<Fallenou>
basically it's something to see if you have interested users who wanna buy your stuff
<Fallenou>
you put a "threshold" for manufacturing
<Fallenou>
and a price
<Fallenou>
when there is enough people in pre-order state, you can "safely" manufacture
<Fallenou>
knowing that you will have buyers
<Fallenou>
you just run the production batch when you reach the threshold
<Fallenou>
so that you can lower the price, (puttin a higher treshold maybe
<Fallenou>
taking less risks
<Fallenou>
that's the idea
<Fallenou>
lekernel: when I said you had an open source fpga toolchain as a work in progress
<Fallenou>
they were amazed :p
<Fallenou>
I hope some of the guys will come here and contribute
<kristianpaul>
Fallenou: (getimade) oh, i'm sorry to say but... those guys really know what _manufacturing_ really is about??
<kristianpaul>
I could not be adam or wolfgang, but that is clear taht manufacuring is not behind a click and preordering.. well i should agree with sebastien, and said that for "blinling" stuff it may work
<kristianpaul>
(anyway personal opinion)
<kristianpaul>
aw_: hi
<kristianpaul>
Had you already talked with vendor about crystal?
<aw_>
kristianpaul, not yet. any news or suggestions?
<kristianpaul>
i'll just about to run a self check here
<kristianpaul>
xiangfu: do you think is posible port that code a rtems command ?
<kristianpaul>
gn8
<kristianpaul>
lekernel: had you noticed  the ADV is ESD sensitive device?
<kristianpaul>
wolfspraul should avoid ESD warning in the mm1 box
<kristianpaul>
i comment ESD, because you'were doing ESD tests if remenber well
<roh>
esd doesnt mean immediate failiure in all cases
<kristianpaul>
comment about**
<kristianpaul>
roh: sure
<kristianpaul>
_just_ _saying_
<roh>
sometimes its just 'damaging' the chip and it dies soon after.. or a year after that
<kristianpaul>
oh, yes !
<kristianpaul>
hmm
<roh>
so i wouldnt give too much of a significance to the chip dying in _one_ single case on the board of a developer who soldered around on the board _a lot_ and used it without case most of the time i guess
<kristianpaul>
;-)
<roh>
when the second one is dead.. please remind me;)
<Schmoove>
hi
<Schmoove>
bye
<lekernel>
aw_: an overtone crystal is a crystal of a lower frequency that resonates on one of its harmonics
<lekernel>
typically 3rd harmonic, but there are sometimes 5th overtones too
<lekernel>
for example, we could use a 9.54MHz crystal in 3rd overtone for generating the 28.36MHz
<lekernel>
(with proper oscillator dynamics, which the adv7181 doesn't have)
<aw_>
hmm..interesting ..now i know your meanings on 'overtone'.
<aw_>
also i discovered it's a R1 50 ohm at range 10MHz <= fo <= 27MHz in foundamental mode. Since it's not sure if its the root cause, but it's good to discover/follow that we at least to use a fundamental mode xtal.
<lekernel>
it's probably a fundamental mode xtal... well, if you're worried, you could check that by sending 9.54MHz to it and check that it doesn't resonate
<aw_>
lekernel, good capture on this mode check. btw, I'll reconfirm with here vendor and request.
<lekernel>
the main point is to choose an appropriate crystal for rc3
<aw_>
yeah...at least to pick that mode first. later I got a new datasheet or doc. I'll let you know. or even I just order CSA309 one.
<lekernel>
btw, the board already had a CSA309 crystal when it failed
<lekernel>
it's the board you shipped me with a non-working crystal and I replaced it right away
<aw_>
uhm? wow..
<lekernel>
currently the symptom is best described as: video chip power supply OK, crystal oscillating and clock running, I2C signals from the FPGA correctly measured at the chip's pins, but no I2C response from chip
<lekernel>
it really sounds like the adv7181 got internally damaged somehow, but this did not break the crystal oscillator
<lekernel>
i'll probably replace it pretty soon
<aw_>
so the I2C now from decoder is no more response at all. it's really bad.
<aw_>
I think that I'll need to check things as: 1, review the four supplies (DVDD/AVDD/PVDD/DVDDIO & their grounds) routes is suitable to apply enough driven power/current to chip inside?
<aw_>
2, does part unqualified at all even it works few times? 3, solderability on chip..hmm...
<lekernel>
yeah, I checked the solders too, I saw no lifted pin
<lekernel>
and I measured signals on the pin, near the package
<aw_>
yeah..good, confirmed those connection by touching the pads on board not probe the package's pin directly is good.
<aw_>
not sure sure current the route's thickness is good enough for chip. since i met problems before like thickness is wider enough. then either I added a more bigger value of decoupling capacitor old board or used a same capacitor value on a new re-layout board with wider thickness routes.
<aw_>
this was rarely happened though.
<aw_>
but our case with a likely internally damaged somehow, it's real strange.
<lekernel>
insufficient power traces wouldn't break i2c imo
<lekernel>
just cause semi-random glitches
<aw_>
umm....don't know what possible accidence would cause from outside or internal. anyway i would review that all ground copper layer/area to keep a well-taken considerations.
<aw_>
during rc3 routing.
<wolfspraul>
lekernel: you got the failure when csa309 was on the board?
<lekernel>
yes
<wolfspraul>
not that I believe there is a connection, but it's good to keep in mind. aw_ - you read that?
<aw_>
yes, i just knew his board with that cas309 already.
<Fallenou>
kristianpaul: the getitmade stuff is not for doing the actual manufacturing, it's for being sure that you have clients
<Fallenou>
getitmade does not manufacture anything
<lekernel>
wolfspraul: vital signals (power supply, clock, i2c, reset) at the adv7181's pins are measured ok... guess the chip is simply dead
<lekernel>
my next step is to replace it
<lekernel>
roh: you have a hot air station, don't you?
<lekernel>
my usual technique for replacing tqfp's with little hardware is to cut the pins, remove the chip body, then lift the pins one by one
<lekernel>
quite aggressive :)
<aw_>
um..to replace it needs hot air station is better.
<lekernel>
interesting: my video camera died as well
<lekernel>
(the one that was connected to the board when it failed)
<aw_>
which color RCA con. on board you connected? green?
<lekernel>
yes, it's cvbs
<lekernel>
but maybe it's just that when dying the camera sent some overvoltage down the video cable that burnt the video chip
<lekernel>
it's powered on 12V so it has the potential for doing that
<aw_>
this could be possible. but you need to recall how/what you did recently on camera or else...to track down any upstreams.
<aw_>
so please think both your camera and m1 board. the only common ground goes through RCA connector. make sure no one else possible existed.
<lekernel>
the camera has its own transformer and only one RCA cable going out
<lekernel>
so no, it's not a ground problem
<lekernel>
maybe let's just add some diodes on the video input as additional protection against broken cams on RC3
<lekernel>
what's the parasitic capacitance of a typical dirt cheap zener?
<aw_>
have you ever tried to use multimeter to measure current. and at that time. has it ever happened that you touched the current ground probe pin?
<aw_>
yeah...we can take consider to add them.
<aw_>
second..
<lekernel>
the board was cased when the failure happened, with no weird experiment going on
<lekernel>
just normal use case
<aw_>
yeah...your camera is dead. but why?!  checking...zener..
<lekernel>
well it's an old relic from the early 90s
<lekernel>
i'm not surprised it died
<wolfspraul>
lekernel: aw_ do I understand correctly that we decided to add additional diodes on video-in for rc3?
<wolfspraul>
(just trying to understand the conclusion from above...)
<lekernel>
yeah let's do that
<lekernel>
i'll just check first that by replacing the video chip it still works
<aw_>
i think we'd better to use bidirection diode (TVS) to protect it...still recall my sch.. before..:-)
<aw_>
with just 3pF could be enough.
<roh>
lekernel: yes
<wolfspraul>
roh: another small feedback. the screws on top and bottom are torx (I think)
<wolfspraul>
ah no, something else, checking...
<wolfspraul>
hex socket?
<wolfspraul>
but the ones to fasten the connectors are Phillips (+)
<lekernel>
I want some of what lattice people are smoking
<Fallenou>
LOOL
<Fallenou>
wtffff
<lekernel>
"that's what I call green technology"
<Fallenou>
:''
<roh>
wtf
<CIA-37>
flickernoise: Sebastien Bourdeauducq master * rf52cbc3 / src/main.c : Mount memory card read only - http://bit.ly/hqxCI9
<kristianpaul>
wtf?
<kristianpaul>
Fallenou: any interesting question from OSHUG talk?
<kristianpaul>
nice pics btw :-)
<Fallenou>
hehe thanks
<Fallenou>
yes a loooooot of questions
<Fallenou>
I managed to answer to most of them
<Fallenou>
for one or two I had to say that I didn't know and redirected the guys to the ML / IRC
<kristianpaul>
What you use mac?? ;-)
<Fallenou>
they were very interested
<Fallenou>
yes :p I apologized for that at the beginning
<kristianpaul>
haha
<Fallenou>
but some of them are using mac too
<Fallenou>
so it was ok ;)
<kristianpaul>
You shoudl use mm1 next time for, i dont see better way of showing the product you're talking about works
<Fallenou>
I would if I had one :p
<kristianpaul>
ah, yes..
<Fallenou>
at least I showed pictures and videos from vimeo
<kristianpaul>
Man you need to by one too ! :-)
<kristianpaul>
just tww left
<kristianpaul>
twO
<Fallenou>
yeah for the moment I am paying my rent in London and I will see what's left of my bank account :)
<kristianpaul>
:|
<kristianpaul>
used credit card to pay mm, so well, i pay bill as anynone does
<lekernel>
my opinions about London and what costs what there could be compared to those on autocrap, so I won't comment on that
<Fallenou>
you're working on your communication skills ? :p
<kristianpaul>
;-)
<kristianpaul>
14:36Â Â * Fallenou is back
<kristianpaul>
14:37 < Fallenou> no no dma
<kristianpaul>
Fallenou: can you exaplaing what was that?
<Fallenou>
scrolling up
<kristianpaul>
lekernel: is the xbar doing something usefull right now?
<kristianpaul>
I saw the comit is from 10/2010
<Fallenou>
sorry I can't scroll up that far  kristianpaul :x
<kristianpaul>
also, can you make a quick explanation of the DMA between TMU and wishbone, i mean how do you avoid a botleneck if TMU is par of the FML
<kristianpaul>
?
<kristianpaul>
Is tha the xbar for? does it work well? how it works? (i really dunno)
<Fallenou>
leaves the office
<kristianpaul>
For save some chatting i'm okay if you point me a documentation to read, i lack good links/books/whatever about  this topic
<lekernel>
does preloading some wallpapers into the m1 qualify as commercial use?
<kristianpaul>
yes
<kristianpaul>
you sell it, thats is "content"
<lekernel>
dunno... it's a bit of a gray area for me
<lekernel>
I can ask for a specific permission anyway
<Fallenou>
just to be sure I guess it's better to avoid anything in nc-by
<lekernel>
I agree, but in the absence of a good !nc !nd alternative it's better than nothing
<lekernel>
for things like the wallpaper, which is just a preloaded file on the FS, we don't have to be super-picky about freedom of the content (as long as we have the permission from the author to use it in this way)... let's get things done