elms has quit [Read error: Connection reset by peer]
elms has joined #scopehal
gruetze_ has joined #scopehal
gruetzkopf has quit [*.net *.split]
azonenberg has quit [*.net *.split]
<_whitenotifier-f>
[scopehal] azonenberg opened issue #277: Change base frequency unit from Hz to mHz or uHz - https://git.io/JUoU4
<_whitenotifier-f>
[scopehal] azonenberg labeled issue #277: Change base frequency unit from Hz to mHz or uHz - https://git.io/JUoU4
<_whitenotifier-f>
[scopehal] azonenberg opened issue #278: Design APIs for spectrum analyzers (can we just treat it as an oscilloscope where X axis unit is Hz? separate controls for RBW etc?) - https://git.io/JUoUo
<_whitenotifier-f>
[scopehal] azonenberg labeled issue #278: Design APIs for spectrum analyzers (can we just treat it as an oscilloscope where X axis unit is Hz? separate controls for RBW etc?) - https://git.io/JUoUo
Degi has quit [Ping timeout: 264 seconds]
Degi has joined #scopehal
azonenberg has joined #scopehal
<azonenberg>
v1.2 probes are ETA tuesday
<azonenberg>
guess there were some delays, i was hoping for shipping thursday and arriving monday but they didnt ship until very late friday
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on pull request #198: Replace `system()` usage in SerializeWaveforms - https://git.io/JUoO1
<azonenberg>
ooh it looks like the tek mso6 has per channel thresholds?
<azonenberg>
for digital
Degi has quit [Ping timeout: 256 seconds]
<_whitenotifier-f>
[scopehal] azonenberg pushed 2 commits to master [+0/-0/±3] https://git.io/JUoCg
<_whitenotifier-f>
[scopehal] azonenberg 41ff1e0 - TektronixOscilloscope: create digital channels for each flex channel. Automatically disabled if no LA probe is connected, but no proper on/off support yet. See #264.
<_whitenotifier-f>
[scopehal] azonenberg 583b909 - TektronixOscilloscope: can turn MSO5/6 LA channels on/off. See #264.
Degi has joined #scopehal
electronic_eel has quit [Ping timeout: 260 seconds]
electronic_eel has joined #scopehal
_whitelogger has joined #scopehal
_whitelogger has joined #scopehal
_whitelogger has joined #scopehal
gruetze_ is now known as gruetzkopf
<bvernoux>
azonenberg, do you have received the v1.2 probes ?
<bvernoux>
Doing it on KiCad and improving different stuff like freq from 50KHz (or less) to 8GHz with better FPGA ECP5 with Ethernet better RF components ...
<bvernoux>
What is nice is this VNA size is only 10cmx10cm ultra small and passive !!
<bvernoux>
to replace my old HP 8753D ;)
<bvernoux>
But there is some challenge to reach 100dB dynamic range from <1MHz up to 8GHz with a price not exceeding 500USD ;)
<bvernoux>
(including the amazing aluminum shielding)
<bvernoux>
The LTC5510 could be replaced by LTC5562
<bvernoux>
to be checked as the isolation of LO/RF of LTC5562 seems bad ...
<bvernoux>
or LO/IF...
katharina has joined #scopehal
<_whitenotifier-f>
[scopehal-apps] nshcat commented on pull request #198: Replace `system()` usage in SerializeWaveforms - https://git.io/JUK87
<bvernoux>
katharina, nice to see you I hope you are better ?
<katharina>
bvernoux: Hi! Yes, much better. Good to be back
<bvernoux>
I'm trying also to let work glscope on Windoz ;)
<bvernoux>
but it freeze so far ;)
<bvernoux>
I was searching a way to debug it with a GUI like VSCode but it is not that easy to use MSYS2 MinGW64 GDB with it ...
<bvernoux>
it works with cli + gdb but it is really awful ;)
<bvernoux>
katharina, if you can share what is your setup to debug it on windoz you are welcome
<katharina>
bvernoux: I am used to just use GDB in the terminal haha, I am sorry to disappoint
<katharina>
i come from a linux background am just used to do it this way
<katharina>
but the freeze is interesting - it must be a recent change, since it did work before i left. would be interesting to find out what goes wrong here
<bvernoux>
azonenberg, has shared some project to test
<bvernoux>
(some scopesession)
<katharina>
ill look into the logs here to find it in a minute
<katharina>
does it crash (like segfault etc) or actually hang?
<bvernoux>
no it hang with full cpu
<katharina>
oh thats unusual
<bvernoux>
25% CPU (on a quad code)
<bvernoux>
so GUI is just frozen
<katharina>
sounds like a infinite loop to me
<bvernoux>
yes
<bvernoux>
I have such effect also with just loading it without any scope
<bvernoux>
Anyway there is tons of wrong things related to u64
<bvernoux>
%lld is not working on Windows it is not C compliant
<bvernoux>
just supported by Linux/GCC
<bvernoux>
even if I doubt it is related to that ;)
<katharina>
do you have gdb installed right now?
<bvernoux>
yes
<katharina>
want me to give you the commands to find out where its stuck? its not complicated
<bvernoux>
I have an uptodate MSYS2 MINGW64 install with GCC 10.2
<bvernoux>
yes we can try
<bvernoux>
I have created a full directory with all dll/exe
<bvernoux>
I just need to launch it with gdb and use glscopeclient spi-compressed.scopesession
<bvernoux>
as exaple
<bvernoux>
it shall decode quadSPI IIRC
<katharina>
great. `gdb ./glscopeclient.exe` and then in the prompt do `set args spi-compressed.scopesession`
<bvernoux>
are you using also
<bvernoux>
export OMP_WAIT_POLICY=PASSIVE
<bvernoux>
as Andrew told me it is mandatory
<bvernoux>
I use it
<katharina>
then you can go `run`, let it run for a bit and do Ctrl-C. it should drop you into gdb's prompt, and a `where` should show you where the loop is
<bvernoux>
I have GDB 9.2
<bvernoux>
but it is latest version with MSYS2 MinGW64
<bvernoux>
it is the main issue ;)
<bvernoux>
as Ctrl C do nothing ;)
<katharina>
ouff
<bvernoux>
I tried that
<bvernoux>
it is why it is awfull ;)
<bvernoux>
let's try again
<katharina>
it doesnt do anything in the terminal? thats odd as fuck
<bvernoux>
woo I have new things displayed now
<bvernoux>
that my GX card does not support GL_ARB_gpu_shader ;)
<bvernoux>
let's run again MINGW64 cmd with my NVIDIA GPU ;)
<katharina>
Ah, laptop user?
<bvernoux>
yes
<katharina>
i had the same issue months ago
<katharina>
it sucks
<bvernoux>
but I have a GeforceGT 650M
<bvernoux>
I have laucnhed it with it now ;)
<bvernoux>
haha what a joke
<bvernoux>
same error
<bvernoux>
I'm pretty sure my GeforceGT 650M support Open GL4.x or more
<bvernoux>
but it is not detected correctly ..
<bvernoux>
same error with GL_ARB_gpu_shader_int64 extension not supported
<katharina>
is it a hard error?
<katharina>
or just a warning
<bvernoux>
even when launching "MSYS2 MinGW 64-bit" with NVIDIA processor ...
<bvernoux>
it is a hard error
<katharina>
mh we have to look into that, might be something to consider working around?
<katharina>
i shall test this on my thinkpad
<bvernoux>
so we go further than before with infinite loop ;)
<katharina>
i wonder if this is GDB fault
<katharina>
like the OpenGL error
<bvernoux>
I was trying to debug it with VM too with Ubuntu but there is no any VM supporting something > than OpenGL 3.x so it is a dead end
<bvernoux>
as it requires OpenGL 4.2 IIRC
<bvernoux>
it will be interesting if you test same file on your laptop
<tnt>
you have to go back to 344.x to only have 4.4 AFAICT.
<bvernoux>
ha yes my current version of NVIDIA Drivers is 344.75 ;)
<bvernoux>
ok so it should work ;)
<tnt>
OpenGL 4.5 release data August 2014. NVidia 344 drivers release date September 2014 ... so yeah, they didn't have the time to implement the brand new standard :p
<bvernoux>
yes and because I nevere do any game or stuff like that I have never updated them ;)
<bvernoux>
I'm always use the HD4000 in fact by default for all things even KiCad ;)
<bvernoux>
as the GT 650M is very hot
<bvernoux>
for nothing ;)
bvernoux has quit [Read error: Connection reset by peer]
bvernoux has joined #scopehal
<bvernoux>
re
<bvernoux>
so very good news all work fine now ;)
<monochroma>
yay
<bvernoux>
the issue appears when we load glscopeclient.exe without parameter
<bvernoux>
it seems the check on OpenGL stuff is not done in all case
<bvernoux>
as my issue was towards old NVIDIA Driver ...
<bvernoux>
I confirm actual code in git works on WIndoz ;)
<bvernoux>
at least scope session
<bvernoux>
need to test with my old Rigol but IIRC USB support is not merged so far
<bvernoux>
I'm curious to see how the %lld stuff are working on mingw64 ;)
<bvernoux>
maybe they are interpreted correctly as mingw64 use GCC behind ...
<bvernoux>
but I have some warning which say they are not ;)
<bvernoux>
I have reproduced the freeze
<bvernoux>
it appears when loading glscopeclient without any parameter
<bvernoux>
and choose default and go
<bvernoux>
then I try Ctrl-C in GDB it just kill glscopeclient ;)
<azonenberg>
bvernoux: the v1.2 probes are still on the way eta tuesday
<azonenberg>
File a bug for this, about 100% cpu hang on windows when run with no args
<azonenberg>
it SHOULD pop up a dialog asking you for a scope to connect to
<azonenberg>
you and katharina will have to debug
<bvernoux>
yes it ask for a scope
<bvernoux>
but after that i choose rigol null
<bvernoux>
and freeze
<bvernoux>
maybe it is because it wait for data from pipe ?
<bvernoux>
do you have same issue on native Linux ?
<bvernoux>
azonenberg, I know how to debug and break it now
<azonenberg>
You specified null transport?
<bvernoux>
I will write a readme for other Windoz user ;)
<azonenberg>
Oh
<bvernoux>
yes null transport for test
<azonenberg>
That's not quite technically a bug then :p
<bvernoux>
I was thinking it will display an error and I could see the main windows ;)
<bvernoux>
just to see options and potentially load a session...
<azonenberg>
The null transport is intended to be used for non-scpi instruments that just need a transport because of the object model. for example when loading a file from disk or something
<azonenberg>
it ignores all writes and never returns any read data
<bvernoux>
so you have same issue with it ?
<azonenberg>
So what's probably happening is the rigol driver sent a "get trigger" request to it
<azonenberg>
And is then busy-polling trying to get a trigger status message back
<azonenberg>
Which never comes because it's reading /dev/null
<bvernoux>
yes it is what I think too
<azonenberg>
The null driver should not show up in the dropdown, i think
<bvernoux>
so not a bug but very strange ;)
<azonenberg>
that is probably the real bug
<azonenberg>
it's only for internal use with non scpi devices
<azonenberg>
the null transport*
<azonenberg>
There is currently no supported way to launch glscopeclient without either connecting to a scope right away, or specifying a file to open on the command line
<azonenberg>
scopehal-apps:#118 and katharina's startup-dialog branch are intended to fix this
<azonenberg>
But it's not donwe
<bvernoux>
ok so there is no hurry on that ;)
<bvernoux>
I was thinking when launching it like that it will provide main windows to load session
<azonenberg>
Yeah. It will
<azonenberg>
But not yet
<azonenberg>
if you specify the scopesession on the command line it works though?
<bvernoux>
I confirm that when launched with your session it works perfectly
<bvernoux>
so it is a very good point ;)
<_whitenotifier-f>
[scopehal-apps] azonenberg opened issue #199: NULL transport should not be displayed in startup dialog because a lot of drivers act weird with it - https://git.io/JUKxY
<_whitenotifier-f>
[scopehal-apps] azonenberg labeled issue #199: NULL transport should not be displayed in startup dialog because a lot of drivers act weird with it - https://git.io/JUKxY
<bvernoux>
also something strange
<bvernoux>
is the scroll ;)
<bvernoux>
it is required to click on bar with ps
<_whitenotifier-f>
[scopehal] azonenberg opened issue #279: Add optional mode to SCPITransport reads to terminate a read with either a semicolon, new line, or both - https://git.io/JUKxi
<bvernoux>
to scroll left/right
<_whitenotifier-f>
[scopehal] azonenberg labeled issue #279: Add optional mode to SCPITransport reads to terminate a read with either a semicolon, new line, or both - https://git.io/JUKxi
<_whitenotifier-f>
[scopehal] azonenberg labeled issue #279: Add optional mode to SCPITransport reads to terminate a read with either a semicolon, new line, or both - https://git.io/JUKxi
<bvernoux>
it is not very natural
<azonenberg>
bvernoux: there are two ways to move the timeline right now
<bvernoux>
i was trying to do that on center window
<azonenberg>
Click and drag the timeline
<azonenberg>
or use the mouse wheel to zoom on the main plot area
<azonenberg>
in the second case it should zoom to wherever the cursor is
<azonenberg>
I plan to also support a modifier key, maybe shift, plus scroll wheel to scroll horizontally without zooming
<bvernoux>
it will be more intuitive to drag center window and not the timeline ;)
<azonenberg>
I don't want to support that because i need to be able to click and drag labels and stuff in the plot area eventually
<bvernoux>
but maybe for future use case you will use click on main window to zoom ?
<bvernoux>
ha yes ok ;)
<azonenberg>
Like fft peaks
<bvernoux>
yes very nice the fft peaks ;)
<azonenberg>
right now those are auto layout only, but i want to be able to move them eventually
<bvernoux>
i tried them
<bvernoux>
with the 10gb session ;)
<bvernoux>
a remark the menu font are very thin
<bvernoux>
could be bigger ;)
<bvernoux>
the decoder is very impressive with zoom in/out
<bvernoux>
A nice feature will be to compress the view when all signal do not change ;)
<bvernoux>
as there is nothing interesting and display it like ...
<azonenberg>
I have a lot of things i want to improve. Menu font is the default GTK font for your system, i dont think i changed that
<azonenberg>
Fonts in the normal display area will be changeable via preferences once ui-dev is merged to mastere
<azonenberg>
master*
<bvernoux>
yes it is not too important so far
<bvernoux>
I'm impatient to try it in live with my old Rigol scope
<bvernoux>
but my DS1102E have only USB no Ethernet ...
<bvernoux>
it is WIP IIRC to have USB for Rigol
<azonenberg>
I believe USBTMC is supported on linux via /dev/usbtmc*
<azonenberg>
But i don't think we have windows usb support yet
<bvernoux>
ha yes it is the famous USBTMC mode
<bvernoux>
shall be done over libusb to be compatible with multiplatform ;)à
<azonenberg>
to my knowledge all windows testing has been over ethernet
<azonenberg>
or just working with saved files
<azonenberg>
Anyway, i'm gonna start working on Tek MSO6 logic analyzer support
<bvernoux>
anyway saved files/session is still very nice to play with it
<bvernoux>
I will push Tek to sell my the Tek6 ;)
<bvernoux>
now I want the MSO64B ;)
<azonenberg>
Lol
<bvernoux>
it is better with 50GS/s ;)
<bvernoux>
I have project with USB 3.0 and some GIgabitETh which shall arrive in 2021 ;)
<azonenberg>
btw not sure if you saw
<azonenberg>
But i'm in talks with lecroy's secondhand sales guy to get a 4 GHz active differential probe for my waverunner next month
<bvernoux>
In the meantime I could also buy a cheap MSO5000 for 1KEuros ;)
<bvernoux>
ha great
<bvernoux>
I imagine it is quite expensive
<bvernoux>
will be interesting to see a teardown ;)
<azonenberg>
Used, it's a bit under 4 kUSD. And i'm not taking it apart lol
<bvernoux>
But I suspect a teardown on such small probe will be destructive ;)
<azonenberg>
Yeah
<bvernoux>
will be interesting to compare it with passive probe
<azonenberg>
yes i will definitely put it on the VNA
<azonenberg>
and compare S11 vs my probe
<bvernoux>
yes to see what it brings vs a "cheaper" passive probe ;)
<bvernoux>
IIRC in old scope such active probe was just populated with a PIC16F or something like that to provide data to the scope ;)
<bvernoux>
with few passive electronic components ;)
<azonenberg>
bvernoux: So, the WaveLink system is a couple of parts
<azonenberg>
each of which come in a couple of variants
<azonenberg>
seems like this is how most active probe systems are built
<azonenberg>
The first is an adapter cable (WL-PBUS2 or WL-PLINK) that goes from the scope's native input connector to some kind of custom circular connector that provides 50 ohm single ended RF, +/- 12V power, and I2C
<azonenberg>
As far as i know there's no active components in that
<azonenberg>
The second component is a differential amplifier module that goes from this circular connector to some kind of probe connector
<azonenberg>
And the third is a solder in or needle based probe tip
<azonenberg>
They also make a version which i am getting (because it's what's available right now) that combines those two into one
<azonenberg>
it has two tip needles, a thumbscrew to adjust the spacing between them, and an amplifier in one handpiece that plugs into the WL-PBUS2/WL-PLINK cable
<azonenberg>
It's cheaper but cannot be used with other tip styles
<bvernoux>
ha yes very interessting
<azonenberg>
So as far as i can tell, the WaveLink tip modules themselves are basically a transmission line probe
<bvernoux>
I was thinking also there is something like a diff amplifier ...
<bvernoux>
as the aim is to amplify a bit the signal ...
<azonenberg>
Yeah. that's what goes between the cable and the probe tip
<bvernoux>
yes to compensate for the cable loss ... I imagine
<azonenberg>
it looks like a blue or orange marker, you can see it in lecroy press materials
<azonenberg>
As far as the tips, they have either two custom binned carbon resistors you solder to the DUT, or something similar behind tip needles
<azonenberg>
and then, as far as i can tell, just a flex PCB inside a protective rubber sleeve going to the amplifier module
<bvernoux>
yes quite interesting for something which cost more than 4KUSD ;)
<bvernoux>
especially the 4GHz or more
<bvernoux>
they cost >10KUSD
<azonenberg>
Yeah i'm only getting this price because it's used
<azonenberg>
I want to build my own active diff probe at some point
<azonenberg>
My plan is to design the AKL-PT2 first (solder in version of the AKL-PT1)
<azonenberg>
Then make a differential tip based on it
<azonenberg>
Then design an amplifier module that the tip mates to
<_whitenotifier-f>
[scopehal-apps] azonenberg pushed 7 commits to master [+4/-0/±10] https://git.io/JU6nr
<_whitenotifier-f>
[scopehal-apps] nshcat c48bd36 - Linux implementation of rm -rf scope_*
<_whitenotifier-f>
[scopehal-apps] nshcat 6419085 - Some smaller fixes
<_whitenotifier-f>
[scopehal-apps] nshcat c9af2ca - Windows implementation for rm -rf scope_*
<_whitenotifier-f>
[scopehal-apps] ... and 4 more commits.
<azonenberg>
electronic_eel: correct
<azonenberg>
annoyingly, channels use one-based indexing CH1...CH4 or CH1...CH8
<electronic_eel>
hehe
<azonenberg>
But the digital lanes on a flex probe are numbered 0 to 7
<azonenberg>
anyway, at this point i still have to support the spectrum view mode, but that isn't a super critical priority. it will also take some time because i have to think about how to represent controls like RBW and window function in the current API and GUI
<azonenberg>
and there is no support for the AWG or DVM options
<electronic_eel>
do they have color coding for the digital interfaces or something else than the D0 to D7?
<azonenberg>
They do have colors. That's a pending TODO
<electronic_eel>
...like lcd panels?
<azonenberg>
it's just rainbow though
<azonenberg>
i dont think there is any way to distinguish CH1_D0 from CH2_D0 visually
<azonenberg>
No LCD
<azonenberg>
They do allow you to tag channels with a textual label on the scope display (not the probe itself
<azonenberg>
i plan to add an API for this and use the scopehal channel name there. MAXWELL will use the same API to write channel names to the pod LCD
<azonenberg>
finally i do not yet support any advanced trigger types. Only edge triggers so far
<electronic_eel>
but just tagging it on the scope is something glscopeclient could easily do on it's own without any support from the scope itself
<azonenberg>
I mean, i already have channel names displayed in glscopeclient
<azonenberg>
But i need an API to push that name to the instrument
<azonenberg>
so that scopes that have a means of displaying the channel name as text can be synced with the name in glscopeclient
<electronic_eel>
I got my new gpu, amd radeon rx 570. it is working now in my old pc, after I got a few power adapter cables
<azonenberg>
Great
<azonenberg>
So at this point i can spend some time working on additional trigger types for the mso6 as time permits
<azonenberg>
as long as i have access to it
<electronic_eel>
but I haven't had the time to play with glscopeclient yet, want to finish a pcb layout first and send it off to fab. plan to play with it next week
<azonenberg>
Sounds good
<azonenberg>
what scope are you going to be using it with?
<azonenberg>
also did you finally get the probe?
<electronic_eel>
rigol mso4000 series
<electronic_eel>
a bit older one, but 500 MHz and proper 50 ohms mode
<azonenberg>
I don't think that's ever been tested so it will be good to test
<azonenberg>
we have mso5000 series and i dont know how similar the scpi protocols are
<electronic_eel>
the parcel with the probes is still in customs. other stuff is stuck there too, so I guess they are a bit overloaded at the moment
<azonenberg>
if nothing else you will probably have to add a check for mso4 to the model detection code
<electronic_eel>
yeah, maybe. I haven't looked too deeply at the scpi section in the manual, just was looking for an easy way to pull screenshots over lan
<electronic_eel>
so maybe I have to take a deeper look now
<electronic_eel>
the built in math section of this scope really sucks, it just works on the graphics buffer, not on the real data.
<electronic_eel>
I hope to not have to use it again and can rely on glscopeclient for this in the future
bvernoux has joined #scopehal
<bvernoux>
great
<bvernoux>
I see news about Tek6 support
<Bird|otherbox>
azonenberg: do we intend to support OpenMP implementations aside from GNUs? I think I figured out (by looking at the libgomp source code) how to achieve the effects of the PASSIVE setting programmatically
<azonenberg>
Bird|otherbox: Yes, we do. But I am OK with a GOMP specific hack as long as you can reliably (via cmake or other means) determine that we are using GOMP
<azonenberg>
and only enable it if we're using GOMP
_whitelogger has joined #scopehal
_whitelogger has joined #scopehal
<_whitenotifier-f>
[scopehal-apps] azonenberg opened issue #200: Automatically create new waveform group for spectrum channels on Tek 5/6 series - https://git.io/JU6At
<_whitenotifier-f>
[scopehal-apps] azonenberg labeled issue #200: Automatically create new waveform group for spectrum channels on Tek 5/6 series - https://git.io/JU6At
<_whitenotifier-f>
[scopehal] azonenberg pushed 1 commit to master [+0/-0/±4] https://git.io/JU6A3
<_whitenotifier-f>
[scopehal] azonenberg d42ccc8 - Early work on support for Tek 5/6 series spectrum view channels. Can turn them on and off but not much else. See #269.
bvernoux has quit [Quit: Leaving]
<_whitenotifier-f>
[scopehal] azonenberg pushed 2 commits to master [+0/-0/±2] https://git.io/JUiey
<_whitenotifier-f>
[scopehal] azonenberg a34c06e - TektronixOscilloscope: very early support for spectrum view. Can display waveforms but not change frequencies etc. Cannot query gain/offset yet. See #269.
<_whitenotifier-f>
[scopehal] azonenberg b2abc4c - TektronixOscilloscope: minor tweaks to internal implementation of spectrum view. See #269.