lordpil's blog


i do it. for the love of race things
also because if im going to be a fuckup best to do it early in the session
i guess sleep by 7, not like i will be able to stay up for post race drama

Posted by renesis at 19:22 | permalink | 0 comments


ya
thats how curve fitting crossover calculators do it
and how can you forget that its named after host of greatest gp ever
other greatest gp in 2 hours
i can go to sleep now
get decent rest, i work late tomorrow
or i can stay up
watch the gp
in typical west coast f1 fan brain fog
get less sleep in middle of the day, totally fuck my schedule through mid week
i has 8am classes all week =\

Posted by renesis at 19:17 | permalink | 0 comments


can you babysit it?
if youre developing for performance, typically youre doing iterative mods and tons of repeated measurements
youre going to measure THD and SNR with spice? it accepts layouts and mechanical dimensions w/ materials?
i mean, if youre just doing freq response for voicing, sure spice great for that
i would do a lot of my signal path rework design in spice
other mfkrs sitting there with a pot hanging 3ft out the chassis, heh

Posted by renesis at 18:25 | permalink | 0 comments


ohsixi: you can get a 2i2 for under $100 and they have low enough THD and flat enough response for measurement

Posted by renesis at 18:17 | permalink | 0 comments


its kind of tradition to throw random low pF cap across feedback resistor, so oscillating amps in audio isnt that rare, doesnt always hurt things

Posted by renesis at 17:31 | permalink | 0 comments


dunno its completely normal for them to misbehave, theyre high gain
60dB guitar amps same animals
lots of stuff will oscillate while doing okay in audio range
the oscillation can cause clipping, or the high speeds can cause assympetric behavior, so like outputs will pump to a rail
like, cell phone signals will get into amp circuits, and they cant reject the high speed common mode so they will clip one rail
and losing audio during the clipping is what youre hearing

Posted by renesis at 17:26 | permalink | 0 comments


all that shit runs off of +/- 15v
almost all those amps are max rated +/- 18v
ohsixi: headroom
and SNR
kostix: ive seen samples of those at work labs, but never in any of our products
almost everything super discrete, crazy cheap, mostly stable but will misbehave if tickled just right
which what
the higher voltages?

Posted by renesis at 17:21 | permalink | 0 comments


shrug, you likely werent measuring for barely audible issues
or youre not the corner case
like, making a powered speaker is easy
making a powered speaker that works from 10F to 150F, not as much
amps like that are why higher voltage

Posted by renesis at 17:16 | permalink | 0 comments


ohsixi: passive attenuator would reduce the brought noise so you can boost earlier in the stage and then not after
tired
kbblsleep

Posted by renesis at 17:11 | permalink | 0 comments


power supply and processor bleeds everywhere
maybe few mV
500uV to 50mV
source: i worked for a condenser mic company
mic preamps are typically 40dB things, some will go out to 60dB, tho those arent the most stable things in the world
and this isnt for mic pres
10mV@94dB is kind of typical mic sensitivity
anyway im not making a hot rod mic pre, i would prob just buy an IC from That or some shit anyway

Posted by renesis at 17:05 | permalink | 0 comments


i havent done anything yet, this is just what i know from working on the shit for like 8 years
rock and pop music have like 10-20dB crest factors, movies are nuts its like 40dB
theres no standard
like, there are but they dont matter
0dBu is 745mV or 775mV
4dBu is supposedly pro audio levels, but no one seems to be sure if thats a nominal or max thing
most shit will go way higher
when does it not push down noise?
especially with modern shitty digital sources

Posted by renesis at 17:00 | permalink | 0 comments


after a buffer introduces the unattenuated DAC clipping issue were trying to avoid in the first place
and 9V with typical amps is prob only going to leave you with like 4Vpp to use, so like maybe 1.5V rms, so anything pro and lots of consumer devices can clip it, which means lower source levels, which means higher noise
so more money for lower performance RRIO shit
so plan is to use battery voltage of like ~14-17 depending on charge

Posted by renesis at 16:21 | permalink | 0 comments


why would you need a pad to control the impedance of a 10k passive attenuator?
most of them dont even have a post pot passive LPF
i know what an l pad is, i dont understand how it does anything to solve this problem
its an amplifier attenuator so the wiper is going into a hiz input
because non inverting
when they do use a post pot LPF, its usually with a 100k resistor or something nuts for audio like 1M so it doesnt matter anyway
the B100K into A taper trick turns the pot into a A15K-A100K pot, its actually a better log curve, but is kinda dirty used unbuffered because z is all over the place

Posted by renesis at 16:16 | permalink | 0 comments


the alternative is playing pong with SNR and headroom performance
you could divide to get any signal you want but there are still sensitivity vs headroom vs SNR vs resulution issues
a pot and ignoring product designer about most of their input device vs pot position asks can fix this
basically when he says he wants his professional DJ mixer and his iphone to both work with the pot at 12:00, you tell him OKAY! but you think FUCK OFF!
oh?
er
ohsixi?

Posted by renesis at 16:11 | permalink | 0 comments


ohsixi: explain divider pad
10k because its a nominal input impedance, low enough for noise immunity and high enough most anything can drive it and typical output impedance wont affect it so much
if you put a resistor in parallel with the low side of the pot to log shape the taper, impedance goes from like 6k to 50k depending on position
also tons of pro gear has passive attenuator input, its the only way to make sure you dont clip high input signals

Posted by renesis at 16:06 | permalink | 0 comments


im sure, theres still factories making shit there
i imagine most of their US orders are prototype quantity, with big mass production orders coming from asia and mexico
i know everything we ordered from digikey for chinese factories came from a chinese digikey warehouse
actually i dont know that, they might have some sort of direct ship from component manufacturer inventory shit happening

Posted by renesis at 15:37 | permalink | 0 comments


oh i wonder if its push switch, it has stepped shaft
alps sells both
yeah that one is push pull, i think digikey has that pot
thats current fallback option
well, single pole version
besides personal stuff, ive ordered a ton from mouser and digikey and almost never had problems
yeah
greymoon getting stuff from digikey to .ca sounds like a random ship time plus expensive
UPS ground usually ends up being like $8 so if you get enough stuff its not so bad
yeah my ebay experiences are random, some great, some just dishonest
neat

Posted by renesis at 15:32 | permalink | 0 comments


gets towards 1k and becomes an opamp research project
wow nice
and weird, why do they have extensions
they wont fit because of those, but yeah ALPs had the part number i wanted, i wasnt seeing it in stock tho
have to look harder later, or call them and beg
no the sleeve on the pins
and then the wide standoff pins too
ive seen those type of pots made to sit flush on the pcb, this looks like some sort of custom job
i think the pins would still be wider, unless those are some sort of extension adapter
heh
woah dpst i think?
thats actually way perfect, prob saves me a bjt
oh its B10K
still neat tho, means it exists as more than a part number
i didnt see any of those DPST dual gangs in stock

Posted by renesis at 15:27 | permalink | 0 comments


im maybe just going to get the linear pot, do the linearizing trick after an input buffer
or just do switch differently
wait until i can afford like 1000 so pot company makes whatever i want
i <3 those submini switches
all those pots you got been okay?
a10k would be ideal, but if i do linearizing after buffer, ya something like 50k probably about right
worst case equiv resistance would be like 7k

Posted by renesis at 15:22 | permalink | 0 comments


maybe i can troll for samples
on/off soft switch for small battery powered thing
yeah but it changes impedance with position
i think the trick is 100K and 15K to get pretty good log matching, elliot sound products has an article
because i want to use it as an input attenuator, so i want a nominal input impedance
also its higher noise, if not, its getting towards over driving the source stage

Posted by renesis at 15:16 | permalink | 0 comments


kostix: sexy
kostix: i am sad because i dont think i can find audio taper 10k dual pot w/ rotary detent switch
that also has to be small
and i wanted to use as unbuffered input attenuator, so cant use resistor in parallel with wiper trick
not so hard if you want new and dont care so much about bushing, knob, gang and switch options
on off detent at full CW
er, full CCW
i think i might be able to find pull switch
but thats kinda goofy on a lighter product

Posted by renesis at 15:11 | 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