Back
[00:25:06] <rue_house> you cant do 460A, the leads melt off
[00:25:20] <rue_house> ad gimic, base the max current on the heat dissipation
[00:26:24] <rue_house> by that token a 1/4w 10M resistor can have 1600V across it
[00:26:37] <rue_house> (cmon, its gonna arc over)
[00:42:34] <Casper> just use 10x1M :D
[00:56:06] <rue_house> ok yay, I have an 8 pin chip that does 2 channels of adc to pwm
[00:56:15] <rue_house> I only need 1 channel, whatever
[00:56:46] <rue_house> I suppose if I wanted 1 channel I could have used a tiny10
[00:56:55] <rue_house> I'd need a new programmer from tom :)
[09:28:36] <[VEGETA]> back to yesterday's talk, I liked this device:
[09:28:38] <[VEGETA]> http://www.semtech.com/images/datasheet/ts30011_12_13.pdf
[09:29:16] <[VEGETA]> the (12) one says "output" current is up to 2A which is great for me
[09:29:38] <rue_house> 1Mhz is insane
[09:29:50] <[VEGETA]> for what reason?
[09:29:55] <rue_house> do you know how small the wire has to be on a 1Mhz inductor?
[09:29:57] <[VEGETA]> I don't want to control it
[09:30:26] <rue_house> if you want 2A, you need a crazy litz wound inductor
[09:31:00] <[VEGETA]> all will be smd parts so i have to search digikey for suggested inductor in the datasheet
[09:31:09] <[VEGETA]> but it won't output 2a all the time
[09:31:25] <[VEGETA]> maybe 1 amp continuous
[09:31:39] <rue_house> thats good, cause you cant max out the rating
[09:31:48] <rue_house> like 640A on that other one
[09:32:10] <[VEGETA]> no one can give 640amps!
[09:32:18] <[VEGETA]> these devices can 1,2, or 3
[09:32:27] <rue_house> do they say what size pads and heatsink you need to output the continious rating?
[09:32:35] <[VEGETA]> but can you explain why i need special inductor
[09:32:49] <rue_house> do you know about the skin effect?
[09:33:05] <rue_house> high frequencies do not penetrate the wire
[09:33:11] <[VEGETA]> i know a little
[09:33:40] <rue_house> so, all the current is on the surface of the wire, you get a lot more resistance from the wire, to get around it you need to increase the surface area by paralleling lots of small wires
[09:33:45] <[VEGETA]> i know that current will "move" on the outer surface/sides of the conductor
[09:33:57] <rue_house> even at 57Khz in crt monitors you see them doing it
[09:34:35] <rue_house> Note 4: Based on Over Current Detect testing
[09:34:38] <rue_house> did you read note 4?
[09:35:05] <[VEGETA]> yes but didn't know what they mean by it exactly
[09:35:12] <veverak> lal
[09:35:14] <veverak> :)
[09:36:01] <[VEGETA]> there is a pcb layout part in the end too
[09:36:58] <rue_house> [VEGETA], you should probably buy a premade power module
[09:37:43] <rue_house> what do you want, 3V? 5v?
[09:37:59] <veverak> hmm
[09:38:05] <[VEGETA]> rue_house I want to design my own pcb which has this switching supply supplying 5v and another 3.3v linear one for small electronics
[09:38:20] <[VEGETA]> I take it as a learning project too
[09:38:37] <veverak> 3 joints, each can report OK, DISABLED, OUT_OF_RANGE, UNFINISHED
[09:38:38] <[VEGETA]> i won't learn if i continue to buy breakout boards
[09:38:57] <veverak> with 3DOF leg it's only around 3^4 combinations?
[09:39:05] <veverak> no
[09:39:08] <veverak> 4^3!
[09:39:19] <veverak> 64 combinations
[09:39:21] <veverak> jaj
[09:39:44] <[VEGETA]> do i need a heatsink for a switching supply?
[09:41:30] <rue_house> arg, running late
[09:41:47] <[VEGETA]> they specify a certain inductor for this 2 amp regulator
[09:41:58] <[VEGETA]> and they draw a nice pcb layout
[09:42:04] <[VEGETA]> for better layout
[09:42:54] <rue_house> I see
[09:43:24] <rue_house> looks like the are assuming a 4 layer board, I dont see why it couldn't be a 2 layer tho
[09:44:24] <rue_house> I wonder if you'll spend $50 to have boards made for 10 of them and then find out they have a flaw and need to either be given an ugly hack or be thrown out....
[09:45:09] <[VEGETA]> for the vias ?
[09:45:14] <[VEGETA]> this is why 4 layer
[09:46:08] <rue_house> the chips are only $2 ea
[09:46:49] <[VEGETA]> yes very cheap
[09:47:26] <[VEGETA]> i thought of using swithcing controllers with external mosfet
[09:47:32] <[VEGETA]> if this won't work
[09:49:08] <[VEGETA]> isn't chinese manufacturers offer cheap prices for 1 board?
[09:49:36] <rue_house> http://www.aliexpress.com/item/5PCS-DC-DC-Buck-Converter-Step-Down-Module-LM2596-Power-Supply-Output-1-23V-30V/2013251353.html
[09:49:37] <rue_house> http://www.aliexpress.com/item/5-pcs-Mini-360-DC-DC-Buck-Converter-Step-Down-Module-Ultra-Small-Power-Supply-Module/32412017021.html
[09:49:46] <rue_house> I'm gonna be interested to see how this works out for ya
[09:51:32] <[VEGETA]> I saw these but i told u, i want to make a custom pcb to do my stuff
[09:51:39] <[VEGETA]> i won't do any breakout boards
[09:51:45] <[VEGETA]> i wanna learn too
[09:52:04] <[VEGETA]> but, isn't that first board is exactly like my circuit?
[09:52:15] <[VEGETA]> i mean...
[09:52:50] <[VEGETA]> one chip like mine + 2 filter caps + 1 output inductor + selection resistor (pot) which i won't use
[09:55:11] <veverak> anybody experience with these
[09:55:13] <veverak> http://www.aliexpress.com/item/Free-Shipping-10PCS-LOT-4-Phase-5V-28BYJ-48-DC-Step-Stepper-Motor-With-ULN2003-Driver/32329316992.html?spm=2114.01010208.3.80.RRMcEG&ws_ab_test=searchweb201556_10,searchweb201602_1_10057_10056_10065_10037_10055_10054_10069_301_10059_10058_10032_10017_10070_10060_10061_10052_10062_10053_10050_10051,searchweb201603_9&btsid=7fbe12bd-0925-4b0f-a476-766b128a62ee?
[09:55:15] <veverak> nah
[09:55:17] <veverak> without ? on the end
[11:25:01] <[VEGETA]> I wonder how can one solder 0402 packages :D
[13:53:49] <jennie> hello
[14:50:59] <femtogram> hey guys, let's say i'm looking at moving obstacles and i have their positions and velocities at time t_o. what's the right way to make a prediction of where the object might be in time t_n? i'm looking basically for a range to know where is safe and where is not. thanks!
[15:55:15] <deshipu> femtogram: well, a conservative estimate is that it will be somewhere within the light code on the original position
[15:55:22] <deshipu> light cone
[15:55:25] <deshipu> sorry
[15:55:50] <deshipu> that is, you know that matter can't move faster than light, so it couldn't have moved outside of that
[15:56:10] <deshipu> otherwise, unless you add more constraints, all bets are off
[15:59:31] <z64555> do it continously
[15:59:58] <z64555> can put in a filter algo that guesses the acceration, and have your controller compensate for it
[16:00:46] <deshipu> that still requires you to assume some limits on the acceleration changes
[16:01:09] <deshipu> wihout more information that's not possible
[16:01:35] <z64555> it's part of the filter, any bogus readings will be junked
[16:01:46] <femtogram> hmm... okay, interesting.
[16:02:14] <femtogram> so basically i should just make a cone based on the velocity and acceleration limits of the potential obstacles?
[16:03:25] <z64555> yup, if you drew it out, it would look like those hurricane forecast maps
[16:03:35] <femtogram> right. makes sense
[16:03:51] <femtogram> what literature is there on this btw?
[16:04:08] <femtogram> i guess the most common obstacle is for people that are walking around
[16:04:19] <femtogram> s/is for people/are people/
[16:04:38] <z64555> that and cars
[16:04:46] <z64555> birds, lots of things that are moving
[16:04:46] <deshipu> and cats
[16:04:52] <deshipu> also ants
[16:04:57] <femtogram> lol
[16:04:58] <deshipu> hundreds of ants
[16:05:12] <femtogram> cars are interesting though because i would imagine they would be easier to model than a lot of things
[16:05:12] <z64555> just crush them under your tank treads of freedom
[16:05:35] <femtogram> also, what should i be google searching to get more info on this?
[16:05:47] <deshipu> you could also just install a flamethrower and only avoid the obstacles that don't run away
[16:05:58] <z64555> google's self-driving cars might have some info on them
[16:06:15] <z64555> there's also a few darpa challenges that experimented with them
[16:06:56] <z64555> having a robot drive a 2-ton dump truck is scary
[16:07:54] <deshipu> femtogram: when it comes to people, there is an interesting thing
[16:08:21] <deshipu> femtogram: when people navigate around other people, they take a lot of very small hints to predict where they will go
[16:08:47] <deshipu> femtogram: for instance, if you wear dark glasses, people will have harder time to pass you on the sidewalk, because they can't see you eye movements
[16:09:03] <femtogram> oh
[16:09:04] <femtogram> interesting
[16:09:08] <femtogram> huh
[16:09:18] <deshipu> femtogram: no, that makes me think, if you built your robot so that it's obvious from looking at it where it will go
[16:09:29] <deshipu> femtogram: people will instinctively avoid crossing its path
[16:09:42] <femtogram> deshipu: yeah, that's part of the design actually
[16:09:52] <z64555> just put a cattle gaurd on the front
[16:10:25] <deshipu> femtogram: even small things, like rotating the head before it turns
[16:10:38] <femtogram> i suppose that's true
[16:11:01] <deshipu> or just generally looking in the direction of its target
[16:11:31] <z64555> turn indicators
[16:12:07] <femtogram> https://www.ethz.ch/content/dam/ethz/special-interest/baug/igp/photogrammetry-remote-sensing-dam/documents/pdf/ess10ijrr.pdf this looks interesting and pretty related
[16:13:23] <z64555> image processing is a very interesting and fascinating subject
[16:14:47] <z64555> With just one camera, you can't detect objects unless it's moving
[16:15:13] <femtogram> yup, that's true
[16:15:17] <femtogram> well
[16:15:26] <femtogram> unless you know what the objects look like
[16:15:32] <z64555> yep
[16:15:32] <femtogram> ie classification and tracking
[16:15:54] <z64555> in specialized areas where you can paint stuff a certain color
[16:16:21] <z64555> but with two cameras, you can compar the images of both cameras to detect objects
[16:16:24] <z64555> *compare
[16:17:25] <femtogram> also if the camera is moving you can use that to compare
[16:17:29] <z64555> I keep meaning to look more into that, but I haven't seen much material on it
[16:17:38] <femtogram> with stereoscopic cameras?
[16:17:49] <z64555> well the whole field of image processing
[16:17:51] <femtogram> it's basically just feature detection, extraction, matching, and triangulation
[16:17:53] <femtogram> ahh
[16:17:55] <femtogram> yup
[16:18:16] <z64555> I know some general concepts but I didn't know enough matrix math to make sense of it in college
[16:18:23] <femtogram> ahh
[16:18:36] <femtogram> yeah, linear algebra is important for... pretty much everything actually
[16:18:58] <femtogram> computer vision, machine learning, robot motion, etc.
[16:25:21] <femtogram> deshipu: you mentioned light cones earlier. what is the general term for that kind of cone to predict where objects may end up? i'm having trouble with my google searches :/
[16:25:36] <z64555> cone of uncertainty
[16:27:27] <femtogram> ahh, thanks
[16:29:04] <deshipu> femtogram: light cone is a term from relativity physics
[16:29:42] <femtogram> hmm.. alright
[16:30:15] <z64555> that's based on the idea of refraction
[16:30:22] <z64555> or something related to it
[16:35:40] <femtogram> z64555: uhh... i'm getting a pretty different gist from the wikipedia article:
https://en.wikipedia.org/wiki/Light_cone the graph there seems to imply basically that given light is emitted from a single point in time, there is a circle that the light hits simultaneously at every other point in time.
[16:36:29] <deshipu> femtogram: and since no object can move faster than light, you will always find it within the light cone of its original position
[16:36:32] <femtogram> deshipu: so the light cone has two spacial dimensions. in truth it should be a cone-like shape in with 3 spacial and one time dimension right?
[16:36:37] <femtogram> yp
[16:36:39] <femtogram> *yup
[16:36:55] <deshipu> femtogram: yes, actually it's a sphere
[16:37:03] <deshipu> femtogram: with a cone-like part in time
[16:37:11] <femtogram> yup
[16:37:13] <deshipu> femtogram: but it's eaiser to visualise as a cone
[16:37:16] <femtogram> that makes a lot of sense
[16:37:17] <femtogram> cool
[16:37:21] <femtogram> very interesting
[16:38:14] <deshipu> so if you have some heuristic about maximum speed or acceleration or acceleration change, you can make similar estimations
[16:38:26] <femtogram> right
[16:38:42] <deshipu> possibly relate the acceleration to size of the object
[16:39:01] <deshipu> then again, doesn't work for cars
[16:41:20] <femtogram> yeah
[16:41:24] <femtogram> hmm...
[16:42:01] <femtogram> so with hurricane prediction cones they use some sort of model to predict where the hurricane might go correct? and they apply basically this same principle to that
[16:42:36] <deshipu> yes, the better restrictions you have, the narrower your predictions
[16:42:53] <femtogram> alright, cool
[16:44:03] <femtogram> what's also interesting about this is that you can also assume non-instant deceleration, so i guess instead of being a cone it would be like a cone cut out of a cone or something
[18:51:51] <anonnumberanon> https://imgur.com/G6j4BBF
[18:52:00] <anonnumberanon> daily reminder that space is hard
[18:57:47] <Snert_> possibly the hardest thing man has ever done.
[18:58:22] <anonnumberanon> After turning into humans from apes.
[18:58:37] <anonnumberanon> That was hard too ;)
[18:59:27] <Snert_> they had more time though.
[18:59:36] <Snert_> we want to get to space yesterday
[19:03:03] <anonnumberanon> I know lol
[19:03:08] <anonnumberanon> so frustrating
[19:03:30] <anonnumberanon> one planet is not enough also, imho
[19:20:19] <mrdata> it's all we will be allowed, until we demonstrate that we can play well with other species
[19:22:11] <mrdata> and not mess up the place
[19:24:42] <Casper> which will be hard with kim hong-un, poutine and thrump...
[19:53:10] <anonnumberanon> Trump will preside we know that now?
[19:55:27] <Casper> I hope he will not
[19:55:39] <Snert_> His vote will me mine
[19:57:28] * Casper throws acid at Snert_
[19:57:48] <Snert_> he's the devil I know. As opposed to the devil I don't know.
[19:58:04] <Snert_> Billary is crooked as the day is long.
[19:59:27] <Snert_> Casper .. and if that's Lysergic acid dyethelimide ... kool :)
[19:59:53] <Snert_> (however one would spell that)
[20:00:30] <Casper> I was more thinking about one of those fluoric acid...
[20:01:42] <Snert_> seriously, there's enuff checks and balances in the system to stop Chump if he were trooly bad.
[21:09:01] <rue_shop3> using the 555 as a logic level converter and fet driver, but it starts to distort at 20Khz
[21:09:11] <rue_shop3> which seems low for it
[21:12:31] <Casper> dosen't surprise me
[21:12:46] <Casper> the 555 is relativelly low current for a fet driver
[21:12:49] <Casper> only ~200mA
[21:13:01] <Casper> good fet driver will go past 1A easilly
[21:17:17] <anonnumberanon> Should you use 555 as an FET?
[21:17:56] <rue_shop3> fet driver for me
[21:17:57] <anonnumberanon> So if we assume killing Bin Laden is good, we should say that Hillary put in a lot of energy to make that happen and therefore that she did a good thing?
[21:18:09] <rue_shop3> at the freqs Im doing, 200mA is fine
[21:18:13] <anonnumberanon> I feel like democrats are kinda good for the country right now...
[21:18:47] <rue_shop3> so I hear the usa is gonna make mexico build a wall along the border? :)
[21:19:26] <orlock> Now canada just nneeds to make the USA do the same thing
[21:19:43] <Casper> rue_shop3: how much CFM would you use as the sole air source for a welding booth?
[21:19:50] <anonnumberanon> ah man Trump went to mexico I havden't yet read what he did there or how many walls he wanted to build
[21:20:33] <anonnumberanon> but if he did build walls that would stop mexicans from mowing lawns and then liberal arts majors would start doing it instead of trying to learn technology and then us real nerds would get more jobs
[21:20:41] <Casper> mexico president said: no way mexico will pay
[21:20:55] <Casper> trump said they will definitivelly pay for it
[21:28:16] <anonnumberanon> lmao
[21:30:50] <rue_shop3> Casper, I have the great outdoors, why would I need a booth?
[21:33:13] <rue_shop3> the most dangerous person is the person who thinks they know what they are doing
[21:36:00] <Casper> the great outdoor is too cold in winter
[21:38:32] <anonnumberanon> rue_shop3, have you ever seen a bear in the wild where you live?
[21:38:42] <anonnumberanon> Canada has bears...
[21:38:55] <Casper> canada have -40°C
[21:39:42] <rue_shop3> well, at 20khz, the 555 switches the IGBT pretty damn good
[21:40:06] <rue_shop3> anonnumberanon, no, just my back yard
[21:40:18] <rue_shop3> pushin down my damned hedges
[21:40:38] <rue_shop3> then the deer eat them
[21:41:13] <rue_shop3> when my cats pass away, I might make my next pet a raccoon
[21:41:29] <rue_shop3> it'll solve a multitude of issues
[21:44:30] <rue_shop3> @$^#%*@# the 555 switches it WAY better than my little transistor totem pole driver
[21:44:57] <rue_shop3> I want a fet driver that comes in a 4 pin to-92
[21:47:00] <Casper> I had one that I made with a totem pole, 100kHz, near squarewave, rounded corner that's it
[21:47:20] <Casper> so offered a very nice turn on and off
[21:47:31] <Casper> ... I forgot what part I used :(
[21:48:16] <rue_shop3> LM386 works ok, 555 works good
[21:50:03] <anonnumberanon> you should try to run servos with a 555, that's your style
[21:50:48] <rue_shop3> WAY too much trouble
[21:50:55] <rue_shop3> tuning a tiny13 is MUCH easier
[21:51:07] <rue_shop3> AND, tiny13, --2 servos--
[21:51:14] <rue_shop3> same 8 pin package
[21:51:37] <rue_shop3> but a tiny13 cant drive a 12V IGBT gate
[21:53:42] <anonnumberanon> I'm waiting for your 18 servos servo driver with negligible jitter.
[21:53:47] <rue_shop3> its taking 15us for my driver to fall, I think I used a resistor :/
[21:54:07] <rue_shop3> why? controllers are so cheap I can just do 8/uC
[21:54:20] <rue_shop3> gang up all the serial protocol
[21:54:50] <rue_shop3> 1 port, 32 servos across 4 boards
[21:55:27] <rue_shop3> THO
[21:55:55] <rue_shop3> with that I'v learned about these TL4950, I could prolly get all 12 bits of resolution for 16 channels ea
[21:56:33] <rue_shop3> totally possable to chain those up
[21:56:55] <rue_shop3> hmm
[21:57:04] <rue_shop3> 4096 counts in 2.5ms
[21:57:13] <anonnumberanon> thats good
[21:57:37] <rue_shop3> the uC can delay the first 0.5ms, so you get full rez on the pwm
[21:57:51] <rue_shop3> using the tl5940's glitch
[21:58:07] <anonnumberanon> calls for potentiometers for human control that have that same res, or should i say the ADCs
[21:58:34] <rue_shop3> whatever!!! it dson't matter, right now I'm making a project that controlls gutted servso that have a direct wire to the motor and pot
[21:59:27] <rue_shop3> its about having a stable signal that the servo hardware dosn't jitter because of
[21:59:39] <rue_shop3> and being able to bump within the deadband
[22:00:20] <anonnumberanon> is there usually gains to be had by making your own servo hardware or not?
[22:00:36] <anonnumberanon> i know a lot of time their pots suck
[22:01:23] <rue_shop3> yea, I need current feedback, position feedback, and to be able to just shut it off
[22:01:40] <rue_shop3> (yaya, most shut down if you stop pulsing them)
[22:02:02] <rue_shop3> the target of my baord is that two can be used to make a force feedback system over the wire
[22:02:24] <rue_shop3> tho, I'm mentally having a chicken-and-egg issue
[22:02:59] <rue_shop3> after I get it working, I want to make one half virtual
[22:03:01] <rue_shop3> :P
[22:05:48] <rue_shop3> damnit
[22:05:57] <rue_shop3> the 555 works better than my fet driver board
[22:06:23] <rue_shop3> how the hell I gonna use these up
[22:06:44] <rue_shop3> (I'm not talking about the drawr of 555's)
[22:08:04] <anonnumberanon> my 555 areburried inside my electronics shop which itself is boxed into several large platic boxes in storage :(
[22:08:44] <rue_shop3> and every minute you spend not working with the stuff you get 1 minute closer to dieing
[22:09:15] <rue_shop3> ** this may not apply to your last minute of life due to round off error
[22:10:41] <anonnumberanon> yeah
[22:10:51] <anonnumberanon> ill be back soon on this
[22:11:03] <rue_shop3> tick tick tick
[22:11:06] <anonnumberanon> the electrons are crying for my return
[22:11:46] <rue_shop3> tommorow you could get killed by an out of control space ship.
[22:12:29] <anonnumberanon> a rogue Falcon 9
[22:12:50] <rue_shop3> or the day after tommorow, you could fall thru the kitchen floor while making supper and die.
[22:13:56] <anonnumberanon> im pretty safe though
[22:14:01] <anonnumberanon> im good at safety
[22:14:16] <anonnumberanon> just wouldn't mind having a backup of my brain
[22:14:39] <anonnumberanon> and of course a new body ready for me
[22:15:03] <rue_shop3> should my speed controller be wire-thru or should I just provide the - output for the motor?
[22:16:07] <anonnumberanon> wait what?
[22:16:33] <rue_shop3> I'm making a speed controller for my power wheelbarrow
[22:17:15] <rue_shop3> a t13 generates the pwm and a 555 drives the IGBT
[22:17:26] <rue_shop3> yea I know, but the 555 wont do 0-100% pwm
[22:18:07] <rue_shop3> bah, I'll just give the collector leg of the igbt
[22:19:57] <rue_shop3> and the t13 is actually doing 2 channels, I'm just using one
[22:20:30] <rue_shop3> I suppose if I used a dual op-amp for the pwm, I would only need to drive, and wouldn't need the level conversion
[22:20:52] <anonnumberanon> YEAH!
[22:20:56] <anonnumberanon> FIND out
[22:21:00] <rue_shop3> oh I know
[22:21:03] <rue_shop3> I have a circuit
[22:21:05] <anonnumberanon> do you own the Bible of Electronics?
[22:21:25] <rue_shop3> no, I have lots of electronics books I'v not looked at in years
[22:22:03] <anonnumberanon> well if that makes you feel better i still have to finish my C book
[22:30:21] <rue_shop3> irfz44 good for 44 or 60V?
[22:30:48] <rue_shop3> 60
[22:30:49] <rue_shop3> hmm
[22:30:55] <rue_shop3> this motor is 48V
[22:31:11] <rue_shop3> 54 top
[22:31:15] <rue_shop3> k
[22:31:23] <rue_shop3> maybe I shoudl just use a z44
[22:31:34] <rue_shop3> the igbt was fun to play with tho
[22:32:21] <rue_shop3> at 600V, I can find a better use later
[22:33:51] <rue_shop3> hey hey
[22:33:56] <rue_shop3> I have a irfL44
[22:34:04] <rue_shop3> maybe I dont need the driver
[22:35:36] <rue_shop3> 48V, 2HP
[22:35:59] <rue_shop3> 31A
[22:36:41] <rue_shop3> ohm -r .028 -i 31
[22:36:41] <rue_shop3> Wattage is: 26.908001
[22:36:41] <rue_shop3> Current is: 31.000000
[22:36:41] <rue_shop3> Voltage is: 0.868000
[22:36:41] <rue_shop3> Resistance is : 0.028000
[23:41:48] <rue_house> so
[23:41:51] <rue_house> force feedback
[23:42:03] <rue_house> you take two pots, connected to two motors
[23:42:18] <rue_house> you put the motors in series
[23:42:28] <rue_house> you make one a control pot and the other a feedback pot
[23:43:02] <rue_house> the servo loop tries to ensure that the pots are at the same position
[23:43:12] <rue_house> the series circuit of the motors ensures the torque is the same
[23:43:22] <rue_house> -- put a serial link in the middle --
[23:44:06] <rue_house> one system transmitts the pot position and recieves the motor current
[23:44:14] <rue_house> the other system
[23:44:22] <rue_house> -- ERROR CHICKEN AND EGG FAULT --
[23:49:45] <rue_house> the other system, takes its pot position, and applies current to its motor to try to match them
[23:49:54] <rue_house> the current is transmitted to the other system
[23:50:25] <rue_house> so, one is acting as a current regulator with position feedback
[23:50:31] <rue_house> -- ERROR MIXED UNITS --
[23:51:22] <rue_house> the other is acting like a servo controller with current reporting
[23:51:48] <rue_house> there must be a way of breaking this up evenly
[23:52:02] <rue_house> maybe into 3 prices
[23:52:06] <rue_house> peices
[23:52:22] <rue_house> so, you have two sides that you apply current to and report position
[23:52:50] <rue_house> you have a 3rd peice that takes the two positions, and sends out the current correction to both
[23:53:44] <rue_house> [stream: input: current.] [stream: output: position.] on each side
[23:54:30] <rue_house> [stream: input: position.] [stream: input: position.] [dual stream: output: current]
[23:54:32] <orlock> hmm
[23:54:51] <orlock> i have a 5x3 2.54mm pin block
[23:55:03] <orlock> unsure of the best way to connect to them
[23:55:09] <rue_house> this seems to indicate that I need to make the current controll really good on this controller
[23:55:12] <orlock> it's 5 sets of +/-/sig
[23:55:18] <rue_house> and I dont have to worry about the acutal servo loop
[23:55:37] <rue_house> 5x 1x3
[23:55:53] <rue_house> or 2.5 sets of 2x3
[23:55:56] <orlock> yeah, dupont style
[23:56:06] <orlock> which annoyingly enough, the local electronics part shop doesnt sell
[23:56:14] <rue_house> china
[23:56:55] <rue_house> http://www.aliexpress.com/item/80pcs-2-Raw-X-40pcs-Female-to-Female-Dupont-wire-cable-Line-1p-1p-pin-connector/32225940970.html
[23:56:58] <rue_house> 80 conductors
[23:57:18] <orlock> yeah i might just use those, other choice is servo extension cables
[23:57:27] <rue_house> http://www.aliexpress.com/item/100PCS-2x3-6-Pin-6P-Dupont-Jumper-Wire-Cable-Housing-Female-Pin-Connector-Header/32704538969.html
[23:57:29] <orlock> which are FUCKING EXPENSIVE for what they are
[23:57:49] <rue_house> http://www.aliexpress.com/item/200Pcs-3P-Dupont-Jumper-Wire-Cable-Housing-Plug-Female-Pin-Terminal-Dupont-Shell-Kit-Connector-2/32530544555.html
[23:58:10] <rue_house> hey, i __have__ a crimper, 80c for 80 conductors is cheap!
[23:58:43] <rue_house> do you know how long it takes to crimp 160 ends on that 80c cable?
[23:59:15] <rue_house> I save it for speacial stuff and usually use premade ones
[23:59:22] <orlock> 3 hours?
[23:59:40] <rue_house> @$80/hr
[23:59:50] <rue_house> = lots bought from china