Current Server Time: 23:29 (Central Europe)

#apertus IRC Channel Logs

2020/09/28

Timezone: UTC


01:30
intrac_
left the channel
01:31
intrac_
joined the channel
02:40
Spirit532
left the channel
02:40
Spirit532
joined the channel
03:02
Bertl
off to bed now ... have a good one everyone!
03:02
Bertl
changed nick to: Bertl_zZ
06:17
BAndiT1983|away
changed nick to: BAndiT1983
08:44
mumptai
joined the channel
10:17
aombk2
joined the channel
10:21
aombk
left the channel
10:53
comradekingu
joined the channel
11:13
comradekingu
left the channel
12:24
Bertl_zZ
changed nick to: Bertl
12:24
Bertl
morning folks!
14:00
BAndiT1983
changed nick to: BAndiT1983|away
15:20
comradekingu
joined the channel
15:20
comradekingu
left the channel
15:35
BAndiT1983|away
changed nick to: BAndiT1983
16:59
se6ast1an
MEETING TIME
16:59
se6ast1an
who is here?
17:00
Bertl
is here ...
17:02
se6ast1an
seems it will be another rather short meeting then :)
17:02
se6ast1an
BAndiT1983: you wanna give us a brief update of the stuff you are currently working on?
17:04
se6ast1an
ok superbrief updates from me:
17:04
BAndiT1983
not much to tell, sorry, just resumed investigating the gerber problem
17:04
se6ast1an
preetimenghwanis text for the website article is still in the review/rework process
17:05
se6ast1an
I continue to work on the prober accuracy issues, new textpattern that Bertl created was printed today and will be installed this evening
17:05
se6ast1an
I did also print on mm paper and noticed that the print dimensions are not 100% accurate, will report with actual images once the pattern is installed
17:06
Bertl
so the printer is skewing the image?
17:07
se6ast1an
over the 150mm length the grid missaligns by around 0.5mm or so
17:07
se6ast1an
more details follow with images then
17:07
Bertl
but it is a scaling issue, not some kind of distortion?
17:07
se6ast1an
yes
17:07
Bertl
okay, that's fine then
17:08
se6ast1an
printed at 100% (no scaling) and its like ~100.5% for real
17:09
se6ast1an
contacted more flying probe service providers to get an understanding of what this costs in europe
17:09
se6ast1an
thats it from my side
17:10
se6ast1an
do you also have news for us Bertl?
17:10
Bertl
a few, but not much
17:11
Bertl
as you already mentioned, new test pattern for the prober have been created
17:11
Bertl
I also investigated the cmv_snap3 performance when writing images
17:12
Bertl
when we visited the observatory (vup, anu3jn and I) we had a discussion about the 'long time' it takes to transfer a captured image over ethernet
17:13
Bertl
and that this is probably a problem with the inefficient data conversion (repacking from the memory layout to the output image)
17:13
Bertl
so I did some tests and looked into it a litte further
17:14
Bertl
basically a dummy output of an image sized data block takes 34ms
17:15
Bertl
the current implementation (repacking) takes about 430ms
17:16
Bertl
and transferring the dummy data over network (via ssh) takes about 1.68s
17:17
Bertl
we transfer 18MByte over a 1Gbit connection
17:18
Bertl
assuming we get half that bandwidth (500Mbit) we should be able to transfer the data in about 290ms)
17:18
Bertl
so my conclusion so far is that the overhead of ssh is quite significant here and we probably should consider some alternatives
17:19
Bertl
I also looked into using SIMD instructions (NEON) for the repacking
17:20
Bertl
but given that the current implementation does take about the same time the network transfer would, it is questionable if that will speed things up significantly
17:20
Bertl
but anyway, I'll investigate this a little further
17:20
se6ast1an
sounds good
17:20
Bertl
that's it from my side for this week
17:20
se6ast1an
any news from the ultra96-adapter PCB?
17:20
se6ast1an
or what else are you currently working or or planning to work on?
17:21
Bertl
no PCB work this week
17:25
se6ast1an
or what else are you currently working or or planning to work on?
17:27
Bertl
not sure what the question is aiming at?
17:30
se6ast1an
next steps
17:31
Bertl
well, as soon as the power board PCBs are tested, I plan to assemble them
17:31
Bertl
I also plan to design the ultra96 interface board
17:32
se6ast1an
great
17:32
se6ast1an
the mini dp adapters for the ultra96 are still on hold because of the male minidp connector part not being available right?
17:33
se6ast1an
as SMD/SMT part
17:33
Bertl
I think for a test we will go for a cable there
17:33
Bertl
it is not the best solution, but it is definitely better than hunting for exotic parts
17:34
se6ast1an
so you cut a minidp cable and solder the wires to the test PCB I assume?
17:34
Bertl
nah, that would be very bad
17:34
se6ast1an
would reduce speed but still show if it works in general?
17:35
Bertl
we will put a minidp receptacle on the breakout and use a minidp cable
17:35
se6ast1an
ah, right
17:36
BAndiT1983
changed nick to: BAndiT1983|away
17:37
se6ast1an
ok, anyone else with topics for the meeting?
17:38
se6ast1an
MEETING CONCLUDED THEN!
17:48
BAndiT1983|away
changed nick to: BAndiT1983
18:08
BAndiT1983
changed nick to: BAndiT1983|away
19:50
BAndiT1983|away
changed nick to: BAndiT1983
23:16
BAndiT1983
changed nick to: BAndiT1983|away
23:46
mumptai
left the channel