Current Server Time: 09:33 (Central Europe)

#apertus IRC Channel Logs

2013/07/01

Timezone: UTC


22:22
jucar1
joined the channel
22:23
jucar
left the channel
22:28
s3bastian
left the channel
07:16
dmj_nova
joined the channel
08:16
se6astian
joined the channel
08:33
se6astian
left the channel
08:33
se6astian
joined the channel
08:40
se6astian
left the channel
08:40
se6astian
joined the channel
09:04
dmj_nova
left the channel
09:12
dmj_nova
joined the channel
10:18
se6astian
left the channel
10:18
se6astian
joined the channel
10:19
dmj_nova
se6astian: thanks for sharing that axiom research doc
10:20
dmj_nova
Looking at the Blackmagic, I suspect that a thorough consideration for ergonomics would set us apart from them.
10:33
se6astian
left the channel
10:33
se6astian
joined the channel
11:08
se6astian
dmj_nova, the doc was not written by me, yes ergonomics should get some priority
11:09
dmj_nova
se6astian: I know it wasn't you, but looking at it did remind me
11:09
dmj_nova
that ergonomics is a key part of why DSLRs do well
11:10
dmj_nova
and it's a weird thing that blackmagic does, not having the device fit well hands/be usable without looking at the screen
11:11
dmj_nova
same sort of form over function thing that Apple likes to do
11:11
se6astian
true, but the specs just sell their product, even if they make it in the most akward ways ;)
11:12
se6astian
can you take a look at https://www.apertus.org/sharing-creation-attitude please
11:12
dmj_nova
sure
11:12
se6astian
a friend of mine said that the text is very hard to follow because its interrupted by the images all the time
11:13
se6astian
he does like the 3 coloumn text layout, but thinks the interruptions of videos/images are a problem
11:13
se6astian
do you think so as well? because I think the images and videos and those interruptions are what makes it so nice ;)
11:15
dmj_nova
it does look quite pretty
11:15
dmj_nova
At the same time I can see what he's talking about
11:18
dmj_nova
too many breaking points in the text
11:18
dmj_nova
and many things pushed together
11:19
se6astian
do you think a solution would be to group the images and text blocks together more
11:19
se6astian
instead of text - image - text - video we do: text - text - image - video
11:19
dmj_nova
what size screen are you viewing on?
11:20
se6astian
22"
11:20
dmj_nova
resolution?
11:21
se6astian
1680x1050
11:21
dmj_nova
More vertical resolution definitely helps with this one
11:22
dmj_nova
it makes the flow more coherent
11:22
dmj_nova
on my 1920x1200 screen the document feels like it flows better than on my 1366x768 one
11:28
se6astian
left the channel
11:28
se6astian
joined the channel
11:29
se6astian
!ping
11:30
se6astian
damn internet connection
11:30
se6astian
I hope its just my ISP
11:30
se6astian
doe apertus.org work for you?
11:42
se6astian
seems like the godaddy nameservers are down!
11:44
se6astian
actually I can ping them but they dont seem to do what they are supposed to
11:46
dmj_nova
!pong
11:46
dmj_nova
se6astian: works for me
11:46
se6astian
hmm, good
11:47
se6astian
I guess I will just wait a bit and hope the problem solves itself
11:55
dmj_nova
ugh, I should sleep
11:58
se6astian
ok, please do ;)
13:06
se6astian
left the channel
13:07
se6astian
joined the channel
13:30
se6astian
Bertl, are you around currently?
14:19
Bertl
good morning everyone!
14:20
Bertl
se6astian: I am now :)
15:09
se6astian
left the channel
15:10
se6astian
joined the channel
16:10
Bertl
I've uploaded two more examples, one implementing a primitive register file via gpio and another connecting the hdmi i2c for use from linux
16:10
Bertl
http://vserver.13thfloor.at/Stuff/AXIOM/reg_file_gpio/
16:10
Bertl
http://vserver.13thfloor.at/Stuff/AXIOM/hdmi_i2c/
16:12
Bertl
the scripts (in hdmi_i2c) can be used to basically configure the hdmi transmitter to a point where e.g. EDID info can be retrieved
16:12
Bertl
there is also an info script which shows various registers and bits
16:17
Bertl
as usual, let me know if you have any problems with the setup or any suggestions/improvements on the code!
17:13
se6astian
Hello!
17:14
se6astian
thats great!
17:14
se6astian
I was hoping you could help me something stuart recommended with the PCB
17:15
se6astian
1. I see a could sensor pins that you use one via to attach to two sensor pins. These need to be separate vias to reduce inductance and lower resistance to the power plane.
17:15
Bertl
parse errror, please rephrase :)
17:16
se6astian
thats what he wrote ;)
17:16
se6astian
I replied: In some areas this is impossible to do, does it help if I connect ground pins together and make several vias like in this image?
17:16
se6astian
http://picpaste.com/gnd-good-or-bad-mwsEPLxA.jpg
17:16
se6astian
and to that Stuart replied: - Better, but the top layer copper pour that's attached to ground will likely take care of alot of this. You might be able to remove the traces in favor of the lower impedance thermal ties to the ground pour. Keep in mind that you want thermal ties to ground for component pins on the top/bottom layer, NOT a completely surrounding connection. The solder process
17:16
se6astian
will only work properly with thermal ties.
17:17
se6astian
I don't really know what that means now ;)
17:17
Bertl
okay, what am I seeing in that picture?
17:18
Bertl
what are the blue dots?
17:18
se6astian
stuarts original comment was about me connecting several sensor pins that are supposed to be grounded together with a single via to the ground plane
17:18
se6astian
I think he suggested to make one GND via for each sensor pin
17:18
Bertl
ah, okay, that's a bad ide
17:18
Bertl
*idea
17:19
se6astian
so I did that everywhere I could
17:19
Bertl
you basically want as many direct connections to the ground plane as possible
17:19
Bertl
but you want to avoid strange loops and runs
17:19
se6astian
but in the area as seen in the image there isnt enough space, so I wanted to know if its better to do it that way (with several vias spread around and the actual pins also connected together)
17:20
se6astian
the blue dots are the sensor pins supposed to go to GND
17:20
Bertl
why are they so tiny?
17:20
Bertl
i.e. how is the socket? supposed to connect there?
17:21
Bertl
or are the vias huge?
17:22
Bertl
and what do the red squares mean?
17:22
se6astian
the socket pins are rather small SMT
17:22
se6astian
the red squares are the edge points of the traces when you draw them
17:22
Bertl
well, they look like half the size of a via on that picture
17:23
Bertl
less than half the diameter actually
17:23
se6astian
they are rather tiny
17:23
Bertl
assuming that your vias are what 0.6mm 0.4mm that would suggest 0.3 or 0.2mm pads
17:23
se6astian
not sure if they are supposed to be that small though ;)
17:23
Bertl
you can never ever solder onto 0.2mm pads
17:23
se6astian
let me check the andon datasheet
17:24
se6astian
andon replied by the way :)
17:24
se6astian
"we have over 5000 products and cannot provide a single pricelist"
17:24
se6astian
so I sent them the part numbers again I originally inquired for
17:24
Bertl
hehe
17:27
se6astian
normal SMT pins: 0.30mm
17:27
se6astian
rollerball socket: 0.47mm
17:28
Bertl
what is the size of your vias? drill and surrounding?
17:30
se6astian
0.026" via, 0.008" hole
17:31
se6astian
its the smallest via available in expresspcb
17:31
Bertl
hmm, yeah, 0.2mm drill and 0.66mm pad
17:32
Bertl
but I'd definitely make the smd pads at least 0.5 as well
17:32
Bertl
is there no footprint available for the sockets?
17:33
se6astian
no, stuart built a custom part
17:35
Bertl
you have the CMOSIS-ANDON 'data sheet' at hand, yes?
17:35
se6astian
yes
17:35
se6astian
email?
17:35
Bertl
no, I have it, just look at it
17:36
Bertl
the -414T4 on the right (middle of the page)
17:36
Bertl
shows a 0.30mm pad where the socket lands
17:36
Bertl
but, directly below you see how the socket is soldered
17:37
Bertl
as you can see there, the solder will roughly fill up to 0.80mm (diameter)
17:37
Bertl
there is no chance in hell to get a reliable contact with a 0.30mm diameter footprint
17:37
se6astian
I am not sure if the "Mounting Options" box at the right isn't just a general example
17:38
se6astian
not specific to any type of pins
17:38
Bertl
yes, of course, but that is how it really looks like
17:39
se6astian
but what happens to the spike of the pin, in the mounting options box it looks like it melted away ;)
17:41
Bertl
http://i236.photobucket.com/albums/ff142/Stocky6409/PICAXE/IMG_0688.jpg
17:41
Bertl
here you see how solder, pads and pins relate in SMD
17:42
Bertl
now consider what would happen when the pad below the pin was just the size of the pin
17:42
se6astian
Ah I see what you mean
17:42
Bertl
the only case where the pad is of equal size is with BGA
17:43
se6astian
and stuart also refered to Thermal pads
17:43
Bertl
where you basically mount the pins on solder balls
17:43
se6astian
in expresspcb that is a via option, creates a cross like non copper area in the pad
17:43
Bertl
yes, those are thermal
17:44
Bertl
you use them to reduce heat propagation
17:45
se6astian
ok, will use those instead for the socket
17:46
se6astian
and which diameter?
17:46
Bertl
personally I'd strongly consider using the thru-hole sockets for the adapter, not will they be a lot simpler to solder on, but they also 'require' sane connections
17:46
Bertl
no, you don't want thermals on the sensor pads
17:46
Bertl
that is complete nonsense :)
17:47
Bertl
s/not/not only/
17:48
Bertl
so all in all, you don't need as many vias (because you already have the holes)
17:48
se6astian
good point
17:48
se6astian
I dont mind using the throughhole version
17:48
se6astian
the price is the same
17:49
se6astian
and there is still space between PCB and socket as I placed a few SMD parts there
17:50
Bertl
how do you know that the SMD parts do not interfere with the socket?
17:50
Bertl
that's one major missing information on my side (i.e. how does the socket actually look like)
17:50
Bertl
we have the size, and the pins, but beyond that, all I see is a rectangular area
17:51
se6astian
according to the datasheet there are 3mm between PCB surface and the socket
17:51
se6astian
the SMD parts are 0.8mm tall if I remember correctly
17:52
Bertl
yeah, but I don't see any information about the actual socket frame
17:52
Bertl
it could be that while the pins have 2.8mm space, the frame directly touches the pcb
17:53
Bertl
and that could happen everywhere inside the rectangular area
17:53
Bertl
after all, something is holding those pins in place
17:54
se6astian
take a look at CMOSIS 06-21-2012.pdf
17:54
Bertl
how this something looks like, I have no idea ...
17:54
Bertl
I don't have that one
17:54
se6astian
I emailed it to you on June 20th
17:54
Bertl
hmm, okay, sec
17:55
Bertl
indeed, sec
17:58
Bertl
well, when I read that datasheet (thanks by the way) correctly, you have exacly 0.68mm space between socket and pcb, in the areas where the socket is present
17:59
Bertl
now the question is, with the window marked 'optional', what does that mean?
18:00
Bertl
I would prefer to get sockets with that window present
18:00
se6astian
to my understanding that there are 2 products to order (with/without window)
18:00
Bertl
(otherwise you won't be able to reach/cool the cmv)
18:01
Bertl
but you can definitely forget about soldering parts near the socket pins
18:01
Bertl
but IMHO there is no reason for doing that either
18:03
se6astian
at first I had most of the capacitors in the socket area on the back of the PCB, but Stuart recommended to move them to the front to reduce via inductance
18:04
Bertl
thru-hole => front/back is the same
18:05
se6astian
true
18:05
se6astian
ok, so to summarize my todo:
18:06
se6astian
-) change image socket pads to bigger pads
18:06
se6astian
-) make them through holes
18:06
se6astian
-) move caps to back side