#robotics Logs
Nov 16 2017
#robotics Calendar
12:49 AM Luminax-Work is now known as LuminaxWk
01:15 AM rue_bed: ok, my table based RC simulator almost works right
01:15 AM rue_bed: but I'm getting good results
01:15 AM rue_bed: :)
01:16 AM rue_bed: I ended up with 2 256 byte tables, one that tells you the v(t) and the other tells you how far to jump in the table to get to the same value of the table when its inverted
01:17 AM rue_bed: you just count time, and do a brief dance when you change from charge to discharge
01:18 AM rue_bed: this might be fast enough to generate the 1 bit virtual adc I need to regenerate the real value with an RC on the output pin
01:18 AM rue_bed: faster than the adc can do
01:19 AM rue_bed: which means I can use the generated analog value on the comparitor to generate pwm for current control
01:19 AM rue_bed: we having fun yet?
01:21 AM mrdata: lots
01:22 AM mrdata: i'm making a jig to wind a coil for a resonant circuit
01:26 AM Jak_o_Shadows: hmmmm.
01:26 AM Jak_o_Shadows: two lookup tables is probably a more sensible way of controlling the brightness pattern of the torch
01:26 AM Jak_o_Shadows: I want a SLOW wind-up - but I also want plateaus. and I want
01:26 AM Jak_o_Shadows: wait
01:27 AM Jak_o_Shadows: Sounds like I want some sort of tree-like thing
01:28 AM Jak_o_Shadows: Yes. I want a tree. Have it split into say, 1 minute blocks. Which then go into 30 second blocks. But only sometimes. and sometimes going all the way down to 1/2 second blocks
01:42 AM rue_bed: mrdata, for power, signal, other?
08:17 AM mrdata: rue_bed, power, actually
08:19 AM mrdata: it is a little circuit to stress test that new EMP-hardened raspberry pi
02:00 PM polprog_ is now known as polprog
03:39 PM To_Aru_Shiroi_Ne is now known as ToAruShiroiNeko
07:03 PM rue_mohr: ah
07:09 PM rue_mohr: I want to make a face tracker that points a spotlight out the window whereever my cat is looking
07:15 PM AchiestDragon: the cat will want to chase the light like some giant laser spot
07:19 PM AchiestDragon: reserch seems to point to the fact that unlike the vision we see cats seem to work on arias of movement , so moving the aria of illumination will cuase the cat to look at the changes of light so it will eather then look towards the edges of the illuminated aira that has changed the most or appears to have from the cats point , so the system would move again you trying to drive the cat nuts
07:20 PM AchiestDragon: i know form observing my cat with a mouse if the mouse is stationary the cat tends to lose interest and in some cases almost fails to see it untill it moves even if its inches away
07:21 PM rue_mohr: I know
07:21 PM rue_mohr: }:]
07:26 PM AchiestDragon: if you want to have an idea of what its looking at brob better with ir floods and a good low light ir cam , and run it though motion detection , if you tie that with actual tracking of the cats direction of view you may find what its looking at
07:27 PM AchiestDragon: although remeber if night and the rooms light it may be looking at reflections from inside that will sort of be confusing it as it know it cant get to them though the glass , and that apparant room is not there during the day for it
07:34 PM rue_mohr: hmm, my RC curve data has a wierd wiggle in it
07:38 PM rue_mohr: ah, yea, ok, its fromt eh integer rounding
07:47 PM AchiestDragon: int 8 or 16 bit ? speed wise 16 bit you can cheat with the math and have some extra precision
07:50 PM AchiestDragon: ie change the pos of the decimal place in your maths so 1 becomes 100 , 2, 200 , so 250 would actualy be 2.5 so you get a smother result without having to switch to floating point
07:54 PM AchiestDragon: but it depends on your min and max range values you are using
07:57 PM AchiestDragon: speed wise pulling a 16 bit value rather than an 8 on a half decent cpu will take the same time , as apposed to a float , if your table is byte alligned
07:57 PM Luminax-Work is now known as LuminaxWk
08:02 PM AchiestDragon: thats on a 16 bit cpu , and a nifty way to get better precision using quiker intiger only maths to have some decimlal acuracy , if doing the maths in hex you can just strip some of the bits to make it a pure int again
08:04 PM AchiestDragon: good for when you have a lot of math to do and not a lot of cpu to do it on and ints alone lead to rough results
08:05 PM Luminax-Work is now known as LuminaxWk
08:12 PM AchiestDragon: do remeber on 16 , 32 and 64 bit cpus the speed factor is that of the databus with to memory , if thats 32bits then theres no diference speed wise between a single 32bit read and a 8 or 16 bit , unless you have chache to to hold the data 4 byte reads still ususaly take 4 memory cylces as would 4 32 bit ones ,, or 4 16 bit reads
08:13 PM rue_mohr: 8
08:13 PM rue_mohr: buuuut the values are close enough
08:15 PM AchiestDragon: :)
09:07 PM rue_mohr: and I wonder,
09:07 PM rue_mohr: would it be close enought o assume a linear integrator and recreate with a RC filter
09:08 PM rue_mohr: cause, the RC compensation is a LOT of work
09:08 PM rue_mohr: and I'm not sure about its ability to maintain a dc level
09:09 PM rue_mohr: being a sigma-delta dac
09:15 PM AchiestDragon: you meaning noise on output like https://www.computeraudiophile.com/uploads/monthly_2015_04/dac_output.png.71d705415ec897ffb78bf0ea98e0c212.png
09:16 PM rue_mohr: well I would expect a type of distortion, but I'm not sure what the nature of it would be
09:19 PM AchiestDragon: well the sigma delta produces a pulse to charge a rc network , repeting the pulse to hold the level , you could stick a lpf (low pass filter on the outout to remove the pulse noise , well reduce it , although if its only a few khs that lpf is going to have quite phisicaly large inductors
09:19 PM AchiestDragon: )
09:19 PM rue_mohr: its better than that
09:20 PM rue_mohr: my virtual 1 bit adc simulates the output RC filter, so it knows what the error is and compensates for it
09:20 PM rue_mohr: but, it cant account for remote dc drift
09:22 PM rue_mohr: 0101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010
09:22 PM rue_mohr: looks like it found its dc level
09:22 PM AchiestDragon: you need to buffer the output from the rc netowrk on the output using an opamp , so that it wont drain the C in the rc network ,
09:23 PM rue_mohr: ..... or I made the C the gate of a fet
09:23 PM rue_mohr: damn thats fast code
09:24 PM rue_mohr: 111111111111111111111111111111111010101010101010101010101010101010101010101010
09:24 PM rue_mohr: ok
09:24 PM AchiestDragon: yea , if you have an opamp use the low noise fet versions , i do anyway
09:25 PM rue_mohr: my goal is to implement a serial controlled, switching current regulator in 8 transistors
09:25 PM rue_mohr: or 20c
09:25 PM rue_mohr: ($0.20)
09:27 PM rue_mohr: 10000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001
09:27 PM rue_mohr: ooo it dosn't like small dc values!
09:27 PM AchiestDragon: kool , atm im trying to implimient my latest robot without having to use an array of 8 pi3's onboard it
09:27 PM rue_mohr: :)
09:27 PM rue_mohr: I pulled off a circuit that does switching current control in 6 transistors
09:28 PM rue_mohr: one of them I could have optimized out tho
09:28 PM rue_mohr: its got an osc, flipflop, and comparitor
09:28 PM rue_mohr: (and buck driver)
09:29 PM AchiestDragon: transistors are voltage controled devices you should be looking at a fet design there current controled like valves/ tubes
09:29 PM rue_mohr: it worked :)
09:29 PM rue_mohr: a 74hc595 is just 5c
09:30 PM rue_mohr: but I'd like to use a tiny13 and do almost everything with it
09:30 PM rue_mohr: BUT I cant set the trigger voltage on the comparitor
09:31 PM rue_mohr: (without a dac that breaks the budget, cause after the t13 there is no budget left)
09:32 PM rue_mohr: over by 30c just with the t13
09:32 PM rue_mohr: there are NO R-2R sip networks
09:33 PM rue_mohr: its made me tink of taking glass, filming them with a carbon layer, and lasering in a R-2R network
09:34 PM rue_mohr: for now, 0806 resistors are taking up the slack
09:34 PM AchiestDragon: http://cds.linear.com/image/11847.png , but cost more
09:34 PM rue_mohr: yea, there isn't a chip that dosnt break it
09:35 PM rue_mohr: besides, I already have that with 6 transistors, I would still need to add a ref for the variable current control
09:35 PM rue_mohr: I dont know how close to ground I can operate a LM393 comparitor
09:36 PM AchiestDragon: the single inducotr designs are no so bad but once you have a transfromer in there tehy become a nightmare , both in getting the circut to function right and 2 in actualy beeing able to souce the ferrite cores to make the transformers from
09:36 PM rue_mohr: the sigma-delta adc sorta just became an interesting distraction
09:37 PM rue_mohr: the 'black box' for what I'm making, is a 0-1A, 8 bit, serial control, current sink
09:37 PM rue_mohr: (regulated)
09:39 PM rue_mohr: the code I have here works, the next question is can I implement a dual-256 byte lookup-table RC simulator on an avr
09:40 PM AchiestDragon: http://www.electronicdesign.com/sites/electronicdesign.com/files/uploads/2015/02/IFD2630_FO_0.gif current limiter for a lazer diode but simple little circut could be helpfull
09:40 PM rue_mohr: if so, I should be able to make a 1 bit adc on a... t13?
09:40 PM rue_mohr: interesting
09:40 PM rue_mohr: referenced from +
09:41 PM rue_mohr: whats R5 for tho...
09:42 PM rue_mohr: oh, setpoint while Q1 is off
09:42 PM rue_mohr: er, no
09:42 PM rue_mohr: I love circuits like that, cause I know I can use a LM393
09:43 PM rue_mohr: huh, Iwant to build it just to check it out
09:43 PM rue_mohr: wait, 5V on a P channel fet?
09:43 PM rue_mohr: must be logic level
09:43 PM AchiestDragon: r5 would ensure some stabliaty in the reading ie keeping the signal from jumping over 5v when power to the laser is halted
09:44 PM rue_mohr: I only have 2N7000
09:44 PM rue_mohr: I dont even know if there is a P equiv
09:48 PM rue_mohr: bs250 is like 50c ea...
09:52 PM rue_mohr: nope thats as good as thruhole P channel mosfets get
09:53 PM AchiestDragon: main issue i had is other than ebay the main source for components i use is farnells , as they have a £20 min order i used to just bulk buy components usualy getting multiple quantity discounts , like resistors where 2p each in 10's (min order quantity) and it worked out at something like 50p more to buy 100
09:53 PM rue_mohr: how any p to a £?
09:54 PM AchiestDragon: 100
09:54 PM rue_mohr: ah
09:55 PM rue_mohr: hey
09:55 PM * rue_mohr looks at the circuit upside down
09:55 PM rue_mohr: there is no inductor
09:55 PM rue_mohr: wtf
09:55 PM AchiestDragon: its a current limiter
09:55 PM rue_mohr: yea, so where is the inductor to level off the current
09:56 PM rue_mohr: waaaait, its not running linear is it?
09:56 PM AchiestDragon: yea
09:56 PM rue_mohr: oh god! get it off me!
09:56 PM rue_mohr: get it off!
09:56 PM rue_mohr: get it off!
09:56 PM rue_mohr: get it off!
09:56 PM rue_mohr: get it off!
09:56 PM rue_mohr: get it off!
09:56 PM * rue_mohr flips it into the trash and calms down
09:56 PM rue_mohr: 8-| LINEAR!?
09:57 PM rue_mohr: WTF ARE YOU ON!?
09:57 PM rue_mohr: *SIGH*
09:58 PM rue_mohr: my 6 transistor design is nice, but I'd like some kinda smt multi-transistor pack to build it with
09:58 PM AchiestDragon: well actualy no ,, you can pulse drive lazers so like pwm as long as the 100% duty cycle current load is constant
09:59 PM rue_mohr: I got some smt transistors to get closer to a tiny circuit
09:59 PM rue_mohr: ok hmm
10:00 PM AchiestDragon: company i used to work for produced kit for in car use , like they just used 7805's with 12v in to get 5v from and complained they got hot
10:00 PM rue_mohr: part of what I suspect will be trouble with what I want to do, is I have a grounded current sense resistor, and I need to use a comparitor against it, but I think its too close to ground
10:01 PM rue_mohr: would yo like to see the 6 transistor circuit that made me drool when it worked?
10:06 PM logstaren: what's a good website to paste videos that won't stay there forever, for quick sharing?
10:08 PM rue_mohr: dunno
10:08 PM rue_mohr: the lm393 datasheet seems to say the input range includes 0V
10:09 PM AchiestDragon: that circut with the laser diode in the psu out would be accross where the laser is , its a fet so its risitance changes to stablaise the current , you could feed the opamp out back to throttle the osc circut thats driving the switching inductor , although prob means using a dual opamp
10:10 PM rue_mohr: AchiestDragon, did you see the transistor pwm circuit I made?
10:10 PM AchiestDragon: no
10:10 PM rue_mohr: ok lets start with that
10:11 PM rue_mohr: http://ruemohr.org/~ircjunk/tutorials/elex/micropwm/p1080966.jpg
10:11 PM rue_mohr: 4 transistors, 0-100% duty 3.4Khz
10:13 PM rue_mohr: amuzed?
10:14 PM rue_mohr: http://ruemohr.org/~ircjunk/tutorials/elex/micropwm/p1080967.jpg
10:14 PM rue_mohr: not sloppy switching either
10:16 PM AchiestDragon: fig 2 ..... https://ssihla.wordpress.com/2014/09/09/avalanche-pulse-generator/ only 3 transistors
10:17 PM rue_mohr: but awefull switching
10:18 PM rue_mohr: those flipflop oscillators are horrid for performance
10:18 PM rue_mohr: so I find
10:19 PM rue_mohr: and you cant go to 0 and 100%
10:22 PM rue_mohr: sorry, I'm having troubles with the scripts I use to upload images
10:22 PM AchiestDragon: he seems to have done that as a (LT1073) replacement
10:22 PM rue_mohr: seems I removed some program I shouldn't have
10:22 PM rue_mohr: yep
10:22 PM AchiestDragon: fig 12 his page
10:23 PM rue_mohr: http://ruemohr.org/~ircjunk/circuits/I_switchingreg_555.jpg
10:24 PM rue_mohr: here is a 555 based switching current reguator
10:24 PM rue_mohr: then I redid with just transistors
10:24 PM rue_mohr: http://ruemohr.org/~ircjunk/circuits/I_switchingreg_q.jpg
10:25 PM rue_mohr: that thing working made me QUITE amused
10:26 PM rue_mohr: :) thoughts?
10:27 PM AchiestDragon: are you just needing current limiting or do you need a stable voltage also
10:27 PM rue_mohr: well those are both fixed, I want a 0-1A adjustable current sink
10:33 PM rue_mohr: http://ruemohr.org/~ircjunk/images/DIY_R-2R.jpg
10:33 PM rue_mohr: :)
10:33 PM rue_mohr: 1206 is TOO BIG
10:33 PM rue_mohr: by 200, to fit in series, across adjacent pads
10:34 PM rue_mohr: AchiestDragon, so "just current"
10:34 PM rue_mohr: tis for a current controlled motor driver
10:35 PM rue_mohr: for force feedback
10:35 PM rue_mohr: which is why I'm trying to do it cheap, I want to make a lot of them
10:36 PM rue_mohr: aaanyhow
10:37 PM rue_mohr: thats the pathetic amount of progress I'v made on anything in about the last 3 weeks
10:37 PM rue_mohr: +- a bit of software
10:38 PM AchiestDragon: https://www.mouser.co.uk/ProductDetail/Panasonic/EXB-28V330JX/?qs=sGAEpiMZZMvrmc6UYKmaNaPSeR7CYkqV1%2fw2YgPq5cc%3d
10:40 PM AchiestDragon: ive seen smaller ones also usualy 4 resistors in a single smd package and about the size of a single 1206
10:41 PM rue_mohr: yes
10:41 PM AchiestDragon: and no easy way to hand mount them
10:42 PM rue_mohr: https://www.aliexpress.com/item/100pcs-10K-ohm-10-Kohm-5-0603-8P-2-4P-SMD-Thick-Film-Chip-Resistor-Network/32372635218.html
10:42 PM AchiestDragon: prpb the reason there only ever stocked as a full reel
10:43 PM rue_mohr: 4x 0603, which, are tooo small
10:43 PM rue_mohr: 1 damned 10 pin sip thats an 8 bit R-2R would be AWESOME
10:43 PM rue_mohr: technically they exist
10:43 PM rue_mohr: and you can find them for like $18 ea
10:44 PM rue_mohr: you can also find a dac for like $5
10:47 PM AchiestDragon: like https://www.mouser.co.uk/ProductDetail/Bourns/4610X-R2R-103LF/?qs=sGAEpiMZZMvrmc6UYKmaNY%252bbSZmNgoil6GbXIi6V0Yo%3d £0.72 in ones
10:49 PM AchiestDragon: £ 0.33 in 500's off
10:50 PM rue_mohr: yea
10:50 PM rue_mohr: they could just make R-2R ladders on them
10:50 PM rue_mohr: oh hey
10:51 PM AchiestDragon: read the pdf they do
10:51 PM rue_mohr: I know if the pdf they do, but when you try to getthem, nobody has them
10:51 PM rue_mohr: ah see, lead time 12 weeks
10:51 PM rue_mohr: wait they have them tho
10:51 PM rue_mohr: AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
10:52 PM rue_mohr: they have a 'choose your country' stuck in my face that I cant close, so I cant see everything
10:53 PM AchiestDragon: lol
10:53 PM rue_mohr: what country are you in?
10:53 PM AchiestDragon: uk
10:54 PM AchiestDragon: you should remeber , although its been about 8 years since i was on irc
10:55 PM rue_mohr: 652-4610x-r2r-103lf
10:56 PM rue_shop3: drat, of course mouser has an $11 shipping fee
10:57 PM rue_shop3: Mouser Electronics has disabled TLS 1.0 to remain in alignment with the latest standards for security and data integrity.
10:57 PM rue_shop3: FFFFFFFFFFFFFFFF thats right, I'm not allowed to view the mouser site
10:58 PM AchiestDragon: i cant view digikey atm
10:58 PM rue_shop3: ah wait, its $0.58ea in volumes of 500
10:58 PM orlock: rue_shop3: If you can't view that, you should update your client...
10:58 PM rue_shop3: 500 of them cost me $289.50
10:59 PM AchiestDragon: https://www.avnet.com/shop/emea/p/4610x-r2r-103lf-3074457345629047595?k_clickid=017dc3ed-739b-4e23-8fd4-e9272e7d041d&cmp=EMEA-Global-PPC-Google-EN-718323071-SW-201609 seem to have them
10:59 PM rue_shop3: orlock, I have another browser, but its a gamble, sometimes it locks up the computer
10:59 PM rue_shop3: AchiestDragon, so, those things are still over $1ea in any reasonable volume for me
10:59 PM rue_shop3: +$11 shipping
10:59 PM orlock: Yeah... Your problems are not the TLS 1.0 being removed :)
11:00 PM rue_shop3: it is, why the f*** can I not view a public site if my browser is not secure?
11:01 PM rue_shop3: AchiestDragon, so, if you compare that against a custom pcb, with smt resistors, + assembly, its damn near cheaper
11:01 PM AchiestDragon: In Stock7600 on avnet 800: €0.4362 each
11:01 PM rue_shop3: yea, thats almost $1cad
11:01 PM rue_shop3: what do 16 0806 resistors cost?
11:02 PM rue_shop3: 1% for that matter
11:02 PM rue_shop3: $0.000.....
11:03 PM rue_shop3: and even a custom pcb thats about 1cm^2?
11:03 PM rue_shop3: even if I DID ship out to get them made
11:03 PM rue_shop3: custom
11:03 PM rue_shop3: its... sad
11:04 PM rue_shop3: they should cost WAY less than ANY dac
11:04 PM AchiestDragon: 5,000: £0.026 each thats a full reel
11:04 PM rue_shop3: thats also $300
11:05 PM AchiestDragon: and 5000 resistors
11:05 PM rue_shop3: ah
11:05 PM rue_shop3: well
11:05 PM rue_shop3: the 0806?
11:05 PM AchiestDragon: yea
11:05 PM rue_shop3: or the ladders?
11:05 PM rue_shop3: k
11:05 PM rue_shop3: small reel
11:06 PM rue_shop3: 312 R-2R ladders
11:06 PM rue_shop3: hah $1 ea
11:06 PM rue_shop3: that didn't work did it
11:06 PM rue_shop3: but wtf
11:07 PM AchiestDragon: no 0. 026 thats 2.6p each
11:07 PM rue_shop3: I bought 500 0603 for $1.25 from china
11:07 PM AchiestDragon: only drops to 2.3p each in quants of 25,000
11:08 PM rue_shop3: $0.04ea per r-2r in resistors
11:08 PM rue_shop3: so I'm at 4c
11:09 PM rue_shop3: my 1k-ohm reel had 10000
11:09 PM rue_shop3: I have a 2kohm reel that is much smaller
11:10 PM rue_shop3: yea, its 5000
11:10 PM rue_shop3: so, point proven
11:10 PM rue_shop3: its insane to buy one of those R-2R for the price they are asking
11:12 PM rue_shop3: https://www.aliexpress.com/item/50pcs-SMD-To-DIP-0805-0603-0402-SMT-To-DIP-Adapter-Board-Capacitance-Resistance-LED-SMT/32785286235.html
11:12 PM rue_shop3: I can use that to make a 7bit R-2R dip
11:12 PM rue_shop3: heh, they bumped up the price after I bought some
11:12 PM AchiestDragon: yea , but as there thin film you can garantee that there all the same value ie if one is 2% lower than the value they all genaraly are , that does matter sometimes on critical stuff
11:13 PM rue_shop3: but I bought 1% resistors
11:13 PM rue_shop3: I'm sure the sips are laser trimmed tho, I'd be surprised if they aren't pretty darn good
11:14 PM rue_shop3: I have yet to test the linearity of the one I made there
11:14 PM rue_shop3: its a mix of 1% and 5%
11:14 PM AchiestDragon: well same , but tighter tolarance , you may find the whole network is .9% out but its ususaly ok when what matters is a 0.01% diference on the ladder itself
11:15 PM rue_shop3: I'm going to run a ramp on it, scope it out
11:15 PM rue_shop3: not right now tho
11:15 PM rue_shop3: right now I'm trying to re-asses my goals
11:16 PM rue_shop3: I think I can use a lm393
11:17 PM rue_shop3: oh, I can use a current source thru a resistor to pick the feedback voltage off ground
11:18 PM rue_shop3: I think an lm7805/tl431 can do that really easy
11:22 PM Triffid_Hunter: R-2R will only give you about 5 usable bits with 1% resistors last time I checked
11:22 PM rue_shop3: rlly?
11:22 PM rue_shop3: I need to scan that one I made
11:22 PM rue_shop3: I'v done it lots of times and not seen any evil errors
11:23 PM rue_shop3: with all 5% resistors
11:23 PM Triffid_Hunter: better to use ΣΔ or so
11:23 PM rue_shop3: yea, but I dont know if I use a 1 bit adc if the dac (rc filter) will hold dc levels
11:24 PM Triffid_Hunter: why wouldn't it?
11:25 PM rue_shop3: it dosn't hit the rail
11:25 PM rue_shop3: its output is drifting on a cap
11:25 PM rue_shop3: in my head, it dosn't add up
11:25 PM logstaren: anybody got experience with ESCs? electronic speed controllers? they make beeps and I'm looking for the buzzers they use so I can get them. They sound good.
11:26 PM rue_shop3: everything I did with 1bit last year or so, was with ac signals
11:26 PM rue_shop3: logstaren, just peizo elements, if that is, your not talking about the beep they make with the motor itself
11:26 PM AchiestDragon: logstaren: they basicaly modulate the motor coils to make the motor buzz
11:27 PM rue_shop3: dont they also have a panic beeper on the controller?
11:27 PM logstaren: From hearing my ESCs on an airplane and my quadcopter, I find that hard to believe.
11:28 PM logstaren: rue_shop3, piezzo yes, i just want to know what the industry use for cheap and loud
11:28 PM logstaren: the ones i played with were never that loud
11:28 PM AchiestDragon: and when yo take one appart theres no buzzer on the pcb
11:28 PM rue_shop3: logstaren, thts cause you didn't have them in a little case
11:28 PM rue_shop3: they aren't loud till you put them in a pipe with a cap w/hole
11:28 PM logstaren: the loudest i know is on a lipo alarm tester, realy really loud, don't know if i can get more than one note out of it though (note, frequency)
11:29 PM rue_shop3: 78L05 wins the parts count for a current source
11:29 PM logstaren: AchiestDragon, yes it's crazy i just watched a few teardowns there is no buzzer on the board!!
11:30 PM Triffid_Hunter: heh I got the hobbyking clone of those imax B6 chargers, the piezo in those is stupid loud.. I had to duct tape some cardboard over it to not feel like I was being stabbed in the ear with an ice pick when it was done charging
11:31 PM AchiestDragon: the motors are brushless so are a bit like steppers you can push pul the coils to cause the motor to vibrate rather than spin , casuing it to make an audiable note
11:31 PM logstaren: i want a really loud one because i will use it for my car
11:31 PM AchiestDragon: they also dont make a sound when theres no motor attached
11:32 PM logstaren: i want it to beep loud when i remote start it
11:32 PM Triffid_Hunter: logstaren: get a piezo horn as used in hifi speakers, those will happily go up to 130dB or so if you stuff enough volts in
11:33 PM Triffid_Hunter: and if you want just one frequency or a narrow range, you can put an inductor on it and form an LC resonator
11:33 PM logstaren: no i want several frequencies
11:34 PM logstaren: like in this video when the computer is turned on, but a LOT louder
11:34 PM logstaren: https://www.youtube.com/watch?v=JQdXDg9PtzY&t=312s
11:34 PM AchiestDragon: was thinking earler to day that those motors spin rather fast and have quite a bit of torque , so was wondering how good they would be driving a bit of threaded rod to make it work as a high speed linear actuator
11:34 PM rue_mohr: heh
11:35 PM rue_mohr: starting torque sucks
11:35 PM logstaren: 30 seconds in
11:35 PM rue_mohr: and there is a min speed
11:35 PM rue_mohr: well, the low-pole ones have poor start consistancy
11:35 PM logstaren: actually i may just resuse the same code than that one for what i want to do
11:35 PM AchiestDragon: as an altrenative to servos in a hexipod , even if i have to use just a small amount of gearing like 2 or 4 to 1
11:36 PM logstaren: except i won't connect it to the remote startere circuit itself, that is way too much of a pain to get under the steering wheel plus i don't want to reverse engineer the circuit
11:36 PM logstaren: it's time i put my own little computer in this car
11:37 PM logstaren: will hook up that trigger to parking lights probably since that turns on when i have the remote starter start the car
11:41 PM AchiestDragon: lol if you desing the input drive to the shaft with 300 degrees of backlash it should be spinning by the time it gets load , or you just reverse it for an instant and again to hammer it arround to get it moving
11:41 PM Triffid_Hunter: AchiestDragon: servos don't really care what sort of motor is inside, brushless ones work fine
11:41 PM * rue_mohr thinks
11:42 PM Triffid_Hunter: rue_mohr: only if you're doing sensorless with an aeroplane ESC
11:42 PM rue_mohr: thats what I was reffering to
11:42 PM Triffid_Hunter: rue_mohr: think about those electric "hoverboards" and unicycles which took the world by storm this past year or two, those have a crapton of torque at near-zero speed because it's necessary for balancing
11:42 PM Triffid_Hunter: and those definitely have brushless motors in 'em
11:43 PM rue_mohr: oh yea
11:43 PM rue_mohr: the airplane ones is what I meant
11:43 PM Triffid_Hunter: but they have sensors, and the controller has current steering rather than basic PWM plus commutation
11:43 PM rue_mohr: I dont think the rc car ones are sensored tho
11:43 PM Triffid_Hunter: need sensored motor for good starting torque
11:44 PM rue_mohr: I was amused when I got one and realized they kick the armature to get it going and need to run a min speed
11:44 PM Triffid_Hunter: AchiestDragon: really high end servos almost certainly use brushless motors - like those crazy dynamixel ones used in darwin-op
11:44 PM rue_mohr: ok already
11:44 PM rue_mohr: geez, you would think I'd said brushless sucks!
11:44 PM Triffid_Hunter: rue_mohr: yeah since they're using back-EMF for commutation sensing, they basically have no feedback at low speed and have to simply guess where the rotor is and hope for the best until they can detect it
11:45 PM rue_mohr: can I put 2x 256 byte lookup tables in a t13?
11:45 PM AchiestDragon: the way i did the big robot was to use pots mounted on the limb joints to detect the angle of the leg joints , and used a pll loop in logic driving the liniear actuators to position , unlike servos liniar actuators dont draw power when positioned , its held by the treaded bar at that
11:46 PM rue_mohr: I'd done a few
11:46 PM rue_mohr: dont like the way bar wobbles tho
11:46 PM rue_mohr: worked out a way to spin the nut instead
11:47 PM rue_mohr: havn't used that yet, just a test jig
11:47 PM Triffid_Hunter: they actually make steppers with a threaded hole instead of a shaft for that exact reason
11:47 PM rue_mohr: yea
11:47 PM rue_mohr: none of the steppers I have got cores large enough to modify :(
11:50 PM AchiestDragon: id squish the nut into the end of a pice of tube , the threaded bar on the inside if it wabbles a bit its not a prob , add another biger pipe over the exposed end to keep the crap out of it , just seems to be a cheap solution to the problem
11:51 PM AchiestDragon: brushless with esc's are about £10 where as linear 4" motors start at about £24 and servose that come close to the same power load are well over £40 each
11:52 PM rue_mohr: I was looking at brushless motors for my cnc spindle
11:53 PM AchiestDragon: also if i wanted o do the terminator style robot to look like the real thing i would need to make linear actuators
11:53 PM rue_mohr: its way cheaper to just keep buying brushed ones as I wear them out
11:55 PM rue_mohr: I have a t13 based pwm speed regulator, uses optic interrupter for feedback
11:56 PM rue_mohr: adc control voltage ->
11:56 PM rue_mohr: opto interrupter ->
11:56 PM AchiestDragon: steppers are too slow , and usualy quite heavy
11:56 PM rue_mohr: -> pwm
11:56 PM rue_mohr: AchiestDragon, with those new stepper drivers you can get about 2000rpm from a stepper
11:57 PM rue_mohr: if your using unipolar, non current controlled drivers, yea, slow as all hell
11:58 PM rue_mohr: no torque at 2k