Current Server Time: 19:39 (Central Europe)

#apertus IRC Channel Logs

2021/03/01

Timezone: UTC


03:06
rektide
left the channel
04:56
Bertl_oO
off to bed now ... have a good one everyone!
04:56
Bertl_oO
changed nick to: Bertl_zZ
06:09
BAndiT1983|away
changed nick to: BAndiT1983
07:07
mumptai
joined the channel
11:02
Bertl_zZ
changed nick to: Bertl
11:02
Bertl
morning folks!
12:24
aombk
left the channel
12:27
aombk
joined the channel
12:52
eppisai
joined the channel
13:22
eppisai
left the channel
14:35
eppisai
joined the channel
17:00
se6ast1an
MEETING TIME! Who is here?
17:01
Bertl
is here ...
17:01
bluez
is here
17:01
metal_dent[m]
is present
17:01
BAndiT1983
here, but nothing to report
17:01
se6ast1an
great, metal_dent[m] do you want to start as usual?
17:02
metal_dent[m]
okay but i don't have much to report
17:03
metal_dent[m]
just a few updates... i was trying to understand an example of the UART communication i found
17:03
metal_dent[m]
then tried to adjust my code accordingly and also according to bandit's suggestions
17:04
metal_dent[m]
i haven't gotten the time to commit it yet though
17:04
metal_dent[m]
that's it from me..
17:05
se6ast1an
many thanks, BAndiT1983 any comments or next steps to add?
17:07
BAndiT1983
not at the momen
17:07
BAndiT1983
*moment
17:07
se6ast1an
ok
17:07
se6ast1an
bluez: your turn!
17:07
bluez
thanks!
17:07
bluez
so last week i was finally able to reconfigure an LUT 'without' fixing it to a particular location
17:08
Bertl
yay!
17:09
bluez
for this i got the slice/lut location through vivado... i then used the database from prjxray, wrote a few scripts, which could get the frame addresses associated with a particular slice/lut location
17:09
bluez
I tested it and it seems to work correctly :)
17:10
Bertl
this is via PCAP frame access or did you generate a bitstream for that?
17:10
bluez
yea via PCAP access
17:10
Bertl
excellent!
17:10
bluez
i generated a partial bitstream, from one of my previous scripts
17:12
bluez
so, now my next task is to find a suitable 'entity' and a suitable manner to update that entity easily through PDR
17:13
bluez
for eg. i need to buffer the previous values of the LUTs so that they dont change with the reconfiguration takes place... and things like that
17:13
bluez
s/with/while/
17:14
bluez
so i'll be working on a test setup with a group of LUTs acting as a 32-bit register... and i'll try to update its value through PDR
17:14
bluez
that's it from me!
17:14
se6ast1an
many thanks! anything to add Bertl (next steps, etc.)?
17:15
Bertl
nah, dynamic register reconfiguration is within reach!
17:15
bluez
yup! :)
17:15
bluez
well..almost
17:15
se6ast1an
great :D
17:15
Bertl
looking forward to replace the many registers we have in the Beta ....
17:16
Bertl
(and thus saving tons of precious resources)
17:17
Bertl
and yes, there is still a lot of work which needs to be done
17:17
Bertl
but it looks very promising
17:17
bluez
indeed :)
17:17
se6ast1an
BAndiT1983 said no news, eppisai also had exams last week and no progress to report
17:17
se6ast1an
quick updates from me:
17:18
se6ast1an
all components for our hardware production run have been sourced, received and handed over today
17:18
se6ast1an
there is just one 90° pin header THT component without ETA
17:18
se6ast1an
but since THT assembly is the second step we stll have time
17:18
se6ast1an
in worst case we hand solder that one later, no big deal
17:19
se6ast1an
Tele still has a few components inbound and said they should arrive by the end of the week or beginning of next week
17:19
se6ast1an
before that happens its hard to tell when we get a slot
17:20
se6ast1an
manfred milled a custom tray from an ESD plastic for one of the last connectors we handed over today
17:20
se6ast1an
so all green light from us there
17:21
Bertl
do we have a plan for depanelization yet?
17:21
se6ast1an
yes, manfred with the cnc mill
17:22
Bertl
yeah, what I meant was, we probably need a bunch of dowel pins for that, no?
17:23
Bertl
(and a base structure to hold them, and probably the PCBs, in place)
17:23
se6ast1an
I thought manfred will mill a plate where the boards are screwed in individually but I will talk to him how he plans it in detail
17:24
se6ast1an
but I am sure he also has a collection of dowel pins if that is the route
17:24
Bertl
excellent, then!
17:24
se6ast1an
I will also talk to him regarding the sensor socket bank insertion depth and tool
17:25
se6ast1an
I asked tele today what they need for THT assembly documentation
17:25
se6ast1an
they said they have the photos and component locations anyway and dont need anything else
17:26
se6ast1an
do we have requirements in that direction, probably a topic for a separate chat/meeting
17:26
Bertl
I think there shouldn't be anything special or unexpected
17:26
se6ast1an
ok, thats it from my side, Bertl, the finisher?
17:27
Bertl
(as long as the parts go on the correct PCB side :)
17:27
se6ast1an
something to make sure the pcie angle is correct?
17:28
Bertl
I spent the weekend (on and off) with building a ZIF frontend and cleaning and testing the parts we got from felix recently
17:29
Bertl
all connectors have to be seated flush and aligned to the board i.e. centered, straight (onot at an angle) etc
17:29
Bertl
but it might help them to get a dual slot plugin for easier registration
17:29
se6ast1an
noted
17:30
Bertl
on the 'power' connectors (i.e. the 2x2 short and long pin headers)
17:31
Bertl
it helps to have a matching power board PCB (doesn't need any receptacles to work as there are holes)
17:32
Bertl
so maybe also provide them with an empty power board PCB for this purpose
17:33
se6ast1an
ok
17:33
Bertl
back to the ZIF SFE, I haven't been able to test it yet, but this should happen in the next few hours
17:33
Bertl
currently adjusting the power board trimmers for the setup
17:33
Bertl
that's it from my side for this week
17:34
se6ast1an
great thanks
17:34
se6ast1an
anyone else who wants to share/report/ask?
17:36
se6ast1an
alright then, meeting concluded! many thanks all participants!
17:37
Bertl
thanks for the moderation!
17:38
se6ast1an
my pleasure as always
17:45
Bertl
off for now ... bbl
17:46
Bertl
changed nick to: Bertl_oO
18:42
eppisai
left the channel
19:07
gcolburn
joined the channel
19:09
gcolburn
Hello. After being inactive for a while on the axiom-beta I wanted to try to find some time for it. I flashed the latest firmware on an SD card but having an issue where the system just keeps rebooting non-stop and not providing the user login prompt. Is anyone around that might be able to help? Thanks!
19:12
BAndiT1983
hi gcolburn, probably vup or anuejn_ can help there
19:12
BAndiT1983
which version have you used? 2.0?
19:13
gcolburn
I would presume so. I tried the two latest here: https://github.com/apertus-open-source-cinema/axiom-firmware/releases
19:13
BAndiT1983
ok, they should fit, so the guys will maybe know what's going on
19:14
gcolburn
The last image I had was from 2019. It booted fine but was running into wireless driver issues. It seemed like updating to the latest build with possibly newer kernel/drivers might fix it.
19:14
vup
how are you connecting? Are you able to provide a boot log? How are you powering the camera? It sounds a bit like maybe your powersupply is not providing enough power and powering up the sensor resets the camera...
19:15
gcolburn
I have the power connected to USB to my laptop. I could try connecting it to an outlet USB connection and see if that fixes it
19:15
vup
how are you powering it over USB?
19:15
vup
do you have a custom cable?
19:16
vup
But yeah thats probably it
19:17
gcolburn
Yeah. Custom cable with a DC to USB connector
19:17
gcolburn
Came with the camera
19:18
vup
Automatically starting up the aquisition pipeline seems to be causing more problems for people than not, maybe we should disable it again? Thoughts Bertl_oO, se6ast1an, anuejn_?
19:19
vup
gcolburn: I see, so yeah your laptop is probably not providing enough power to power up the sensor, which is now done automatically on boot of the camera. You can either try a beefier power supply (~15 Watts should work), or (atleast if you are using a OS / have a setup that can access a ext4 partition) disable the automatic startup of the aquisition pipeline.
19:19
se6ast1an
yes, I was also about to ask if its a crash caused by kick/start script being run automatically
19:20
se6ast1an
hi gabe, welcome back :D
19:20
se6ast1an
there might be minor changes required in the FW 2.0 image depending on specific hardware versions of your boards
19:21
se6ast1an
Bertl_oO would be able to tell that
19:21
gcolburn
It seems to restart around after it reports started User Login management or Network Management
19:21
gcolburn
No obvious errors
19:21
se6ast1an
how do you power it currently?
19:22
eppisai
joined the channel
19:22
vup
To disable the automatic startup, mount the second partition on the sdcard, under `etc/systemd/system/multi-user.target.wants` there should be a link called `axiom_start.service`. Remove that. (Make sure you are looking at the etc folder on the sdcard and not the one of your own system)
19:22
se6ast1an
vup is right, currently it sounds like a power issue
19:22
gcolburn
For convenience I plugged it into a USB port on my mac. Its actually through a USB C adapter
19:23
se6ast1an
the beta draws up to 20-30W, no laptop can supply that
19:23
gcolburn
Gotcha. Let me switch power source :)
19:23
se6ast1an
also the cable has two USB connectors right?
19:23
gcolburn
It was at least convenient while I tried to get wifi working
19:23
gcolburn
Yes it does. It also has an external DC power in
19:23
gcolburn
But it may not supply enough even with that
19:23
se6ast1an
the two usb connectors power different things in the beta
19:24
gcolburn
Oh. I thought you were referring to on my USB hub
19:24
gcolburn
The cable only has one USB connector for power
19:24
se6ast1an
hmmm
19:24
se6ast1an
we have no usb power cable with one connector...
19:24
gcolburn
I connect a separate USB cable for the UART
19:24
se6ast1an
can you share a picture?
19:25
gcolburn
Sure. Send it to your emaili?
19:25
se6ast1an
https://paste.pics/ ?
19:26
se6ast1an
here is the connector illustration: https://wiki.apertus.org/images/7/70/Side-01.jpg
19:27
gcolburn
There is a power cable that goes to the power adapter (v0.2) R1.1
19:28
se6ast1an
yes, so you use the dedicated power supply?
19:29
gcolburn
Yeah. I'm not powering the camera through the standard USB/UART port (not the JTAG)
19:31
se6ast1an
ah right, all good then
19:31
se6ast1an
then please follow vups advice
19:31
gcolburn
https://paste.pics/BQTLM
19:31
se6ast1an
uart to mac should be fine
19:32
se6ast1an
regarding wifi support
19:32
gcolburn
Yeah it seemed to connect no problems :)
19:32
se6ast1an
https://wiki.apertus.org/index.php/AXIOM_Beta/Manual#WiFi_Access_Point_Setup
19:32
se6ast1an
we collected a few sticks that work
19:32
gcolburn
Yeah. I've had the Edimax for a while, which I've occasionally been able to get working, but it has been finicky
19:37
se6ast1an
off for a bit, vup would you be so kind and assist gabe if any further steps need to be taken
19:39
gcolburn
I made it further to the login prompt. Then it reboots. I'll try another power supply
19:42
vup
yeah that all sounds very much like a power supply issue.
19:42
vup
I threw up a commit that disables the automatic start of the sensor and stuff, however it will take a moment for the CI to run through to get you a new sdcard image
19:46
gcolburn
Thanks. I could see arguments either way about the defaults on that.,
19:48
vup
Yeah, I guess as long as this is more of a developer image, not starting it automatically is probably fine and what most people actually want
19:55
gcolburn
I've tried connecting the power to an outlet with a USB adapter, as well as a 12V power bank and still get a restart around when the login prompt shows up
19:55
gcolburn
I'll try the new image
20:02
eppisai
left the channel
20:14
eppisai
joined the channel
21:33
rektide
joined the channel
21:34
vup
gcolburn: the new image it ready by now: https://github.com/apertus-open-source-cinema/axiom-firmware/releases/download/nightly%2Fc26ca6c4/axiom-beta-c26ca6c4.img.xz
21:35
gcolburn
Thanks. Already flashed it and just booted it up. I'm able to log in now. Thanks!
21:35
gcolburn
The wifi seems to be hosting in AP mode already with no configuration on my part
21:37
Bertl_oO
hey gcolburn! how's going? LTNS!
21:38
gcolburn
Hey Bertl! Yeah I know its been a while! Life has been very busy. How have you been?
21:41
Bertl_oO
considering covid-19 and everything going on nowadays, I'd say I had it quite good so far ...
21:42
gcolburn
That's good!
21:43
vup
gcolburn: nice, so that makes me 99% sure that your powersupply cannot handle running the sensor
21:43
gcolburn
What are the current main projects on-going with the axiom-beta?
21:43
Bertl_oO
raw data transfer out of the camera
21:43
Bertl_oO
dynamic system reconfiguration
21:44
Bertl_oO
gateware rewrite in nMigen
21:44
Bertl_oO
just to name a few :)
21:44
gcolburn
what are the changes being worked on the raw data transfer?
21:44
BAndiT1983
changed nick to: BAndiT1983|away
21:45
Bertl_oO
mostly new hardware solutions based on USB3 and SATA
21:45
Bertl_oO
(plus the inevitably required gateware and software of course)
21:45
gcolburn
vup: what is the recommended power supply? Also, once logged in should I manually activate the service you disabled, or use the axiom_start.sh script?
21:46
gcolburn
Bertl: Is the goal easier interface compatibility, or speed?
21:46
Bertl_oO
do I see that correctly from your image link that you are trying to power the Beta from a single USB port?
21:47
Bertl_oO
gcolburn: mainly bandwidth ... we need to move a lot of data ...
21:47
gcolburn
Gotcha
21:47
Bertl_oO
so USB3 is mainly for PC/Laptop connections, SATA for direct storage
21:47
vup
gcolburn: I have been using various usb-c power bricks with variyng success. USB PD ones almost never work, but the "dumb" ones that just give 3A have been working fine so far. Usually the developer kit includes a suitable power brick though.
21:48
Bertl_oO
gcolburn: we currently use a power supply which is rated 6A at 5V and provides a little more than 5V but roughly within USB power limits
21:49
vup
Enabeling the service will probably make the camera reboot again. But of course leaving it stopped means you can not do anything with the sensor. Currently doing something like `systemctl start axiom_start` and `axiom_start.sh` do the same, but in the future, `systemctl start axiom_start` is probably the better choice
21:49
Bertl_oO
(that is the power brick which usually comes with the Beta)
21:49
vup
*Enabling
21:49
gcolburn
Yesterday I had the camera connected as follows: Mac -> USB-C to USB HUB -> 2 connections to Axiom beta (power adapter with USB + USB/UART). That worked fine but I had problems with the wifi drivers. The last image I had on the beta was over a year old. I flashed the latest version, and it kept rebooting. Today I tried connecting the power adapter
21:49
gcolburn
to a wall outlet with USB power port, as well as a battery and it made it further before rebooting
21:49
Bertl_oO
you will effectively require peak power in the range of 2.5-3A
21:50
vup
gcolburn: were you using the sensor with the old setup?
21:51
Bertl_oO
what we do in the dual USB power connector se6ast1an mentioned is to power each of the two main power rails (Microzed and Beta) separately from one USB power power connector/adapter
21:51
vup
(was the blue led on the beta blinking)
21:51
gcolburn
I have a red led blinking
21:51
Bertl_oO
heartbeat I presume
21:52
Bertl_oO
this means the Linux is running
21:53
Bertl_oO
when you start the image pipeline, the blue led should start flashing (or at least glowing) as well
21:53
Bertl_oO
the green one will be always on as long as there is power
21:54
gcolburn
Do you know if power usage has gone up with newer firmware? I didn't have these issues in the past and even ran it in the field off a battery (using the same battery right now and it is rebooting)
21:54
gcolburn
Could be the battery is getting old though
21:54
Bertl_oO
power usage might have gone up because more stuff is running
21:55
Bertl_oO
also the wifi stick directly draws (usually a lot of) power from the Beta rail
21:55
Bertl_oO
do you still have your original power brick that came with the Beta?
21:56
gcolburn
Yeah I just need to find it. It wasn't in the case with the camera. I'll go find it :)
21:56
Bertl_oO
if so, it should be easy to verify whether the power supply is to blame or not
21:56
vup
Maybe, but I think the more stuff running is negligible. Maybe booting + starting up the sensor at once has increased the peak power consumption a bit
21:56
gcolburn
Looks like my wall USB only puts out 1Amp!
21:57
vup
1A is cutting it close, I think it got it working on 1.5A one before, but that was very fragile
21:58
Bertl_oO
1.5A is pure luck :)
21:59
gcolburn
haha. I'll make sure to get it more power
21:59
gcolburn
Are there any scripts for reporting power usage?
22:02
Bertl_oO
yes, the pac1720_info has that
22:02
Bertl_oO
but it doesn't report peak usage, just some average
22:04
Bertl_oO
but usually you can see power problems in the voltages reported there ... when they are very low compared to the nominal 5V, then it is likely that the power supply cannot keep up
22:09
gcolburn
ok thanks
22:10
gcolburn
do you remember the brand of the power adapter? I think I remember what it looks like, but not the brand
22:12
Bertl_oO
it should be an YU0506, black brick, blue LED, white text on the back with a yellow frame around it
22:12
gcolburn
thanks
22:12
Bertl_oO
https://shop.apertus.org/image/cache/catalog/dc-power-supply-800x400.jpg
22:13
Bertl_oO
https://www.apertus.org/sites/default/files/images/AXIOM_Beta_Developer_Kit_Package_Contents_2018.jpg
22:22
Bertl_oO
in case you can't find it, it doesn't need to be this specific power supply, but you should look for one which can supply at least 5A (and 5V of course) and has the proper polarity (shold be on the tiny power adapter PCB :)
22:24
Bertl_oO
https://wiki.apertus.org/images/c/c6/00-PA-001-_AXIOM_Beta_Power_Adapter_Board_V0.2R1.1_Show_sm.png
22:24
Bertl_oO
(it was conveniently photographed from the side where you can't see the information :)
22:27
Bertl_oO
http://vserver.13thfloor.at/Stuff/AXIOM/BETA/power-adapter-02.jpg
22:27
Bertl_oO
here you go :)
22:28
Bertl_oO
so 5V/5A plus in the center, minus on the outside
22:38
gcolburn
Thanks!
22:40
Bertl_oO
np
22:49
gcolburn
found the power supply and its up and running now :)
22:53
gcolburn
The Axiom Beta overlay on the HDMI is new since I last used it!
22:53
Bertl_oO
yes, it is! :)
22:54
gcolburn
What's the status of putting an IR filter in front of the sensor and not on the lens? I remember seeing a video evaluating some sample filters
22:54
Bertl_oO
this will be part of the 'enclosure'
22:55
Bertl_oO
i.e. it has a filter mount where you can put filters directly in front of the sensor
22:55
gcolburn
Gotcha
22:55
Bertl_oO
https://www.apertus.org/axiom-team-talk-15-4-optical-filters-firmware-2-article-june-2020
22:56
Bertl_oO
if you scroll down a little, you can see the filter mount
22:56
Bertl_oO
(or filter holder as it is called there)
22:57
gcolburn
Great. Yeah that's the video I saw
22:58
gcolburn
What's the current status of the compact enclosure?
23:00
Bertl_oO
prototypes are working so far and it shouldn't be too long till it is actually available
23:00
gcolburn
Great
23:01
Bertl_oO
there are still some open points and we need to test it with the new power board which was adapted for the enclosure
23:01
Bertl_oO
(mainly to allow higher input voltages)
23:01
gcolburn
gotcha
23:08
eppisai
left the channel
23:11
mumptai
left the channel
23:28
gcolburn
I thought of a new application for the axiom beta recently... high resolution film scanning by stitching multiple images behind an enlarging lens
23:31
Bertl_oO
yeah, why not
23:32
gcolburn
I have large format slide film to scan and they don't make new scanners that can extract all of the detail out of it. The standard in quality is a drum scan but those are expensive and the machines are hard to find parts for and maintain. I think this could work well. I've got the HDMI feed from the back of a large format camera focused on the film
23:32
gcolburn
on a light box
23:33
gcolburn
I should be able to do some sample stitching with manual translations, but Ideally you would have a 2 axis (possibly 3 for focus) mount for the camera to automate it
23:34
Bertl_oO
there are nice tools for the stitching ...
23:35
Bertl_oO
http://hugin.sourceforge.net/
23:35
gcolburn
Definitely. and stitching behind the lens would be trivial since you don't have to correct the lens distortion
23:35
Bertl_oO
yes, scanning the film directly has a lot of advantages
23:35
gcolburn
Looks like I can resolve the grain of the film, which isn't possible with an Epson flatbed
23:37
gcolburn
Some people are trying to use a DSLR to digitize film, but its hard to do well. This would work much better.
23:38
gcolburn
On 4x5 film the resolution would easily be hundreds of megapixels. The rough mental math was on the order of 500-600. I'd have to do more rigorous calculations, but that was ball park assuming 12MP/image determining the # of images required for an APS-C to sample a 4"x5" area
23:40
Bertl_oO
sounds like a nice project
23:42
gcolburn
Yeah. I've have stepper motors and a controller but need a good 2 axis linear slide setup
23:43
Bertl_oO
you neither need precision nor repeatability for that
23:44
Bertl_oO
so personally I would put the camera on rails and just move it around
23:45
Bertl_oO
if you have single slides (not a film) then in both directions, otherwise, only up and down and transport the film
23:46
Bertl_oO
as long as you make sure that there is some reliable overlap, you should be fine
23:49
gcolburn
I'll definitely start out that way, but eventually want it automated. The front standard of the camera can be moved up/down/left/right manually
23:50
gcolburn
But as you mentioned it doesn't need to be very precise, so cheaper parts should work well