<DocScrutinizer05>
I think that sign means "fire in case of urinating"
Coyo has joined #qi-hardware
Coyo is now known as Guest6321
<DocScrutinizer05>
which my relatives explained to me applies even for hay shelter
<wpwrak_>
your relatives had urine-activated explosives ? this explains a few things :)
<DocScrutinizer05>
yes, they been farmers
<whitequark>
oh interesting
<DocScrutinizer05>
and I bet in the 70s they even had ammonia nitrate ferilizer in that barn
<whitequark>
I bet farmers today have ammonia nitrate fertilizer
<DocScrutinizer05>
fertilizer
<whitequark>
it's not like you can easily replace that
<DocScrutinizer05>
no, today this is lame stuff passivated with at least 30% innert materials
<whitequark>
ah
<whitequark>
well, you can still reprecipitate it
<DocScrutinizer05>
apropos ammonia nitrate: most amazing explosions been those where ferilizer fabs went BOOM, I think the detonation been audible a 120km away
<DocScrutinizer05>
and the crater been 80m wide and 30 deep, or sth
<whitequark>
oh, yes, I'm aware. it's responsible for a bunch of the biggest non-nuclear explosions
<DocScrutinizer05>
yup
<whitequark>
I've had that stuff taught by a bunch of high school and uni teachers
<whitequark>
taught to me*
<DocScrutinizer05>
25km away from the fab roofs lost their tiles
<DocScrutinizer05>
561 fatalities
<DocScrutinizer05>
125m*90m crater, 19m deep
<DocScrutinizer05>
the idiots mixed ammonianitrate with ammoniasulfate to create a non-explosive fertilizer, but it got solid as concrete in the silos. So the used... **explosives** to portion it, After 200,000 explosions they once day didn't get the mixture of sulfate to nitrate right
<DocScrutinizer05>
4500 tons of fertilizer went partytime
<wpwrak_>
not bad. playing russian roulette and winning 200'000 times may get you a false sense of immortality.
<wpwrak_>
by the way, i wonder if there are battery-less photovoltaic designs with AC output. there ought to be better ways to distribute the energy over 2 (?) cycles than with a frail battery.
* DocScrutinizer05
googles frail
<wpwrak_>
"easily broken" :)
<DocScrutinizer05>
and which 2 cycles?
<DocScrutinizer05>
sure you usually can feed changing amount of energy/power into an AC grid. No problem
<DocScrutinizer05>
whitequark: similation of a really minor scale tiny atomic bomb
<DocScrutinizer05>
simu*
<whitequark>
I know
<DocScrutinizer05>
4kT is about as small as atomic devices get
<DocScrutinizer05>
I think the backpacker bomb is adjustable from 5..15kT
<DocScrutinizer05>
dunno about those 12" atomic projectiles
pcercuei has quit [Quit: dodo]
Guest6321 has quit [Ping timeout: 240 seconds]
<DocScrutinizer05>
o.O
<DocScrutinizer05>
The amount of energy released by fission bombs can range from the equivalent of just under a ton of TNT, to upwards of 500,000 tons (500 kilotons) of TNT.[8]
<DocScrutinizer05>
/quote
Coyo has joined #qi-hardware
Coyo is now known as Guest61216
<JewFro297>
Whatabout antimater bombs?
ffio has quit [Quit: WeeChat 0.4.1]
<DocScrutinizer05>
sounds pretty silly
<DocScrutinizer05>
only military can invest money into this
xiangfu has joined #qi-hardware
ffio has joined #qi-hardware
xiangfu has quit [Ping timeout: 240 seconds]
xiangfu has joined #qi-hardware
<nicksydney>
Debugging network stack is like walking on a road of broken glass
<nicksydney>
especially if the device you are working on doesn't have serial port :(
<nicksydney>
found a good site that will calculate and draw up schematics for LED array http://led.linear1.org/led.wiz ... good for newbie like myself :)
kristianpaul has quit [Ping timeout: 252 seconds]
kristianpaul has joined #qi-hardware
kristianpaul has quit [Ping timeout: 240 seconds]
kristianpaul has joined #qi-hardware
kristianpaul has joined #qi-hardware
JewFro297 has quit [Quit: Leaving]
JewFro297 has joined #qi-hardware
JewFro297 has quit [Ping timeout: 240 seconds]
xiangfu has quit [Ping timeout: 246 seconds]
fire has joined #qi-hardware
fire is now known as ffio_
ffio has quit [Quit: WeeChat 0.4.2]
dos1 has joined #qi-hardware
<DocScrutinizer05>
nicksydney: 12, 3.8, 200, 9 -- then do same with *11V*, 3.8V... !
<DocScrutinizer05>
led.wiz is fine, but not easy to use
<DocScrutinizer05>
nicksydney: for the meanders - they are for delaying signals. E.g you want t data-ready signal to arrive later at a chip than the data in the data lines
<DocScrutinizer05>
and as whitequark said, you want all data bits to arrive same moment, thus the traces have to have same propagation delay, even when they are maybe different length. So you add length and thus delay to the fast short ones
<DocScrutinizer05>
in designs beyond maybe 100MHz this starts to become relevant
xiangfu has joined #qi-hardware
dos1 has quit [Remote host closed the connection]
dos1 has joined #qi-hardware
<DocScrutinizer05>
btw EE are awesome, they design that stuff without any tests. Usually you do all that propagation delay matching based only on calculations and experience, no test equipment and measurement and metering tools like scopes included
pcercuei has joined #qi-hardware
pcercuei has quit [Ping timeout: 240 seconds]
pcercuei has joined #qi-hardware
jekhor has joined #qi-hardware
xiangfu has quit [Ping timeout: 246 seconds]
jekhor_ has joined #qi-hardware
xiangfu has joined #qi-hardware
jekhor has quit [Ping timeout: 240 seconds]
<wpwrak_>
chances are that measuring would be a PITA anyway ;-)
jekhor_ has quit [Remote host closed the connection]
viric has quit [Remote host closed the connection]
viric has joined #qi-hardware
jekhor has joined #qi-hardware
panda|x201 has joined #qi-hardware
<nicksydney>
DocScrutinizer05: "12, 3.8, 200, 9 -- then do same with *11V*, 3.8V... !" ... not sure what those numbers are
<nicksydney>
wpwrak_: had a first crack at using autorouter on kicad :)
<nicksydney>
wpwrak_: when i tell the auto router not to use the bottom layer it keeps on trying to autoroute on the front layer and it could not stop goes in a loop ... :)
<nicksydney>
30C3: Hardware Attacks, Advanced ARM Exploitation, and Android Hacking (EN) --> http://www.youtube.com/watch?v=2fYPtByBY-s ... hardware hacking for sofware people..... this is easy for people like myself to follow :)
<wpwrak_>
nicksydney: the kicad autorouter is worse than useless :)
<DocScrutinizer05>
and for a hypothetical car rear light
<DocScrutinizer05>
wpwrak_: HAHA
<nicksydney>
wpwrak_: it's a good start for me to learn how to make it more neat .. i'm messy when it comes to laying out tracks
<nicksydney>
when people mention Windows 7 i always almost puke .. don't know why
<wpwrak_>
nicksydney: well, kicad also has an auto-placer. that one is even worse than the autorouter.
<DocScrutinizer05>
nicksydney: get eagle free trial version
<nicksydney>
DocScrutinizer05: i have eagle..but i want to go the full open source way like wpwrak_ :)
<DocScrutinizer05>
autoplacer ROTFL
<wpwrak_>
DocScrutinizer05: just because you're trapped in proprietary purgatory (not quite hell, but close), doesn't mean others have to choose that path :)
<DocScrutinizer05>
nicksydney: please don't try to learn stuff from kicad autorouter!
<DocScrutinizer05>
kicad hardly can teach you *anything* about correct EE business
<wpwrak_>
i find routing has the following phases: 1) confusion: all the parts are somehow where you don't want them, the signals are on the wrong side of things, and so on
<DocScrutinizer05>
it's hard enough to make it do the right thing when you already know what#s decent EE work
<wpwrak_>
2) a hope for order: when you move things to better places, reassing signals, etc.
<nicksydney>
wpwrak_: for the 0805 LED I have reuse the qi library 0805 and make the pad just a tad bigger so it's easier for me to solder :)
<nicksydney>
i still have a LOOONG way to learn on doing routing correctly
<wpwrak_>
3) spaghetti hell and gradual claustrophobia: while a few things route nicely, lots of others get in the way. also, when you're about to add that last trace at a difficult spot, you find there's not enough room.
<nicksydney>
when become better will move to a better tool
<nicksydney>
wpwrak_: (point 3) AGREE !
<DocScrutinizer05>
nicksydney: one fndamental rule I can offer for you: contemporary PCB layout style is not always the really smartest thing you can find. For example there's usually NO reason to keep all traces in a sub-millimeter width
<wpwrak_>
4) the zone: all of a sudden it all falls into place. you see with perfect clarity how that board is meant to be routed. do not sleep, do not eat, do not shower. get it done before the moment is lost.
<nicksydney>
DocScrutinizer05: so than why a lot of design are with sub-mm traces ?
<nicksydney>
wpwrak_: (point 4) ... if i don't shower ..mrs going to kick me out of the house :)
<DocScrutinizer05>
because nowadays designers are used to do that
<wpwrak_>
5) perfectionism: you find small imperfections and mistakes, often only after printing, that you then correct. by doing that, you learn your design even better and will be able to return to the zone later on with only a small effort.
<nicksydney>
oh another trick i learn now is to print out on paper and try the component on it :)
<nicksydney>
i was thinking of gluing the small component to see how it looks :)
<DocScrutinizer05>
and particularly, tools like eagle and kicad have that default
<nicksydney>
DocScrutinizer05: yeah.....even printing on paper to look at those traces i have to take off my glasses....damn small traces :)
<DocScrutinizer05>
it's usually silly to make traces smaller than "needed"
<DocScrutinizer05>
there's nothing wrong with a trace 5mm wide
<DocScrutinizer05>
unless you don't have enough space to do that
<wpwrak_>
printing is good. by the way, fped's postscript output includes a 1:1 version for just that purpose: that you can place components on the footprint and see whether they fit. useful especially for complicated electromechanical parts
<nicksydney>
combination of auto-routing and manual
<DocScrutinizer05>
your design could use traces 5 times as wide as they are
<DocScrutinizer05>
easier to print, to etch, to solder
<nicksydney>
DocScrutinizer05: yeah...i want to make it smaller to practise soldering with small traces so i can do it for anelok later on
<wpwrak_>
DocScrutinizer05: (wide trace) toner transfer sometimes acts up if you have large copper areas. but yes, chances are people simply use the parameters they've used the last time, without thinking much
<DocScrutinizer05>
the latter not always applies
<nicksydney>
one thing i was really surprised that i was able to do with kicad was selecting a pad as a module and assigning it's NET as 5v or GND I can use it as a power pad and connect the trace too :)
<nicksydney>
small thing like that make me smile..hahahaha
<nicksydney>
always excited to find something new to learn
<DocScrutinizer05>
but particularly for hand soldering you worry less about thermal isolation pads and heatsink to huge planes of copper, you rather worry about sub-mm wide traces coming off when you solder on them
<wpwrak_>
careful with traces under LEDs: did you check you can actually do that ? some have asymmetric pads that block the middle (e.g., my beloved LTST C190-Kxxx)
<nicksydney>
DocScrutinizer05: yup...soldering with those traces is a skill would like to hone if possible :)
<nicksydney>
wpwrak_: not sure what you mean by block the middle ?
<DocScrutinizer05>
btw all your pads look terribly optimized for high density layout and reflow soldering
<DocScrutinizer05>
almost impossible to solder by hand
<nicksydney>
DocScrutinizer05: good point ... in screen it looks decently big but didn't think about that ... let me print it out and see how it small it is
<DocScrutinizer05>
generally you *always* want pads to be way "longer" than the SMD component
<wpwrak_>
DocScrutinizer05: they're optimized for reflow but still with manual soldering in mind. so there's that extra 0.1 mm here and there that makes it doable
<DocScrutinizer05>
nicksydney: the rationale being: you should easily touch the *pad* with your solder tip
<DocScrutinizer05>
0.1mm is stingy
<wpwrak_>
hmm, i'd touch both parts, as close ot the joint as possible
<DocScrutinizer05>
make that 1.0mm and we talk
<wpwrak_>
yes, i didn't leave a lot of room. the spacing is meant to be efficient.
<wpwrak_>
(layout) seems that you could have a ground ring in the middle. and if you turn the Rs by 90 deg, the 5V connection would just be a straight line
<DocScrutinizer05>
wpwrak_: seems one best practice is to coat the pad with solder, then place the component on top and heat the pad (and only the pad) until solder melts and component gets fixed
<wpwrak_>
agreed on the coating. i tend to try both items. well, unless things are so small that already the energy of the solder i'm adding will be enough to heat them up properly. 0402 basically doesn't need "full contact"
<wpwrak_>
that's what an autorouter who's very concerned about job security does :) "see, it's sooooo complex. only i could do that."
<DocScrutinizer05>
LOL yeah
<nicksydney>
it is damn small .. DocScrutinizer05:
<nicksydney>
DocScrutinizer05: yeah it is dizzying :)
<DocScrutinizer05>
why isn't that one ring GND, with a dozen or some radials consisting of R and D, and an outer ring of V+ around that
<DocScrutinizer05>
wait a minute
<DocScrutinizer05>
nah, too lazy to do that in eagle now
<DocScrutinizer05>
had to create a new project for it
jekhor has quit [Read error: Operation timed out]
<DocScrutinizer05>
and why the heck are D6 and D9 *parallel*?
<wpwrak_>
do it in kicad :)
<nicksydney>
DocScrutinizer05: those were drawn by autorouter
<DocScrutinizer05>
rather in dig
<DocScrutinizer05>
nicksydney: that is a bug in circuit, definitely. You never ever parallel Diodes
<DocScrutinizer05>
(unless it#s a varicap)
<DocScrutinizer05>
diodes have physical properties that forbid connecting them in parallel
<nicksydney>
DocScrutinizer05: will check the schematics
<nicksydney>
the printer is acting up again .. it's print 1/2 the actual size :( damn it
<nicksydney>
+
<DocScrutinizer05>
printing 600dpi data to a 1200dpi laserprinter?
<DocScrutinizer05>
IOW your printer driver thinks the printer does draft, but the printer wants to do a decent job and is using fine resolution 1200dpi rendering?
<DocScrutinizer05>
that's what you get when each app has its own printerdriver, and the printer doesn't reset parameters between print jobs
<DocScrutinizer05>
or by sharing one printer between two computers or OSes
<DocScrutinizer05>
or - more generally - by simply not using the correct driver for the printer
<DocScrutinizer05>
when the driver is for a model that has no 1200dpi mode, the driver won't bother to set mode according to what the data it sends to printer
<DocScrutinizer05>
so you can switch printer to 1200 mode manually or with a nifty app that knows how to do that, but the standard driver has no clue about 1200 mode even being supported
<DocScrutinizer05>
much fun for weeks to come
<DocScrutinizer05>
;-P
* DocScrutinizer05
recently curses gray level dithering rendering nearly every of the modern gray in gray websites illegible on paper printout - THANK YOU webdesigner-turdheads!
<nicksydney>
as per what the wizard has outlined..but seems like something is off in my schematics
<nicksydney>
duh ! i made the series connection to parallel connection for the LED ... hahahaha...
<DocScrutinizer05>
and you defintely should go for solution1, version0 is idiotic
<DocScrutinizer05>
to see what I mean, use 4 LEDs
panda|x201 has quit [Ping timeout: 246 seconds]
<DocScrutinizer05>
then make it calculate the schematics and dimensions for 8.5V Vbatt
<nicksydney>
yeah solution1
<nicksydney>
what about 9v battery ?
<DocScrutinizer05>
you don't want to run 4 LEDs in series, with a series R that only has <=1V to do its regulation job
<DocScrutinizer05>
since the 2.0V for LED can vary
<DocScrutinizer05>
and when all 4 diodes cool down and go to maybe 2.2V out in the winter cold, you have 0.2V left for resistor
<DocScrutinizer05>
also same time your battery drops to 8V due to the cold outside
<wpwrak_>
it'll still kinda work :)
<nicksydney>
so use semi-realistic value i guess
<DocScrutinizer05>
for the 3 LED series these changes are not critical
<DocScrutinizer05>
as a rule of thumb: at very least 20% of the Vbat should go for series resistor, max 80% for feeding the diodes. When you need to keep diodes at a constant brightness, make that 50%/50%
<DocScrutinizer05>
obviously the solution0 4LED-series design has only 1V of 9V across the resistor, while the rudimentary 2-diode-series add-on has 5V of 9V across the resistor
<nicksydney>
DocScrutinizer05: will keep in mind that
<DocScrutinizer05>
so when your ambient temp changes, or your battery depletes, the 2-LED branch will stay same brightness while the 4-LED branches change brightness *massively*
<DocScrutinizer05>
toldya this calculator is simple to use but hard to use correctly
<DocScrutinizer05>
first approach use worst case voltage for Vbat, in that calculator
<DocScrutinizer05>
avoid assymetric designs like solution0
<DocScrutinizer05>
thus I suggested you use 8V for Vbat and see what this calculator suggests then
<DocScrutinizer05>
when you want to operate LEDs at maximum (IOW when 20mA is absolute maximum rating for your LED type) then you should use maybe 17.5mA and 8V instead of 9V, in that calculator
<DocScrutinizer05>
the reduction in forward current from 20 to 17.5 compensates for the battery voltage possibly going to 9.5V instead of the 8V you use in the calculator
<DocScrutinizer05>
also LEDs or diodes in general may have a negative U/temp characteristics, meaning when the diode heats up, the forward voltage drops from 2.0V to maybe 1.9V or even lower
<DocScrutinizer05>
that's the reason why you can't parallel diodes
<DocScrutinizer05>
the hotter diode will draw more current due to lower V(fwd), and thus heat up even more, compared to the cooler diode
<nicksydney>
cool...make sense now DocScrutinizer05
<nicksydney>
DocScrutinizer05: i believe human explanation than a calculator :)
<nicksydney>
wpwrak_: is it possible to change the track width for all the traces ?
<nicksydney>
in kicad i meant
<DocScrutinizer05>
nicksydney: well, better than before, as far as I can see
<DocScrutinizer05>
though sth fishy arounf D13, R3, D3
<DocScrutinizer05>
and I honestly would place/number the LEDs strictly monotonic in clockwise counting from D1 .. D16
<DocScrutinizer05>
and make D1, D2, D3 one series couple, then D4, D5, D6 the next. Connect D1, D2, D3 in series to R1. D4,5,6 in series to R2, and so on
<nicksydney>
ok will try how the trace will look like if i put that way...initially was going to do that but i thought the auto trace is going to have a hard time laying the track
<DocScrutinizer05>
nah
<DocScrutinizer05>
it looks like a bike wheel, with spokes, each spoke is a series of 1 R and 3 D
<DocScrutinizer05>
the hub is GND, goes to 6 R, then the Rs connect to 3 D each. The Ds are 3 next to each other, doing zigzag series connects. Then from most "outer" LED on each spoke it connects to the rim/ring which is V+
<DocScrutinizer05>
obviously the ring is outside of all Ds and Rs
<wpwrak_>
rjeffries: hehe, that sounds a like a good motto :) happy year of anelok, too ! ;-)
<rjeffries>
wpwrak how do things stand re power consumption of anelok?
ffio_ has quit [Quit: WeeChat 0.4.1]
arielenter has joined #qi-hardware
<wpwrak_>
i'm measuring a few things. alas, it seems that my scope no longer wants to take screenshots :-(
<wpwrak_>
ah, i should try with a memory stick. maybe that still works ...
<wpwrak_>
of course, that would eb alot easier of memory sticks weren't made of elusium ... or is that ephimerium ?
<wpwrak_>
grnbl. not one of the bastards in sight anywhere.
arielenter has quit [Ping timeout: 272 seconds]
arielenter has joined #qi-hardware
jekhor has quit [Ping timeout: 245 seconds]
heberth has joined #qi-hardware
rz2k has joined #qi-hardware
jekhor has joined #qi-hardware
jekhor has quit [Ping timeout: 265 seconds]
jekhor has joined #qi-hardware
arielenter has quit [Ping timeout: 252 seconds]
jekhor has quit [Ping timeout: 260 seconds]
jekhor has joined #qi-hardware
heberth has quit [Ping timeout: 272 seconds]
wolfspraul has quit [Ping timeout: 246 seconds]
heberth has joined #qi-hardware
wolfspraul has joined #qi-hardware
<wpwrak_>
hmm. can't save screenshots to the USB stick either. this is increasingly ungood ...
arielenter has joined #qi-hardware
<zrafa_>
wpwrak_: windows?
<wpwrak_>
rigol :(
<zrafa_>
wpwrak_: which of your machines is? :(
wej has quit [Ping timeout: 260 seconds]
<wpwrak_>
the scope
<wpwrak_>
refuses to send screenshots as usb device. and writes garbage as usb host. seems that the communication subsystem is broken :(
<zrafa_>
:(
wej has joined #qi-hardware
<wpwrak_>
seems that i've finally gotten a good excuse for spending a ridiculous amount of money on a new scope
<wpwrak_>
now i just need the money :)
pcercuei has quit [Read error: Connection reset by peer]
<larsc>
make a kickstarter campain
<wpwrak_>
the "get werner a new scope" campaign :)
<larsc>
no no, "I want to build a super cool scope" campain
<wpwrak_>
ah yes, i want to do that, too. but just not yet :)
<viric>
it's only for the money
<larsc>
exaclty, you don't have to build it now, just send out a couple of t-shirts
<wpwrak_>
hehe :)
<viric>
right. Maximum contribution, $10, and you get a t-shirt.
<viric>
but be prepared. Maybe it will be hundreds of t-shirts
valhalla has quit [Ping timeout: 252 seconds]
valhalla has joined #qi-hardware
pcercuei has joined #qi-hardware
<kristianpaul>
:notsureif:
<nicksydney>
morning all
<derRichard>
larsc: i see the -110 (timeout) xbboot issue also on my old thinkpad :-\
<larsc>
noy good
<derRichard>
is "sudo xbboot -u 0x80600000 /PATH/TO/zImage.bin" correct? i took the command from the wiki
<derRichard>
maybe the memory location is wrong
<derRichard>
or do i need special stage1/2 files?
<pcercuei>
shouldn't you use jzboot?
<derRichard>
i have no idea. all i want is booting a kernel on my nanonote without flashing it to nand
<derRichard>
larsc: updated xburst-tools-firmware. now it works :D
dandon has joined #qi-hardware
dandon_ has joined #qi-hardware
dandon has quit [Ping timeout: 272 seconds]
dandon_ is now known as dandon
<larsc>
good
<derRichard>
larsc: are you aware of the fact that sound/soc/jz4740/ does not even build in mainline? look like you broke with commit 25ce6c (MIPS: jz4740: Remove custom DMA API)
<derRichard>
stuff in sound/soc/jz4740/ still uses the custom dma crap