gyros rock
Posted by renesis at 17:03 | permalink | 0 comments
lordpil's blog
it just has to be light, and you have to up all the ranges
you have to drive SAS
tho itll land on moons fine
F16 style
like, SAS is really good at planes
thats maybe it
its okay at space but its not greay
right it wont do that
mechjeb is ok but its picky with planes
like if you up speed and shit isnt bolted down itll just rip shit apart
yeah i forgot what i did to make it okay
Posted by renesis at 16:58 | permalink | 0 comments
ziegler-nicholds pid tuning worked on this shit and its unstable as fuck
i think automating pid tuning on something like a heat controller would be pretty simple
so you keep upping P gain until shit oscillates
that gain becomes Pu, and the oscillation period is Tu
then they figured out stable coefficients to relate Kp to Ku, and then Kd and Ki to Kp based on Tu
right i could probably tune that if i can time the oscillation
because shit in KSP needs to be pretty stable to ge off the ground
yeah
i was really picky about symmetry and bracing
right
tho you can still get light fighter jet performance
*get like
Posted by renesis at 16:53 | permalink | 0 comments
i wanted full power stall without nuisance blows
for example i have to do impulse test
and yeah have to push P waaay up and run it a long time to get shit that hot
its pretty fuckin scary
Posted by renesis at 16:46 | permalink | 0 comments
like i smelled the motor burning and my first thought, from testing speakers, was that ive broken it in
Posted by renesis at 16:25 | permalink | 0 comments
nunchuck robot would be pretty sweet
you need like, a motor and rotating disc to get them spinning, and an armpit simulator
blackmoon: also i totally trust the motor burning completely
even a few mm away from the battery
shorted batteries through blown h bridges, not at all
Posted by renesis at 16:20 | permalink | 0 comments
kid who wasnt in the class was like YOUR SPINNY SHIT WAS TIGHT
some kid was broadcasting it
and like, i changed to just P controller on the fly, shit tried for a few seconds than just gives up and smashed the table
im pretty sure short beam is the scary part
it can move it faster
like, short fat beams would be scariest
instructor was like, I FOUND A STOOL LETS MAKE IT DO 360!
im like woah woah there are kids
its designed specifically for that tho
its effectively elbow balancing sword tech
Posted by renesis at 16:12 | permalink | 0 comments
yeah that sucks
the teeth mostly looked fine
like, when i spin them alone in the body they feed great
but both sets together, and they bounce
so there was a washer between the nylon gears and the motor face
so i moved it to between the top of the brass planets and bottom of the nylon carrier
no like the bottom of the carrier is a platform with holes for pins and moving grease
and it was leaning because worn nylon
so kind of pivoting on the center gear below it
like, motor is tight shit gotta be straight or its crashing
ha
anyway with the moved washer theres more damping, it still feels detentend
but it doesnt make clicks
and the pid is handling everything else
yeah shit works
i was told i was snapchat famous for a few minutes
Posted by renesis at 16:07 | permalink | 0 comments
it was more to protect the drives
the PID will protect the motor!
it actually runs pretty cool, tuned
i had it doing the 200rpm sustained freakout in class, handles it
so washer hack works
robotshop got back to me like, k checking with manufacturer, re nylon gears when datasheet says metal
i hit them up again tomorrow
but looks like i gotta try freq response test with janky motor
or i could be a total badass, and linearize the platform movement with the gyro chip on the discovery board
its soaked in grease
and mad torquing
anyway its causing the planet carrier to lean and hit the planets underneath
haha
Posted by renesis at 16:02 | permalink | 0 comments
i havent blown a fuse yet
fuse prob too big
heh, h bridge rated at 30A continuous, fused at 25A, motor is 20A stall
ive only smoked the coils twice, that i know of
Posted by renesis at 15:51 | permalink | 0 comments
23:11:55 <+BlackMoon> Yes calculate <insanely long equasion> for all values of x.
right but im pretty sure its taking advantage of it being repeated data, so a lot of the equation kind of folds in on itself
Posted by renesis at 14:42 | permalink | 0 comments
but you laplace transform, and suddently calculus drops down in order to normal algebra
fourier is like, sample version of that
x''[mass] + x'[damping] + x[spring] = F
' is a derivative, so x is position, x' is velocity, and x'' is acceleration
so you fucking with polynomials, but x is actually a function, not a value
like, time domain is fucked
Posted by renesis at 12:44 | permalink | 0 comments
so it seems like when you get to non 2^n number of samples, theres a lot more methods
but all the algorithms ive read, its basically lining up terms in buffers and multiplying them all by each other in the right sequence
which is exactly what matrix math is to solve systems of polynomial equations
iunno
like, it takes in a time function, and pops out sines and phases, so roots of negative numbers and imaginary domain happens
like, you have these complex waves that are sine/cos/e^tx components
and modeled they involve derivatives and derivatives of derivaties, because mass and springs with movement
Posted by renesis at 12:39 | permalink | 0 comments
i think someone figured out the shortcuts in the matrix math
thats what the algorithms seem like
so like, setup a bunch of dependent equations, and cross multiple the fuck out of all the things
japan got old
An FFT rapidly computes such transformations by factorizing the DFT matrix into a product of sparse (mostly zero) factors.
heh, the 'put zeros in as many places as possible and kill all the terms' method
Posted by renesis at 12:28 | permalink | 0 comments
i think you have to?
nite
just gets worse
grades prob not as good still i stopped all nighters but fuckit im too old for that shit
Posted by renesis at 10:32 | permalink | 0 comments
also ive gotten mass spring dampers and RLC shit from both perspectives
in time and frequency domain
so from the transducer geek perspective, degree is perfect
like, im going to make a sinusoidal pwm to angle mechanical plot, and its going to come out like an passive filter electrical plot
= sex
Posted by renesis at 10:16 | permalink | 0 comments
like, the phd who do this for the software companies make $$$
those are the guys who use all the shit everyone says you never use
that said, having your face rubbed into math for years makes some shit go a lot quicker
which was kind of the point of all of this
some of the things in some of the white papers just didnt make sense
now they do
mostly
Posted by renesis at 10:10 | permalink | 0 comments
prob better than the average engineer in the field because fresh
but not as good as i should be
like, we dont do calc 2 or linear algebra
but we have to figure it out because our classes use it
bit fucked
mechatronic program is a bit nuts
i dont even know what goniometrics are
oh
shrug, get used to it
circles = everything
it usually gets easier when its applied later one and you actually know what youre going to be using
math courses are like 80% corner cases compared to irl stuff
yup
i know how to look up the algorithm
most theoretical shit is handled by software, you need to be able to check it and know why it does weird shit at its limits
Posted by renesis at 10:05 | permalink | 0 comments
they make all the MECAs and MECHs do a shop class, ton of manual machine processes, bit of mill and lathe time, little mig welding
also a prototyping class w/ PLC controllers and some goofy project
mine was sorting batteries based on voltage
mechanical engineers, mechatronic engineers
EE at this school kind of sucks
just physics and maybe a controls class
eh youre fine
do you have to take differential equations?
hardcore
so i understand a lot of the mechanical shit because im pretty intuitive with filters
no
well kind of
but thats calc 1 here
diff equations is doing second order polynomials with functions
i kind of suck at math
like, im way better now
Posted by renesis at 10:00 | permalink | 0 comments
haha
fuck working in those shoes
70s runners ftmfw
dunno, probably lots
jero32: yes
i dont think thats close to the worse thing that can happen with belts
what are those things by the ceiling on the right?
with the coily cables
sprinklers?
there is some closer to the middle, theyre by pipes
Posted by renesis at 09:38 | permalink | 0 comments
rab: everything works in openshot except flac
pretty neat
i was going to do commentary track with snowball in reaper, but snowball just disappeared after a few minutes
like, bitches im not going to buy a yeti
is just cold here
is that kid smoking?
i think its a wheel handle from another machine
that sucks
Posted by renesis at 09:33 | permalink | 0 comments
heh nice
did she drive it in combat?
regardless, girls in power armor #1
rab: loctite got into the grease
i rebuilt w/ ptfe machine oil because i dont have grease, after a bit of steel wool, so now feels fucking awesome
like, i let go of the handbars and lean and it turns instead of trying to put my face in the ground
Posted by renesis at 09:25 | permalink | 0 comments
woooo, bike isnt scary anymore
can ride without hands again
its missing two balls =\
Posted by renesis at 09:03 | permalink | 0 comments
ha
blackmoon: headset bearing cage smells like loctite
i think it mixed with the grease
oh wait i think it was rab ans sculptor here when i was bitching about this
because ive smelled it a lot?
its just super specific
like CA, nothing else really smells like it
anyway, mtn bike geek says he uses loctite for bunch of stuff at bike shop he works at
so i used small parts stuff and it works great, but i think some dripped out into the bearing race
because my steering is damped and its scary as fuck
jero32: loctite is like 20 diff thingds
Posted by renesis at 08:32 | permalink | 0 comments
if companies i worked for didnt have an IT department or they were remote, the solution was always 'the kid'
geek in khakis and a polo shirt
mostly they run anti virus
i bet they still defrag all the ssd because fuckit\
Posted by renesis at 06:03 | permalink | 0 comments
nice
federal aid expected family contribution: $0
i remember feeling awesome when it was below $10k
kind of doesnt matter because i will hit my lifetime credit max in first semester of year
which is both awesome and sad
Posted by renesis at 05:51 | permalink | 0 comments
wtf i google maps 'beauty supply' and airgas pops up
Posted by renesis at 03:36 | permalink | 0 comments
Top | Add to Technorati Favorites
© 2007 lordpil. XHTML 1.0! CSS! Site design by GNAA Blog Engine by pbx | MULTI2 | ian hanschen | lolwat