they do it on production automobiles
the VW DSG system isnt unreliable
and its crazy fast
Posted by renesis at 20:45 | permalink | 0 comments
lordpil's blog
nm its a turbo
and yeah its dealing with lag
so pretty much what i said, it just cant make the engine response with the torque instantly
so it puts the elctric motor on the engine
yeah shifts in F1 are like prob under 100ms now
i dont think theyre trying to remove momentum from shit in creative ways
the old kers system was flywheel based on some cars
during braking it spooled up a flywheel, then they pushed a button to couple it
some shit like that
yeah but i think theyre dual clutch
so they already have the other clutch matched
VW golf has a dual clutch
fords now too but theyre a joke
Posted by renesis at 20:34 | permalink | 0 comments
right but that dioesnt make an engine behave a certain way at a certain rpm
if you give it more gas to get the shit faster youre essentially running out of power
this just puts more power in a way that makes the engine response linear
instead of just dumping more fuel
your method is basically just put more gas
timing can only do so much
i guess because in certain conditions, more gas and timing dont help instantaneously
like, you tell the car you want 90% of torque
and the only way it can do that is dumping as much fuel as you can think itll burn
which might not instantly be as much torque as you asked for
because who knows, the shit just doesnt have the power there
so it puts the motor
until the gas engine can make the requested power and it backs off
i dont think the mclaren is a turbo
The electric motor can be deployed manually by the driver or left in automatic mode, whereby the car's ECUs 'torque fill' the gaps in the petrol motor's output, which is considered turbo lag.
Posted by renesis at 20:29 | permalink | 0 comments
snap oversteer shit when coming off throttle to brake
right
but youre on the accel right before that
like, if your car is balanced turning in under breaking, then a sudden change in power to the rear can knock you loose
and the drivers dont really control it that much
drivers are mostly about fast reflexes now, the shifting is tactile switches now
throttle is totally programmed and variable
mclaren has an electric motor in their new road car
they use it to fill power gaps
to make torque response linear with pedal input
like, thew motor is there to make the motor deliver power more smoothly, its not just there to put more
well, kinda
but that seema subtractive
like, you can pull power in spots to make it linear
you cant really add power unless you allowed headroom the whole time
headroom isnt racecar
mclaren roadcars are more racecars than most racecars
Posted by renesis at 20:24 | permalink | 0 comments
mclaren + honda + turbos = domination
what difference can 30 years make
heh
but yeah, im gonna buy the overpriced hat
pretty sure you have to get creative to not just have it eat up your boost
pretty cool way to get over throttle lag tho
maybe runs powered at spool up, open during boost, and regen during spool down
i guess not so complicated
but i imagine if transistions arent smooth you drive the driver nuts
Posted by renesis at 20:19 | permalink | 0 comments
i guess its not kinetic, just a motor/gen on the turbo shaft
heh @ active spooling the turbo to reduce lag
blackmoon: haas automation is going to run a team next year
american formula 1 team, its going to use ferrari drivetrain tho
like, we make the best mass production CNC in the world, but we cant make an engine
mikk0: development wise, probably
they would have had to bought f1 engineers and debug themselves
i dont care so much because mclaren-honda is a thing again in 2015
Posted by renesis at 20:14 | permalink | 0 comments
14:57:46 <@BlackMoon> not allowed active suspensions, fuel injection, turbochargers, traction control, yadayadayada
theyre running turbochargers on v6 this year
theyre using the turbo mass as part of a kinetic energy recovery system
push to pass, boost button
blackmoon: the most powerful era of F1 was the v6 turbos in the 80s
and the most uncompromised racecars, like not fucked by rules, were prob late 90s cars
right before they introduced grooved tires
cars are still slower than the late 90s cars
http://www.formula1.com/inside_f1/understanding_f1_racing/8763.html
MGU-H is an energy recovery system connected to the turbocharger of the engine and converts heat energy from exhaust gases into electrical energy. The energy can then be used to power the MGU-K (and thus the drivetrain) or be retained in the ES for subsequent use. Unlike the MGU-K which is limited to recovering 2MJ of energy per lap, the MGU-H is unlimited. MGU-H also controls the speed of the turbo, speeding it up (to prevent turbo lag) or slowing it do
Posted by renesis at 20:09 | permalink | 0 comments
tho rab is supposed to hate it since it is laying waste to his home
Posted by renesis at 19:01 | permalink | 0 comments
in solidworks it would bounce the dong
at like 1hz to show deflection
also how are you holding it?
prob by the balls somehow but are you modeling how it reacts when being held, when just hanging out, lots of stuff to figure out before we fea
haha man im all paranoid like FUCK YOU GOTTA START YOUR HOMEWORK OR YOULL BE A GUILTY MESS
im half done
yay weed
=\
blackmoon: do you follow f1 at all?
#electronics just doesnt care about f1
Posted by renesis at 18:52 | permalink | 0 comments
your cock to bals interface is horrible
some sort of table for the scene?
better be up for F1
Posted by renesis at 18:03 | permalink | 0 comments
i think either something solidworks is patched, or it messes with your card ID
judging from some of the complains about behavior after the hack, my guess is its changing non-cad card ID into cad card ID
like, my ATI drivers have been behaving for awhile now im not trying to poke them
timecop: do you solidworks yet?
timecop: say yes and tell me how well it does in vm
my guess is it doesnt
Posted by renesis at 17:56 | permalink | 0 comments
let me buy one
ill pay you next year
youre not using both
bullshit
they will parallel process?
do they need a stupid link cable?
oh its not a hardware thing
they dont show up as one card like a crossfire setup
oh they do
did know they did that with their cad cards
*didnt
i rabbed
maybe i sell these and get a quadro
there will be a new fallout eventually, tho
dude i want shiny solidworks parts
isnt
and yeah i tried the patch thing, prob just gave myself a trojan, shit dont work
didnt look deep enough into it to see what the actual hack is
i dont think its just reg editing
Posted by renesis at 17:51 | permalink | 0 comments
and within a month, they were below $350 again
is that nvidia?
nvidia sucked for mining
ati cores were just way better, it didnt have a lot to do with graphics processing capability
right i wish i had one
i cant render in solidworks
yeah serious
a pair of 280x
i have more gpu mem than system mem right now
because the only shit i havent upgraded is mobo/ram/cou
and sadly shit is doing well enough i dont really need to
so sick of this core
fucking crap out already
this is the problem with building good pc out of good parts and knowing wtf a chipset is
your pc last until youre sick of them
zeeshan: yeah i couldnt get that working
i had driver cert disabled and the app seemed to run but i still coouldnt render
its because they dont want it to lock up solidworks
because a lot of rendering would just get standard gpu stuck
sounds right for a quadro
Posted by renesis at 17:46 | permalink | 0 comments
that article, which is old, is just the courtroom outing of the situation
hashing random numbers against ID data to produce a certain number of leading zeros
well, similar, i think thats scrypt algorithm i dont know is the sha256 one is the same
zeeshan: kind of
it scales difficulty depending on how much pwoer is being thrown at it
to keep coin generation a consistent rate until all coins are allocated
right because the amount of power just increases
its not a consistent increase
it would get less if people abandoned mining
you cant
its over
there was a point earlier this you where you could mine alt coins and convert to bitcoin for profit
but it was the edge of profit
and it was a small window
gone now
and only if you were on multimining pools
that basically raped alt coins
literally
there was a few months where those would sell
i got into it because the hardware i needed to mine was what i needed to upgrade anyway
so like, even i didnt break even who cares
i didnt buy 20 GPU i bought two
prices of gpu that could mine went from like $350 to 500 in a few days at one point
Posted by renesis at 17:41 | permalink | 0 comments
custom logic
speed, versatility, simplicity
but usually you want custom silicone
and cant afford it or dont have time
so you fpga
at a huge impact in cost
right
because people scamming buttcoin dont have time or contacts or resources to do real silicone
so they kick out some fpga shit
after premining with their hardware while its still profitable
yeah
but they still premine
because it will never make monetary sense to ship what you can make more money on
heheh @ burn-in tests
WHY DID YOU USE TEST NETWORK!??!!?
witness: because boss said we dont make money using test net
right because theyre not making enough units to make custom fab of IC realistic
so they charge you guys a fuckton for custom logic
fpga should be fast if youre doing it right
cpu have overhead that make them versatile
work in a lot of situations
a burnt FPGA is basically a prototype ASIC
asic is application specific ic
custom for your system
FPGA is prototype version, shipping FPGA is kind of like shipping prototypes
in terms of cost and turnaround time
its not my idea its their idea
everyone knows they were doing that
Posted by renesis at 17:36 | permalink | 0 comments
hmm fuck
so my calc book had like 8 steps to figure out this total area of split pig pens with 750ft total fence problem
and i did it with math, which is what it usually wants, and got the right answer
fuckin book is never happy
its 14062.5 ft^2
its using ft not me man
im pretty sure instructor would mark it wrong if i gave the correct answer in meters
tho
i kind of want to try
but yeah book it like, draw some pictures and guess
draw a diagram and label
have some tea and estimate things
dunno that doesnt seem like a fuckton to me, shrig
i guess they come with a lot of peripherals so you dont have to use up gates on everything
heh
i dont think anyone really specs them like that
maybe marketing types
but who knows where they got their answer
i dont care how many gates the counter peripher took
well all logic is gates
basically
Posted by renesis at 17:31 | permalink | 0 comments
ha, i know some linux sucking swedes who follow your blog
theyll post links and im like, heh i know that guy
Posted by renesis at 16:15 | permalink | 0 comments
they have full ceramic heaters now
i dont pay that bill
srsly
still broke as fuck
gonna have to start prioritizing which bills to let drop, ha
next day prob felt great!
we lived in apartments so no one really came to get candy
so we mostly kept what we bought for trick or treaters
(apartment kids usually go to house neighborhoods or malls because better candy)
population density doesnt make up for it
Posted by renesis at 15:01 | permalink | 0 comments
def an 8b problem
the 420 thing really only is clever with the msp430
i kind of miss my flower vape
last thing had an ir remove, 3 fan speeds plus off, temp set and display, sleep timers, stupid little beeping status indicator, backlit LCD
think i kind of liked the adjustable dimmer vape better
Posted by renesis at 14:56 | permalink | 0 comments
one just dropped almost $100 on a fancy vibe
i dont remember which, i trust that its awesome tho because shes an ME
zeeshan: why did you get a backwards FD
macegr: seen people walking around with digital mod vapes?
the electronics usually take up about 2" past the 18650 battery compartment, plus atomizer fitting is like 1/2"
theyre basically light sabers
mechanical mods not quite
msp doesnt mean anything to 99.9% of stoners
it means something to me and it kind of makes it less desirable to own
Posted by renesis at 14:51 | permalink | 0 comments
other side of the battery?
like it has a little missile launch pad stand?
thats pretty funny
i know least two girls would prob be down to test
Posted by renesis at 14:43 | permalink | 0 comments
timecop: you find solution to button problem?
most of those things have UL certs?
pic?
the two diode thing?
pic.
but sounds like maybe it works
my first designs i thought the input was edge detecting for some reason, idle state was different
timecop: yeah usually
test pcb would prob need most the shit on the real one in your case
zeeshan: both times ive banned him, channel gets back to productive engineering convo within seconds =)
Posted by renesis at 14:38 | permalink | 0 comments
jezus fuck every other day YOURE NOT ENGINEERING I AM ENGINEERING
ya!
do you masure the hammer and everything on what youre measuring first?
because if not, is not engineering
actually timecop you kind of are
i cant do that
whenever i strike my numbers stamps twice i get little shadows
thats not how it works zee
there is an anvil and punches and a big hammer
and thats just how you do in plumas lab
also i think we just have mechanical presses
and one odd electric flywheel punch
both all loaded up with tooling
i cant, not in the lab
zeeshan: same argument can be made both ways, i just been here longer, and thats how irc works
less fighting
i just made fighting go away
Posted by renesis at 14:33 | permalink | 0 comments
right but do you measure all dimensions of the large slong?
or just the ones that matter
theres two people ohsixi
you and zeeshan
you ship back?
theyve never asked for shit back when we show them fuckups
ohsixi: would you measure every vein?
all the complex radius
and reject it because it wasnt your vein even tho it was a good vein to everyone else
would you even go back and match the drawing to the new vein?
right
but if you can use it and you meet schedule because of it
why reject it
figure out something, make them fix it if you want, pay less
source the part from somewhere else for next build
again, this doesnt mean you measure everything
09:31:08 < ohsixI> it is like zeeshan's car parts, it will not mate with other parts
you dont need to measure everything to make sure shit fits
thats kind of the whole point of GD&T
zeeshan: cool
change of convo anyone?
Posted by renesis at 14:28 | permalink | 0 comments
what?
dude the shit in the program is not the manufactured board
timecop: neat
right so is optical inspection
well, because design rules
what do you mean not because, thats how the software is going to keep your shit in spec
youre dumb if youre setting up design rules after
zeeshan: its like .008 for decent precision, .004 for high end precision, .012 if you want to be able to use shit board houses
thats kind of the standard
space/trace
timecop: yeah thats not inspection
thats just design based on a datasheet drawing
zeeshan: well, capacitance is a big issue
this is an inspection convo i think
well also another convo where ohsix is the authority on engineering
and if youre not doing 100% inspection youre not engineering
and your products arent engineered
even tho they work 1000% okay and meet critical dimensioned spec'd
Posted by renesis at 14:23 | permalink | 0 comments
no im saying it works that way in the shit i do
and i said not everything is measured
only critical dimensions is not everything
#cars flys your shit timecop
i think that kind of kills you only pros statement
um
i started a convo
about a specific thing
and then you decided to tell me how todo it another way
which turned into another convo
wait so the EE shit people do in here isnt engineering?
because i know for a fact people in here arent measuring every dim
what?
so you saying the pcb is the gerber?
and you dont need to check?
timecop: exactly
timecop: so you measure everything on your boards every time you fab?
in the sense that ohsix and zeeshan say
you measure everything
Posted by renesis at 14:18 | permalink | 0 comments
from dc to light
dude no one measures everything
sorry, it just doesnt happen
you measure critical dimensions
usually you stick those on the drawings
that barely anyone looks at because everyone has the 3d models
how is 80% close to 100%
right so you measure those
well thats ford, they have lots of money
well yeah thats kind of critical
thats looks easy to inspect
09:17:13 < ohsixI> you did not qualify your earlier statements with 'critical', i'll accept that i was also always talking about critical measurements and we never disagreed
okay so now youre saying youre right because when you said everything, you really meant everything critical?
but yet products are made
Posted by renesis at 14:13 | permalink | 0 comments
dude cars ship a lot of crap
ohsixi: yeah every time
but thats how the shit gets made
and in general its pretty cool, and the industry makes progress
but yeah, saying shit is fucked up doesnt make the project stop
thats kind of when you earn your paycheck
shit is what it is, you make the best decision, move on
ohsixi: it depends on context
obviously your pissed
but do you need to ship tomorrow and does 10% work?
honestly people just spec 5% by default
because its usually what the factories have
yeah
for lots of audio shit in signal path, 1% is just default unless its super not critical
then it becomes 5%
even when 50% would have worked
where do youget 50%? who knows
but by your arguments, if 50% works it should be spec'd
if they say they are out of what you spec'd that may happen
timecop: 3d printed trash is trashy
i made an argument for it being great to test fit and position on complex shapes
but ohsix says all of audio gear is basically shit
and the musicians are stupid
not sure i totally disagree
exactly
Posted by renesis at 14:08 | permalink | 0 comments
and a schedule, how do you get to the end
measuring all of everything every time isnt gonna get you there
hell a lot of everything, dimensions arent critical
why throw some shit away when it measures wrong when all your parts fit
who cares if the brace is 2mm off, if they says thats how they come out, change the drawing, make sure the variance observed doesnt kill you, hold them to it, move on
wtf @ measure everything, because its possible, and nothing else is engineering
um, you dont always know when specifying it
if they fuck it up, you can either roll with it or make them do it your way
schedule usually dictates
its def not gonna go down the same every time
sometimes you dont reject, you just deal with it
yeah well sure
in a liability sense
that aint gonna help you hit your dates
failed suppliers are not uncommon in china
its what competition dictates
this is the state of a lot of industries
ohsixi: when schedule dictates
okay well thats nice maybe in yours its diff
but a lot of people using overseas CM will work with a failed CM rather than reject, move to another CM
it just takes too much time sometimes
09:08:16 < ohsixI> you've already fucked up if you are off schedule with a bullshit part on hand
oh so you just give up?
shits fucked up so you just say fuck this project?
well it makes customers happy
Posted by renesis at 14:03 | permalink | 0 comments
well, design is but you have to verify you got what you designed somehow before pulling the trigger
its a lot of lengths
this is ID driven shit with a lot of dumb curves and shit
attacking shit with calipers is duymb compared to walk up to the thing and screw it down with the gasketing in place and see if shit interferes
ohsixi: end of line QC with finished parts
you reject shit that dont pass and they will make future ones within spec more often
because people dont like working when they dont get paid
this was a problem fitting electromechanical controls into machined wooden guitar bodys
and a problem with fitting plastic functional bits into speaker cabinets
zeeshan: you cant measure everything
its just not going to happen
you have a box with complex bracing and shit hitting at random angles with a lot of hidden features
its not that you cant
its that at a certain point you just take a few point measurements, and check fit and placement of all parts
and engineering isnt 100% every time
engineering is, you have given resources available, and a goal
Posted by renesis at 13:58 | permalink | 0 comments
or stratasys to make stratasys
homebrew 3d printers are like stratasys on a really bad day
ohsixi: right im saying theyre not a useful simulation of finished ABS parts
ABS printed parts
theyre a joke in engineering labs
you try and drill them and half the prototypes just fall apart
its not even easy to get the support material off without damaging the part everytime
this is top of the line gear
most of the time you need to wait for off tool or machine finish material parts before making decisions anyway
3d printing is good for checking fit and placement thats about it
we werent doing stupid shit with them thats what you do with them
ohsixi: it totalluy depends on the scale
but they hold decent tolerance within their horrible resolution
ohsixi: yeah wood cabinets dont always meet drawn spec
so you get a cabinet sample and stick the 3 printed part in the complex cnc'd pocket and see if it interferes somehow
how is CAD gonna help with that?
seriously, thats real life manufacturing
Posted by renesis at 13:53 | permalink | 0 comments
3d printing is good for visual models
and for making $500 captured gear demos
yeah except current shit probably isnt economically viable
shit will be better in 10 to 20 years but right now its pretty crappy
in a lot of cases arguments for 3d printing in manufacturing is a fucked solution for very poor DFM and DFA
what do you mean the alternative is printing nothing
right
which produces very usable prototypes
3d printed parts are a joke in prototyping, unless SLA, which is pretty brittle too
right im saying its questionable at the moment
its not going to be used in mass production manufacturing of functional components on the level of injection molded plastics, cast and machined metals
08:53:28 < ohsixI> building a 3d printer involves 3d printing and is a hobby to itself
im pretty sure they dont use objets to make objets
Posted by renesis at 13:48 | permalink | 0 comments
08:37:31 < ohsixI> that is a matter of how slowly you print and with what material
fdm is shit compared to SLA and machined parts
SLA is crazy expensive
and UV cured printed sahit is just a sticky mess
some people seem to get rashes from some of the materials
so yeah, compared to just getting shit machined, 3d printing is convenient and cheap but sucks in just about every other wayy
you cant even drill FDM without worrying about the tip catching and shattering your part
and its not like you can hold the shit tight in a vice
Posted by renesis at 13:43 | permalink | 0 comments
3d printing is ugly and not structurally sound and is difficult to machine =\
Posted by renesis at 13:35 | permalink | 0 comments
i think he mentioned feb
FCS was next year
which is ~3 months, funding to ship?
sounds like a finished project to me
Posted by renesis at 13:12 | permalink | 0 comments
dominos pizza tracker got new barry white sounding theme
oh because slow prep
because pan pizza
hmm i dont know if its cool that a pizza tracker is using this much of my subwoofer
they showed a girl making it in the vid, so im down
Posted by renesis at 13:00 | permalink | 0 comments
07:06:16 < zee-Lathe> cause i guess reversing the polarity moves the plunger in wrong direction
y
zee-lathe: diode bridge could make it non polarity sensitive
but it wont test like a normal diode in either direction
Posted by renesis at 12:21 | permalink | 0 comments
gpf: that seems expensive again
but yeah so much things
right
the $70 with almost nothing was more impressive, heh
that just seems sane
Posted by renesis at 11:27 | permalink | 0 comments
hingston: sounds like learning cnc programming
real programmers crash tools
1st semester of the cnc program at last school, youre just drawing on a notepad with an old bridgeport
felt pen loaded into a sprung loaded tool
crashing that shit was fun
ink everywhere, shattered pen barrel
knee was set so that it couldnt crash the pen holder
like, instructor would get mad because he only had so many $2 felt pens
drawn with a cnc?
maybe for a makerspace
Posted by renesis at 11:22 | permalink | 0 comments
right you can manipulate buffered state data
i imagine this is literally the only procoessing that happens
aside from maybe in peripheral modules
gpf: just because its harder to recognize output state for a user?
that actually makes sense
ohsixi: the grey keyboard looks pretty cool
i wonder if they will connect to logitech usb radios
Posted by renesis at 11:17 | permalink | 0 comments
all the rungs run in paralle, so i can put something at the beginning that causes runs to either run or skip, maybe?
if i can do that i feel like i can get shit done
hmm then i dont understand
right
but someone said its not sequential, it just goes left to right
i think zeeshan
ok thats what i thought first
but it scans all rungs before updating
basically all data is imported and exported to unbuffered space at the end/beginning
and thats it
maybe thats what he meant by parallel
hingston: right because relays
got that
instructor didnt explain, i feel bad for the kids who are actually kids (i act like one but am in 30s)
Posted by renesis at 11:12 | permalink | 0 comments
thats neat, i like parallel solutions
direct line to product?
link
in ladder code, no?
so to do a state machine, i set a value at the end of each rung?
or what
right because it just uses the values that exist at the scan start
right so buffered regs
so i can write a rung, at the end i can set some sort of state variable
and then the beginning of each rung checks for state
and i can get conditional state machine action, yes?
wat
Posted by renesis at 11:07 | permalink | 0 comments
image processing seems like a dead end to maintain in some instances
i know of two companies just seem lost trying to figure out wtf earlier people were doing
and it kind of seems like theyre discovering a lot of the processing isnt doing shit
youre going to spy on yourself with a drone
ohsixi: prob already exists?
coin on toss on weather the code is open
i would look for it tho
hingston: so you basically use the micro to preprocess something and feed table in the plc?
Posted by renesis at 11:02 | permalink | 0 comments
honestly for my own intentions, it would be me programming them or i would be around
if the project gets sold or i die, why do i care how they maintain it
friend burn out some servo interface modules
hingston: heh
actually i believe that
WE GOT THE GERMAN SHIT FOR $900 OBVIOUSLY WE CARE THE CHINA SHIT IS $80
ohsixi: okay thanks
Posted by renesis at 10:56 | permalink | 0 comments
ohsixi: so what about programming a plc in c
gpf: ohsixi says never
gpf: it doesnt even begin to meet the demands of industry and all my assumptions are fallacious
so i want to know about programming industry standard equipment in the most used language ever
well plc are hardened
but you can opto isolate a micro pretty cheap
like, basically when i open a project on the school PLC< it lists programming languages
and C was default we had to change to ladder
=(
heh yeah
well you can usually kill anything some way
reverse a voltage on a 78 series reg, eheheh...
i like the modularity of the PLC
and the fact theyre built systems ready to go, but $900 was dumb for this
under $200 seems super useful i would just want to be able to write code in basic or c
Posted by renesis at 10:51 | permalink | 0 comments
if your factory cant find a dude who can figure simple c code, you need to get out of 1985
are they strictly ladder?
fuck even basic would be better
ohsixi: to be maintainable
eventually the industry has to move forward
some tech can hook up software and edit code
ohsixi: youre mistaking me not getting it for me not caring that much
05:48:22 < ohsixI> there's no moving forward dude
yeah ill just ignore that, i guess
ohsixi: k
Posted by renesis at 10:46 | permalink | 0 comments
so we can maintain factory stuff
but to me, EE background, this is a $10 problem
to fix forever
with as much control as i need
this is a prototype
and AVR programming isnt exactly a rare skill
http://www.automationdirect.com/adc/Shopping/Catalog/Programmable_Controllers/CLICK_Series_PLCs_(Stackable_Micro_Brick)/PLC_Units/C0-00DD1-D
thats decent shit?
thats rather fair
program the PLC in c
Posted by renesis at 10:41 | permalink | 0 comments
if i cant get 10ms accurate timing, im blown away and slightly scared of industry
i just need to drive actuators, measure a voltage, and read some optical switches
these are B&R
seem easy enough to get code on
http://www.br-automation.com/en/
likely donated to the school
or on sale very very cheap
or easy to license lots of licenses
er, lots of seats
dude its $900
and i can do this with under $10 of microcontroller
8 pin AVR, some unclad perf, some artistic use of solder
right
i have a $2600 BOM, i think im at $1900 with like $400 of ohshit headroom cost as a line item
i cant
you can program these things in c, but im not allowed
has to be these controller, has to be ladder logic
because its mechanical engineering and automation engineering students
so they want to force the ladder logic
Posted by renesis at 10:36 | permalink | 0 comments
and im in a two person group and other dude is gone half the time
most people in 3 or 4 per group
but like, i dont this a few times, so...
*done
boms and fast prototyping models and shit with weird functionality is nothing new
its an awful idea?
as opposed to rolling
?
gravity and consistent placement
also it comes from the hopper idea
which was discarded
the hopper was the only part i was worried about
now its should do instead of must do, so im like yeah fuck all that
my shits gonna work
let the rest of the class try and sort shit out
they have to cycle at 3 per second
mine can prob do like 1 per second
i think a lot of the random orientation hopper ideas are going to jam even approaching 3 per second
just voltage
i figure quarter to half second there
i dont know what kind of averaging or delay the plc has
i dont think you need one for anything versus a microcontroller
gpf: i can get this one to clock a counter at 4mhz i guess
but i dunno if ill be able to use that for timing or what
if i cant get like ms accurate timing, im pissed
Posted by renesis at 10:31 | permalink | 0 comments
but yeah, as soon as they dumped the any orientation part of the design spec, i just took the hopper bullshit off
no one had good idea for that
in the speaker one?
it hops it
until it fits into the hole
it wouldnt work well
no ones ideas would
not with random batteries
gpf: just mix them with high pressure air?
ohsixi: again, these are batteries stored in buckets
no taped ends
my shit is considerably safer
oh
gpf: i wanted bounce more than vibe
its a polycarbonate tube
well the platform is a live terminal
and everything else is insulated somehow
the only thing it can touch is the pillar behind the foam or rubber pad, when it wears
and thats in electrical contact with the platform live terminal, so...
you had to do with one
go with
so i has the woofer based hopper
and the 3 actuator one
Posted by renesis at 10:26 | permalink | 0 comments
also using shit we have in class, yours is a stepper design
it adds $500+ to bom
we dont have servos and we dont have dc motors big enough
you do with the shit we have in the lab
right you need a motor with decent torque and speed control
how many actuators do you need for two bins
gpf: hand fed
originally they wanted any orientation
check this out...
http://imgur.com/eHlkqkl,OsFOg1t
hehe, i like the vented woofer hopper
but too expensive school doesnt reimburse for that much
Posted by renesis at 10:21 | permalink | 0 comments
dude if bad feeds can happen your shit can get stuck
ohsixi: you cant have the shit get stuck?
anyway, three penumatic actuators is cheap
ohsixi: ?
so your shit will have enough power to push them one way or another in a jam?
gpf: heh cool
ohsixi: guess what it doesnt matter
design is submitted, its not changing
all i wanted to know what to do to detect end condition in one edge case
moon fixed it, thanks
right and keep them low as possible
i think its like $90 in the bom for the sensor and i have like $600 left
then it wont work because the platform is the lower contact
Posted by renesis at 10:15 | permalink | 0 comments
so my problem
that i was asking for help with, is what to do when the thing decides its done, because nothing in feed tube
it pulls feed block back to push the tested battery into a bin
and some asshold drops one battery into it
we cant damage batteries, heh
see thats why little rubber bumper!
and im going to have seperate pressure regs for the ejectors and the feed actuators
blackmoon: hmm
its inconvenient because switch needs to be where aluminum extrusion currently exists
but that works
or i can put more optical sensors on the side of the feed block
i think there is space
ohsixi: what do you want me to do with that
dude its not going to happen, design is submitted
save it for when you take the class
and you way needs a stepper
or some limit switch jankiness
anyway
the actuators are cheap
Posted by renesis at 10:10 | permalink | 0 comments
im prob going to have delrin or nylon plates on top of an acrylic core
i have a fuck ton of 1" acrylic
so only issue i can see with mine is when positive ends are together
and i think with D cells the tube will be tight enough to make that not matter
and with C cells, its maybe loose enough to work but it doesnt have to work with c cells
you would have to adjust the height of the terminal block anyway
what wears?
ohsixi: so itll go from perfectly aligned to aligned enough
dude you have a revolver style magazine that feeds into terminals
thats nice im sure it would work, maybe i would have come up with something similar if we had a complete project spec 2 months ago
but this is the design
also we have to use mostly whats in the lab already
ohsixi: i dont think the inside of a polycabonate tube is going to wear that much with betteries going down it
and i dont think wear on the inside of the feed block battery holder will hurt anything
prob helps a bit
make the drop smoother
Posted by renesis at 10:04 | permalink | 0 comments
so problem happens if you drop a battery in after it decides there is no battery and moves the feed block back
why not
so
theyre 1.5V batteries
they have bins of them
this is considerably safer
and why wont it feed correctly
ohsixi: this is pretty easy to adjust
yeah
itll be pretty tight fit
and i want offset of hopper to feed block to be few mm at most
i already said, measure battery voltage and bin depending on voltage
gpf: harder to get into set of terminals
yeah that seems prone to jamming
also stepper drives add ridiculous cost to bom
Posted by renesis at 09:59 | permalink | 0 comments
assuming feed block is kept perpendicular to tube
well obviously the mechanical design would have to take that into consideration
not a huge issues, its about how high the feed block iss to make sure batteries cant stick too far out of tube in idle state
anyway
stop asking specifics about timing thats obviously something to deal with when a prototype exists
issue is what to do to make sure someone didnt drop a better down the thing when its pushing batteries into bins
yeah
sec i post image
i should make a gif
http://imgur.com/pzyllKA,nfd5svN,AclTPK2,1jpRK9a,PW0utcR,CFVnLzQ#3
design is a bit diff now
ones images are caches you can slideshow to animate
Posted by renesis at 09:54 | permalink | 0 comments
heh wtf
one plc?
gpf: why not
gpf: im wrong to assume you have fuck tons of experience programming and interfacing plc?
anyway, battery falls down a tube, is stopped on top of a feed bock, optical sensor knows its there
feed block pulls back (pneumatic) and battery falls down, feed block shoves into some terminals so battery voltage can be measured
and then feed block pulls back so one of the ejector rams can push it into a good or bad bin based on voltage
when feed block pulls back, it uses optical sensor to know if a new battery drops down
but i figure this thing processes slow and stupid so i want to do some sort of end condition check so it makes sure theres no battery there when it resets to idle state
whats a bat
oh, it cant really
Posted by renesis at 09:49 | permalink | 0 comments
k
oh neat
this is from zee germanz
the software is a bit sensitive and takes minutes to load the splash screen after double click
ha, instructor took longer than a lot of the students to get the things up and spinning from old instructors directions\
but i can get buttons on the front to blink LED status stuff on the back
so can run code but this shit seems dumb
can do this with a 53 micro setup
not really
guys an unprepared lab student
he was just gonna let students try and figure shit out, no direction, before another instructor walked in and was a bit wtf
its only 3 weeks in this class because instructor couldnt figure it out first week
yeah the manufacturing class instructors impressed with my parts im like, yeah for my AS we spent a course on what you guys spend a week on
youre GPF
Posted by renesis at 09:44 | permalink | 0 comments
then can i make the delay variable
zeeshan: they dont execute sequentially?
wait wtf
shit just goes left to right
jezus fuck
how deep can i go, left to right
http://www.br-automation.com/en/products/software/
were using this shit, pretty sure it can sim?
not sure
is cx-programmer free?
Posted by renesis at 09:39 | permalink | 0 comments
how hard is it to do conditional looping and sub programs
okay so its maintanable by techs
are rungs ID'd?
i think i noticed a number
what do you mean you dont do that
i need timers and counters
and somewhat not dumb logic
dude if i could use microcontrollers this shit would be done
if i could use c for the plc, itd prob be fun
we have to use ladder logic
xanix: who knows, maybe dead
he had health issues
he was around after 2007 i think
but not often
and not long
fuck that guy
so youre saying i cant just load a counter and have a sequence run on the overflow
i dont want static delays
where shit is just sitting there
Posted by renesis at 09:34 | permalink | 0 comments
months and months, i think maybe because the smoke particles dissolved in the water, shrug
ohsixi: observant, since bong has been said like 5 times in the convo
or just test
try both ways and see whart makes it stick out?
if they used diodes under rated for reverse polarity theyre assholes
its a coil pushing on a magnet
thank you, he likely has me on ignore
ohsixi: do you know anything about ladder logic
why is it dumb =(
Posted by renesis at 09:29 | permalink | 0 comments
no idea
my down stem clip broke
fuckin china
modern bongs are glass on glass taper fittings, from the pullout carb bowls to the stem, and from the stem to the bong
naw it works
easy to break down to clean, and water in the ground taper seals airtight real well
but sometimes you pull on the bowl to carb, and the downstem pulls out instead
so they make clips that hold the down stem fitting to the bong fitting
so when you pull the bowl the down stem doesnt pull out instead
theyre $3
so you dont expect them to last forever
i got a clear one and it was all crazed
should have just got a colored one
i used to do dental rubber tubing for seals
before glass on glass stuff
tho the latex will eventually degrade if you leave in water all the time
Posted by renesis at 09:24 | permalink | 0 comments
so wait until im recovering from a coma or something
new era brando
whatever man
kilmer = real genius, willow, top secret, top gun, tombstone
also he went to chatsworth high school, hometown mofo
Posted by renesis at 09:18 | permalink | 0 comments
macegr: woah, val kilmer was KITT?
fucked, now i feel like i gotta watch it
Posted by renesis at 08:57 | permalink | 0 comments
yeah those are his instructions, the end is all crazy shit like that
at one point hes like 'aim for the rock [lifts finger points at roof] NOW DONT BE A HERO'
ac-130u: the drivers drive mostly by sight and crodrivers feed them turns
codriver losing his place is very bad
i think most hill climb stuff is without codriver, so hillclimb is kind of one step more insane
i guess because weight and also lots of those cars dont have two seats
because its not circuit tracks
yeah long stages, would be hard to remember everything
hillclimb is as nuts as rally
blackmoon: doesnt even lap, they usually dont end where they start
theyre not continuous either
Posted by renesis at 06:48 | permalink | 0 comments
RIGHT 1 OVER CREST at 2:40 is insanity
haha BIG CREST JUMP LEFT was literally when you land be pointed left
haha LEFT OVER CREST BE ON THE ROAD
Posted by renesis at 06:43 | permalink | 0 comments
fans are as insane as the drivers
fuck there was some video few weeks ago, codriver has 'BE BRAVE' and 'DONT BE A HERO' in his notes
best codriver
http://oppositelock.jalopnik.com/be-brave-1648162861
rally.
Posted by renesis at 06:38 | permalink | 0 comments
well, they aim for the inside edges of the corners
so its not like thats the greatest place to be
on the outside well before the apex, or on the inside well after the apex is prob safest places
most these spectators have no concept of apex
so i dunno you gotta make drawings or something
with red places and green places
Posted by renesis at 06:29 | permalink | 0 comments
would be sad if he got hurt
you could keep racing after that crash
sucks when fan gets hurt and drivers pull out of series
rally is nuts
Posted by renesis at 06:07 | 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