lordpil's blog


oh, it likes the extension *.cnc
okay it imported my helix
flat because it doesnt seem to have a z axis
helix is probably thousands of little segments
this is fucked
im going to see if i get the teach pendant to do a straight line between two XY points on diff Z planes
but im not sure points have Z data

Posted by renesis at 02:03 | permalink | 0 comments


the software is kind of like venctor graphics workflow, like laser cutters, label cutters
*vector
but it outputs into some binary format
then that needs to be converted into some other binary format
then you can feed that to a teach pendant
im not sure this thing will understand things in 3d space, it seems XY plane oriented...
like lasers and label cutters
of course fren wants to ink fill on a curved surface
he also wants to use it to dispense solder paste, but im not sure how to like, dwell on point and dispense, the software is setup to do paths with layers/colors for certain nozzles
like, he could have just got random CNC and we would have stuck the dispenser syringe on it instead of normal cutting headstock
the machine itself seems cool, physically

Posted by renesis at 01:58 | permalink | 0 comments


definitely
me and other hacker got laid off
we just caused trouble
nice
heh this esr stuff
friend bought a chinese 'cnc' machine for fluid dispensing
software is so trash
trying to import gcode file, it wont show it because i dont know the magic extension and it doesnt have all files .* type

Posted by renesis at 01:53 | permalink | 0 comments


it never occurred to them to use three resistors, one for each LED
we tried to tell them, but they didnt understand :(
this was the same place where an "EE" accused me of inventing ESR when his crossover didnt work
'is this just some theory ren has, or did we test it'
'i put the capacitors in some circuit ren made, and the capacitors are different on the audio precision measurement'
you know you are fucked when the crossover designer calls the single pole RC circuit 'some circuit that ren made'
what happened was they used expensive polypropylene caps in prototypes, shit laying around in the labs
then had production made with the cheapest bipolar electrolytics we had
well, that china had
when i measured them on an LCR, the ESR was higher then the resistors and speakers they were in the circuit with
jero32: bipolar electrolytic ESR can be in the ohms
resistors in crossovers ten to be in the low ohms
because transducers tend to be like, 2 to 16 ohms

Posted by renesis at 01:48 | permalink | 0 comments


heh @ raw doggin the LED
me and other hacker knew an EE like that
we knew someone else that put LEDs directly in parallel behind a logo plate
wondered why all the speaker logos lit up diff
he blamed the china factory
I TOLD THEM TO PUT THEM IN PARALLEL THEY NEVER DO ANYTHING RIGHT
we are like, dotdotdot
there was enough but some LEDs would shunt out others
kind of random of course
they had one resistor for like, three directly parallel LEDs
actually
you could see they had tried shit to get it working
like they used two resistors, one high side one low side, then very high watt resistors
those ideas came from the american engineering tech
this guy was later hired as a manufacturing engineer at another company
:(
like, their concept was sound at a top level
they wanted them in parallel so if one 'burnt out', because this is obviously a huge fucking problem with leds, wtf, the others would stay lit

Posted by renesis at 01:43 | permalink | 0 comments


heh

Posted by renesis at 01:19 | permalink | 0 comments


to connect the sample dots
jan-: if lookup and interpolate doesnt work, you can maybe curve fit in pieces
oh no my solidworks student license gonna expire

Posted by renesis at 01:10 | permalink | 0 comments


thats all of calculus
like, figure out how to convert equations for curves can be drama
but like, doing that in a fixed sample rate realm is pretty straight forward
jan-: curve fits
thats what most scientists do
'oh that data curve looks like this math curve'

Posted by renesis at 01:05 | permalink | 0 comments


yeah i get that
like, math courses fuck me up, i dont really like it until its applied to something
like, i need word problems or my mind kind of rejects it by default
like, im find with calculus, diff equations kind of fuck me up, but spring/mass/damper problems helped
calc class can be fucked, but integrals and dirivatives are like, anything time related
laplace is magic
its hard not to for some stuff
running averages and sample to sample change is calculus
integrals are area under a curve, and a derivative is the angle of a point on a curve

Posted by renesis at 01:00 | permalink | 0 comments


which was clue i needed
like a third this country thinks learning shit turns people into heretics

Posted by renesis at 00:54 | permalink | 0 comments


https://www.youtube.com/watch?v=-y1Ly4hdVrU
i wish i had taken DSP classes
no time in mechatronic curriculum
yeah still
jero32: that track, the better the speakers, the worse it sounds
theres something clipped in the reverb
and the low end can totally fuck the mids
yeah i need to read up
oh
so when i was babysitting speaker tests in abandoned buildings...
synth geek had that book, we would read it to deal with boredom, great book
when CTO saw it when he came to help babysit, he basically said that book was how he started this DSP guitar amp shit (company was about 15 years ahead of the industry)
we think the asshole jazz musician kid on contract, friends of upper management, stole the book
great fucking book
CTO is like, 'yeah once i figured out F was normalized to sample rates, everything was easy easy'

Posted by renesis at 00:49 | permalink | 0 comments


however, if you have a very low distortion, very flat source, people will take few minutes to 'break in' and accept it, and then they are typically very impressed
people like being able to hear detail in familiar music
also speakers dont break in, people do, this is pretty much established on transducer measurement gear, shit breaks in in seconds
the opposite is somewhat true, good speakers will make all recordings sound bad
you will hear noises you didnt notice on consumer junk
like there is an amp buzz in SRVs little wing, its not in your face unless youre listening on expensive monitor grade speakers
yeah but its not secret or anything
building awesome speakers that arent very very small or very very cheap is not hard now
because transducers are lower distortion and DSP is awesome

Posted by renesis at 00:44 | permalink | 0 comments


rab: i should start saying the taig steps at under 7 microns, sounds more impressive that way
between tweet and woof?
um
yeah hes never done this before
burny: yeah but typical crossover setup, its going to be half or one phase shift and theyre going to be way far apart in the freq domain
this is kinda settled through listening tests
people dont notice
they notice voicing if its a few dB up or down
and *no one* likes distortion above 1%
speakers are voiced, general public doesnt like flat freq response
yeah, typically people want 'smiley face' or 'rock' eq

Posted by renesis at 00:39 | permalink | 0 comments


jero32: they have sigmastudio D amps too, 100w and 2x50w
i need a bass guitar amp, pretty sure im making one now
even tho i really wanted a 15" ampeg
like, i have to figure out how to control parameters over the spi/i2c port so im not limited on controls, but 4 pots and broken out gpio is pretty good to start with
looks like sigmastudio has more algorithms since i played with it
i thought $187 for adau dev board was cheap
i guess it is when its big companys money

Posted by renesis at 00:34 | 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