Current Server Time: 05:54 (Central Europe)

#apertus IRC Channel Logs

2014/10/10

Timezone: UTC


00:05
Bertl
via IRC :)
00:07
Rebelj12a
Well kudos I know how hard it is to organize a team over the internet. Its not easy, and that wasnt even an actual product just a service.
00:11
Bertl
well, we'll see how it scales, I think the mailing list and bug/feature tracker will soon gain importance
00:12
Rebelj12a
Well, I do have something that might help if you can get it set up.
00:12
Rebelj12a
You have a web dev right?
00:14
Bertl
"web dev" is somebody developing web based applications?
00:14
Rebelj12a
well developer
00:14
Rebelj12a
as in someone who knows the ins and outs of server management, ruby, php all that?
00:17
Rebelj12a
Either way heres some links but ill include it in the email im sending you bertl. http://www.discourse.org/ and http://kandanapp.com/
00:18
Rebelj12a
Kandan is the good one I thought, reminded me sort of Google Wave (which i loved) but desktop notifications, live team cheat and drag and drop file sharing which is convenient for larger documents.
00:19
philippej
joined the channel
00:26
Rebelj12a
*chat not cheat lol
00:27
Bertl
no problem
00:27
Bertl
from the first glance, discourse looks interesting
00:28
Rebelj12a
Indeed Kandan I like it because of the different rooms, and filesharing plus notifications. I get distracted easily so hah
00:28
Bertl
kadanapp doesn't convice me judging from the page
00:28
Rebelj12a
Yeah and their demo is down which is concerning.
00:29
Bertl
the 'rooms' part we have with IRC as well, file sharing is typically done via git, ftp or paste servers
00:29
Bertl
anyway, thanks a bunch, we'll look into it
00:30
Rebelj12a
yeah, I used google drive back in the day. Although Git is a bit more precise, ( dont get how it works somewhat but whatever I dont use it that often either )
00:30
Rebelj12a
Hey no problem.
00:30
Bertl
(might want to send your email to team at apertus dot org
00:30
Bertl
)
00:30
Rebelj12a
hah ok
00:43
Rebelj12a
Yeah ive got a checklist of things to send. Probably will get time to this weekend. Have a client this saturday. Prepare prepare. Testing out the DJI ronin on a rope line hopefully. If i can make it strong enough.
00:44
philippej
stay tuned for the phabricator install :-)
00:45
Rebelj12a
oooh phabricator
00:45
seku
now whats that :)
00:46
philippej
I'm testing it, it looks really cool
00:46
philippej
http://phabricator.org/
00:46
seku
ooooh i see
00:46
Rebelj12a
Hey if it looks cool and does something cool well, then its a success
00:46
Rebelj12a
Oh yeah that is nice
00:47
seku
it does look nice.
00:47
Rebelj12a
HAH i like their headline partway down. Review Code or just stare at it O.O
00:47
philippej
it's in use here for example : https://developer.blender.org/
00:48
Rebelj12a
Oooh blender good people, looks awesome. Much better than what I had for examples.
01:10
theuberkevlar_
joined the channel
01:11
theuberkevlar_
\nick theuberkevlar
01:11
theuberkevlar_
how do I do it?
01:11
theuberkevlar_
haha
01:11
Rebelj12a
hahaha ( /msg NickServ
01:12
Rebelj12a
or /msg NickServ help
01:12
Rebelj12a
Gives a rundown
01:12
theuberkevlar_
huh
01:12
theuberkevlar_
it says it's already in use
01:12
Rebelj12a
â /msg NickServ identify?
01:13
Rebelj12a
theres a reclaim command in theres somewhere i think
01:14
theuberkevlar_
haha
01:14
theuberkevlar_
oh well
01:14
theuberkevlar_
no big deal
01:15
theuberkevlar_
Yeah, phabricator looks interesting
01:15
Rebelj12a
Well if I ever have a project or team again Definitely going to use that. Would have helped so much.
01:16
theuberkevlar_
Their site is super funny! " ...a dragon who guards your hosted repositories."
01:27
Rebelj12a2
joined the channel
01:28
Rebelj12a2
Crap in a hat I locked myself out D:
01:30
philippej
see you later everyone :-)
01:30
Rebelj12a2
Bah of the days annoyances.
01:30
Rebelj12a2
Bye!
01:30
philippej
left the channel
01:32
Rebelj12a2
Important things, keys, eating, drinking right don't forget.
01:36
Rebelj12a2
0 for 3 so far, great... So for the beta because of the modularity of the front mount and sensor, you could really attach any lens to it?
01:37
Bertl
if the lens system matches the sensor and a lens mount is available, then yes
01:40
Rebelj12a2
left the channel
01:42
Rebelj12a
hmm awesome awesome.
01:43
Rebelj12a
Although it might be too close for the flange distance, built in NR filters would be nice indeed.
01:44
Rebelj12a
Plus no limit really to mount compatibility is defintely a plus. Well as if people needed any more reason to keep buying up all the good older lenses for cheap.
01:46
Rebelj12a
not flange mount distance but whatever
02:19
Rebelj12a
All is quiet on the western front.
02:19
Bertl
yeah, I guess I'm off to bed now ... have a good one everyone!
02:19
Bertl
changed nick to: Bertl_zZ
02:22
Rebelj12a
night
02:29
Rebelj12a
Blasted colorchecker I know I put you somewhere I just dont know where....
02:30
Rebelj12a
Already checked the fridge now im just out of options.
02:36
Rebelj12a
Hey does anyone know the â¦. well i suppose its Shutter life of the sensor for the beta?
02:38
Rebelj12a
I dont know if a mirrorless camera has a shutter life, no wait it has to⦠the A7s has to have one its a dslr.
02:38
Rebelj12a
I smell a photo version of the beta. Smells like truimph and napalm. I know i have a few photographer friends who would be all in for an open source dslr
02:45
dmjnova
left the channel
02:45
Rebelj12a
granted as long as the shutter life is long enough to appeal to photographers
02:47
Rebelj12a
Although iâd sit on that till gamma from a marketing perspective. Do a dual launch news type thing, with beta updates for photo software controls. Appeases both gamma investors and beta investors somewhat.
02:56
sb0
left the channel
02:56
Rebelj12a
ill add it to the list of things to email
03:03
Rebelj12a
Heading home talk to you guys later
03:03
Rebelj12a
left the channel
03:09
AndroUser2
joined the channel
03:15
theuberkevlar_
left the channel
03:44
AndroUser2
left the channel
03:45
Rebelj12a
joined the channel
03:45
Rebelj12a2
joined the channel
03:45
Rebelj12a2
Blah
03:46
Rebelj12a2
left the channel
03:46
Rebelj12a2
joined the channel
03:46
Rebelj12a2
Blast whatever
03:48
Rebelj12a
left the channel
03:48
Rebelj12a2
Aha ok all fixed
03:49
Rebelj12a2
left the channel
03:49
Rebelj12a
joined the channel
03:50
Rebelj12a
There
04:16
Rebelj12a
http://www.engadget.com/2014/10/09/samsung-tlc-v-nand-ssd/?ncid=rss_truncated good news everyone, although I' doubt we will see price drops...
04:30
intracube
left the channel
04:57
Gegsite
joined the channel
05:28
danieel
left the channel
05:29
danieel
joined the channel
06:24
daFred_
joined the channel
06:25
daFred_
left the channel
06:32
Gegsite
left the channel
06:50
niemand
joined the channel
06:56
Even
joined the channel
06:57
Even
good morning.
06:58
Even
and happy friday! =)
06:59
seku_
joined the channel
07:00
seku_
moarnings
07:01
Even
moin seku
07:04
seku_
moinmoin
07:04
seku_
way too early to feel awake
07:06
Even
go an put some coffeine into your system. :)
07:06
se6astian|away
changed nick to: se6astian
07:08
seku_
doing so right now
07:08
se6astian
good morning
07:09
seku_
hi seb
07:11
Even
good morning se6bastian
07:11
Even
http://38.media.tumblr.com/30786e8c297a40dbb7267607c709096b/tumblr_nd4koj6t3z1swr8wlo1_1280.png =)
07:21
se6astian
any native speaker around, I need a shor text reviewed before I post it
07:23
seku_
sorries, not native
07:24
se6astian
damn those timezones :)
07:24
Even
i guess my english isn't better than yours but i can give it a look if you want me to
07:24
se6astian
why not :)
07:24
seku_
not many UK backers then :)
07:25
seku_
sure, can also have a look
07:37
niemand
left the channel
07:49
philippej
joined the channel
07:53
se6astian
https://apertus.org/axiom-beta-after-campaign-end-article-2014
07:57
wescotte
still need an editor se6astian?
08:00
se6astian
its published already
08:00
se6astian
but please take a look
08:01
se6astian
if you find an error I can fix it
08:01
wescotte
https://apertus.org/axiom-beta-after-campaign-end-article-2014 <-- assume it's this one then?
08:25
daFred
joined the channel
08:25
daFred
good morning
08:26
se6astian
wescotte: yes
08:26
se6astian
hi daFred
08:29
daFred
hi se6astian, what's on the roadmap? any meetings planned for developers?
08:30
se6astian
would you like to participate in one?
08:32
daFred
yes, regarding mechanics, I sent you the images of the 3d printed parts you know...
08:33
philippej
btw there is also a very interesting cooling research doc that was posted, let's keep track of that
08:34
daFred
from anton, I've already bookmarked it...
08:34
regmac
Need anyone to edit tech docs? I was editor of "develop" at Apple. I wouldn't want to do full edit, but readability and structure, sure. I'm pretty good at it.
08:37
regmac
philippej: Where was the cooling doc, if yo don't mind.
08:38
philippej
regmac, that's rpoblem, somewhere on the irc logs
08:38
philippej
we'll soon have a propoer system to collect all this (we have the wiki by the way : wiki.apertus.org , but it's not used enough, people just like the convenience of irc in those rushed days :-))
08:38
daFred
you mean this one http://octoray.co.uk/axiom/ ?
08:39
philippej
yes :-)
08:40
philippej
please add it to the wiki
08:40
philippej
in general, the wiki needs a lot of love
08:40
philippej
most important, a proper table of content
08:41
daFred
it's from __anton___ but i can set a link on the wiki...
08:41
regmac
I don't have this client set to log channels. Is there a log somewher?
08:42
philippej
irc.apertus.org is your friend :-)
08:43
philippej
or enemy, depending on the time you want to allocate on reading logs :-)
08:45
regmac
http://octoray.co.uk/axiom/ ?
08:46
regmac
Ah. yes. Did not recognize it the firt time :-)
08:55
daFred
I put the link of __anton___'s document to the wiki https://wiki.apertus.org/index.php?title=Heat_management
08:56
daFred
__anton___ can you put your document there?
08:57
philippej
left the channel
08:57
se6astian
daFred: Its you da_manfred - I am having a hard time keeping track of all the nicknames/real names :D
08:59
se6astian
regmac: the wiki could definitely need some restructuring, collecting of stuff together
08:59
se6astian
http://wiki.apertus.org/
09:00
daFred
I'm Manfred yes...
09:00
se6astian
hurray :D
09:20
wescotte
concerns related to plan B1: "if we put a little more of thermal compound we may end placing to the lens slightly closer to the lens than we wanted to" <--- lens slightly closer to the lens is a typo of some sort..
09:24
Bertl_zZ
changed nick to: Bertl
09:24
Bertl
morning folks!
09:24
se6astian
good morning!
09:39
daFred
morning bertl, are there any alternatives to the andon sockets to get more space between sensor and board?
09:39
Bertl
we probably don't want more space, but we want less socket
09:41
Bertl
one option we already investigated but with little luck so far, which is missing in __antons__? cooling review is to put a heat plate (copper/aluminum) under the sensor, right around the pins
09:42
Bertl
i.e. have a plate which is almost as high as 'the socket' with some isolation to the PCB side and thermal grease on the sensor side
09:42
Bertl
the "socket" consists only of socket pins, no fiberglass or similar holding them in place
09:43
Bertl
how can this be done? well, there are so called Peel-away sockets available at certain manufacturers, which basically have to perforated flexible PCB attached to the socket pins, which can be removed after soldering
09:45
Bertl
the plate should be screwed against the lens mount, fixing the sensor between itself and the lens-mount
09:46
daFred
less socket sounds good... what is the distance at the moment? thinking of a ceramic part with holes and/or peltier...
09:47
daFred
afk for a moment...
09:54
Bertl
the distance is the andon socket height plus a small gap from the base of the sensor pins
09:55
Bertl
I have no board/sensor ATM, one is at MQ the other with sebastian
10:14
wescotte
left the channel
10:17
ThatCantBe
joined the channel
10:37
aombk2
changed nick to: aombk
11:22
seku_
left the channel
11:31
philippej
joined the channel
11:37
philippej
left the channel
11:39
mikea
joined the channel
11:43
mikea
left the channel
12:37
swah
joined the channel
12:40
AndroUser
joined the channel
12:42
AndroUser
changed nick to: __anton
12:44
__anton
Hi guys I can move my article to wiki if you dont mind such a large mind dump. I may put each option on its own page if that sounds ok
12:46
Bertl
that would be great
12:46
Bertl
please check the option I explained above and if possible add that as well (maybe with a nice/fancy graphic)
12:47
__anton
Bertl: my Plan B1 (full metal jacket) was actually my fantasy inspired by your plan with peel away sockets
12:49
Bertl
ah, okay, but that did't sound like what I had in mind
12:49
__anton
Bertl: I happy to make a graphics of your plan but I need to understand it better
12:49
Bertl
so maybe we need to refine that or break it down somehow
12:49
Bertl
sure, glad to explain/discuss it
12:50
Bertl
the idea is to have a solid plate, with drill holes where the socket pins are
12:50
Bertl
which is electrically and thermally isolated from the PCB
12:50
__anton
In B1 that metal plate is part of camera front panel so heat is dissipated by that, and no chance for a peltier
12:51
__anton
Wow, a metal socket!
12:51
Bertl
the plate can extend over the PCB (top/bottom) or go to the side after having a small knee
12:51
Bertl
with either heat finns or peltiers attached
12:51
Bertl
it can also connect to the lens mount if desired
12:54
__anton
So Peltiers would attached to camera sides?
12:54
__anton
Where would the pins face? Inside of the cam?
12:55
Bertl
what pins?
12:55
Bertl
I had two concepts for the peltier, one which uses a single peltier on top of the camera
12:56
Bertl
which is directly attached to the plate (which would make a bend at the top-front edge
12:56
Bertl
and the second one is with two smaller peltiers, attached on the sides, which connect to the front side of the case
12:57
Bertl
so they would basically sit above the plate
12:57
Bertl
should I make a drawing to illustrate?
12:58
se6astian
gotta go
12:58
se6astian
bbl
12:58
Bertl
cya
12:58
se6astian
changed nick to: se6astian|away
13:02
__anton
Typing on my mobile, lots of typos
13:02
__anton
Wanted to ask where the fins face not pins
13:04
__anton
Understood about peltiers. Btw do you already know what sizes are available?
13:05
__anton
What would the thermal isolation from the pcb look like?
13:12
Bertl
in the simplest case it would be a layer of mylar or so
13:12
Bertl
regarding fins, they would face away from the camera
13:14
__anton
BTW have you been considering smth like this? http://www.mpe-connector.de/index.php?lang=en&menu=4&product_group%5B0%5D=1&action=Search&change_view=true&list_view=true no fancy peel away tape but you may be able to create a socket
13:17
__anton
These peel aways do they actually come with isolation which would separate them from the metal plate with drilled holes?
13:23
__anton
Bad link from me, they are wrong pitch.. we need 1.27mm...
13:23
Bertl
not the ones we found so far, but I'm not too worried about this problem
13:24
Bertl
because we can either cut out a slightly larger area (if we have precision problems)
13:24
Bertl
or orient the plate very precisely on both sides, with holes in the sensor and the PCB
13:24
Bertl
s/with/via/
13:25
Bertl
but I think for the first approach a cutout around the pin-group(s) would be more than sufficient
13:30
__anton
Btw 1.27mm pitch is also known as 0.50
13:30
Bertl
yes, that's inch
13:31
Bertl
well, actually 0.05"
13:47
__anton
Not that hard to find 3 row 1.27 pitch sockets, but they are for pins 0.4mm in diameter while cmv12000 seems to have 0.30 mm
13:52
Bertl
the problem is, we can't use 3 row sockets
13:52
Bertl
they won't align well enough to actually get the sensor in during reflow
13:53
Bertl
(btw, we need SMD/SMT sockets, but that shouldn't be a problem)
13:54
Bertl
so we definitely need something like those peel-away sockets
13:55
Bertl
of course, we could use a dummy sensor during reflow, but that would complicate things a lot I guess
13:56
Bertl
http://www.advanced.com/products/peel-a-way-removable-carriers
14:03
intracube
joined the channel
14:03
Bertl
wb intracube!
14:09
__anton
I wonder if calling their sales rep or several of them might help to get a response from Advanced
14:09
Bertl
I thin a few reqyests might already help, so feel free to send an email with a description of the sensor(s)
14:10
intracube
afternoon Bertl :)
14:11
__anton
Cool will do, I am in UK so I will try one of those. What size of an order should I speak about?
14:15
__anton
Well I will try anyway.. I understand we need several hundred for crowdfunding and then possibly a lot more
14:16
__anton
...and firt there need to be samples (ideally)
14:26
swah
left the channel
14:38
Bertl
__anton: precisely
14:47
danieel
Bertl: why would you put the peltiers that far from the sensor? if you want the best efficiency, the elements shall be directly on the sensor package, no?
14:48
danieel
btw congrats to the campaign :)
14:49
__anton
left the channel
14:52
daFred
what about removing the sensor? isn't it a bit dangerous to break a pin from the board when using SMD/SMT sockets. as it's for development it should not be a single use solution. danieel: i'm with you!
14:56
danieel
if you use smd socket, you can break the socket/pcb connectivity ruining the board, my preference is tht socket or no socket at all
14:56
danieel
smd socket is usable at one time assembly when the sensor is precise mounted to a mechanical piece
14:58
daFred
again with you danieel
14:58
danieel
and to remove the sensor... you need the proper tool, the ceramics is fragile so a hard screwdriver is not the best (as we figured out the hard way)
14:59
intracube
is peltier tech being considered for cooling?
15:00
intracube
danieel: ^
15:01
daFred
maybe to migrate the sensor to the gamma apertus can offer a service to remove the chip with propper tool...
15:02
intracube
I thought peltier was too inefficent for portable applications
15:02
daFred
intracube: i think it should be an option
15:05
intracube
peltier efficiency seems to be very roughly in the 5% to 10% range
15:06
intracube
I'm trying to find a proper formula but this seems to suggest it wouldn't work off batteries
15:11
__anton
joined the channel
15:11
__anton
Well its just 2Wt of power sensor generates. With 10% efficiency the peltier will consume 20Wt.. perhaps ok for portable
15:12
intracube
__anton: what about FPGA cooling?
15:12
dmjnova
joined the channel
15:13
intracube
that might be 10-20w range
15:13
intracube
https://tetech.com/Peltier-Thermoelectric-Cooler-Module-Calculator/?mode=1&dtmax=70&heatLoad=10&hotSideTemp=40&coldSideTemp=20&potted=0&emailsent=0
15:13
__anton
Peltier doesnt have to be on sensor.. on wiki theres a story about a guy who had peltir outside cam and drove sensor sub zero
15:13
intracube
isn't this normally reserved for special uses like astro photography?
15:14
intracube
do existing cinema cameras (ARRI, Sony, Red) use peltier?
15:15
__anton
Another important role of peltier is to stabise keep temp constant
15:15
__anton
On sensor
15:15
__anton
Zynq we need to cool in other ways
15:15
__anton
I think high end cams may use peltier
15:16
__anton
Zynq doesnt need to be stable temp
15:16
__anton
You want to avoid thermal damage and avoid heating sensor
15:17
__anton
Thats all needed for zynq
15:17
__anton
Sensor needs to be stable ideally
15:18
intracube
__anton: yep, I just searched and it looks like ARRI indeed uses peltier for the sensor
15:18
__anton
My thinking peltier could be an option
15:18
intracube
yep
15:19
__anton
E.g. not everybody would need it but it is nice to be able to put one in optionally
15:21
intracube
yes. if Arri has done the research and decided on peltier, it's a safe bet
15:22
intracube
from the Alexa manual: "ALEXA has a Peltier element that keeps the image sensor at a stable temperature. This is important to achieve constant image quality."
15:27
dmjnova
left the channel
15:32
__anton
Guys if we use a peel away socket but we put a "metal socket" on top of it maybe removing the sensor would still be possible? This metal plate might work as an extraction tool? Hmmm our sensor legs are just 2.5mm.. we would need tall pins on a peel of socket
15:36
dmjnova
joined the channel
15:38
daFred
yes, metal socket with four threads to push it off the pcb. good idea...
15:41
dmjnova
left the channel
15:50
dmjnova
joined the channel
15:53
__anton
left the channel
15:57
slikdigit
left the channel
15:57
slikdigit_
joined the channel
16:46
Bertl
off for a nap ... bbl
16:46
Bertl
changed nick to: Bertl_zZ
17:19
__anton
joined the channel
17:33
se6astian|away
changed nick to: se6astian
17:41
ff_
joined the channel
17:43
ff_
left the channel
17:46
Rebelj12a
Hm I don't know if it's feasible what about a magnetic socket combined with removable fitting or is that still too big? And not feasible for a sensor?
17:46
Rebelj12a
http://www.google.com/patents/US7758349
17:47
Rebelj12a
http://www.google.com/patents/US8406007
17:51
Rebelj12a
http://m.asianproducts.com/?op=product&item_id=P13177237084709599
17:53
danieel
Rebelj12a: what are you trying to solve?
17:55
Rebelj12a
I think I did that wrong
17:55
Rebelj12a
left the channel
17:55
Rebelj12a
joined the channel
17:55
surami
joined the channel
17:55
danieel
now you did it wrong :)
17:55
Rebelj12a
Yes yes I did
17:56
Rebelj12a
Sensor mounting.
17:56
danieel
solder it or socket it, thats not hard
17:56
Rebelj12a
I was just thinking rather than risk both sensor and underneath board by requiring the user to solder both together, it might be better to only have the user risk one component of the camera as opposed to two.
17:57
danieel
you can solder it to a board which has a b2b connector..
17:57
danieel
there is no risk if you choose the right components
17:58
Rebelj12a
Well yeah, I might have been misundeerstanding the conversation
17:58
surami
good evening
18:01
Rebelj12a
Good afternoon XD
18:02
Rebelj12a
Oh yeah danieel it seems they were trying to get it closer or come up with options for doing so.
18:02
danieel
closer to?
18:03
Rebelj12a
The logs for Friday has it. Alternatives to andon sockets
18:04
Rebelj12a
Peel away sockets or something
18:05
Rebelj12a
And peltier
18:05
Rebelj12a
I don't know I'm not an engineer I'm just throwing things in the pool. In just a filmmaker :b with a lot of tech experience. xD
18:08
Rebelj12a
Well I'm off, gotta scout out a site before a job. See if we can rig up some cables for a BCam talk to you guys later.
18:09
Rebelj12a
left the channel
18:20
surami
bye
18:20
surami
left the channel
18:27
aombk
redesigning a website is worse than creating a new one
18:33
__anton
Hi, after lots of googling I've seen lots of puctures of through hole mounted peel away sockets and strips. However that would require drilling 240 holes per sensor board. Too expensive? However I just saw one mentioning of surface mounted peel away female sockets and it was "contact the factory" - e.g. it was referred to as smth unusual. Now I am wondering if that would actually work.. Smth has to hold the terminals straight up during so
18:37
__anton
However if we agree to drilling 240 times then aligning two separate 4 row sockets is no longer an issue and we can go scout for suppliers of 30*4 1.27 pitch mount through hole sockets?
18:37
derWalter
joined the channel
18:45
Bertl_zZ
changed nick to: Bertl
18:45
Bertl
back now ...
18:45
Q_
I'm a little confused.
18:47
Q_
Why would you want to have something like peeling away the sensor?
18:47
Bertl
__anton: SMD/SMT is not really unusual
18:48
Bertl
__anton: and drilling might be an option for the CMV12000 but not for the CMV2000/4000
18:48
__anton
Yeah, but a SMT peel away socket?
18:48
Bertl
shouldn't be a problem, IIRC, they had them as well
18:49
Q_
Isn't it important that you can properly allign the sensor with the lens?
18:49
Bertl
Q_: we want to peel-away the thing which _holds_ the socket pins together
18:49
Bertl
check out the url I pasted some time ago, it should make things clearer
18:50
__anton
Q: its about a replacement for that 90 eur socket for cmv12000
18:50
Rebelj12a
joined the channel
18:50
Rebelj12a
Oh hey it worked
18:51
__anton
We dont like it cause its expensive and doesnt allow to put a metal strip under cmv12000
18:52
__anton
The metal strip is to take away the heat
18:52
Q_
Bertl: It doesn't make things more clear.
18:52
Q_
(And I already found that.)
18:53
__anton
Q_: do my words make sense? :-)
18:53
Bertl
okay, the point is, that we want to cool the sensor, from the back, yes?
18:53
Bertl
we do not want to have large holes in the PCB, because that isn't good for the traces and components on the other side :)
18:54
Q_
Oh.
18:55
Bertl
so we want to get a metal plate right under the sensor
18:55
Q_
I didn't know that the sensor would require cooling.
18:55
Bertl
it doesn't but it increases image quality by reducing noise
18:56
Bertl
so it's not essential, but it is something we would like to have, if possible
18:56
__anton
Bertl: btw, 3 words: metal core pcb
18:57
Bertl
yes, we could use aluminum PCBs I already considered that
18:57
Rebelj12a
What is optimal temperature for image quality and noise reduction?
18:57
Bertl
but they are expensive and not really easy to get if you want to build your own AXIOM Beta
18:57
Bertl
Rebelj12a: as low as possible
18:58
Bertl
there are some problems once you go below a certain temperature because of the water in the air, you have to switch to other cooling environments like liquid nitrogen or so :)
18:59
Bertl
but for our purposes, around 20° C would probably be good enough
18:59
Rebelj12a
God I don't think I know how to pm whatever let me check my archives for tech related to this.
19:01
Q_
Bertl: So I think there are standard mount products for those, like you see for CPUs?
19:02
Bertl
well, if you find something which can be used for the CMV12000, please let us know
19:03
__anton
Imagine all Beta could be mounted on one side of 6mm alum pcb! Yean Zynq would heat up CMV12000 but what a degree of mechanical simplicity! Then make the lens mount one piece with front cover milled from a brick of alum and voi la!
19:03
Rebelj12a
Ah blast well price is an issue. Of course
19:03
Bertl
Q_: the only manufacturer producing a socket specifically for the CMV12000 is Andon, and they charge a lot for a primitive socket
19:04
Q_
Bertl: We buy special sockets too, but we ussually only buy 1, and they tend to be very expensive.
19:05
__anton
You could not really go one sided so you'd need an aux pcb attached on connectors but you would only place components there that do not need cooling. What a nice fantast camera :)
19:05
Bertl
Q_: okay, yeah, that probably isn't an option folks would like very much :)
19:06
Rebelj12a
So you want the sensor as close as possible. There's microcooling PCB however I have a feeling it will be expensive.
19:06
Q_
Bertl: But they're the type of sockets they don't keep much stock form and ussually have to make a batch when we order one.
19:08
Rebelj12a
http://arstechnica.com/gadgets/2008/01/startup-shrinks-the-peltier-cooler-and-puts-it-inside-the-chip-package/
19:09
__anton
Bertl, could we be wrong to avoid connection cables? Is there a cable already for the connectors on Microzed?
19:13
__anton
Then mechanically Beta becomes solid and simple. CMV 12000 is attached to and cooled via front metal panel, zynq is attached to and cooled via back metal panel. The middle is there to thermally insulate one from another.
19:14
__anton
You no longer have to have io shields on the front, you can have them in the middle
19:15
__anton
And connection sockets on the sides as many cameras do
19:21
Bertl
cables are usually problematic because of shielding/impedance
19:23
Rebelj12a
Ok so you are trying to get the board as close to the PCB as possible and cool it with fairly easy interchangeable functionality?
19:25
__anton
Rebelj12a: the words "board" and "pcb" in your sentense are synonyms aren't they?
19:28
__anton
Bertl: I understand all hi bandwidth communication on those cables would be LVDS.Now i am a total noob but doesnt that remove problems with shielding?
19:28
__anton
Like on twisted pair?
19:29
Bertl
if you twist and shield them, then yes :)
19:29
Bertl
i.e. cat7 would work :)
19:30
Bertl
FPC/FFC with shielded strips works as well, if you put a ground lane between pairs
19:30
aombk
forever funding. scary!
19:31
Bertl
that's their name, it will end in less than a fortnight
19:31
__anton
Mental experiment: I replace the cable with a straight aux pcb 10 cm long running from front to back of the cam, then cam body is made of tin to shield from external interference. Do i still have a problem?
19:32
Bertl
depends on the PCB and the connectors
19:32
__anton
Well and i have two of them running in parallel because microzed got two sockets
19:32
Bertl
10cm shouldn't be a big problem
19:34
daFred
the sensor surface has a distance of 1.8mm +/- 0.2mm from de back of the sensor. So we need the back of the sensor as a reference area for the flange focal distance. This surface should be plain with the back of the lens tube. anyway we should be able to adjust the lens tube this +/- 0.2mm
19:34
Bertl
__anton: but as I said, requires a proper PCB, i.e. 4 layer, and connectors which allow for high frequency
19:35
Bertl
daFred: we just need to make shure that it isn't too long, the rest can, if necessary, be done with shims
19:36
daFred
and we have to do this with 450 cams :-)
19:37
Bertl
which makes it essential that we devise a good testing setup :)
19:37
daFred
jo
19:38
Bertl
we should have the feature/task/bug tracker up and running after the weekend (I hope)
19:41
daFred
maybe we get the position of the sensor area from the manufacturer or we can measure it to before assembling and sort tubes and sensors together...
19:41
Bertl
but I think you have a flaw in the logic here
19:42
Bertl
you are referring to the tolerances from back to front, yes?
19:42
Bertl
we do not really care about them, as the sensor will be screwed against the lens mount
19:43
Bertl
so the only tolerances we care about are between the front of the sensor and the image area
19:45
__anton
left the channel
19:47
__anton
joined the channel
19:47
__anton
left the channel
19:48
daFred
looking at the datasheet i think the back of the sensor is the saver surface regarding the position tolerance..
19:48
daFred
but i may be wrong
19:51
__anton
joined the channel
19:54
__anton
changed nick to: __anton_
19:57
__anton__
joined the channel
19:57
Rebelj12a
Is there a cad drawing or schematic I could get so I can get an idea of how big it is and fits togwther.?
19:57
Bertl
what exactly?
19:57
Rebelj12a
3d model
19:58
Rebelj12a
Housing not internals.
19:58
__anton_
left the channel
19:59
_anton
joined the channel
19:59
Bertl
yes, se6astian should know where you can get it
19:59
_anton
changed nick to: ___anton
19:59
Rebelj12a
Ok thanks
20:00
daFred
https://github.com/apertus-open-source-cinema/beta-hardware
20:00
daFred
inventor drawings
20:00
Bertl
ah, thanks
20:01
__anton__
left the channel
20:01
___anton
Rebelj12a: in github there's some stuff. Mechanics is in autodesk inventor format and pcbs are in Eagle format. There are several pdf files as well
20:03
Rebelj12a
If I use github improperly I apologize in advance. Still getting the hang of it.
20:04
daFred
please don't delete the internet!!!
20:05
___anton
left the channel
20:05
Rebelj12a
I broke the internet once I'll try not to again. xD
20:06
daFred
you have to log in to do bad things on github, and then sebastian will know...
20:07
Rebelj12a
left the channel
20:08
Rebelj12a
joined the channel
20:08
Rebelj12a
Hah I know
20:09
daFred
ohh i thought you deleted yourself
20:11
Rebelj12a
Tried to close the wrong window. Pile that on mobile internet in the boonies is not a good combo
20:15
danieel
the distance of the mount/sensor is not critical, what is critical is the coplanarity of the 2 things... so you do not end up with unfocusable picture (at 1 time only one corner will focus properly)
20:15
danieel
for that you have mechanical adjustment in every pro camera
20:16
Rebelj12a
That's what I thought. I mean if it interrupts form factor then yes, however on the beta with the interchangeable mount enclosure that should not be a problem
20:16
danieel
others (mostly dslrs) solve it at manufacturing calibration, with a 3 point mount and variable sized washers
20:17
Rebelj12a
Pie in the sky idea, moving sensor for tilt shift with fixed lens. /end madness
20:18
danieel
??
20:19
Rebelj12a
Don't ask it probably doesn't make sense ignore my insanity.
20:19
___anton
joined the channel
20:19
Rebelj12a
Its not feasible for a commercial product anyways.
20:19
___anton
changed nick to: _anton__
20:23
_anton__
left the channel
20:24
_anton__
joined the channel
20:24
daFred
pee in the sky idea: we order 450 pcb's shaped as our custom socket and 106650 terminals and make our own sensor sockets.... Including thick copper for cooling...
20:25
Rebelj12a
Don't we want the sensor to be interchangeable?
20:25
Rebelj12a
I mean will Apertus sell parts as well individually?
20:27
Rebelj12a
Don't need thick copper for cooling, liquid cooling just seems cool though. Haha
20:28
daFred
I thought this pcb is socket and cooling area in one part but we have to build a machine to shoot the terminals in ...
20:28
_anton__
daFred: order 450 6mm aluminium core pcbs 11*7cm, have the manufacturer drill 240 holes in each, get 450 peel away sockets - but you still need a 0.7mm thick copper between your cmv12000 and your alum core pcb. That alum core pcb is also your front panel
20:30
danieel
_anton__: how do you prevent the pins from shorting each other? :)
20:30
daFred
this is evolution ...
20:32
_anton__
Instead of a whole 240 terminal peel away socket you can use two 120 terminal peel.away stripes or even bigger number of peel away stripes... maybe smaller stripes will be easier to get... inserting all 240 terminals manually sounds too difficult
20:33
_anton__
danierl: I am talking of metal core pcbs, you can order them, alum is isolated on them and in the vias it is isolated too
20:33
danieel
never seen vias on metal core pcbs
20:33
daFred
http://www.ats.net/wp-content/uploads/2012/10/Aluminiumkern_web_03.jpg
20:35
danieel
plating over elox? hard to tell how likely would be that you cause a short by mechanical force
20:39
_anton__
left the channel
20:45
designbybeck_
joined the channel
20:45
designbybeck__
joined the channel
20:48
Rebelj12a
left the channel
20:48
Rebelj12a
joined the channel
20:48
Rebelj12a
I long for civilization D:
20:57
danieel
left the channel
21:04
subdonline
joined the channel
21:04
Even
left the channel
21:05
subdonline
left the channel
21:06
jucar
left the channel
21:07
jucar
joined the channel
21:09
intracube
left the channel
21:16
intracube
joined the channel
21:17
aombk2
joined the channel
21:18
aombk
left the channel
21:22
aombk2
changed nick to: aombk
21:22
Rebelj12a
left the channel
21:24
Rebelj12a
joined the channel
21:25
derWalter
left the channel
21:26
Rebelj12a
left the channel
21:27
danieel
joined the channel
21:58
Gegsite
joined the channel
22:01
designbybeck__
left the channel
22:01
designbybeck_
left the channel
22:35
daFred
left the channel
22:54
DirdyW
joined the channel
22:56
DirdyW
left the channel
23:02
se6astian
if anyone has 3d printer experience please share it here: https://wiki.apertus.org/index.php?title=3D_Printer_Evaluation
23:10
__anton__
joined the channel
23:10
__anton__
left the channel
23:10
__anton__
joined the channel
23:15
__anton__
Hi, we talked a lot about peel-away-sockets. I looked at this PDF http://www.advanced.com/pdf/AIC_DIP_Sockets_Peel-A-Way_16A_rev0.pdf The terminals shown here do not look suitable to form a CMV12000 socket. CMV12000 requires 1.27mm pitch. However the diameter of the terminals in the thickest part in this datasheet is 1.47 minimum.
23:16
__anton__
If advanced interconnections don't have smaller diameter terminals for peel-away sockets then they can not help us..
23:16
se6astian
changed nick to: se6astian|away
23:25
danieel
__anton__: check also the insertion and removal force. putting 300+ contacts in parallel quite increases the force
23:26
__anton___
joined the channel
23:27
__anton___
danieel: thank you for helpful comments. What is the max insertion force that would be okay for us? CMV12000 datasheet doesn't seem to mention this
23:28
__anton___
Or what is it for the sensor you're working with?
23:28
danieel
i work with kac+andon sockets
23:28
__anton__
left the channel
23:28
danieel
its quite hard to push it in, would prefer next time a softer socket
23:28
liwanma
joined the channel
23:31
__anton___
danieel: I see. BTW I've seen your photo with a hole under the sensor. What do you do with it? What goes into that hole?
23:31
danieel
a peltier, but have not yet made a regulated psu for it (needs low voltage, high amperage)
23:32
__anton___
so a peltier is quite close to the sensor. and on the other side of peltier? a radiator?
23:32
__anton___
which is cooled by internal flow of air inside the camera?
23:32
__anton___
a heat pipe?
23:33
danieel
thought of a heat pipe to place the radiator to a more convenient place, but also an option to put the radiator right on the other end of the peltier... depends on the mechanical design really
23:34
danieel
in a narrow head, it would be just connected to outside metal casing
23:34
Bertl
the andon sockets are quite tight, they are designed for one time insertion I guess
23:35
danieel
need to check the mating cycles... sec
23:36
__anton___
we seem to have a serious development dead end here, at least I do. The sensor is firmly attached to something in order to maintain a fixed distance to lense mount. But we can not
23:36
__anton___
rely on the peltier to be of a calibrated thickness. So it becomes unclear how the peltier can be pressed against the outside metal casing
23:37
Bertl
http://www.advanced.com/products/ic-socket-adapters/pga-sockets-adapters/documentation
23:38
danieel
forces which i personally thing are over the limit: insertion=1.1oz max, withdrawal 0.3oz min (@ 267 pins)
23:38
Bertl
I don't think there is any point to 'press' it to the outside, there are better solutions for heat conduction
23:38
danieel
there was a discussion where to put the peltier
23:39
danieel
my idea is to put it as close to the sensor... to have the minimal cooled volume
23:39
danieel
there is no sense in cooling something more than the chip...
23:40
danieel
__anton___: no need to rely on calibrated thickness, just pust a thermal conductive "paste" there
23:41
Bertl
I don't think it is a good idea to have the peltier directly attached to the sensor for several reasons
23:42
Bertl
first, you need to make it fit perfectly. secondly it will be 'in the way' most of the time
23:42
__anton___
danieel: I have found this document quite interesting. These guys seem to know what they are talking about: http://www.customthermoelectric.com/TECmounting.html
23:44
__anton___
danieel: they advocate the use of a "extender" or "spacer" block of copper/alum between the object being cooled and the Peltier. My understanding of their reasonin is: we're going to have a large hot radiator on the other side of Peltier. Better have it in a centemeter or two away from the object being cooled so that hot air does not carry the heat back
23:45
danieel
read it again, the extender is optiona and shall be of minimal size (not larger than the TEC)
23:46
__anton___
danieel: I do not insist, just though you might feel intereseted. Yes it's optional. My understanding of size it needs to have the same "footprint" as the peltier and be between 0.25" and 0.75" in thickness
23:46
__anton___
danieel: I mean this is how I read their doc.
23:46
danieel
if needed...
23:47
danieel
(in the picture)
23:47
Bertl
the distance/placement is not that relevant to the cooling because of the thermal conductivity differences
23:48
danieel
but placing the extender might help to average the "pwm" operation on the TEC...
23:48
Bertl
do you plan to switch it on a per second basis?
23:49
Bertl
i.e. one second on, one second off?
23:49
__anton___
danieel: I'm a total total compleet noob.. will not PWM cause interference with sensor signalling?
23:49
danieel
not that slow.. but making a fine regulated supply at those voltages is quite a challenge
23:49
danieel
so better to make one fixed voltage and pulse it
23:50
Bertl
you should really look a little into physics
23:50
danieel
which part?
23:51
Bertl
the part about heat transfer and propagation in metals and air
23:51
danieel
where is the bug?
23:51
danieel
any extra volume will act as a capacitor, for heat.. so i do not see why that is a problem
23:52
Bertl
well, first, it's not really that relevant where you put the peltier, because of the big difference between thermal conductivity of air and e.g. aluminum or copper
23:52
danieel
that is not true.
23:52
danieel
it can be true, if the cold end would be totally isolated (vacuum, no touching...)
23:53
Bertl
second, if you use PWM, which might not be the best idea near a sensor, it will be way faster than the heat transfer
23:53
Bertl
so, the metal of the peltier will keep a constant temperature
23:53
Bertl
no need to 'even it out' or so
23:53
danieel
so another reason to not use any extension...
23:54
Bertl
did you read what I wrote?
23:54
danieel
it will even out inside the tec, what else did you wrote?
23:54
Bertl
it doesn't matter if you use an extension or not for the cooling part
23:54
danieel
it does
23:54
Bertl
so be it ... end of discussion
23:55
Bertl
__anton___: you might be right about the peel-away being too large for our purpose
23:55
danieel
you might achieve the goal, yes, but with different efficiency, as you will cool the extra stuff you would not need to cool
23:56
Bertl
http://upload.wikimedia.org/wikipedia/commons/1/1e/Thermal_conductivity.svg
23:57
danieel
what is the y axis for?
23:59
Bertl
no idea, I don't think it has a special meaning