come up as inductance
08:51 rue_bed I = LV/ Δt
08:51 rue_bed iirc, and I just woke up
08:52 rue_bed na, must be Vdt/L
09:16 rue_house the current will determine the magnetic strength, but no the part of the current that goes into the coil resistance
09:16 rue_house t
09:18 rue_house so if you have a motor with 5mH of inductance
09:29 Triffid_Hunter rue_house: the inductance caps di/dt which is relevant to commutation, but as the motor spins up, the driver provides more V which increases di/dt, no?
09:32 rue_house its current regulated
09:32 rue_house it tries to keep the current at a constant level... which is determined by the microstep postiion
09:32 Triffid_Hunter microstep? we were talking DC brushed motors
09:33 rue_house ooooh
09:33 rue_house well the resistance still turns into useless heat
09:34 rue_house but the inductive effects are more complex because of the commutator
09:34 rue_house using constant current control of dc motors simplifies the math greatly
09:35 rue_house the motor turns into a constant torque device
10:55 malinus I was considering just buying gears and using some motors I can find. But you can buy something like this http://www.ebay.com/itm/DC-6V-30RPM-Mini-Metal-Gear-Motor-Gearwheel-Model-N20-10mm-Shaft-Length-/391091018132?pt=LH_DefaultDomain_0&hash=item5b0ed75994
10:56 malinus I really doubt I can make a better gearing that that (it's probably like 1:500 or so)
11:08 deshipu yeah, those are quite common in robots
11:09 deshipu I especially like this one http://letsmakerobots.com/robot/project/pibot-a
11:10 malinus Cute, I want one that is much bigger though
11:16 SquirrelCZECH malinus: recently I wanted to make proper gearing
11:16 SquirrelCZECH but found out that it's usually easier to avoid doing so :)
11:16 SquirrelCZECH "proper gearbox"
11:18 deshipu malinus: then those motors might not be enough
11:18 malinus deshipu: not even 4 of them? With small wheels(5cm diameter)?
11:19 malinus I'm talking ~2kg
11:23 deshipu malinus: hard to say
11:23 deshipu malinus: as I said, depends a lot on the surface, the wheels, the speeds
11:24 deshipu malinus: that website has a lot of examples of robots, maybe you can find some that use the same motors?
11:34 SquirrelCZECH fu
11:34 SquirrelCZECH yeah
11:35 SquirrelCZECH because it makes so much sense selling motor and telling only the RPM
11:35 SquirrelCZECH :/
11:35 SquirrelCZECH not torgue/wattage or whatever
11:37 SquirrelCZECH nah, at least wattage/rpm on manufacturer website
11:45 malinus okay, thanks, I'll check it out
13:38 Hyratel MISTER TORGUE
13:45 malinus haha this website
13:45 malinus http://letsmakerobots.com/robot/project/ultimate-roboscoot
13:47 blib I'm looking to get a set of sensors that are usually in a smart phone. Does anyone know of a sensor set that one can buy ?
13:47 blib apart from the camera that is.
13:48 malinus so a gyro/acceleormeter combo?
13:48 malinus I suggest the msp6050
13:48 malinus *mpu
13:48 malinus mpu-6050
13:49 malinus blib: you can get them as as nice boards with pinouts and all necessary components.
13:49 malinus or rather *on nice boards.
13:50 malinus As of light sensing, I would assume that they use photoresistors in phones? Or at least, you can use a photoresistor, to get the same effect :)
13:51 deshipu malinus: they actually use tof distance sensors
13:52 deshipu malinus: I have a couple of them
13:52 malinus deshipu: are those also able to detect light levels?
13:53 malinus ah nvm. reading about it now :)
13:53 blib https://www.sparkfun.com/products/11028 - this?
13:55 blib doesn't have light/proximity/magnetic/... many sensors?
13:55 malinus blib: wow there, if you are not in a hurry, I would suggest china
13:55 malinus (costs ~$2 there)
13:56 blib malinus: is there a better/bigger set (of sensors) on one board that one can buy?
13:58 malinus Probably not, if you want everything on a single pcb, you probably best off designing one yourself.
13:58 blib then maybe 2 boards? :)
13:59 malinus I don't know, but if there is, it costs too much (compared to what you get).
14:02 blib malinus: can you point me to the $2 chip?
14:03 malinus blib: just search ebay. Just remember that shipping time can be long (or short) :)
14:05 blib malinus: maybe I should get an imu?
14:05 malinus imu?
14:06 malinus deshipu: they seem pretty neat. I don't think you get any of those IC's in a nice package though (re: tof).
14:08 malinus blib: what's imu?
14:08 deshipu malinus: https://www.sparkfun.com/products/12785
14:09 malinus oh yeah, they get them on boards. But imagne trying to solder them yourself :|.
14:09 malinus *imagine
14:10 blib malinus: http://www.sensonor.com/gyro-products/inertial-measurement-units/stim300.aspx - what's the accuracy comparison between these and those?
14:10 deshipu malinus: I have a bunch of naked chips
14:10 deshipu malinus: I didn't use them anywhere because they need weird voltage
14:11 deshipu I would have to muck about with a voltage regulator
14:11 malinus deshipu: something like 1.2V?
14:11 deshipu 2.8
14:11 malinus yuck
14:11 deshipu I have my robots running on 3.3
14:11 deshipu so yeah
14:12 malinus yeah, everything runs 3.3V in this age
14:12 malinus (almost)
14:12 malinus blib: haha I wonder what the price is
14:12 deshipu I have to try one with 3.3 and see if it burns or gives wacky readings
14:13 malinus blib: also that looks like some industry-standard stuff, haha.
14:13 malinus spoiler: probably more than $3 ;)
22:37 Hyratel http://www.jaycar.com.au/p/KJ8918
23:30 rue_house !?50!